Digestibility and nitrogen and water balance in horses fed rhizoma peanut hay

Abstract

Rhizoma peanut (RP, Arachis glabrata) hay has the potential to meet horses’ crude protein (CP) requirements with less nitrogen excretion than other legumes. This study aimed to evaluate nutrient intake, apparent digestibility, and nitrogen balance of RP “Florigraze” hay compared with alfalfa (ALF, Medicago sativa L. “Legendary XHD”) and bermudagrass (BG, Cynodon dactylon L. “Coastal”) hays when fed to maintenance horses at 2% body weight/d on a dry matter (DM) basis. We hypothesized that nutrient intake would be comparable between the legume hays and lesser for BG and that RP would result in reduced nitrogen excretion compared with ALF. Six mature Quarter Horse geldings (593 ± 40 kg; mean ± SD) were randomly assigned to one of the hays in a replicated 3 × 3 Latin square with 21-d periods. A 14-d adaptation phase was followed by a 3-d total fecal and urine collection. Days 18 to 21 were used for a companion study. Intake of nutrients is reported on a DM basis. Digestible energy (DE) intakes from ALF (29.91 Mcal/d) and RP (29.37 Mcal/d) were greater (P < 0.0001) than BG (20.78 Mcal/d). CP intake was greater (P < 0.0001) for ALF (2.5 kg/d), followed by RP (1.9 kg/d) and BG (1.5 kg/d). All hays exceeded maintenance requirements for DE, CP, Ca, and P. Apparent digestibility of DM and CP was greatest (P < 0.0001) for ALF (69% and 84%), intermediate for RP (61% and 72%), and least for BG (46% and 64%). Apparent digestibility of neutral detergent fiber did not differ (P = 0.2228) among hays, while digestibility of acid detergent fiber (P = 0.0054) was least for RP but similar for ALF and BG. Water intake (kg/d) for ALF (57) was greater (P = 0.0068) than RP (45) and BG (41). Greater (P = 0.0271) water retention (kg/d) was observed for ALF (13.5), followed by RP (10.8) and BG (7.5). There was a difference in nitrogen excretion, with greatest urinary nitrogen excretion for ALF (P < 0.0001) and greatest fecal nitrogen excretion for BG (P = 0.0001). Total nitrogen excretion was greater (P < 0.0001) for ALF (278 g/d), followed by RP (211 g/d) and BG (179 g/d). Nitrogen retention was greater (P = 0.0005) for ALF when represented as g/d (ALF: 129, RP: 86, and BG: 57 g/d) but similar (P = 0.0377) to RP when presented as percent of nitrogen intake (ALF: 32%, RP: 29%, and BG: 24%). Results indicate that RP hay is a suitable legume for horses by meeting DE and CP requirements and having a significant reduction in nitrogen compared with ALF.

Introduction

Rhizoma peanut (RP, Arachis glabrata Benth) is a warm-season, perennial legume proven to combine high nutritive value and high persistence under a wide range of environmental conditions and management practices (Ortega-S et al., 1992; Quesenberry et al., 2010; Muir et al., 2011). It is primarily cultivated as a hay crop for livestock, and its nutritive value is commonly compared with alfalfa (ALF) when fed to ruminants (Gelaye and Amoah, 1991). However, protein concentrations are usually lower in RP (14% to 17% CP) than in ALF (18% to 19% CP) (Gelaye and Amoah, 1991; Terrill et al., 1996). Research on the nutritive value of RP hay for horses as compared with ALF (Medicago sativa L.) and other common forage sources, such as bermudagrass (BG, Cynodon dactylon L.), is scarce. To our knowledge, only two studies (Lieb et al., 1993; Eckert et al., 2010) have evaluated the nutritive value of RP for horses, and only dry matter (DM), crude protein (CP), and neutral detergent fiber (NDF) intake and their digestibilities were reported.

ALF is a widely grown and popular forage legume fed to horses due to its nutrient profile. It is well accepted by horses and its voluntary intake is typically 1.2 to 1.6 times greater than for grass hays (Cymbaluk, 1990; LaCasha et al., 1999; Rodiek and Jones, 2012). However, the high soil fertility demands of ALF cause it to be a high-cost hay. Additionally, some horse owners are reluctant to feed ALF due to its excess nutrients, potentially resulting in excessive weight gain. In the southern United States, BG hay (especially “Coastal”) is a less expensive, locally grown forage that is well accepted by horses, but its nutritive value is often lesser than legumes and cool-season grass hays (LaCasha et al., 1999; Sturgeon et al., 2000; Eckert et al., 2010). Additionally, horse owners are often hesitant to feed Coastal BG hay due to its perceived association with colic; however, there is little evidence of a significant effect of feeding this warm-season forage on the incidence of colic (Hudson et al., 2001; Little and Blikslager, 2010).

Horse owners desire quality forage sources that provide adequate nutrient intakes for the horses being fed. Forages are usually selected based on their concentration of digestible energy (DE) and CP. Legume forages are generally greater in DE (2.44 ± 0.13 Mcal/kg DM basis; mean ± SD) and CP levels (21 ± 2.6% DM basis) compared with grass hays (2.07 ± 0.24 Mcal/kg DM basis and 10.9 ± 3.9% DM basis, respectively) (Dairy One, 2021). Although appealing, the intake of CP in excess of requirements leads to increased nitrogen losses to the environment, contributing to environmental impacts, such as groundwater contamination and decreased air quality (Knowlton and Cobb, 2006). Therefore, selecting forages that meet the required nutrient intakes but minimizing excess protein intake may help reduce nitrogen excretion. To our knowledge, Eckert et al. (2010) was the only study to report data on nitrogen balance of horses fed RP hay compared with BG, but RP has never been compared with ALF hay.

RP’s high productivity and persistence in various management and environmental conditions, along with its relatively lower CP levels compared with ALF, make RP a possible alternative source of nutrition that combines high nutritive value with potential to reduce environmental impacts. We hypothesized that RP would be as digestible as ALF and that BG would be less digestible than the legume hays. Moreover, we hypothesize that the intermediate level of CP in RP would result in lesser nitrogen excretion compared with ALF but greater than BG.

Materials and Methods

All animal protocols were approved by the University of Florida Institutional Animal Care and Use Committee (201710092) under the Federation of Animal Sciences Societies Guide for the Care and Use of Agricultural Animals in Research and Teaching (FASS, 2010).

Horses and dietary treatments

Six mature Quarter Horse geldings (10 ± 4 yr and 596 ± 39 kg of body weight [BW]; mean ± SD) were housed at the University of Florida, Institute of Food and Agricultural Sciences Equine Sciences Center located in Ocala, Florida (29°17′46″N, 82°9′53″W) from December 2019 to February 2020. During the study, the mean temperature was 16 ± 4 °C and relative humidity was 82 ± 10%. Before the study, the horses received routine deworming and vaccinations, and dental floating if needed. During the study, the horses were individually housed in 4 × 5 m stalls bedded with wood shavings and turned out in pairs in 10 × 30 m outdoor grass-free paddocks with sand footing for 4 h each day for voluntary exercise. Horses had unlimited access to fresh, clean water while in stalls or in paddocks.

Dietary treatments consisted of three commercially available hays (Table 1): RP cv Florigraze, ALF cv Legendary XHD, and BG cv Coastal. All hays were harvested in 2019 and purchased from local commercial hay dealers. Hays were selected based on recommended harvesting intervals of 4 wk for ALF and 6 wk for RP and BG. ALF hay was grown and shipped in from Idaho and RP and BG were grown in Florida.

Table 1.

Nutrient composition of the alfalfa, rhizoma peanut, and bermudagrass hays fed to mature horses (dry matter [DM] basis)¹

	Hay
Item	Alfalfa	Rhizoma peanut	Bermudagrass
DM	90.9	90.1	93.5
Organic matter	88.9	91.1	92.7
Digestible energy, Mcal/kg DM	2.58	2.53	1.94
Crude protein	22.0	16.0	13.6
Ether extract	2.4	2.4	2.0
Neutral detergent fiber	32.9	40.4	66.4
Acid detergent fiber	25.7	30.0	37.4
Hemicellulose	7.2	10.4	29.0
Cellulose	19.8	22.7	32.2
Lignin	5.9	7.3	5.2
Ethanol-soluble carbohydrates	8.1	7.7	3.7
Water-soluble carbohydrates	10.1	8.8	6.9
Starch	1.3	1.3	0.9
Nonstructural carbohydrates	11.4	10.1	7.8
Calcium	1.64	1.34	0.50
Phosphorus	0.26	0.30	0.28
Calcium:Phosphorus, ratio	6.3	4.5	1.8

¹Nutrients are reported as percent (%) unless otherwise specified. In addition to the hay diet, horses also had ad libitum access to a mineral–vitamin supplement during the study containing a minimum of 15% calcium, 8% phosphorus; 12% sodium chloride, 1% magnesium, 1,000 ppm copper, 45 ppm iodine, 4,000 ppm manganese, 19 ppm selenium, 2,950 ppm zinc, 198,416 IU vitamin A/kg, 19,841 IU vitamin D/kg, and 1,653 IU vitamin E/kg.

Dietary treatments were fed at 2% BW (DM basis) into three meals representing 30%, 30%, and 40% of daily DM fed at 0700, 1500, and 2200 hours, respectively. Hays were offered to horses in 66-L rubber-polyethylene containers with a 46-cm diameter opening and 36 cm depth. Clean, fresh water and a commercial mineral supplement (Mineral Equalizer; Sweetlix, Mankato, MN) were provided ad libitum. Waste hay was removed, and stalls were cleaned daily before each morning feeding.

Experimental design and sample collection

Hays were evaluated in a replicated 3 × 3 Latin square design experiment. Horses were weighed using a livestock platform scale (PS2000; Salter Brecknell, Fairmount, MN) before starting the study and separated into two groups (squares) with similar BW. Within each group (square), horses were randomly assigned to one of the three hay treatments. After each experimental period, all horses were group-housed in a bahiagrass (Paspalum notatum Flüggé) pasture with free access to BG round bales hay during a 4-d washout. At the beginning of each experimental period, the assigned hay was introduced gradually over a period of 4 d (25%, 50%, 75%, and 100%) by mixing with BG hay. Each experimental period lasted 21 d and consisted of a 14-d dietary adaptation (day 1 to 14), followed by a 3-d total fecal and urine collection (day 15 to 17). The last 4 d of every period were used for a companion study that evaluated hay preference and feeding behavior(Vasco et al., 2021).

Before study initiation, 20 bales of each treatment hay were randomly cored and immediately shipped to a commercial forage testing laboratory (Equi-Analytical, Ithaca, NY) to determine nutrient composition, which was used for calculations of nutrient intake. During the first experimental period, horses were fitted and adapted to a harness specifically designed to collect voided feces and urine separately (Stablemaid, Melbourne, AUS) on days 5, 9, and 11, for 30, 60, and 120 min, respectively. The harness design allowed horses to lie down and move comfortably in the stall.

On day 14 of each period at 2200 hours, the harnesses were placed on all horses for the collection phase. On day 15 at 0600 hours, feces and urine deposited in the harness overnight were discarded and stalls were stripped of bedding and swept clean to formally begin the sample collection phase. Fecal and urine samples were subsequently collected at 8-h intervals (0700, 1500, and 2200 hours). At each collection interval, total feces voided were weighed, recorded, homogenized, and a 10% representative subsample was temporarily stored at 4 °C. Urine was weighed and volume excreted was determined using a graduated cylinder. Urine was homogenized, filtered through a double layer of cheesecloth, and a 200-mL subsample was acidified with hydrochloric acid (12 M) at 2% (vol/vol) and temporarily stored at 4 °C. For each 24-h collection cycle, the three fecal subsamples (1500, 2200, and 0700 hours) were composited by horse and homogenized, and a 200-g daily sample was stored at −20 °C for further analysis. Urine subsamples were processed similarly, and a 100-mL daily sample was stored at −20 °C for posterior analysis. When horses were relieved of their harnesses to permit sample collection, horses were turned out in grass-free paddocks for 15 min of daily self-exercise. While turned out without the harness, any voided feces or urine was collected or caught by bucket. These samples were included in the daily measurements for total excretion and chemical analysis.

The amount of hay offered at 0700, 1500, and 2200 hours was recorded to measure daily intake and digestibility of DM and nutrients. Additionally, total refusals (orts) were collected immediately prior to the 0700-hours feeding of the following day. Orts were collected, homogenized, subsampled, and stored at 4 °C for subsequent analysis of nutrient composition for calculations of intake and digestibility of DM and nutrients. Water intake was measured at 8-h intervals using the water disappearance method. Briefly, water was offered into three 19-L rubber-polyethylene containers with a 46 cm diameter opening and 36 cm depth. Full containers were placed into the stall, and water disappearance was recorded 8 h later. Containers were subsequently rinsed, refilled, and returned to the stall. Daily water intake was recorded in kilograms, where 1 kg is equivalent to 1 L, and determined as the sum of water disappearance measured at 1500, 2200, and 0700 hours.

Chemical analysis and calculations

Orts and fecal samples were thawed at 4 °C for 48 h, and representative samples from each 24-h collection were dried in a forced-air oven at 60 °C until a constant weight was achieved. Fecal samples were dried in duplicates and the DM concentration was used to report fecal DM. Duplicates were subsequently combined and processed for analysis of nutrient composition. Samples were ground to pass a 1-mm sieve using a Wiley mill and sent to a commercial forage testing laboratory (Equi-Analytical BW, Ithaca, NY) for analysis of nutrient composition. Hay, orts, and fecal samples were analyzed by wet chemistry using the analytical procedures described in the Dairy One’s September 2020 update (https://dairyone.com/download/forage-forage-lab-analytical-procedures). DM was determined using the procedure AOAC 930.15 (AOAC, 2010), and DE was calculated using the equation developed by Pagan (1998). CP was determined by multiplying the percentage of nitrogen by 6.25 using the procedure 990.03 (AOAC, 2010). NDF, acid detergent fiber (ADF), and lignin were calculated using the ANKOM Technology methods 15, 14, and 9, respectively (ANKOM Technology, 2017a, 2017b, 2020a). Starch, water-soluble carbohydrate (WSC), and ethanol-soluble carbohydrate (ESC) were measured using techniques described by Hall et al. (1999). Nonstructural carbohydrates were calculated as the sum of WSC and starch. Ether extract (EE) was determined using the ANKOM Technology method 2 (ANKOM Technology, 2020b). Ash was determined using the method 942.05 (AOAC, 2010). Organic matter (OM) was calculated as 100 minus the percentage of ash. Mineral concentrations were determined using the Thermo Jarrell Ash IRIS Advantage HX Inductively Coupled Plasma Radial Spectrometer (Thermo Instrument Systems Inc., Waltham, MA) after microwave digestion using the Microwave Accelerated Reaction System (CEM, Mathews, NC). Hemicellulose was determined as the difference between NDF and ADF, and cellulose was calculated as the difference between ADF and lignin.

Dry matter intake (DMI) was measured as the difference of the daily amount of hay offered (kg DM) minus the daily amount of orts (kg DM). Nutrient intake was calculated using the following equation:

$${\displaystyle \begin{array}{ll}{\displaystyle \mathrm{I}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{k}\mathrm{e}\left(\frac{\mathrm{k}\mathrm{g}}{\mathrm{d}}\right)=}& [\mathrm{D}\mathrm{M}\mathrm{I}\left(\mathrm{k}\mathrm{g}\mathrm{D}\mathrm{M}\right)\times \mathrm{\%}\mathrm{n}\mathrm{u}\mathrm{t}\mathrm{r}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{i}\mathrm{n}\mathrm{h}\mathrm{a}\mathrm{y}]\\ & -[\mathrm{O}\mathrm{r}\mathrm{t}\mathrm{s}(\mathrm{k}\mathrm{g}\mathrm{D}\mathrm{M})\times \mathrm{\%}\mathrm{n}\mathrm{u}\mathrm{t}\mathrm{r}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{i}\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{t}\mathrm{s}]\end{array}}$$

Daily intake of mineral supplement was not controlled or measured. Therefore, intakes of Ca and P reported in the current study are exclusively from hay intake.

Apparent digestibility of DM and nutrients was calculated using the daily intake of DM or nutrient and daily DM or nutrient fecal output, in kg DM, using the following equation:

$${\displaystyle \begin{array}{ll}& {\displaystyle \mathrm{D}\mathrm{i}\mathrm{g}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{i}\mathrm{b}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{t}\mathrm{y}\left(\mathrm{\%}\right)=}\\ & \frac{(\mathrm{I}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{k}\mathrm{e}\mathrm{o}\mathrm{f}\mathrm{D}\mathrm{M}\mathrm{o}\mathrm{r}\mathrm{n}\mathrm{u}\mathrm{t}\mathrm{r}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{t}-\mathrm{F}\mathrm{e}\mathrm{c}\mathrm{a}\mathrm{l}\mathrm{o}\mathrm{u}\mathrm{t}\mathrm{p}\mathrm{u}\mathrm{t}\mathrm{o}\mathrm{f}\mathrm{D}\mathrm{M}\mathrm{o}\mathrm{r}\mathrm{n}\mathrm{u}\mathrm{t}\mathrm{r}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{t})}{\mathrm{I}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{k}\mathrm{e}\mathrm{o}\mathrm{f}\mathrm{D}\mathrm{M}\mathrm{o}\mathrm{r}\mathrm{n}\mathrm{u}\mathrm{t}\mathrm{r}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{t}}\times 100\end{array}}$$

Urine samples were sent to the Forage Laboratory at the UF/IFAS North Florida Research and Education Center located in Marianna, Florida, for analysis of total nitrogen by dry combustion using an automated analyzer (Vario Micro cube, Elementar, Langenselbold, DE) coupled to an isotope ratio mass spectrometer (IsoPrime 100, IsoPrime, Cheadle, UK).

For water balance, water intake (kg/d) was calculated as the daily amount of water offered (kg) minus the daily amount of water remaining (kg). For water excretion (kg/d), urinary water excretion (kg/d) was measured as the total urine excreted in the 24-h collection. Similarly, fecal water excretion (kg/d) was calculated as the amount of feces excreted (as-is basis) multiplied by the percent of fecal moisture. Total water excretion (kg/d) was calculated as the sum of urinary and fecal water excretion. Water retained (kg/d) was calculated as the water intake minus total water excretion.

For nitrogen balance, nitrogen intake (g/d) was calculated as the intake of CP divided by 6.25, and intake of digestible nitrogen (DN, g/d) was calculated as daily intake of nitrogen (g DM/d) multiplied by the CP apparent digestibility. For nitrogen excretion (g/d), urinary nitrogen excretion was calculated as the daily urinary water excretion (kg/d) multiplied by the concentration of urine nitrogen. Fecal nitrogen excretion (g/d) was calculated as the daily fecal output (kg DM) multiplied by the concentration of fecal nitrogen. Total nitrogen excretion (g/d) was calculated as the sum of urinary and fecal nitrogen excretion. For nitrogen retention (or balance), absolute nitrogen retention (g/d) was calculated as the nitrogen intake minus total nitrogen excretion, and relative nitrogen retention (% of nitrogen intake) was calculated as absolute nitrogen retention divided by nitrogen intake.

Statistical analysis

Nutrient intake and digestibility variables were calculated by day, and statistical analysis was performed on the 3-d average for each collection within a period. Data were tested for normal distribution using the Shapiro–Wilk test and submitted to analysis of variance using the GLIMMIX procedure in SAS (SAS Inst. Inc, Cary, NC). The model included square, period, horse(square), and hay treatment as fixed effects. A Tukey’s post hoc test was used to determine differences between the means. Additionally, Pearson’s correlation coefficients were obtained using the CORR procedure in SAS to test the relationship between nutrient intake and water balance and nutrient intake and digestibility. Results were considered significant at P ≤ 0.05. Data are presented as least square means ± standard error of the mean (SEM).

Results and Discussion

Intake and digestibility

Daily nutrient intake is presented in Table 2. Hay diets were provided at 2% of BW and daily DMI was similar (P = 0.098) across when expressed as kg/d. The DMI in this study averaged 11.8 kg DM/d and is within the expected range for mature horses consuming all-forage diets (NRC, 2007; Eckert et al., 2010; Staniar et al., 2010; Miyaji et al., 2014). Except for the cell-wall components and P, intakes of DE, CP, EE, ESC, WSC, starch, NSC, Ca, and Ca:P were greater (P < 0.0011) for legume hays compared with BG. Intakes of NDF, ADF, hemicellulose, and cellulose were greater (P < 0.0001) for BG compared with both legume hays. Because the DMI expressed as kg/d did not differ across hay types, the differences in DE and nutrient intake can be explained by differences in forage nutrient composition. Legumes are generally greater in CP, DE, lignin, and Ca than warm-season grasses (Van Soest, 1994; Buxton et al., 1995; Moore et al., 2004; Blevins and Barker, 2007; Hansen et al., 2019). Concentration of NDF in legumes is typically lower than grasses due to more pectin and less hemicellulose in the cell wall of legumes, which explains the lower intake of hemicellulose for the legumes compared with BG. Greater concentrations of CP in ALF compared with RP hay found in this study were consistent with previous studies (Romero et al., 1987; Terrill et al., 1996). The greater cell-wall content in RP compared with ALF may be a function of the high-temperature tropical climate where RP usually grows, which directs a greater proportion of the photosynthesis products to structural components (Fales and Fritz, 2007). Furthermore, stage of maturity and harvesting management are important factors affecting fiber concentration of the forage (Lloyd et al., 1961; Burns et al., 1997; Palmonari et al., 2014).

Table 2.

Intake of nutrients by mature horses fed alfalfa, rhizoma peanut, and bermudagrass hays¹

Daily intake	Hay			SEM	P-value
	Alfalfa	Rhizoma peanut	Bermudagrass
Dry matter	11.8	11.8	12.0	0.05	0.0939
Organic matter	10.4	10.6	10.1	0.14	0.0718
Digestible energy, Mcal/d	29.91a	29.37a	20.78b	0.35	<0.0001
Crude protein	2.54a	1.86b	1.48c	0.025	<0.0001
Neutral detergent fiber	3.86c	4.68b	7.18a	0.090	<0.0001
Hemicellulose	0.84c	1.22b	3.15a	0.038	<0.0001
Acid detergent fiber	3.00c	3.47b	4.07a	0.050	<0.0001
Cellulose	2.32c	2.62b	3.53a	0.042	<0.0001
Ether extract	0.28a	0.28a	0.20b	0.010	0.0011
Ethanol-soluble carbohydrates	0.94a	0.90a	0.41b	0.010	<0.0001
Water-soluble carbohydrates	1.17a	1.03b	0.71c	0.010	<0.0001
Starch	0.15a	0.15a	0.08b	0.003	<0.0001
Nonstructural carbohydrates	1.31a	1.17b	0.80c	0.015	<0.0001
Calcium, g/d	189a	155b	53c	1.9	<0.0001
Phosphorus, g/d	30b	35a	31b	0.4	<0.0001
Calcium:Phosphorus, ratio	6.3a	4.4b	1.7c	0.01	<0.0001

¹Intake is expressed as kg/d unless otherwise specified.

^a–cWithin row, means without a common letter differ based on a Tukey’s test (P ≤ 0.05).

Greater concentrations of NSC, WSC, ESC, and starch usually found in legumes may be partially explained by their decreased growth rates compared with BG. The accelerated growth rate of warm-season grass results in greater usage of NSC for the development of seed heads (Longland and Byrd, 2006). Additionally, time of the day at which forages are harvested and environmental and agronomic conditions where forages are gown may play a role in the NSC levels of the hays and, consequently, the NSC intake (Longland and Byrd, 2006; Shewmaker et al., 2006). The NSC intake as percent of total DMI observed for ALF (11.1%) was above the maximum suggested (10%) for horses with metabolic disorders, such as insulin resistance, while NSC intake for RP was at the maximum level suggested (Geor and Harris, 2009; Frank, 2011). Nonetheless, either hay type may be suitable for metabolically challenged horses if NSC is below 10% and NSC analysis of either hay is recommended before feeding.

All hays exceeded nutrient requirements of DE, CP, Ca, and P of the horses in the current study (Figure 1). Based on concentrations of DE and CP of the hays used in the current study, a 596-kg horse at maintenance could have both requirements met when fed at 1.3%, 1.3%, and 1.7% of BW of ALF, RP, and BG hays, respectively (NRC, 2007). In this scenario, however, the legume hays would not meet the minimal DMI of 1.5% BW when provided as the only source of nutrients but would be above the absolute minimum of 1.25% BW recommended for maintaining equine health and ethological needs (Harris et al., 2017). This suggests that ALF and RP may be suitable ingredients within the diet of horses with higher nutritional requirements. The ratio of calcium to phosphorus (Ca:P) must be taken into consideration when balancing horse diets. Adequate intake of P allows Ca:P ratio of up to 6:1 in the diet of most classes of horses (NRC, 2007). All hays used in the current study met both Ca and P requirements.

Figure 1.

Depiction of the percentage of the recommended nutrient requirements (P ≤ 0.05) met by alfalfa, rhizoma peanut, or bermudagrass hays fed to the maintenance horses in this study. Numbers within bars represent the percent of nutrient requirement met by each hay. Dashed line represents 100% of requirement. ^a–cWithin nutrient, means without a common letter differ based on a Tukey’s test (P ≤ 0.05).

Apparent digestibility of nutrients is presented in Table 3. Greater DM digestibility (DMD) (P < 0.0001) was observed for legume hays compared with BG, and for ALF compared with RP. Greater DMD for legume hays compared with warm-season grasses is commonly reported in the literature (LaCasha et al., 1999; Sturgeon et al., 2000; Eckert et al., 2010; Hansen et al., 2019). Forage chemical composition and leaf anatomical structure play an essential role in plant digestibility (Wilson and Kennedy, 1996). Increased DMD is commonly associated with increased CP levels and decreased levels of NDF, with NDF concentration being reported as the best predictor of DMD (Edouard et al., 2008; Hansen and Lawrence, 2017), which may explain the superiority of ALF over RP hay in DMD. The greater DMD of ALF compared with RP may be explained by the decreased concentration of the cell-wall components, NDF, ADF, and lignin, and increased concentration of CP observed in the ALF hay compared with RP used in the current study. Apparent digestibility of OM and CP was greatest for ALF (P < 0.0001), intermediate for RP, and least for BG. The OM digestibility mirrored the differences in DMD, which is consistent with published data for ALF, RP, and BG hays (LaCasha et al., 1999; Eckert et al., 2010; Hansen et al., 2019). Greater CP digestibility for RP and ALF compared with BG has been previously reported in the literature (LaCasha et al., 1999; Sturgeon et al., 2000; Eckert et al., 2010). Compared with legume forages, warm-season grasses generally have higher concentrations of cell-wall components which are known to negatively impact CP digestibility (Glade, 1984; Van Soest, 1994). Therefore, the increased CP digestibility of legume hays compared with BG in the current study was expected. Additionally, the digestibility of CP is influenced by the level of CP in the diet (Gibbs et al., 1988; Oliveira et al., 2015), which supports the CP digestibility results found in this study.

Table 3.

Apparent digestibility of nutrients in maintenance horses fed alfalfa, rhizoma peanut, and bermudagrass hays

Digestibility, %	Hay			SEM	P-value
	Alfalfa	Rhizoma peanut	Bermudagrass
Dry matter	69.2a	60.7b	46.2c	0.5	<0.0001
Organic matter	70.0a	62.0b	46.2c	0.6	<0.0001
Crude protein	83.6a	71.7b	63.8c	0.3	<0.0001
Neutral detergent fiber	36.6	35.6	38.5	0.9	0.2228
Acid detergent fiber	37.8a	29.8b	36.0a	1.1	0.0054
Ether extract	50.2	53.2	51.8	1.8	0.5630
Ethanol-soluble carbohydrates	97.7a	96.3a	90.5b	0.7	0.0006
Water-soluble carbohydrates	94.2a	89.7b	82.8c	0.6	<0.0001
Starch	86.0a	87.3a	63.8b	1.7	<0.0001
Nonstructural carbohydrates	93.1a	88.7b	79.0c	0.4	<0.0001

^a–cWithin row, means without a common letter differ based on a Tukey’s test (P ≤ 0.05).

Digestibility of starch and ESC for RP was greater (P < 0.0001) than for BG and similar to ALF, while NSC and WSC digestibilities were greatest (P < 0.05) for ALF and least for BG. The intake of NDF was negatively correlated (P < 0.0001) with digestibility of WSC (r = –0.91), ESC (r = –0.86), starch (r = –0.88), and NSC (r = –0.94). Therefore, the daily amount of fiber ingested may have affected the digestibility of the soluble carbohydrates in the hays. There was no difference among hays for digestibility of EE (P = 0.5630) and NDF (P = 0.2228). Similar EE digestibility between ALF and BG hays fed to maintenance horses has been reported by Sturgeon et al. (2000). Similar NDF digestibility between legumes and grasses is consistently reported in the literature (Crozier et al., 1997; Sturgeon et al., 2000; Eckert et al., 2010; Hansen et al., 2019). Both the concentration and type of cell-wall components may have altered NDF digestibility in the current study. Hansen et al. (2019) reported that high-fiber forages have longer mean retention times than low-fiber forages. The longer retention time of high-fiber forages in horses results in longer exposure of the cell-wall components to microbial fermentation in the hindgut, which compensates for the slower rate of digestion of warm-season grasses such as BG (Koller et al., 1978; Miyaji et al., 2014; Hansen et al., 2019).

Digestibility of ADF was similar for ALF and BG hays and the least (P = 0.0054) for RP. The reduced ADF digestibility for RP could be a consequence of the increased lignin concentration in RP hay (7.3%) when compared with ALF (5.9%) and BG (5.2%) hays. In legumes leaves, lignin is exclusively located in the vascular bundle, while it is evenly distributed throughout the tissues of warm-season grasses (Akin, 1989; Wilson and Kennedy, 1996). Therefore, lignin in legumes is of less consequence to digestibility compared with grasses, which might explain why the increased lignin concentration in RP affected ADF digestibility but not DMD.

Water and nitrogen balances

Differences among hays (P < 0.05) were observed for water intake, urinary water excretion, fecal water excretion, total water excretion, and water retention (Table 4). Water intake was greatest (P = 0.0068) when horses were fed ALF, but similar between RP and BG. The effect of DMI on water intake is well established in the literature, with greater water intake associated with greater DMI (Hintz, 1994). However, in the current study, DMI (kg DM/d) was not correlated (P = 0.1991; r = 0.33) with water intake. Therefore, the level of CP in the diet may have played a role in the water intake, as reported by Oliveira et al. (2015). Results in this study show that water intake was significantly correlated to CP intake (P = 0.0037; r = 0.66) as has been reported by Hyslop (2004).

Table 4.

Water balance in maintenance horses fed alfalfa, rhizoma peanut, and bermudagrass hays

Item	Hay			SEM	P-value
	Alfalfa	Rhizoma peanut	Bermudagrass
Water intake, kg/d	56.7a	44.7b	41.5b	2.2	0.0068
Urinary water excretion, kg/d	27.9a	14.1b	8.4c	1.4	0.0001
Fecal water excretion, kg/d	16.6c	21.0b	26.5a	0.8	0.0002
Total water excretion, kg/d	44.5a	35.1b	34.9b	1.4	0.0060
Water retained, kg/d	13.5a	10.8a,b	7.5b	1.1	0.0271
Fecal dry matter, %	18.0	17.9	18.1	0.4	0.9651

^a–cWithin row, means without a common letter differ based on a Tukey’s test (P ≤ 0.05).

Differences in urinary water excretion (kg/d) between hays followed a pattern similar to water intake, where urinary water excretion was greatest (P = 0.0001) when horses consumed ALF, intermediate for RP, and least for BG. Excess nitrogen is primarily excreted through urine; therefore, increased urinary water excretion is expected when CP levels are in excess of their requirement and horses have free access to water, resulting in increased water intake and hence urinary excretion (Cymbaluk, 1989; Eckert et al., 2010; Oliveira et al., 2015).

Fecal water excretion was greater (P = 0.0002) for BG, followed by RP, and lesser for ALF. The pattern of water excretion from urine vs. feces changed when horses were fed ALF vs. BG, with RP being intermediate. Fecal water excretion is highly influenced by the daily fecal output and is negatively correlated with DMD (Fonnesbeck, 1968). Therefore, the least DMD for BG, intermediate for RP, and greater for ALF may explain the greater fecal water excretion for BG, followed by RP, and the least for ALF. Nonetheless, total water excretion was greatest (P = 0.0060) for ALF-fed horses and lowest for horses fed BG and RP. Despite the increased total water excretion, horses fed ALF had greater (P = 0.0271) water retention (kg/d) compared with BG-fed horses. The level and type of fiber in the diet are positively related to the volume of water retained in the hindgut affecting the content of water excreted via feces and hence fecal DM (Cymbaluk, 1989; Cuddeford et al., 1992). As the fiber components are fermented, the retained water is gradually released, which can lessen the effect of fiber on fecal DM (Brøkner et al., 2012).

Fecal DM (%) was similar (P = 0.9651) across all hays (18 ± 0.4%; mean ± SE). Therefore, the similarity in fecal DM across hay diets and increased water retention for ALF followed by RP could be explained by the type of fiber and extent of fiber digestion. The similarity in fecal DM could be attributed to the similar NDF digestibility across all hays, which might have attenuated the effect of the different fiber levels in the hays used. Concurrently, the increased level of pectin in legumes may have played a role in increasing the water retained in horses fed ALF and RP. Pectins alter the viscosity of the digesta and are highly related to the water-holding capacity of the digesta in the hindgut (Eastwood, 1992). Considering the similar fecal DM across hays, daily fecal output (as-is basis) from horses fed BG (32.3 kg feces/d), RP (25.6 kg feces/d), and ALF (20.1 kg feces/d) is directly related to the level of fecal water excreted on each diet. The greatest fecal output from horses consuming BG and least from ALF, combined with similar DMI, supports the least DMD for the BG hay and greatest for the ALF hay.

The effect of hay on nitrogen balance is presented in Table 5. As expected, nitrogen intake reflected the CP concentration of the hays, with greatest (P < 0.0001) nitrogen intake for ALF and least for BG. Urinary nitrogen excretion was greater (P < 0.0001) in ALF-fed horses and lesser in BG-fed horses. In contrast, fecal nitrogen excretion in RP- and BG-fed horses was similar and greater (P = 0.0001) than ALF-fed horses. The hay-dependent shift in nitrogen excretion from feces to urine observed in the current study may be explained by the differences in CP intake and DN of the hays evaluated. Increased DN intake is reported to cause greater urinary nitrogen excretion, while a lesser DN intake causes greater fecal nitrogen excretion (Graham-Thiers and Bowen, 2011). In the current study, ALF hay provided the greatest (P < 0.0001) intake of digestible N, while BG hay provided the least. Increased fecal nitrogen excretion from BG-fed horses can also be explained by more fiber fermentation resulting in more microbial protein synthesis and excretion via feces (Meyer, 1983). In support of this, NDF intake was greater for BG, while NDF digestibility was similar across hays. Total nitrogen excretion paralleled nitrogen intake, with greater (P < 0.0001) excretion observed in horses fed ALF and lesser in horses fed BG. The excess of nitrogen excreted may contribute to environmental impacts such as nitrate leaching to groundwater sources and decreased air quality (Bott et al., 2016). The sources of nitrogen excretion (urine vs. feces) differ in their potential of nitrogen pollution of the environment, where urinary nitrogen is more susceptible to losses than fecal nitrogen (Bussink and Oenema, 1998; Weir et al., 2017). Nitrogen mineralization in fecal organic matter is a slower process, which decreases its susceptibility to losses compared with urinary nitrogen (Waldrip et al., 2015). Therefore, feeding diets that provide reduced nitrogen input into the environment and a shift in nitrogen excretion from urine to feces are pivotal for more sustainable management practices.

Table 5.

Nitrogen balance in maintenance horses fed alfalfa, rhizoma peanut, and bermudagrass hays

Item	Hay			SEM	P-value
	Alfalfa	Rhizoma peanut	Bermudagrass
Nitrogen intake, g/d	407a	297b	236c	3.9	<0.0001
Digestible nitrogen, g/d	339a	213b	151c	2.7	<0.0001
Urinary nitrogen excretion, g/d	211a	127b	94c	4.7	<0.0001
Fecal nitrogen excretion, g/d	67b	84a	85a	1.4	0.0001
Total nitrogen excretion, g/d	278a	211b	179c	5.0	<0.0001
Nitrogen retention, g/d	129a	86b	57c	6.2	0.0005
Nitrogen retention, % N intake	31.7a	29.2ab	24.2b	1.6	0.0377

^a–cWithin row, means without a common letter differ based on a Tukey’s test (P ≤ 0.05).

When horses were fed RP, absolute nitrogen retention (g/d) was lower (P = 0.0005) than ALF and greater than BG; however, as a percentage of nitrogen intake, relative nitrogen retention from the RP diet did not differ from either ALF or BG. Both absolute and relative nitrogen retentions were greater (P < 0.05) when horses were fed ALF than BG. The relative and absolute nitrogen retention values for RP and BG in the current study were similar to those reported in maintenance horses by Eckert et al. (2010). ALF hay fed to mature horses was previously reported to result in less absolute nitrogen (68 g N/d) than observed here (129 g N/d) but similar relative nitrogen retention (31% vs. 31.7% N intake) was found in the current study (Potts et al., 2010). The percent CP in the ALF hay fed was similar to the present study; therefore, the difference in absolute retention is related to protein fed in excess of the requirement (293% vs. 343% of the requirement). At maintenance, nitrogen intake beyond nitrogen requirements should increase nitrogen losses, maintaining zero nitrogen balance (Slade et al., 1970; Olsman et al., 2003). It is unlikely that there were depositions of 129, 86, and 57 g N/d in ALF-, RP-, and BG-fed maintenance horses used in the current study. Therefore, it is possible that the retained nitrogen was overestimated in the current study due to underestimation of endogenous nitrogen losses, such as sweat, hair, skin cells, and ammonia in expelled gases (Slade et al., 1970; Trottier and Tedeschi, 2019; Mok and Urschel, 2020). The integument (hair and skin) losses are independent of the amount of nitrogen or DM consumed and are estimated to be between 15 and 166 mg of nitrogen/kg BW (Coenen, 2004; Trottier and Tedeschi, 2019). In the current study, nitrogen lost via skin and hair could have ranged from 9 to 98 g of nitrogen per day; however, calculating the exact integument nitrogen losses is unfeasible thus far.

Conclusions

The results of this study indicate that RP is a high-quality legume hay for horses providing nutrient intake and digestibility intermediate between ALF and BG. The results partially supported our hypothesis that RP would provide intake and digestibility comparable to ALF and greater than BG and showed that most nutrients are less digestible in RP than in ALF. Nonetheless, the nutrients provided by RP hay meet the nutritional needs of a horse at maintenance, while resulting in less nitrogen excretion compared with ALF. Compared with RP, the excess CP intake provided by ALF causes more nitrogen to be excreted into the environment. Facing current environmental concerns and the encouragement of implementing best management practices, RP should be prioritized in lieu of ALF hay. Lastly, although BG resulted in the least nitrogen excretion, it requires high rates of nitrogen fertilizer for hay production (~100 to 200 kg N/ha), while ALF and RP require no nitrogen application. To compare environmental impacts, future studies need to investigate the whole cycle of nitrogen beginning with fertilization of crops and ending with excretion from animals.

Glossary

Abbreviations

ADF

acid detergent fiber

ALF

alfalfa

BG

bermudagrass

BW

body weight

CP

crude protein

DE

digestible energy

DM

dry matter

DMD

dry matter digestibility

DMI

dry matter intake

DN

digestible nitrogen

ESC

ethanol-soluble carbohydrates

EE

ether extract

NDF

neutral detergent fiber

NSC

nonstructural carbohydrates

OM

organic matter

RP

rhizoma peanut

WSC

water-soluble carbohydrates

Conflict of interest statement

We wish to confirm that there are no known conflicts of interest associated with this publication.