Effects of feeding juniper as a roughage on feedlot performance, carcass measurements, meat sensory attributes, and volatile aroma compounds of yearling Rambouillet wethers,

Abstract

The majority of U.S. lambs are born during late winter or early spring, which can create downstream variability in carcass quality if commercial lamb harvest is to be relatively constant throughout the year. Flavor is an important quality determining characteristic of sheep meat and is influenced, in part, by animal age at harvest. However, management practices to mitigate the risk of objectionable flavors in meat from old crop lambs or yearlings are not well known. Yearling (16.8 ± 0.14 mo) Rambouillet wethers were assigned to 1 of 3 treatment groups, which consisted of feeding a 20% ground sorghum-sudangrass hay diet for 40 d (JUN0; n = 10), a 20% ground juniper diet for 40 d (JUN40; n = 10), or a 20% ground hay diet for 20 d followed by a 20% ground juniper diet for 20 d (JUN20; n = 10). Wethers were harvested on day 41 and a whole bone-in loin and a boneless inside leg roast were fabricated from one side each of carcass. After grilling (loin chop) or convection air roasting (leg roast), trained sensory panel evaluation and measurement of aroma volatiles by gas chromatography/mass spectrometry were performed. Treatment diet did not affect (P ≥ 0.17) wether feedlot performance, dressing percentage, or loin eye area. However, wethers fed JUN0 tended (P = 0.06) to have greater back fat depth than wethers fed JUN20 or JUN40. No trained sensory panel trait of loin chop samples was affected (P > 0.10) by treatment. Leg roasts from JUN0 and JUN20 wethers had greater (P = 0.01) lamb identification sensory score than JUN40. Benzaldehyde, 1-heptanol, and 1-octanol concentrations were greater (P < 0.05) and decanal and nonenal concentrations were less (P < 0.05) in loin chops from JUN0 compared with JUN40 wethers. Additionally, the terpenes cedr-8-ene, gamma muurolene, and widdrene tended to be greater (P < 0.07) in loin chops from JUN20 and JUN40 than JUN0 wethers. The 2-pentyl-furan concentrations were greatest (P = 0.03) in leg roasts from JUN40 wethers. Like the loin chops, cedr-8-ene, gamma-muurolene, toluene, and widdrene were greater (P < 0.05) in leg roasts from wethers fed either of the juniper diets compared with JUN0. Yearling wethers can be finished on a feedlot diet containing 20% juniper for up to 40 d prior to harvest with no impact on feedlot performance, carcass characteristics, nor negative impact on sensory attributes or volatile compounds of either grilled loin chops or roasted legs.

INTRODUCTION

The estimated per capita retail disappearance of lamb and mutton in the United States declined 47% from 1977 (0.68 kg) to 2017 (0.36 kg; USDA ERS, 2018a). Still, recent reports in the United States and abroad have indicated that flavor is an important factor in lamb consumer acceptance (Pleasants et al., 2005; Pethick et al., 2006; Hoffman et al., 2016). Since sheep are most reproductively active in the fall, the vast majority (>85%) of U.S. lambs are born from January to May (USDA APHIS, 2014). However, commercial lamb harvest remains fairly constant throughout the year (USDA ERS, 2018b). This can contribute to downstream variability in animal age at harvest and carcass quality; principally older lambs or yearlings finished beyond optimal end-point harvest weights. It is well known that harvest age and finishing diet can affect sheep meat flavor (Duckett and Kuber, 2001; Elmore et al., 2000; Luciano et al., 2013). However, alternative feeding to mitigate potential off-flavors in meat from old crop lambs or yearlings have not been extensively investigated.

There is evidence that certain plant secondary compounds (PSC) can enhance flavor attributes of sheep meat (Vasta and Luciano, 2011). Whitney and Smith (2015) reported that substituting ground juniper for oat hay in lamb diets enhanced sensory panel traits and partially attributed the results to meat fatty acid composition and PSC in juniper. Kerth et al. (2018) found that out of the 81 volatile aroma compounds identified in grilled loin chops, only 7 were affected by feeding 4 different species of juniper to lambs. Certainly, more information is needed to fully describe flavor components, especially of yearling mutton from animals fed alternative forages. Therefore, the objective of this study was to determine whether finishing yearling wethers on diets containing juniper would affect feedlot performance and carcass characteristics and enhance meat sensory attributes.

MATERIALS AND METHODS

Animals

The live animal experimental protocol was approved by the Montana State University (MSU) Agricultural Animal Care and Use Committee (2016-AA17). Thirty yearling (16.8 ± 0.14 mo) Rambouillet wethers located at MSU’s Fort Ellis Research Farm (Bozeman, MT) were randomly selected. Wethers were sheared, vaccinated against clostridial disease (Covexin 8; Merck Animal Health, Intervet Inc., Madison, NJ), orally drenched with an anthelmintic (Valbazen; Zoetis Inc., Kalamazoo, MI), and fitted with an electronic identification ear tag. Animals were managed in drylot (14.5 × 54 m) and adapted to GrowSafe bunks (GrowSafe Systems Ltd., Airdrie, AB, Canada) for 2 wk with ad libitum access to a ground, 20% sorghum-sudangrass hay diet. Following the adaptation period, a 12-h fasted BW was recorded for 2 consecutive d for each wether.

Treatment groups were allocated to separate pens each containing 2 GrowSafe bunks. GrowSafe feeders, or other technology that measures feed intake, can increase labor efficiency while improving statistical power since experimental units (i.e., animals) can be managed in group pens rather than individually. Average starting BW and sire were balanced between treatment groups consisting of wethers fed a 20% ground hay diet for 40 d (JUN0; n = 10), 20% ground hay diet for 20 d followed by 20% ground juniper diet for 20 d (JUN20; n = 10), or 20% ground juniper diet for 40 d (JUN40; n = 10). Nonfasted BW was recorded on days 19 and 20, after which wethers in JUN20 were switched from the 20% ground hay diet to the 20% ground juniper diet. Fasted BW was recorded at the end of the trial on days 39 and 40. Average daily gain and DMI were calculated from days 1 to 20 and 21 to 40.

Feed Collection and Processing

The entire aboveground biomass from Juniperus ashei and Juniperus pinchotii trees were cut and allowed to air-dry in the pasture for approximately 2 mo. The majority of the trees were between 2 and 4.5 m in height. Trees were ground (Rotochopper, Inc., Model MC266, St. Martin, MN) and hammermilled (Bliss 4440, Ponca City, Oklahoma) to pass a 4.76-mm sieve. Each treatment diet was sampled once per day and combined into one sample per diet for later proximate feed analysis (Midwest Laboratories Inc., Omaha, NE; Table 1).

Table 1.

Ingredient and nutrient composition (DM basis, %) of treatment diets

Item1	Hay-based diet	Juniper-based diet
Ground hay	20.0	0.0
Ground juniper	0.0	20.0
DDGS	40.0	40.0
Sorghum grain, rolled	31.5	31.5
Molasses, cane	4.0	4.0
Limestone	2.0	2.0
Ammonium chloride	1.0	1.0
Mineral premix	1.5	1.5
Nutrient composition
CP, %	28.6	26.6
Crude fat, %	3.4	4.4
ADF, %	12.9	18.0
Ash, %	9.2	8.2
S, %	0.72	0.73
Ca, %	1.22	1.19
P, %	0.77	0.73
K, %	1.54	1.30
Mg, %	0.34	0.32
Na, %	0.43	0.40
Fe, ppm	369	311
Mg, ppm	57.2	48.9
Cu, ppm	8.5	8.3
Zn, ppm	189	158

Yearling Rambouillet wethers were assigned to a treatment group consisting of feeding a 20% hay-based diet for 40 d (JUN0), 20% hay-based diet for 20 d followed by a 20% juniper-based diet for 20 d (JUN20), or a 20% juniper-based diet for 40 d (JUN40).

¹Ground hay = sorghum-sudangrass; Ground juniper = entire above-ground biomass was hammermilled to pass through a 4.76-mm screen; DDGS = corn-dried distillers grains with solubles were a byproduct of corn ethanol production; Mineral premix = NaCl, KCl, S, MnO, ZnO, vitamins A, D, and E, CaCO3, cottonseed meal, cane molasses, and animal fat.

Chemical Analyses

Sub-samples were ground through a 2-mm screen (Wiley mill, Arthur H. Thomas Co., Philadelphia, PA) and dried at 55 °C in a forced-air oven for 48 h, ground through a 1-mm screen, and analyzed for nutrient composition. Nitrogen was analyzed by a standard method (Method 990.03; AOAC Int., 2006); CP calculated as 6.25 × N. Feed ADF was analyzed according to procedures of Van Soest et al. (1991), which were modified for an Ankom 2000 Fiber Analyzer (Ankom Technol. Corp., Fairport, NY). A standard method was used to evaluate crude fat (Method 945.16; AOAC, 2006) and ash (Method 942.05; AOAC, 2006). A modified method (Method 985.01, AOAC, 2006) was used to evaluate individual minerals concentrations; samples were digested with a Microwave Accelerated Reaction System (MARS6; CEM, Matthews, NC) and then analyzed by a Thermo Jarrell Ash IRIS Advantage HX Inductively Coupled Plasma Radial Spectrometer (Thermo Instrument Systems, Inc., Waltham, MA).

Condensed tannins (CT) in the ground juniper were assayed for extractable, protein-bound, and fiber-bound fractions by methods described by Terrill et al. (1992). Samples were oven dried and standards prepared as recommended by Wolfe et al. (2008), using CT extracts purified on a Sephadex LH-20 (GE Healthcare Bio-Sciences Corp, Piscataway, NJ) and lyophilized to recover purified CT. Subsamples were also analyzed for protein precipitable phenolics (PPP; a measure of readily bioactive CT), amount of protein bound by CT per kg of juniper material (PB), and protein binding capacity of CT (PB/PPP; CT potency) according to procedures reported by Naumann et al. (2014). Additional juniper subsamples, not mechanically dried, were steam distilled to determine total volatile oil yield as adapted by Koedam and Looman (1980) and Adams (1991).

Carcass Evaluation

Wethers were humanely harvested on day 41 (Pioneer Meats, Big Timber, MT). Final BW and hot carcass weight were used to calculate dressing percentage (DP). Whole carcasses were chilled for 24 h, transported to MSU, and processed 5 d postharvest. Back fat depth (BF) and loin eye area (LEA) between the 12th and 13th ribs were measured on each carcass. Additionally, a whole bone-in loin and a boneless inside leg roast were fabricated from one side of each carcass. The Longissimus lumborum and the semimembranosus muscle were removed from the loin and leg, respectively, vacuum-packaged individually as a whole muscle, and frozen. These samples were shipped to Texas A&M University and held at −10 °C until flavor compound and sensory analysis. Samples from the frozen loin were cut into 2.54cm-thick chops and repackaged individually. Each sample received a random 3-digit code for sensory analysis.

Sensory Panel Evaluation

This research was approved by Texas A&M University IRB (IRB2017-0618M). For each analysis, individual loin chops and leg roasts were selected and thawed in refrigerated (4 °C) storage for 12 to 24 h. Loin chops were cooked on a 2.54-cm-thick flat top Star Max 536TGF 91 cm Countertop Electric Griddle with Snap Action Thermostatic Controls (Star International Holdings Inc. Company, St. Louis, MO) set to 232 °C. Loin chops were placed on the grill, turned when the internal temperature reached 37 °C, and removed when the internal temperature reached 71 °C (medium degree of doneness). Internal temperatures were monitored by iron–constantan thermocouples (Omega Engineering, Stanford, CT) inserted into the geometric center of the loin chop. Temperatures were displayed using an Omega HH501BT Type T thermometer (Omega Engineering, Stanford, CT). Leg roasts were cooked in a Hobart (model HGC502, Hobart Corp., Troy, OH) convection oven set at 177 °C. Internal temperatures of leg roasts were monitored by iron–constantan thermocouples as described for loin chops. Leg roasts were removed from the oven when an internal temperature of 71 °C was reached, wrapped in foil, and held (65 °C) for no longer than 20 min.

After cooking, grilled loin chop samples were cut into 1.3-cm × 1.3-cm × chop thickness cubes. For leg roast samples, a 2.5-cm slice was removed from across the center of the roast and then cut into 1.3-cm cubes. Two cubes per sample were served in clear, plastic soufflé cups tested to assure that they did not impart flavors on the samples. Samples were identified with random 3-digit codes, arranged in random serving order, and cut and served immediately to assure they were approximately 37 °C upon time of serving. During evaluation, panelists were seated around a benchtop where they evaluated each sample individually. After panelists evaluated a single sample, a consensus score was agreed upon, which was used in subsequent statistical analyses. To prevent taste fatigue, each evaluation day was divided into 2 sessions (6 samples per session), with a 10-min break between sessions and samples were served 4 min apart as described by Wall et al. (2019).

Loin chop and leg roast samples were evaluated by a 5-member, expertly trained flavor descriptive attribute panel with over 200 h of training and 10 yr of experience. The panel was trained on 38 basic flavors and 3 texture attributes adapted from the beef lexicon (0 = none and 15 = extremely intense; Adhikari et al., 2011) with the addition of attributes for lamb identity, mutton, and juniper flavors as shown in Table 2. After training, panelists were presented 12 samples per day, divided into 2 sessions. Prior to the start of each trained panel evaluation day, panelists were calibrated using 1 orientation or “warm up” sample that was evaluated and discussed orally. After evaluation of the orientation sample, panelists were served the first sample of the session and asked to individually rate the sample for each beef flavor lexicon attribute along with the lamb identity, mutton, and juniper flavors. Double distilled water, unsalted saltine crackers, and fat-free ricotta cheese were available for cleansing the palette between samples.

Table 2.

Trained panel flavor attributes, definitions, and reference standards with their intensities where 0 = none and 15 = extremely intense adapted from Adhikari et al. (2011)

Attribute	Definition	Reference
Bitter	The fundamental taste factor associated with a caffeine solution	0.01% caffeine solution = 2.0 0.02% caffeine solution = 3.5
Bloody	The aromatics associated with blood on cooked meat products. Closely related to metallic aromatics	USDA Choice strip steak = 5.5 Beef brisket = 6.0 Boneless pork chop, 57 °C = 2.0
Brown	A round, full aromativ generally associated with beef/lamb suet that has been broiled	Beef suet = 8.0 80% lean ground beef = 10.0 Pork Fat, cooked and browned = 3.0
Burnt	The sharp/acrid flavor note associated with over- roasted beef muscle, something over-baked, or excessively browned in oil	Alf’s red wheat puffs = 5.0
Buttery	Sweet, dairy-like aromatic associated with natural butter	Land O’Lakes unsalted butter = 7.0
Cardboardy	Aromatic associated with slightly oxidized fats and oils, reminiscent of wet cardboard packaging	Dry cardboard = 5.0 Wet cardboard = 7.0
Chemical	The aromatics associated with garden hose, hot Teflon pan, plastic packaging, and petroleum- based product such as charcoal liter fluid	Clorox in water = 6.5
Cooked milk	A combination of sweet, brown flavor notes, and aromatics associated with heated milk	Mini Babybel original Swiss cheese = 2.5 Dillon’s whole milk, cooked = 4.5
Dairy	The aromatics associated with products made from cow’s milk, such as cream, milk, sour cream, or butter milk	Dillon’s reduced fat milk (2%) = 8.0
Fat-like	The aromatics associated with cooked animal fat	Hillshire farms Lit’l beef smokies = 7.0 Beef suet = 12.0
Fishy	Odor associated with aged fish	Canned StarKist Tuna = 12.0
Floral	Sweet, light, slightly perfume impression associated with flowers	Welch’s white grape juice = 5.0
Green	Sharp, slightly pungent aromatics associated with green/plant/vegetable matters such as parsley, spinach, pea pod, fresh cut grass, etc.	Fresh parsley water = 9.0
Green/Hay-like	Brown/green dusty aromatics associated with dry grasses, hay, dry parsley, and tea leaves	Dry parsley in ~30-mL cup = 6.0
Heated oil	The aromatics associated with oil heated to a high temperature	Lay’s potato chips = 4.0 Microwaved Wesson’s vegetable oil (high 3 min) = 7.0
Juniper	Tart and sharp, with a resinous, piney flavor and hints of citrus	Five dry juniper berries = 12.0
Lamb identity	Amount of lamb flavor identity in the sample	Grain-fed ground lamb (80% lean) = 9.0 Grass-fed ground lamb (80% lean) = 12.0
Leather	Musty, old leather (like old book bindings)	2,3,4-Trimethoxybenzaldehyde = 3.0 (aroma)
Liver-like	The aromatics associated with cooked organ meat/ liver	Beef liver = 7.5 Oscar Mayer Braunschweiger liver sausage = 10.0 Pork liver, 71 °C = 15.0 Chicken liver, 71 °C = 9.0
Metallic	The impression of slightly oxidized metal, such as iron, copper, and silver spoons	0.10% potassium chloride solution = 1.5 USDA choice strip steak = 4.0 Dole canned pineapple juice = 6.0
Muttony	Flavor associated with older sheep	Mohair from 3 yr old in-tact male = 6.0
Overall sweet	A combination of sweet taste and sweet aromatics. The aromatics associated with the impression of sweet	Post-shredded wheat spoon size = 1.5 Hillshire farms Lit’l beef smokies = 3.0 SAFC ethyl maltol 99% = 4.5
Painty	Aromatic associated with oxidized oil, similar to the aromatic of linseed oil and oil-based paint	Wesson oil 14 d at 100 °C 8.0
Petroleum	A specific chemical aromatic associated with crude oil and its refined products that have heavy oil characteristics	Vaseline petroleum jelly = 3.0 (smelled)
Rancid	The aromatics commonly associated with oxidized fat and oils. These aromatics may include cardboard, painty, varnish, and fishy	Microwaved Wesson vegetable oil (3 min at high) = 7.0 Microwaved Wesson vegetable oil (5 min at high) = 9.0
Refrigerator stale	Off-flavor associated with a product that has absorbed odors from the refrigerator	Reheated ground lamb that has been exposed to the elements of the refrigerator over night = 4.5
Roasted	A round, full aromativ generally associated with beef/lamb that has been broiled/roasted
Salty	The fundamental taste factor of which sodium chloride is typical	0.15% sodium chloride solution = 1.5 0.25% sodium chloride solution = 3.5
Smoky charcoal	An aromatic associated with meat juices and fat dripping on hot coals, which can be acrid, sour, burned, etc.	Wright’s Natural Hickory seasoning in water = 9.0 (smelled)
Smoky wood	Dry, dusty aromatic reminiscent of burning wood	Wright’s Natural Hickory seasoning in water = 7.5 (smelled)
Sour	The fundamental taste factor associated with citric acid	0.015% citric acid solution = 1.5 0.050% citric acid solution = 3.5
Sour aromatics	The aromatics associated with sour substances	Dillon’s buttermilk = 5.0
Sour dairy	Sour, fermented aromatics associated with dairy products such as buttermilk and sour cream	Laughing Cow Light Swiss Cheese = 7.0 Dillon’s Buttermilk = 9.0
Spoiled putrid	The presence of inappropriate aromatics and flavors that is commonly associated with the products. It is a foul taste and/or smell that indicates the product is starting to decay and putrefy	Dimethyl disulfide (10,000 ppm) = 12.0
Sweet	The fundamental taste factor associated with sucrose	2.0% sucrose solution = 2.0
Umami	Flat, salty, somewhat brothy. The taste of flutamate, salts of amino acids, and other molecules called nucleotides	0.035% accent flavor enhancer solution = 7.5
Warmed over	Perception of a product that has been previously cooked and reheated	Reheated ground lamb = 6.0
Myofibrillar tenderness	The ease in which the muscle fiber fragments during mastication	Select eye of round steak cooked to 70 °C = 9.0 Select tenderloin steak cooked to 70 °C = 14.0
Connective tissue amount	The component of the muscle surrounding the during mastication	Cross cut beef shank cooked to muscle fiber that will not break down 70 °C = 7.0 Select tenderloin cooked to 70 °C = 14.0
Overall tenderness	Average of muscle fiber tenderness and connective tissue amount when connective tissue amount is 6 or less	If connective tissue amount is 12 to 15, then overall tenderness = the value of muscle fiber tenderness; if connective tissue amount is less than 12, then overall tenderness is the average of connective tissue amount and muscle fiber tenderness

Gas Chromatography/Mass Spectrometry

After loin chops and leg roasts were cooked, all external fat was removed and each chop was cut into pieces as was done for sensory panel evaluation (1.3-cm × 1.3-cm × chop thickness cubes). Two to 3 pieces were placed in a 473-mL glass jar with a Teflon lid and submerged in a water bath held at 60 °C to approximate normal holding temperature for sensory analyses. After equilibrating for 20 min, a solid-phase micro-extraction (SPME) Portable Field Sampler (Supelco 504831, 75-μm Carboxen/polydimethylsiloxane [PDMS], Sigma-Aldrich, St. Louis, MO) was inserted through the lid and the headspace above each meat sample in the glass jar was collected for 2 h. Upon completion of collection, the SPME was removed from the jar and manually injected into the injection port of a gas chromatograph (GC; Agilent Technologies 7920 series GC, Santa Clara, CA) where the sample was desorbed at 280 °C for 3 min. The sample was loaded onto a multidimensional gas chromatograph into the first column [30 m × 0.53 mm ID/BPX5 (5% phenyl polysilphenylene-siloxane) × 0.5 μm, SGE Analytical Sciences, Austin, TX] and then the second column [30 m × 0.53 mm ID (BP20—polyethylene glycol) × 0.50 μm, SGE Analytical Sciences] using helium as the carrier gas. The GC temperature started at 40 °C and increased at a rate of 7 °C/min until reaching 260 °C. The GC column then went to a 3-way valve split to 2 olfactory ports (for detecting an aroma event during which the volatiles were identified and kept for analysis) and a third to a mass spectrometer (MS; Agilent Technologies 5975 series MS, Santa Clara, CA) for relative quantification and identification, using the NIST Chemical Library and to published linear retention indices as appropriate for the given column.

Statistical Analyses

GrowSafe feeders enabled DMI to be evaluated for each wether; thus, wether was considered the experimental unit for all analyses. Body weight, ADG, and DMI were analyzed as repeated measures using the MIXED procedure of SAS (v. 9.4; SAS Institute Inc., Cary, NC) with fixed effects of treatment (JUN0, JUN20, or JUN40), trial day [days 0, 20, or 40 (BW); days 1 to 20 or 21 to 40 (ADG and DMI)], and their 2-way interaction. Akaike Information Criterion (AIC) was used to determine that an unstructured covariance matrix best fit the data. Dressing percentage, LEA, and BF were analyzed using the GLM procedure of SAS with the fixed effect of treatment.

Data for sensory and gas chromatography variables were analyzed with JMP version 13.0 (SAS Institute Inc., Cary, NC) using ANOVA for a completely randomized design with treatment as a fixed effect (loin chops and leg roasts were analyzed separately). Volatile compounds that were below the level of detection or were absent in a sample were analyzed as zeroes in the statistical analysis and reported as not detectable (nd). Sensory order of serving to the panelists was included in the model as a random effect in the trained sensory panel analysis. Least squares means were generated and separated using Student’s t-test when a significant (P < 0.05) F-test was indicated.

RESULTS

Feedlot Performance and Carcass Characteristics

No treatment × trial day/period interaction effect was observed (P ≥ 0.48) for BW or ADG (Table 3). As expected, BW increased throughout the trial (P < 0.001) across all treatment groups, though change in BW was numerically greater during the first 20 d than last 20 d. This resulted in ADG (across all treatment groups) being greater during the first 20 d than last 20 d (P < 0.001). However, over the entire 40-d trial, no differences (P ≥ 0.17) in BW or ADG were observed among treatment groups. The treatment × trial period interaction (P = 0.04) for DMI was due to wethers fed JUN40 increasing DMI from period 1 (days 0 to 20) to period 2 (days 20 to 40), whereas DMI of wethers fed JUN0 or JUN20 remained similar in both periods. Still, least-squares means of DMI between treatment groups were not different (P ≥ 0.42) within either period. There were no differences (P ≥ 0.28) in DP or LEA among treatment groups (Table 3). Back fat depth tended (P = 0.06) to be greater in wethers fed JUN0 than wethers fed JUN20 or JUN40.

Table 3.

Effects of treatment on yearling Rambouillet wether feedlot performance and carcass measurements

	Treatment				P 3
Item1	JUN0	JUN20	JUN40	SEM2	T	D	T × D
BW, kg					0.87	<0.001	0.48
Day 0	50.8	51.0	51.5	1.15
Day 20	60.5	59.2	59.6	1.32
Day 40	62.2	60.4	60.9	1.51
ADG, g/d					0.17	<0.001	0.60
Days 0 to 20	485.8	406.3	404.1	38.1
Days 20 to 40	83.8	63.7	65.9	25.7
DMI, kg					0.56	0.09	0.04
Days 0 to 20	2.27	2.25	2.23	0.10
Days 20 to 40	2.28	2.22	2.52	0.11
DP, %	53.5	52.7	52.5	0.5	0.42
LEA, cm2	14.7	14.3	15.1	0.3	0.28
BF, cm	0.48	0.34	0.36	0.04	0.06

Yearling Rambouillet wethers were assigned to a treatment group, which consisted of feeding a 20% ground sorghum-sudangrass hay-based diet for 40 d (JUN0; n = 10), a 20% hay-based diet for 20 d followed by a 20% ground juniper-based diet for 20 d (JUN20; n = 10), or a 20% juniper-based diet for 40 d (JUN40; n = 10).

¹DP = dressing percentage; LEA = loin eye area between the 12th and 13th ribs; BF = back fat depth at 12th rib.

²SEM = greatest standard error of treatment group means.

³T = treatment effect; D = trial day/period effect.

Sensory Panel Evaluation

Although fat-like, overall sweet, sweet, salty, sour aroma, and juniper sensory traits in the loin chops showed trends (P = 0.05 to 0.10; Table 4) for treatment effects, numerical differences between treatments were extremely low (<0.2 unit difference in sensory trait) and therefore showed no meaningful differences in those traits because of finishing diet. No other sensory trait in the loin was affected (P > 0.10) by treatment. Leg roasts from wethers finished on JUN0 or JUJN20 scored greater (P = 0.01) for lamb identification sensory score (lamb ID) than leg roasts from wethers finished on JUN40. The JUN20 leg roasts tended (P = 0.06) to score greater for bloody compared with the other 2 diets and leg roasts from wethers finished on JUN20 or JUN40 tended to have greater (P = 0.07) juniper flavor scores, but only by a score of 0.1, compared with wethers finished on JUN0. Other than a 0.2 to 0.4 difference for lamb ID, no differences in flavor descriptors were found indicating that feeding juniper for 20 or 40 d before harvest did not significantly affect leg roast flavor.

Table 4.

Effects of treatment on trained sensory panel descriptive analysis traits of the grilled loin chop and roasted inside leg roast¹

	Loin chop, treatment					Leg roast, treatment
Item2	JUN0	JUN20	JUN40	SEM3	P > F	JUN0	JUN20	JUN40	SEM	P > F
Lamb identity	8.9	8.7	8.8	0.13	0.59	10.0a	9.8a	9.6b	0.18	0.01
Brown	4.1	4.3	4.1	0.15	0.35	11.3	11.4	11.5	0.22	0.65
Roasted	8.0	7.9	8.1	0.24	0.70	10.0	10.0	10.0	0.20	0.95
Bloody	2.6	2.3	2.4	0.15	0.61	1.5	1.8	1.4	0.16	0.06
Fat-like	2.0	2.0	2.0	0.07	0.84	2.2	2.3	2.1	0.07	0.21
Metallic	2.7	2.4	2.5	0.12	0.25	2.1	2.1	2.1	0.04	0.99
Liver-like	0.7	0.6	0.8	0.11	0.59	0.2	0.3	0.4	0.10	0.42
Umami	4.0	3.7	3.7	0.12	0.07	5.1	5.1	5.0	0.11	0.86
Overall sweet	1.3	1.3	1.3	0.05	0.57	1.8	1.8	1.9	0.04	0.49
Sweet	1.5	1.4	1.5	0.06	0.34	1.9	2.0	2.0	0.03	0.22
Sour	2.6	2.6	2.5	0.11	0.61	2.1	2.1	2.1	0.04	0.88
Salty	2.1	2.0	2.0	0.05	0.07	2.1	2.1	2.1	0.03	0.54
Bitter	2.4	2.3	2.5	0.06	0.16	2.1	2.1	2.1	0.05	0.47
Sour aroma	0.2	0.2	0.3	0.06	0.30	0.1	0.1	0.1	0.04	0.97
Muttony	1.8	2.1	1.9	0.10	0.11	1.2	1.2	1.1	0.14	0.32
Cardboard	0	0	0	0	–	1.4	1.3	1.6	0.12	0.32
Juniper	0.1	0.1	0.0	0.05	0.75	0.0	0.1	0.1	0.03	0.07
Juiciness	9.9	9.7	9.6	0.37	0.79	9.3	8.8	9.3	0.44	0.20
Myofibrillar tenderness	10.7	10.6	10.6	0.22	0.96	11.4	10.8	11.4	0.44	0.18
Connective tissue amount	10.9	10.8	10.5	0.22	0.79	12.1	11.4	11.8	0.40	0.22
Overall tenderness	10.7	10.6	10.4	0.22	0.92	11.3	10.9	11.4	0.45	0.35

¹Yearling Rambouillet wethers were assigned to a treatment group, which consisted of feeding a 20% ground sorghum-sudangrass hay-based diet for 40 d (JUN0; n = 10), a 20% hay-based diet for 20 d followed by a 20% ground juniper-based diet for 20 d (JUN20; n = 10), or a 20% juniper-based diet for 40 d (JUN40; n = 10).

²All attributes were scored on a 16-point scale where 0 = none and 15 = extremely intense according to descriptors outlined in Table 2.

³SEM = greatest standard error of the means.

Gas Chromatography/Mass Spectrometry

Total ion count (TIC) of volatile aroma compounds for loin chops are reported in Table 5. No volatile acid compounds were found to be affected by treatment (P = 0.38). The TIC of 2-decenal (a fatty, earthy aroma according to Burdock, 2010) tended (P = 0.07) to increase as the duration juniper was fed increased. The decanal (orange, citrus; Kerth and Miller, 2015) TIC for JUN40 chops tended (P = 0.06) to be greater than for loin chops from wethers fed the other 2 diets. Nonenal (cucumber, melon aroma) TIC values were greatest (P = 0.004) for JUN40 loin chops compared with the other 2 treatments. Undecanal, having a soapy, metallic aroma, and TIC was greater (P < 0.05) for JUN40 loin chops compared with JUN20.

Table 5.

Least squares means and SEM for volatile aroma compounds reported as total ion count (TIC) area under the peak for diet effects on grilled loin chops of yearling wethers¹

		Treatment
Item2	LRI3	JUN0	JUN20	JUN40	SEM4	P > F
Acid
Acetic acid	8675	8097	nd	nd	4675	0.38
Aldehyde
Acetaldehyde	3755	nd	nd	1528	1021	0.58
2-methyl-propanal	5695	5357	nd	nd	3092	0.38
3-methyl-butanal	6336	16646	40976	16763	16582	0.50
2-methyl-butanal	6416	10689	2849	13886	7085	0.53
Pentanal	6836	37675	104723	45800	25574	0.15
Hexanal	7896	594795	1328113	510275	351763	0.21
N-heptanal	8916	249971	610514	436378	179965	0.38
2-heptenal	9546	nd	1627	7067	3007	0.24
Octanal	9875	243937	501509	388221	101362	0.22
Benzaldehyde	10097	699978	719104	680622	109722	0.97
Nonanal	10865	536987	817834	892087	154036	0.25
Phenyl acetaldehyde	10965	nd	nd	3117	1806	0.41
Nonenal	11567	4871b	ndb	69152a	14760	0.004
Decanal	11877	30029	32549	65652	11217	0.06
3-ethyl-benzaldehyde	11925	nd	nd	3853	1486	0.13
2-decenal	1263ac	18924	27657	74661	17693	0.07
Undecanal	13076	3350gh	ndb	14565a	4134	0.05
2,4-decadienal	1336ac	nd	nd	3995	1373	0.12
2-undecenal	1363ac	18627	12749	32210	10132	0.69
3-dodecen-1-al	13665	nd	nd	7532	3559	0.25
Tetradecanal	13985	3516	nd	5264	2674	0.44
Hexadecanal	14335	nd	nd	nd	–	–
Alcohol
1-heptanol	9798	ndb	2416b	19156a	4126	0.005
1-octen-3-ol	991bc	1910	nd	8075	4791	0.47
1-Octanol	1076bc	3326b	12527gh	32098a	7937	0.04
2-(hexyloxy)-ethanol	11165	26990	13161	87776	23908	0.08
2-cyclohexen-1-ol	12525	nd	nd	nd	415	0.38
Alkane
Heptane	6437	14301	32455	7724	11757	0.32
4-methyl heptane	7225	9933	15282	nd	5721	0.18
Octane	7597	58027	95611	25545	39283	0.46
4-methyl-octane	8235	6291	8837	nd	5113	0.46
3-ethyl hexane	8305	2639	nd	nd	1523	0.38
Benzene
Methyl-benzene	7545	14118	27844	14098	11476	0.63
Furan
2-ethyl-furan	6705	nd	nd	nd	–	–
2-pentyl-furan	966bc	nd	8622	16057	5560	0.14
Ketone
2-butanone	5675	22171	48051	16775	14250	0.27
2,3-butanedione	59910	8957	35137	5807	17395	0.44
2-pentanone	6725	nd	nd	nd	531	0.38
3-hydroxy-butanone	7945	43151	93695	2941	39880	0.29
4-heptanone	8545	nd	nd	nd	–	–
2-heptanone	8739	2406	4170	nd	2141	0.40
2,3-octanedione	968de	nd	nd	6018	2323	0.13
2-decanone	11728	nd	nd	5434	2121	0.13
Pyrazine
Methyl-pyrazine	8508	7004	4510	4548	3927	0.88
2,5-dimethyl-pyrazine	9248	53287	47880	41444	22502	0.93
Trimethyl-pyrazine	10028	11199	24165	33643	15808	0.63
2-ethyl-3,5-dimethyl-pyrazine	10698	14928	25867	24364	2723	0.67
2-ethyl-6-methyl-pyrazine	16195	nd	4332	1525	2307	0.42
S-containing
Methanethiol	3775	1780	2969	1359	893	0.43
Sulfur dioxide	4465	3692	10795	nd	5966	0.44
Carbon disulfide	4895	7148	11146	17220	12541	0.85
Terpene
1-octene	7495	nd	2834	nd	1636	0.38
Toluene	7527	8484	34139	18488	16954	0.57
Styrene	8935	nd	nd	16314	8551	0.31
Cedr-8-ene	14215	ndb	32406a	8012b	7549	0.014
Gamma-muurolene	14335	ndb	ndb	9626a	2360	0.012
Widdrene	14415	nd	44334	30474	13485	0.08
Thujopsene	14445	nd	39029	11510	16972	0.26

¹Yearling Rambouillet wethers were assigned to a treatment group, which consisted of feeding a 20% ground sorghum-sudangrass hay-based diet for 40 d (JUN0; n = 10); a 20% hay-based diet for 20 d followed by a 20% ground juniper-based diet for 20 d (JUN20; n = 10); or a 20% juniper-based diet for 40 d (JUN40; n = 10). Means within a row without a common superscript differ (P < 0.05).

²Volatile compounds identified through mass spectrometry and reported as total ion count area under the curve (nd indicates that a volatile was not detected or was below the threshold of detection or less than 1000 TIC area under the curve).

³Linear retention index on a multidementional GC using a BPX5 30m column followed by a BP20 30m column.

⁴SEM = greatest standard error of the means.

⁵Mass spectrum NIST library identification only.

⁶ Elmore et al. (1997).

⁷ Madruga et al. (2000).

⁸ Elmore et al. (1999).

⁹ Byrne et al. (2002).

¹⁰ Shellie et al. (2002).

^abcLeast squares means in a row with a different superscript differ (P < 0.05).

Loin chops from JUN40 wethers were greater (P < 0.05) in 1-heptanol (oily, fatty aroma; Burdock, 2010) and 1-octanol (orange-rose aroma) compared with the other 2 diets. Treatment did not affect (P > 0.10) any of the alkanes, benzenes, ketones, pyrazines, or S-containing compounds. Within the terpenes, loin chops from JUN20 wethers had greater (P < 0.02) TIC for cedr-8-ene (woody, cedar), whereas JUN40 alone had greater (P < 0.02) levels of gamma-muurolene (no descriptor). Widdrene (no descriptor) tended to be greater (P = 0.08) in treatments that contained juniper.

Treatment did not affect (P > 0.12) acids, aldehydes, alcohols, alkanes, benzene, ketone, or S-containing volatile organic compounds found in leg roasts (Table 6). The TIC for 2-pentyl-furan, with a green, to caramellic aroma (Kerth and Miller, 2015) was greatest (P < 0.05) in JUN40 compared with the other treatments. Because the cooking method of roasting does not allow for the Maillard reaction to occur, for the most part, no pyrazines were found in the leg roasts. Cedr-8-ene was greatest (P = 0.01) in JUN40 leg roasts followed by JUN20 and then JUN0 leg roasts. Gamma-muurolene TIC was greatest in the JUN40 leg roasts (P < 0.05) and toluene TIC tended (P = 0.05) to be greatest in JUN20 leg roasts.

Table 6.

Least squares means and SEM for volatile aroma compounds reported as total ion counts (TIC) area under the peak for diet effects on oven roasted leg roasts of yearling wethers¹

		Treatment
Trait2	LRI3	JUN0	JUN20	JUN40	SEM4	P > F
Acid
Acetic acid	8675	5738	9500	nd	4107	0.31
Aldehyde
Acetaldehyde	3755	nd	nd	nd	–	–
2-methyl-propanal	5695	3690	4274	8151	4562	0.76
3-methyl-butanal	6336	0	1353	1421	1133	0.61
2-methyl-butanal	6416	0	1353	1421	1133	0.61
Pentanal	6836	22989	60086	114729	30901	0.13
Hexanal	7896	398934	615132	2048428	582237	0.11
N-heptanal	8916	380762	684722	728329	246359	0.56
2-heptenal	9546	nd	nd	6088	3515	0.38
Octanal	9875	257337	352928	444445	98494	0.42
Benzaldehyde	10097	369706	446326	439864	79988	0.76
Nonanal	10865	475471	546396	747648	128981	0.31
Phenyl acetaldehyde	10965	nd	nd	nd	–	–
Nonenal	11567	12637	19103	22543	12062	0.84
Decanal	11877	9860	25625	19559	8843	0.46
3-ethyl-benzaldehyde	11925	nd	1130	nd	754	0.57
2-decenal	1263ac	5028	nd	14779	6388	0.27
Undecanal	13076	nd	nd	nd	–	–
2,4-decadienal	1336ac	nd	nd	nd	–	–
2-undecenal	1363ac	nd	nd	nd	–	–
3-dodecen-1-al	13665	4216	nd	15099	5628	0.17
Tetradecanal	13985	20547	nd	nd	11862	0.38
Hexadecanal	14335	24062	8689	nd	10739	0.29
Alcohol
1-heptanol	9798	nd	9010	8118	4288	0.28
1-octen-3-ol	991bc	nd	nd	nd	–	–
1-Octanol	1076bc	5507	2808	8464	3621	0.55
2-(hexyloxy)-ethanol	11165	8567	22200	50750	15129	0.15
2-cyclohexen-1-ol	12525	nd	nd	1598	922	0.38
Alkane
Heptane	6437	1956	nd	6694	2569	0.19
4-methyl heptane	7225	4576	14021	8495	5581	0.50
Octane	7597	8600	7643	39130	14916	0.28
4-methyl-octane	8235	4893	6221	nd	3708	0.47
3-ethyl hexane	8305	1208	nd	nd	697	0.38
Benzene
Methyl-benzene	7545	19213	5137	21263	11073	0.54
Furan
2-ethyl-furan	6705	nd	nd	6230	2257	0.09
2-pentyl-furan	966bc	331b	ndb	40804a	11218	0.03
Ketone
2-butanone	5675	nd	6488	7211	5600	0.61
2,3-butanedione	59910	7263	39720	27414	12587	0.20
2-pentanone	6725	nd	nd	1253	723	0.38
3-hydroxy-butanone	7945	32894	nd	21804	20965	0.54
4-heptanone	8545	nd	nd	nd	707	0.61
2-heptanone	8739	nd	nd	4642	2680	0.38
2,3-octanedione	968de	nd	nd	9687	4047	0.17
2-decanone	11728	nd	nd	nd	–	–
Pyrazine
Methyl-pyrazine	8508	nd	nd	nd	–	–
2,5-dimethyl-pyrazine	9248	nd	nd	nd	–	–
Trimethyl-pyrazine	10028	nd	nd	nd	–	–
2-ethyl-3,5-dimethyl-pyrazine	10698	nd	nd	nd	–	–
2-ethyl-6-methyl-pyrazine	16195	nd	nd	nd	–	–
S-containing
Methanethiol	3775	383	2832	2537	1292	0.36
Sulfur dioxide	4465	nd	nd	894	516	0.38
Carbon disulfide	4895	26914	13788	4375	9852	0.28
Terpene
1-octene	7495	nd	nd	nd	–	–
Toluene	7527	4080b	26620a	3730b	7160	0.05
Styrene	8935	nd	3134	9414	5729	0.51
Cedr-8-ene	14215	ndb	11024gh	22835a	4864	0.01
Gamma-muurolene	14335	ndb	ndb	4979a	1501	0.04
Widdrene	14415	nd	29420	26475	11188	0.14
Thujopsene	14445	nd	5704	28919	11626	0.20

¹Yearling Rambouillet wethers were assigned to a treatment group, which consisted of feeding a 20% ground sorghum-sudangrass hay-based diet for 40 d (JUN0; n = 10), a 20% hay-based diet for 20 d followed by a 20% ground juniper-based diet for 20 d (JUN20; n = 10), or a 20% juniper-based diet for 40 d (JUN40; n = 10). Means within a row without a common superscript differ (P < 0.05).

²Volatile compounds identified through mass spectrometry and reported as total ion count area under the curve (nd indicates that a volatile was not detected or was below the threshold of detection or less than 1000 TIC area under the curve).

³Linear retention index on a multidementional GC using a BPX5 30m column followed by a BP20 30m column.

⁴SEM = greatest standard error of the means.

⁵Mass spectrum NIST library identification only.

⁶ Elmore et al. (1997).

⁷ Madruga et al. (2000).

⁸ Elmore et al. (1999).

⁹ Byrne et al. (2002).

¹⁰ Shellie et al. (2002).

^abcLeast squares means in a row with a different superscript differ (P < 0.05).

DISCUSSION

The lack of negative effects on wether growth performance due to replacing ground sorghum-sudangrass hay with ground juniper was expected, even though the juniper-based diet had greater structural fiber than the hay-based diet (18.0% vs. 12.9%, respectively). The juniper-based diet in the current trial contained 20% juniper which appears to be an optimal level of inclusion in lamb feedlot diets. Whitney et al. (2014) reported that using up to 24% juniper in a lamb feedlot diet, similar to what was used in the current trial (e.g., 40% DDGS), did not negatively affect growth performance, but actually increased BW, DMI, and ADG. These authors suggested that the enhanced growth performance was mainly due to differences in the chemical and physical characteristics (e.g., particle size and buoyancy) of the juniper compared with the hay.

Plant secondary compounds such as CT and volatile oil have been reported to enhance or reduce DMI and animal performance. The juniper used in the current trial contained low concentrations of CT (2.7%) and volatile oil (1.6%; data not shown), whereas PPP, PB, and PB/PPP were 2.6 mg/g, 57.3 mg/g, and 22.0 g/g, respectively. Furthermore, it is apparent that the juniper used in the current trial did not negatively affect DMI or animal performance. Others have reported that animal growth performance increased when fed diets that contained CT (Moore et al., 2008) or volatile oil (Min et al., 2012; Whitney et al., 2014).

Kerth and Miller (2015) have described the human physiological mechanisms of detecting aromas and their importance to the eating experience especially as it pertained to flavor. Hornstein and Crowe (1963) reported that flavor precursors in lean lamb were low molecular weight, water soluble compounds that produce meaty aromas upon heating. Additionally, Mottram (1998) indicated that flavor and aromas from meat products are derived from 2 distinct mechanisms: lipid degradation and the generation of water-soluble aroma compounds. Still others (Calkins and Hodgen, 2007) have described some of the aroma chemical compounds derived from the Maillard reaction and lipid degradation, and reported the importance of animal diet on meat flavor.

Sheep meat has a distinctive flavor that makes it recognizably different from beef or pork (Pearson et al., 1973), but its flavor is largely dependent on the age of animal and whether lean or fatty cuts are evaluated (Duckett and Kuber, 2001). The impact of diet and feeding regimens in sheep backgrounding and finishing vary greatly when measuring flavor (Watkins et al., 2013). Although some have reported that diet has direct impacts on flavor (Almela et al., 2010; Resconi et al, 2010), others have reported no effect of diet on lamb flavor (Young et al., 1994; Fraser et al., 2004; Pethick et al., 2006).

It is clear that the type and length of feeding period prior to harvest can determine lamb sensory characteristics. Sensory results from the present study agree with Kerth et al. (2018) who reported that no differences in lamb loin chop–trained sensory panel scores were found when chops were taken from lambs finished on any of 4 different species of juniper. Whitney and Smith (2015) reported that substituting ground oat hay with juniper did not affect off-flavor, but enhanced juiciness, tenderness, and flavor intensity. They attributed the greater sensory characteristics to the fact that ground juniper contains condensed tannins and terpenes (Stewart et al., 2015). Growth performance and sensory data in the current study are unique to previous investigations feeding juniper to sheep as yearlings (~18 mo at harvest) were utilized instead of lambs (~7 mo at harvest; Whitney et al., 2014; Whitney and Smith, 2015; Kerth et al., 2018). It is interesting to note, in the present study, that sensory scores in leg roasts were as much as 1.2, 7.4, and 2.1 units greater than loin chops for lamb ID, brown, and roasted sensory scores, respectively. These differences between loin chops and leg roasts are likely caused by the differences in cooking methods and the lack of high heat temperatures applied to the samples. Additionally, leg roasts samples were taken from the interior of the muscle and did not include any exterior surfaces like the loin chops.

Condensed tannins can alter rumen biohydrogenation, thus alter meat fatty acid composition (Vasta et al., 2009), which can then affect sensory characteristics (Melton, 1990). Condensed tannins can also reduce skatole in adipose tissue (Young and Baumeister, 1999; Vasta and Luciano, 2011) and thus potentially reduce its associated negative odors and flavors (Young et al., 1997; Young et al., 2003). Additionally, volatile compounds (e.g., terpenes) are deposited in adipose tissue (Serrano et al., 2007) and have been reported to affect flavor (Vasta and Priolo, 2006; Resconi, et al., 2013). However, sensory characteristics of yearling mutton were not enhanced in the current trial by replacing hay with ground juniper. This may have been partially due to removing all external fat and only serving the interior of the cut to the sensory panel. It may have also been due to the yearling wethers in the JUN0 treatment not having much aversive lamb flavor or volatile compounds. Hornstein and Crowe (1963) reported that the basic meaty flavor is constituted by the water-soluble fraction, whereas species-specific flavors are found in the lipid fraction. In a review, Duckett and Kuber (2001) noted that varying the type (pasture vs. grain) and duration of finishing diet affects lamb flavor, but more research was needed to determine optimal nutritional schemes to enhance flavor in a given production environment.

Many reports have described the impact of volatile aroma compounds on the flavor profile of lamb (Bueno et al., 2011, 2014; Frank et al., 2017). In fact, Bueno et al. (2014) developed a model explaining lamb flavor using GC/MS techniques and found that lamb flavor is positively dependent on concentrations of volatile fatty acids and several dimethylpyrazines and is negatively associated with different alkenals and alkadienals. In the present study, feeding juniper for up to 40 d had no impact on lamb flavor identification score of grilled loin chops. It is important to note that even though a small decrease in lamb flavor identification score of leg roasts was observed when juniper was fed for up to 40 d, all of the treatments had nearly a 1-unit greater sensory score compared with grilled loin chops. Furthermore, the positive sensory attributes of brown and roasted were 7.0 and 1.9 units greater in leg roasts compared with loin chops, respectively. This can likely be attributed to the cooking method and the fact that more of the volatile aroma compounds are lipid-derived rather than the pyrazines found in the loin chops.

Conclusions

Replacing ground sorghum-sudangrass hay with juniper in a yearling wether finishing diet did not negatively affect growth performance, dressing percentage, or loin eye area. Even though treatment diet affected some volatile aroma compounds from cooked meat, sensory attributes of yearling mutton were not enhanced as initially hypothesized. It is apparent that published literature disagrees on the impact of diet and feeding duration on sensory attributes and volatile aroma compound concentrations. Although the effect of finishing diet on sheep meat sensory attributes have been shown to differ widely, many studies, including this one, observed no effect. Thus, juniper is a suitable alternative roughage feed ingredient that does not affect sensory characteristics of yearling mutton.