Effects of breed and test duration on feed efficiency and feeding behavior in growing goats using GrowSafe feeding technology

Abstract

This study aimed to determine the effects of breed and test durations on feed efficiency and feeding behavior in growing goats. Nineteen female goats (10 Alpine and 9 Spanish) were used in an 84-d feeding trial. Goats were housed in pens equipped with a GrowSafe feed bunk and received a pelleted diet twice daily with ad libitum access to feed, water, and mineral blocks. Feed intake, feeding behavior, and body weight were monitored throughout the trial to estimate average daily gain (ADG), metabolic body weight (MBW), expected dry matter intake (DMI), and residual feed intake (RFI) at 42, 56, 70, and 84 d. Data were analyzed using a repeated-measures model. The breed × test duration (TD) interaction was not significant for any trait except head-down (HD) duration and HD duration: bunk visit duration (BVD). Alpine goats exhibited greater ADG (0.233 vs. 0.191 kg/d) and DMI (1.80 vs. 1.44 kg/d) than Spanish goats (P ≤ 0.03). A tendency for a breed effect was observed for RFI (P = 0.09), with Alpine goats exhibiting greater RFI values than Spanish goats. However, feed efficiency indices and most feeding behavior traits were not affected by breed (P > 0.05). Feeding behavior changed over time, with greater feeder activity early in the trial and higher eating rates later in the trial (P ≤ 0.004). These findings indicate that breed influences growth performance and feed intake but has limited effects on feed efficiency and feeding behavior in growing goats. The greater growth performance observed in Alpine goats appears to be associated with higher feed intake rather than differences in feed utilization efficiency. Further research involving larger populations, additional breeds, and diverse production systems is needed to validate these findings and assess their broader applicability.

Graphical Abstract

Graphical Abstract.

Check out our latest study using automated feeding technology to uncover breed differences in growth, feed efficiency, and feeding behavior between Alpine and Spanish goats

Introduction

Productivity and financial sustainability in goat production are influenced by multiple input factors, with feed being the most critical (Goetsch et al. 2010; Brzáková et al. 2021). Feed costs often account for more than 70% of total variable operating expenses (Muktiani et al. 2020), and recent increases in feed prices have created substantial challenges for goat producers worldwide (Muktiani et al. 2020). Beyond feed price volatility, overall production costs and profitability in goat farming are strongly affected by the efficiency with which animals convert feed into marketable products (Volden and Gustafsson 2011). Consequently, improving feed efficiency has become a central objective for goat producers, achieved by selecting genetically superior animals and adopting management practices that increase output per unit of feed input (Joy et al. 2020).

Feed efficiency, defined as an animal’s ability to convert feed into meat, milk, or fiber (Cammack et al. 2005; Berry 2008; Connor et al. 2019; Scholtens 2023), is recognized as an economically important trait (Scholtens 2023) and is increasingly used in breed improvement programs (Bezerra 2013). Research in cattle, pigs, and sheep has demonstrated that selection for improved feed efficiency provides substantial economic benefits (Archer et al. 2002; Redden et al. 2011; Saintilan et al. 2013; Tedeschi et al. 2021) and contributes to environmental sustainability by reducing feed consumption and greenhouse gas emissions (Arthur and Herd 2008; Johnson et al. 2019; Ferreira et al. 2025).

Several metrics have been developed to assess feed efficiency across livestock species, including feed conversion ratio (FCR) in poultry (Aggrey and Rekaya 2013), goats (Tesema et al. 2024), and beef cattle (Davison et al. 2023); partial efficiency of growth (PEG) in cattle (Grion et al. 2014); Kleiber ratio (KR) in cattle (Köster et al. 1994) and sheep (Mahala et al. 2023); and residual feed intake (RFI) in cattle (Koch et al. 1963; Herd et al. 2003; Arthur and Herd 2008), goats (Kirkham 2025), sheep (Li et al. 2026), and poultry (Aggrey and Rekaya 2013; Ramankevich et al. 2025). These metrics differ in their biological interpretation but are commonly used to characterize an animal’s ability to convert feed into growth and production. Among these measures, RFI has gained widespread acceptance as a reliable indicator of feed efficiency (Knott et al. 2008; Bezerra et al. 2013; Mahala 2020; Johnson et al. 2022). Residual feed intake represents net feed consumption and is defined as the difference between an animal’s actual feed intake and its expected intake based on maintenance and growth requirements (Koch et al. 1963; Do et al. 2013). Expected intake is typically estimated using linear regression of dry matter intake (DMI) on average daily gain (ADG) and mid-test metabolic body weight (MBW^0.75) (Koch et al. 1963; Archer et al. 1997; Basarab et al. 2013; Khanal et al. 2024). Because RFI is phenotypically independent of growth rate and body size, it is particularly useful for identifying more efficient animals without compromising performance.

Accurate estimation of RFI depends on precise measurement of individual feed intake over extended feeding trials. Traditional intake measurement methods, which often involve individual housing and manual data collection, are prone to measurement error (Dickson et al. 2019) and are labor-intensive, time-consuming, and costly. Advances in precision livestock farming (PLF) have enabled the development of automated feeding systems that accurately record individual feed intake in group-housed settings (Berckmans 2014; Tedeschi et al. 2021). These systems combine modern sensor technology, radio-frequency identification (RFID), and wireless connectivity to improve data capturing and reduce labor requirements (Goetsch et al. 2010; Mendes et al. 2011; Tedeschi et al. 2021).

One such technology is the GrowSafe Feeding System (GrowSafe Systems Ltd., Airdrie, Alberta, Canada; now Vytelle), which has been widely adopted for monitoring feed intake and feeding behavior in beef cattle (Dini et al. 2019), sheep (Redden et al. 2011, 2013), and horses (Dickson et al. 2019). However, limited information is available on breed differences in feed efficiency and feeding behavior in growing goats evaluated using automated feeding systems (Gipson et al. 2006; Hu et al. 2012; Chavez 2008). As a result, estimating RFI and other feed efficiency metrics has been more challenging in goats than in other livestock species (Cannas et al. 2019), limiting progress in genetic selection and management strategies aimed at improving efficiency (Nkrumah et al. 2007).

Feeding behavior traits, including feeding frequency, duration, and feeding event patterns, are both phenotypically and genotypically associated with residual feed intake (RFI) and overall feed efficiency (Nkrumah et al. 2007; Lancaster et al. 2009; Parsons et al. 2020). These relationships highlight the importance of characterizing feeding behavior when evaluating feed efficiency. Therefore, the objective of this study was to determine the effects of breed and test duration on feed efficiency, feeding behavior, and performance traits in growing goats using GrowSafe technology.

Materials and methods

Ethics and compliance

All animal handling and experimental procedures were approved by the Prairie View A&M University Institutional Animal Care and Use Committee (IACUC; Protocol No. AUP‑2019) and were conducted in accordance with the United States Department of Agriculture (USDA) Animal Welfare Act and the Guide for the Care and Use of Agricultural Animals in Research and Teaching (USDA 2023).

Experimental design and animal management

A feeding trial was conducted in growing goats at the International Goat Research Center (IGRC), Prairie View A&M University, Prairie View, Texas, USA. Initially, 20 female goats (10 Alpine and 10 Spanish) were randomly selected from a pool of 210 weaned kids born between February and April 2020. Goats were transferred from the yearling barn to the GrowSafe feeding facility four weeks before the start of data collection. Within each breed, ten goats per pen were housed in a shaded barn and assigned to two non‑adjacent pens (6 × 9 m) for a 14‑day adaptation period. Each pen had a concrete floor, partially covered by an iron slatted surface, and was equipped with two electronic GrowSafe feed bunks. At the beginning of the adaptation period, goats were fitted with electronic identification (EID) ear tags (RU‑HDXOTP982‑O/GSMO‑50; Amflex Plastic Inc., Oceanside, CA, USA) containing radio‑frequency identification (RFID) chips.

All goats were dewormed against internal parasites using a combination of Cydectin, Valbazen, and Prohibit, and were vaccinated against tetanus and Clostridium perfringens types C and D. During the adaptation phase, goats were fed a total mixed ration (TMR) ad libitum twice daily (0900 and 1600 h) and had unrestricted access to clean drinking water and a mineral block. This adaptation period allowed goats to acclimate to the GrowSafe feeding system and group‑housing conditions prior to the initiation of the feeding trial.

GrowSafe system calibration and verification

The GrowSafe feeding system (Model 6000E; GrowSafe Systems Ltd., Airdrie, AB, Canada) used in this study was originally designed for cattle. To adopt the system for goats, a custom wooden feed trough (100 × 80 × 15 cm; 0.12 m³; 50-kg capacity) was constructed and mounted on an iron angle frame attached to the original GrowSafe bunk, approximately 50 cm from the base of the feed bunk. To ensure that only one goat could access the feed at a time, the gate opening between the neck bars was narrowed using a fitted plywood panel. Additionally, a concrete step (40.6 cm high × 55.9 cm deep × 91.4 cm wide) was installed beneath each entry gate to facilitate easier access to the feed trough. Detailed information about the modification is given elsewhere (Kidane, personal communication).

The load cells on each GrowSafe feed bunk were calibrated using a 22.7 kg (50-lb) dummy weight (Lake Weighing Systems, Model 55NA86 #12839, USA) at the beginning of the experiment and every 2 weeks thereafter. The feed bunk sensors (responders) were verified using a Test Wand (Model GS0-11828), which was positioned over the center of the bunk for at least 2 seconds or until 20 audible beeps were recorded, confirming proper sensor functionality.

Experimental animal management

Before data collection, goats underwent a 14-day pre-trial adaptation period. The initial body weight (BW) and age of Alpine goats averaged 37.6 ± 2.8 kg and 364 ± 24 days, respectively, whereas Spanish goats averaged 34.7 ± 2.8 kg and 353 ± 31 days, respectively. Within each breed, goats were randomly assigned to one of two non‑adjacent pens, with five goats per pen (Pen #1 and Pen #3 for Alpine goats; Pen #2 and Pen #4 for Spanish goats).

In each pen, goats had access to a single GrowSafe feed bunk and were offered a pelleted total mixed ration (TMR) ad libitum. Feed was delivered twice daily at 09:00 and 16:00 h. To minimize periods of bunk emptiness and maintain at least 3 kg of feed in each bunk, as recommended by GrowSafe user protocols, feed was provided in excess at each feeding. One Spanish goat was removed from the study 2 days after the start of the experiment due to health-related issues, leaving 19 goats for data collection.

Individual feed intake was continuously monitored over 24‑h periods using the GrowSafe system. When substantial fluctuations in intake were observed, whether as abrupt increases or decreases, feed allowances were adjusted in accordance with established GrowSafe operating procedures. Feed bunks were vacuum‑cleaned biweekly while goats were temporarily removed from pens for body weight measurements and were subsequently refilled with fresh feed upon the goats’ return. Throughout the 84‑day experimental period, goats had unrestricted access to fresh, clean water and a trace mineralized rock salt block (Ranch House; United Salt Hockley, LLC, Hockley TX, USA).

Feed composition

Goats were offered a pelleted TMR composed of 32.6% roughages, 25.0% processed grain byproducts, 12.5% forage products, 10.5% soybean meal, 10.4% grain products, 5.0% mineral mix and salt, 3.05% molasses and pellet binders, 0.75% ammonium chloride, and 0.20% vitamin premix. The ingredients and their respective proportions used to formulate the TMR are presented in Table 1. The diet was formulated in-house at the International Goat Research Center (IGRC) to meet or exceed the nutrient requirements of growing goats, in accordance with the Nutrient Requirements of Small Ruminants (NRC 2007) and was manufactured by the Producers’ Cooperative Association (Bryan, TX, USA).

Table 1.

Ingredient composition of the pelleted TMR formulated for growing goats.

Ingredient	% of TMR
Roughages	32.60
Processed grain byproducts	25.00
Forage products	12.50
Soybean meal	10.50
Grain products	10.40
Mineral mix and salt	5.00
Molasses and pellet binders	3.05
Ammonium chloride	0.75
Vitamin premix	0.20
Total	100.00

Goats had ad libitum access to the TMR in a group-feeding environment using the GrowSafe automated feeding system. Feed samples were collected biweekly and stored at 4.5°C until chemical analysis. Samples were composed by collection period and analyzed for nutrient composition at Servi-Tech Laboratories (Amarillo, TX, USA). The analyzed chemical composition and nutrient content of the TMR are presented in Table 2.

Table 2.

Analyzed chemical composition and nutrient contents of the TMR fed to growing goats over the 84-day experiment period.

Nutrient	DM basis
Dry Matter (DM), %	87.40
Crude Protein (CP), %	18.20
Acid Detergent Fiber (ADF), %	29.30
Neutral Detergent Fiber (NDF), %	33.60
Total Digestible Nutrients (TDN), %)	63.30
Net Energy for Maintenance (NEm), Mcal/kg	1.41
Net Energy for gain (NEg), Mcal/kg	0.84
Net Energy for lactation (NEl), Mcal/kg	1.43
Metabolizable Energy (ME), Mcal/kg	2.29
Digestible Energy (DE), Mcal/kg	2.80
Calcium (%)	1.24
Phosphorus (%)	0.42
Magnesium (%)	0.27
Potassium (%)	1.43
Sulfur (%)	0.23
Sodium (%)	0.23
Zinc (mg/kg)	140
Iron (mg/kg)	297
Manganese (mg/kg)	107
Copper (mg/kg)	23.0

Chemical analysis of feed samples was performed by Servi-Tech Laboratory, Amarillo, Texas. All nutrient concentrations are expressed on a dry matter (DM) basis unless otherwise indicated. Dry matter (DM) content is reported on an as-fed basis.

Live weight measurements

Body weight (BW) was recorded for each goat at the start (day 1), every 14 d thereafter, and at the end of the trial (day 84), resulting in seven BW measurements per goat. A 14-d weighing interval has been shown to provide accurate estimates of growth and feed-efficiency traits in cattle, with minimal loss of accuracy relative to weekly weighing during feed-efficiency testing (Archer et al. 1997).

Weights were obtained using a standard weighing chute (Rice Lake Weighing Scale, Model 107102A) equipped with a load cell and digital recorder (Tru-Test ID 5000, Model 503028, New Zealand). Before each biweekly weighing session, the scale was calibrated using a 22.7 kg cast-iron dummy weight. Each goat was individually weighed by stepping onto the chute, during which its electronic identification (RFID) tag was scanned using a Tru-Test RFID reader, and the recorded BW was automatically transferred to a desktop computer for data storage and analysis. All BW measurements were conducted before the morning feeding (between 9:00 and 10:00) to minimize gut-fill variation, with no prior restriction on feed or water.

Feed intake and feeding behavior measurements

Data on individual feed intake and feeding behavior were collected continuously over 84 days using the automated GrowSafe feeding system. Daily individual feed intake was automatically calculated as the total amount of feed consumed within 24 hours. The GrowSafe system also recorded feeding behavior parameters, including the frequency of bunk visits (BV), duration of each visit, head-down (HD) duration, and total feeding duration (FD) throughout the experimental period for each goat.

Data verification and auditing

Daily feed intake records for individual goats were screened by plotting feed intake across each day of the 84-day trial. Total daily feed disappearance from each node (scale) was deemed acceptable when the assigned feed disappearance (AFD) rate per bunk was ≥ 90%, consistent with GrowSafe guidelines for acceptable system accuracy (GrowSafe Systems Ltd. n.d., p. 113).

Based on this criterion, data from days 14 and 16 for Pens 1 and 3 and from days 11 and 23 for Pens 2 and 4 were excluded from the analysis. To validate individual animal intake records, a threshold of ± 4 standard deviations from the mean was applied, resulting in the exclusion of 10 intake records from 8 goats. Because the specific causes of these data anomalies were not systematically documented, the excluded records were attributed to data quality issues that may have arisen from power interruptions, environmental interference (e.g., excessive wind-induced noise), or other equipment-related factors.

Feeding behavior traits were evaluated based on the frequency of bunk visits (BV) and the duration of each BV. Definitions of the feeding behavior traits are in Table 3. Bunk visit frequency was calculated as the total number of BV events recorded over 24 hours, regardless of whether feed was consumed. Bunk visit duration was defined as the cumulative length of all BV events during the same period (Dickson et al. 2019; Parsons et al. 2020). Bunk visit event recordings commenced when an animal’s RFID tag was detected at the feed bunk and concluded when either the RFID tag of another animal was detected at the bunk or the original animal’s tag was identified at a different bunk (Mendes et al. 2011; Dickson et al. 2019; Parsons et al. 2020). To evaluate the intensity of feeding patterns, both head-down (HD) duration and the ratio of HD duration to BV duration were calculated. As detailed by Dickson et al. (2019), HD duration represents the total number of times a RFID ear tag was detected within a day, multiplied by the scanning frequency of the GrowSafe system. The HD duration per BV duration (measured in minutes per hour) was computed as the ratio of HD duration to total BV duration at the feed bunk, as described by Mendes et al. (2011). Bunk visit (BV) eating rate (g/min) was calculated as the average feed intake per bunk visit divided by the duration of the corresponding bunk visit (Golden et al. 2008). The day-to-day variation in feeding behavior traits was determined as the root mean squared error (RMSE) from linear regression of each trait on the day of the trial as described by Parsons et al. (2020) and applied by other researchers (Irvin et al. 2023; Kidane et al. 2024).

Table 3.

Definitions of feeding behavior traits evaluated in the study.

Trait	Definition
BV frequency, events/d	Number of BV events recorded each day
BV duration, min/d	Sum of the durations of BV events recorded each day
HD duration, min/d	Number of RFID recordings each day multiplied by the scan rate of the GrowSafe system
HD duration per BV duration	Ratio of HD duration to BV duration

BV, bunk visit; HD, head down; RFID, radio frequency identification.

Residual feed intake calculations

Average daily gain (ADG) and mid-test metabolic body weight (mid-test BW^0.75) were calculated from body weight (BW) measurements collected over the 84-day experimental period, following the methods of Basarab et al. (2013). The ADG for each goat was estimated as the slope of the linear regression of BW on test duration (TD). Mid-test metabolic body weight (mid-test BW^0.75) was calculated separately for each TD as the predicted BW at the midpoint of that test duration, obtained from the regression of BW on days on test and raised to the 0.75 power.

Expected feed intake for each goat was estimated by regressing dry matter intake (DMI) on mid-test metabolic body weight (mid-test BW^0.75) and ADG, using the subsequent multiple growth model described by Kirkham et al. (2025) for goats:

$$Y=\beta 0+{\beta 1}^{*}(\mathit{mid}-{\mathit{test} BW}^{0.75})+{\beta 2}^{*}ADG+\epsilon$$ ([1])

where Y is expected daily feed intake (DM kg/d), β0 is the Y intercept, β1 is the partial regression coefficient of mid-test metabolic body weight (Mid-test BW^0.75), and β2 is the partial regression coefficient of average daily gain (ADG), ε is the error term. Feed intake (FI) was adjusted for liveweight (mid-weight from the measurement period), and ε is the error term.

The residual feed intake of an individual goat was calculated as observed feed intake minus expected feed requirement using the linear model below.

$$\mathit{RFI}=\mathit{Observed} \mathit{DMI}-\mathit{Expected} \mathit{DMI}$$ ([2])

Statistical analysis

Statistical analyses were performed using JMP software (Version 9.4; SAS Institute Inc., Cary, NC, USA), with individual goats serving as the experimental units. To evaluate the effect of test duration (TD) on DMI, feed efficiency, feeding behavior, and performance traits in growing goats, progressively longer periods were examined at 14-d increments during the 84-d trial. The four TDs were 0–42, 0–56, 0–70, and 0–84 d. Trait values were calculated separately for each TD using observations available within that period. Consequently, BW and derived performance measures varied among TDs because estimates were based on different observation lengths. Estimated initial BW also varied because BW parameters were estimated independently from observations available within each TD. Thus, these differences reflected differences in data availability rather than actual changes in initial BW. TD was treated as a repeated categorical factor with four levels: 42, 56, 70, and 84 d. Repeated-measures analyses were used to evaluate the effects of TD on BW, feed intake, feed efficiency, and feeding-behavior traits, with individual goats specified as subjects. Data were analyzed separately for the shorter testing periods (42, 56, and 70 d) and the full 84-d dataset using a repeated-measures mixed model (PROC MIXED) with an unstructured (UN) covariance structure.

Pen was initially evaluated as a random effect to account for potential environmental variation among housing units. However, because the pen effect was not significant (P > 0.05), pen was excluded from the final model. The final statistical model included breed, test duration (TD), and their interaction (breed × TD) as fixed effects. Differences among least squares means (LSMeans) were evaluated using the Tukey-Kramer adjustment for multiple comparisons in the multivariate platform of JMP (SAS Institute Inc., Cary, NC, USA). Relationships among continuous variables were evaluated using simple linear regression. Linear models were fitted using the least squares method, and model fit was assessed using the coefficient of determination (R²). Pearson correlation coefficients (r) were calculated to quantify the strength of associations among variables. Statistical significance was declared at P < 0.05.

Results

Effect of breed on ADG and DMI

Breed differences in performance traits are presented in Table 4. Nineteen goats (10 Alpine and 9 Spanish) completed the 84-day test period. At the conclusion of the study, Alpine and Spanish goats were 448 and 437 d of age and had BW of 51.7 ± 5.0 kg and 46.5 ± 4.3 kg, respectively. The ADG and DMI were 0.232 ± 0.019 kg/d and 0.191 ± 0.02 kg/d, and DMI of 1.80 ± 0.31 kg/d and 1.44 ± 0.18 kg/d for Alpine and Spanish goats, respectively. There were no breed × TD interactions (P > 0.05). However, there was a significant difference in initial BW between breeds (P < 0.039), with Alpine goats having higher initial BW than Spanish goats. Similarly, a significant breed effect (P = 0.033) was observed in final BW. Average daily gain also differed significantly (P = 0.030) between breeds, with Alpine goats showing higher ADG than Spanish goats throughout the experiment. ADG increased slightly from day 42 to day 56 and then declined through day 84 in both breeds. The similar temporal pattern observed in both breeds is consistent with the absence of a breed × day interaction (P = 0.108), whereas the higher overall ADG of Alpine goats reflects a significant breed effect (P = 0.03).

Table 4.

The effects of breed on performance traits, feed intake, and feed efficiency in goats fed a TMR diet (n = 19).

	Breed			P-values
Trait	Alpine	Spanish	SE	Breed	TD	Breed ×TD
Performance traits:
Initial BW, kg	37.6	34.7	0.931	0.039	<.0001	0.110
Final BW, kg	51.7	46.5	1.650	0.033	<.0001	0.497
Mid-test  metabolic BW kg0.75	17.3	16.1	0.356	0.026	<.0001	0.251
ADG, kg/d	0.232	0.191	0.019	0.030	<.0001	0.108
DMI, kg/d	1.80	1.44	0.100	0.006	<.0001	0.379
DMI RMSE,  kg/d	0.383	0.315	0.068	0.546	0.0351	0.50
Feed Efficiency traits
F: G	8.21	7.99	0.890	0.890	<0.0002	0.212
G: F	0.128	0.131	0.008	0.730	<.0001	0.24
RFI, kg/d	0.0568	−0.0632	0.0486	0.090	1.000	0.92

ADG, average daily gain; DMI, dry matter intake; DMI RMSE, day-to-day variation in DMI during the trial; MBW, metabolic body weight; RFI, residual feed intake; F: G, feed to gain ratio; G: F, gain to feed ratio; TD, test duration.

The DM intake of goats differed significantly between breeds (P = 0.006). Alpine goats consumed an average of 0.366 kg/d more dry matter than Spanish goats (1.80 vs. 1.44 kg/d), representing approximately 26% higher DMI in Alpine goats. The magnitude of the breed difference was relatively consistent across all measurement periods. In both breeds, the mean DMI remained relatively stable from days 42 to 56 and declined slightly from day 70 to day 84. The similar temporal pattern observed for Alpine and Spanish goats is consistent with the absence of a significant breed × day interaction (P = 0.379), indicating that breed differences in DMI were maintained across the experimental period. Feed efficiency, expressed as feed: gain ratio (F: G), was not affected by breed (P > 0.73), and breed did not affect G: F (P = 0.89). The breed × day interaction was not significant for either trait (P = 0.21).

Effects of test duration on performance traits

Results on the effects of test duration (TD) on performance traits are presented in Table 5. There were no breed × TD interactions (P > 0.05) across the testing period. Initial BW differed significantly (P < 0.05) across TD. Goats’ weight increased with test duration, reaching the highest value on day 84, followed by day 70, and the lowest on days 42 and 56 (Table 5). Final BW also differed significantly (P < 0.05) across test durations, with the highest value on day 84 and the lowest on day 42. DMI differed significantly (P < 0.05) across test durations, with the highest DMI on days 42 and 56, followed by day 70, and the lowest on day 84. Similarly, average daily gain differed significantly with test duration (P < 0.05). The highest ADG was observed on day 56, followed by day 42, and the lowest on day 84 (Fig. 2). The F: G ratio differed across test days (P < 0.001). The breed × TD interaction was not significant for either trait (P = 0.21). Mean F: G increased from approximately 7.3 at d 42 to 9.4 at d 84, whereas G: F decreased from approximately 0.142 to 0.111 over the same period. Conversely, G: F differed significantly (P < 0.05), with higher values on days 42 and 56, intermediate on day 70, and lower on day 84. These findings indicate that a decline in feed efficiency as the goats progressed through the trial was consistent across days, likely reflecting reduced growth rates as body weight increased.

Table 5.

The effects of test duration on performance traits, feed intake, and feed efficiency in goats fed a TMR diet (n = 19).

Trait	Test duration (TD), d				SE	P-values
	42	56	70	84		Breed	TD	Breed × TD
Performance traits:
Initial BW, kg	35.6a	35.5a	36.5b	37.2c	0.672	0.039	<.0001	0.110
Final BW, kg	45.5a	49.0b	50.3c	51.7d	1.29	0.033	<.0001	0.497
Mid-test metabolic  BW kg0.75	16.1a	16.6b	16.9c	17.2d	0.265	0.026	<.0001	0.251
ADG, kg/d	0.236a	0.240a	0.197b	0.173c	0.016	0.130	<.0001	0.108
DMI, kg/d	1.65a	1.65a	1.62a	1.56b	0.072	0.006	<.0001	0.379
DMI RMSE, kg/d	0.330	0.341	0.359	0.367	0.050	0.546	0.0351	0.500
Feed efficiency traits
F: G	7.33a	7.19a	8.46b	9.42c	0.725	0.890	0.0002	0.212
G: F	0.142a	0.144a	0.122b	0.111c	0.006	0.730	<.0001	0.240
RFI, kg/d	−0.003	−0.003	−0.003	−0.003	0.040	0.091	1.000	0.920

^a-cWithin a row, means without a common superscript letter differ (P ≤ 0.05). BW, body weight; ADG, average daily gain; DMI, dry matter intake; DMI RMSE, day-to-day variation in DMI during the trial; F: G, feed to gain ratio; G: F, gain to feed ratio; RFI, residual feed intake; TD, test duration.

Figure 2.

Bunk visit frequency (BVF; visits/d) of Alpine (n = 10) and Spanish goats (n = 9) throughout the 84-d test duration. Values represent least-squares means ± SEM. Alpine goats consistently visited the feeder more often than Spanish goats at each measurement period (P < 0.05). Bunk visit frequency decreased steadily throughout the trial in both breeds. Although feeder visits became less frequent over time, the decline was more pronounced in Alpine than in Spanish goats (15.7% vs 9.3%), but the overall pattern was similar between breeds.

Residual feed intake

For each goat, residual feed intake (RFI) was calculated as the difference between actual dry matter intake (DMI) measured during the 84-d testing period and expected DMI. Expected DMI was strongly correlated with actual DMI (r = 0.82; P < 0.001), explaining approximately two-thirds of the variation in observed intake (R² = 0.67; Fig. 1). Residual feed intake ranged from −0.252 to 0.380 kg/d and tended to differ between breeds (P = 0.09). Mean RFI was lower in Spanish than Alpine goats (−0.063 vs. 0.057 kg/d), with a mean breed difference of 0.12 kg/d. Spanish goats consumed 0.073 ± 0.13 kg DM/d less feed than expected, whereas Alpine goats consumed 0.066 ± 0.18 kg DM/d more feed than expected.

Figure 1.

Relationship between the expected dry matter intake (DMI) and the observed dry matter intake (DMI) of Alpine and Spanish goats fed for 84 experimental days (n = 19). Expected DMI was strongly associated with observed DMI (r = 0.82, P < 0.001). The regression of observed DMI on expected DMI yielded the equation: $\text{Observed} \text{DMI}=0.007+0.995\times \text{Expected} \text{DMI},$with an R² of 0.667. Thus, approximately 67% of the variation in observed DMI was explained by ADG and metabolic body weight, while the residual variation was used to calculate RFI. The regression slope did not differ substantially from unity (0.995; 95% CI = 0.634 to 1.356), supporting the validity of the prediction equation used to estimate expected feed intake.

Effects of breed on feeding behavior

Table 6 summarizes breed differences in feeding behavior. The frequency of bunk visits (BV) did not differ significantly between breeds (P = 0.77); however, it was numerically higher in Alpine goats (22.8 visits/day) than in Spanish goats (19.2 visits/day) (Fig. 2). Similarly, the duration of these visits did not differ significantly (P > 0.05), with Alpine goats spending 3.2 min/visit compared to their Spanish counterparts, spending 3.7 min/visit. Total feeding time did not significantly differ between the two breeds (P > 0.05), with Alpine goats spending approximately 74 minutes per day and Spanish goats 71 minutes at the feed bunk. The eating rate was also not significantly different between breeds (P > 0.05). Head-down duration was comparable for both Alpine and Spanish goats (P = 0.87), averaging 33.7 and 32.5 minutes per day, respectively. The HD duration-to-BV duration ratio was also comparable between the breeds, with values of 0.46 for Alpine goats and 0.45 for Spanish goats. However, a significant breed × TD interaction was observed for HD duration (P < = 0.05), suggesting differences in temporal patterns across breeds, with Spanish goats showing slightly lower HD duration values over time. Notably, less than 10% of the total time spent at the feed bunk was dedicated to actual eating, while total time spent within the detectable range of the feed bunk sensors ranged widely (48 to 113 min/d), and eating rates among individual goats varied significantly (16.2 to 38.1 g/min).

Table 6.

Effect of breed on feeding behavior traits in goats fed a TMR diet (n = 19).

Trait	Breed		SE	P-values
	Alpine	Spanish		Breed	TD	Breed × TD
Bunk Visit (BV) traits:
BV frequency, events/d	22.8	19.2	2.1	0.415	<.0001	0.0793
BV duration, min/d	74.0	71.8	5.2	0.77	<.0001	0.317
BV eating rate, g/min	26.6	21.8	1.9	0.244	0.0040	0.407
Intensity traits:
HD duration, min/d	33.7	32.5	4.3	0.87	<.0001	0.007
HD duration per BV duration	0.456	0.450	0.045	0.93	0.287	0.0004
Day-to-day variation of traits:
BV frequency RMSE, events/d	3.37	4.01	0.66	0.046	0.009	0.120
BV duration RMSE, min/d	19.3	19.9	1.76	0.79	0.262	0.72
BV eating rate RMSE, g/min	8.26	7.03	0.83	0.39	0.0015	0.174
HD duration RMSE, min/d	11.9	11.6	1.01	0.81	0.104	0.402

BV, bunk visit; HD, head down; RMSE, day-to-day variation of the traits during the trial; TD, test duration.

Effects of test duration on feeding behavior

Bunk visit traits were determined on each test period (days 42, 57, 70, and 84) using pooled data from both breeds. Results on the effects of test duration on feeding behavior are presented in Table 7. Test duration significantly affected bunk visit frequency (P < 0.001). Goats made the most frequent visits on day 42, followed by day 56, with a significant decline by day 84 (Fig. 4). Bunk visit frequency declined throughout the testing period in both breeds. Alpine goats reduced feeder visits from 24.9 to 21.0 visits/d (15.7% decline), whereas Spanish goats decreased from 20.4 to 18.5 visits/d (9.3% decline) between d 42 and d 84. Despite the decline over time, Alpine goats consistently exhibited a higher bunk-visit frequency than Spanish goats across all measurement periods. Similarly, BV-duration differed significantly (P < 0.001) across trial days, being longest on the shortened days (days 42 and 56), intermediate on day 70, and shortest on the long-duration day (day 84). Head-down duration showed a similar pattern, differing significantly with trial length (P < 0.001). Goats showed higher HD duration values on days 42 and 56 and lower values on days 70 and 84. In contrast, the HD duration to-BV-duration ratio (P = 0.287) was not affected by trial length. However, except for HD duration, breed × TD interactions did not significantly affect most behavioral traits (P > 0.05). HD duration showed a significant breed × TD interaction (P = 0.0072), indicating that Alpine goats showed clear differences across trial days, whereas Spanish goats did not. Specifically, Alpine goats showed higher HD duration at shorter testing periods (42 and 56 days) than at longer ones (70 and 84 days). Furthermore, a significant breed × day interaction was detected for the HD duration-to-BV duration ratio (P < 0.001), indicating that the effect of breed on these traits varied across the study period, even though goats spent half of their BV duration engaged in eating or pretending to eat.

Table 7.

Effect of test duration on feeding behavior traits in goats fed a TMR diet (n = 19).

Trait	Test duration				SE	P-values
	42	56	70	84		Breed	TD	Breed × TD
Bunk Visit (BV) traits:
BV frequency, events/d	22.6a	21.4b	20.2c	19.6d	1.5	0.415	<.0001	0.0793
BV duration, min/d	74.7a	75.2a	71.8b	69.9c	3.8	0.77	<.0001	0.317
BV eating rate, g/min	23.9ab	23.8a	24.5b	24.5ab	1.4	0.244	0.004	0.407
Intensity traits:
HD duration, min/d	34.3a	34.1a	32.4b	31.7b	3.0	0.87	<.0001	0.007
HD duration per BV duration	0.457a	0.452a	0.449b	0.452b	0.032	0.93	0.287	0.0004
Day-to-day variation of traits:
BV frequency RMS, events/d	5.45ab	5.29a	5.01b	5.01ab	0.57	0.046	0.009	0.120
BV duration RMSE, min/d	18.4	18.8	20.4	20.9	1.7	0.790	0.262	0.72
BV eating rate RMSE, g/min	6.99a	7.21a	7.93b	8.44b	0.70	0.390	0.0015	0.174
HD duration RMSE, min/d	11.1	11.8	12.1	12.1	0.8	0.810	0.104	0.402

^a-cWithin a row, means without a common superscript letter differ (P ≤ 0.05).

With each trial day, there were no significant differences between breeds.

HD, head down, RMSE = day-to-day variation of the traits during the trial; TD, test duration.

Figure 4.

Average daily gain (ADG) of Alpine (n = 10) and Spanish goats (n = 9) across the 84-d test duration. Values represent least-squares means ± SEM. Average daily gain (ADG) was significantly higher in Alpine than in Spanish goats across test durations (P = 0.001). ADG was greatest at 42 and 56 d and lower at the longer test durations in both breeds. The breed × test duration (TD) interaction was not significant (P = 0.108), indicating that TD’s effect on ADG was similar across breeds.

Discussion

Feed is a major economic input in livestock production; therefore, improving feed utilization is important for reducing production costs across livestock species (Arthur and Herd 2008; Muktiani et al. 2020). Feed efficiency, including RFI, has been evaluated in cattle (Arthur and Herd 2008), sheep (Knott et al. 2008), poultry (Aggrey and Rekaya 2013), and goats (Hu et al. 2012; Muktiani et al. 2020; Kirkham et al. 2025). Accordingly, efficiency-related traits are of considerable interest because they offer a means to improve production efficiency and profitability. In goats, these traits are particularly important for sustainable and economically viable production systems (Tesema et al. 2024). Understanding the relationships among feed intake, growth rate, and efficiency is fundamental for identifying animals with superior biological performance (Cockrum et al. 2013). However, these traits are influenced by other factors, including feeding behavior, breed, and animal age, sex, and season (Urge et al. 2004; Tesema et al. 2024). Despite the economic importance of feed efficiency, limited information is available regarding feed efficiency and feeding behavior in Alpine and Spanish goats, particularly under automated feeding systems. Accordingly, the following discussion evaluates breed-related differences in growth performance, feed efficiency, and feeding behavior in Alpine and Spanish goats.

Effects of breeds on DMI and ADG

The present study found that DMI differed significantly between breeds, with Spanish goats consuming less feed (kg d⁻¹) than Alpine goats (Fig. 3). These findings are consistent with previous studies by Urge et al. (2004) and Lu and Potchoiba (1990), who reported breed differences in DMI, with Alpine goats exhibiting greater intake than Spanish and Nubian goats. The clear differences in DMI between Alpine and Spanish goats in the present study may be partly attributed to differences in body size. Despite both breeds being born at the same time (within a 1–2‑week interval) and being at similar developmental stages during the experiment, Alpine goats were 10% bigger than Spanish goats.

Figure 3.

Average daily dry matter intake (DMI) of Alpine (n = 10)and Spanish goats (n = 9) across the 84-d test duration. Values represent least-squares means ± SEM. DMI was higher in Alpine than in Spanish goats across test durations. The breed × TD interaction was not significant (P = 0.379), indicating that TD had a similar effect on DMI across breeds. The breed difference (0.366 kg/d; 25%) was consistent across test durations.

This interpretation is supported by the DMI trends observed throughout the experimental period. In both breeds, DMI increased during the early to mid-experimental phases (days 0–57) and declined toward the end of the trial (days 70–84), indicating a similar temporal pattern of eating. Similar intake trends were reported by Brand et al. (2017) in Boer goats and by Gipson et al. (2006) in Boer × Spanish goats. The authors observed a gradual decrease in DMI over the 82‑day feeding period, with a further decline towards the end.

Based on average body weight during the trial, dry matter intake averaged 4.0% and 3.5% of BW for Alpine and Spanish goats, respectively. These rates were within the typical 2–4% range recommended by the NRC (2007) for growing goats. However, Alpine goats consumed 0.367 ± 0.133 kg more DM than Spanish goats, corresponding to approximately 26% greater DMI per unit of BW, reflecting higher feed intake and growth performance throughout the study. Interestingly, feed intake expressed as a BW% was higher in the present study than previously reported in the literature. Using reported DMI and BW data from Cameron et al. (2001), expressing the rate of DMI relative to BW resulted in approximately 2.2% BW for Spanish goats, 2.2% BW for Boer × Spanish goats, and 2.5% BW for Boer × Angora goats. Thus, Spanish goats in the present study consumed substantially more feed per unit of BW. Similarly, Alpine goats consumed approximately 4.0% BW, which exceeded the 3.0% BW reported for intensively managed Alpine goats by Ribeiro et al. (2023). The intake rates reported in the present study indicate that both breeds received adequate nutrients to support normal physiological functions and growth (Luo et al. 2004). However, at similar ages, Alpine goats achieved significantly higher body weights than Spanish goats, reflecting higher intake and, in turn, a corresponding higher ADG across test durations (Fig. 4). Overall, the DM consumption rate observed in the present study (3.5–4.0% of BW) was near the upper end of the range commonly reported for growing goats, likely reflecting the high quality and digestibility of the pelletized diet used during the experiment.

As noted for DMI, ADG showed a similar trend across both breeds, with the highest growth rate observed during the early phase of the testing period, followed by a gradual decline toward the end. The temporal pattern was nearly identical to that of average DMI, although differences between breeds were significant. Previous studies have shown that growth rate is influenced by both feed intake and diet composition, particularly dietary energy (Brand et al. 2017; Malik et al. 2023). In the present study, all goats were offered the same total mixed ration (TMR); therefore, differences in growth between breeds are attributable to differences in DMI rather than diet composition. The higher DMI observed in Alpine goats may reflect increased maintenance requirements associated with greater body weight, resulting in less available energy for growth compared with Spanish goats (Brand et al. 2017). This interpretation is consistent with findings by Cammack et al. (2005), who reported that body weight significantly influences feed intake in growing sheep. Similarly, Brand et al. (2017) reported that intake increased with live weight to meet the energy demands of growing Boer goats.

These relationships are further supported by the significant positive correlation observed between DMI and ADG across all goats and testing periods (r = 0.797; P < 0.001) (Fig. 5). Alpine goats with greater feed intake generally exhibited greater growth rates, and DMI accounted for approximately 64% of the variation in ADG. Thus, these findings indicate that feed intake plays a vital role in the growth performance of goats. However, Brand et al. (2017) warned that higher dry matter intake does not necessarily translate into an increased growth rate, as differences in maintenance energy requirements and energy partitioning among breeds influence growth efficiency.

Figure 5.

Relationship between dry matter intake (DMI) and average daily gain (ADG) in Alpine and Spanish goats fed for 84 experimental days. Dry matter intake was positively associated with average daily gain (r = 0.80, P < 0.001). Linear regression (y = 0.1503x − 0.0319; r = 0.80, P < 0.001) showed that DMI explained 63.5% of the variation in ADG (R² = 0.635; P < 0.001). Average daily gain increased by 0.150 kg/d for each 1 kg/d increase in DMI, with a 95% confidence interval of 0.092 to 0.209 kg/d.

Feed intake to gain efficiency

No statistically significant differences in F: G or G: F were observed between the breeds. However, Spanish goats had more favorable values than Alpine goats. Both F: G and G: F results indicate that breed differences in feed efficiency were inconsistent with those observed for DMI and ADG. However, in a previous study, Urge et al. (2004) reported lower gain efficiency in Spanish goats than in Alpine goats. The lower gain efficiency observed in Alpine goats, despite higher DMI, may reflect greater maintenance energy requirements in dairy‑type breeds compared with meat‑type breeds (Luo et al. 2004; Urge et al. 2004). In other words, a larger proportion of energy in Alpine goats is directed toward maintenance, whereas less is allocated to growth (Luo et al. 2004). In contrast, for each kilogram of feed consumed, Spanish goats showed a numerically higher efficiency in converting feed into body weight, although the difference was not statistically significant.

Furthermore, Gipson et al. (2006) indicated that goats’ feeding behavior can influence feed utilization efficiency. In the current study, Alpine goats made more frequent bunk visits and spent more time at the bunk than Spanish goats, potentially leading to increased non‑eating activity and higher maintenance energy requirements. These findings suggest that Alpine goats with more activity tended to convert feed into body weight less efficiently than those with less activity, although the differences were not statistically significant.

Residual feed intake

Because published studies evaluating residual feed intake (RFI) in goats remain limited, comparisons were made with findings from other ruminant species. In the present study, RFI ranged from −0.252 to 0.380 kg/d, indicating substantial variation in feed efficiency among goats. This range exceeded that reported for growing Saanen goats (−0.16 to 0.17 kg/d) by Kirkham et al. (2025) but was within the range observed during lactation (−0.38 to 0.50 kg/d). Similarly, Bezerra et al. reported that RFI variation in sheep commonly ranges from −0.20 to 0.30 kg/d. Considerable animal-to-animal variation in RFI has also been documented in other ruminants. For example, Knott et al. (2008) reported significant variation in RFI among Scottish Blackface sheep, while Cockrum et al. (2013) observed that efficient rams consumed less feed than inefficient rams. Likewise, Berry (2008) demonstrated that cattle with negative RFI required less feed than animals with positive RFI to achieve similar rates of gain. Arthur and Herd (2008) further reported that low-RFI cattle consumed 10–20% less feed than high-RFI cattle while maintaining comparable production levels. Collectively, these studies highlight the potential for genetic and phenotypic selection of animals with superior feed efficiency, as measured by RFI.

Spanish goats tended to exhibit lower RFI values than Alpine goats, with a mean difference of 0.12 kg/d. Although the breed effect was not statistically significant (P = 0.09), the lower RFI observed in Spanish goats suggests a tendency toward greater feed efficiency. Variation in RFI among animals may be influenced by metabolic and physiological factors (Martin et al. 2021). Although Spanish goats had a numerically lower average RFI, the difference between breeds was not statistically significant. Thus, the biological mechanisms underlying the numerical variation in RFI between breeds remain unclear. Despite higher feed intake and growth rates, Alpine goats did not show improved feed conversion efficiency, as reflected by similar RFI, F: G, and G: F values between breeds. These findings suggest that increased feed intake and growth do not necessarily correspond to improved feed efficiency.

In addition to differences in maintenance requirements, feeding behavior may contribute to variation in feed efficiency. Previous studies have reported associations between feeding behavior traits, including feeding frequency, duration, and feeding event patterns, and feed efficiency in ruminants (Nkrumah et al. 2007; Lancaster et al. 2009; Parsons et al. 2020). In the present study, Alpine goats exhibited a greater frequency of bunk visits than Spanish goats, which may have contributed to their higher feed intake and numerically greater RFI values. Although the mechanisms underlying these relationships remain unclear, the results suggest that feeding behavior traits may contribute to variation in feed efficiency and could serve as useful indicators of RFI in growing goats. Similar conclusions were reported by Redden et al. (2011), who demonstrated substantial variation in feed intake and RFI among growing lambs using the GrowSafe feeding system and highlighted the utility of automated feeding technologies for identifying efficient and inefficient animals.

Overall, the tendency for lower RFI in Spanish goats suggests a potential advantage in feed efficiency; however, additional research involving larger populations and multiple production environments is needed to confirm breed differences and further elucidate the biological mechanisms underlying variation in RFI among goats.

Effects of breed on feeding behavior

Although most feeding behavior traits were not significantly affected by breed, Alpine goats had higher feeder visit frequency and longer feeding duration than Spanish goats throughout the trial. Alpine goats averaged 13 to 36 visits/d and 49 to 115 min/d of feeding activity, whereas Spanish goats averaged 13 to 30 visits/d and 49 to 91 min/d. These values generally exceeded those reported by Gipson et al. (2006) for Boer crossbred wethers housed in automated feeding systems, particularly for Alpine goats, which also tended to exhibit greater feeder occupancy per visit. The greater feeding activity observed in Alpine goats is consistent with their higher DMI and ADG, suggesting that increased feeder use may have supported their greater nutrient intake requirements.

Numerically, Spanish goats made 16% fewer feeder visits and spent less time (3 min/d) at the feed bunk than Alpine goats. Similar behavioral patterns have been reported in cattle, in which more feed-efficient animals tended to visit the feeder less frequently and spend less time feeding (Parsons et al. 2020). However, because breed effects on most feeding behavior traits were not significant in the present study and breed differences in RFI were only suggestive, these observations should be interpreted cautiously. Nevertheless, the tendency for lower feed intake and more favorable RFI values in Spanish goats may indicate differences in feeding strategy or feed utilization between breeds.

Previous studies have suggested that feeding duration traits are more closely associated with DMI than visit-frequency traits (McGee et al. 2014). In the present study, Alpine goats spent more time at the feeder and consumed more feed than Spanish goats, suggesting that feeding duration may better reflect intake behavior than feeder visit frequency alone. Alpine goats also exhibited greater day-to-day variation in feeding activity and feed intake. Parsons et al. (2020) suggested that variation in feeding behavior may be associated with differences in animal activity patterns and feeding strategies, thereby contributing to individual variation in performance and feed utilization.

The ratio of HD duration to BV duration averaged approximately 46% in Alpine goats and 45% in Spanish goats, indicating that both breeds devoted nearly half of their time at the feeder to active feeding. Consequently, a substantial portion of bunk occupancy likely involved activities other than feed consumption, such as feed investigation, feed sorting, bunk licking, or short pauses between feeding bouts. Although the overall HD: BV duration ratio was similar across breeds, a significant breed × TD interaction indicated that the proportion of time spent actively feeding differed over the course of the study for Alpine and Spanish goats. This finding suggests that the temporal patterns of feeder utilization varied between breeds despite similar overall feeding intensity.

Few studies have simultaneously evaluated feeding behavior and feed efficiency in dairy- and meat-type goats using automated feeding systems. Therefore, the present study provides additional insight into the relationships among feeding behavior, feed intake, growth performance, and feed efficiency in Alpine and Spanish goats under controlled conditions using GrowSafe technology. While breed-related differences in feeding behavior were generally limited, the observed variation in feeder utilization patterns highlights the complexity of feeding behavior and its potential influence on nutrient intake and animal performance. These findings contribute to a better understanding of behavioral factors associated with feed utilization and demonstrate the usefulness of GrowSafe technology for characterizing feeding behavior in goats.

Effects of test duration on feeding behavior

Goats showed significant changes in both BV frequency and duration across test periods, with the highest visit frequency observed during the early phase (days 0–56), followed by a gradual decline (Fig. 6). Similarly, goats spent the longest time at the feed bunk during the early phase of the test (days 42 and 56) and the shortest time at the feed bunk toward the end of the test period (day 84). Time spent eating varies widely among animals and depends on feed quality and feeding conditions, particularly stocking rate (Gipson et al. 2006; Brand et al. 2017). Notably, Gipson et al. (2006) reported that increasing the number of Boer crossbred wethers (from 6, 8, 10 to 12 goats/feeder) assigned to an automated feeding unit influenced feeder utilization patterns due to greater competition for feeder access. In the present study, however, goats were offered a TMR diet on feed bunks at a stocking rate of 5 goats per automated feeding unit, which is lower than the stocking rate used by Gipson et al. (2006). Therefore, the steady decline in BV and duration after 56 days during the testing period may reflect goats’ adaptation to the diet and to feeding bunks over time, rather than competition for the bunk occupancy. Studies have shown that ruminants modify feeding patterns through learning and increased familiarity with dietary and management conditions (Schwartzkopf-Genswein et al. 2002; Forbes 2007). DeVries et al. (2003) demonstrated changes in feeding patterns as Holstein dairy cows adapted to feeding management and feed availability.

Figure 6.

Mean BV frequency (visits/d) of goats (n = 19) across test durations of 42, 56, 70, and 84 d. Error bars represent SEM. Means without a common superscript letter differ significantly among test durations (P < 0.001). Data represent pooled observations from 19 goats (n = 19). Error bars represent SEM. Test duration significantly affected bunk visit (BV) frequency (P < 0.001), with the highest frequency at 42 d, followed by 56 and 70 d, and the lowest frequency at 84 d.

Despite fewer bunk visits and shorter feeding durations later in the trial, the eating rate increased over time, suggesting that goats consumed feed more rapidly as they became accustomed to the feeding environment. Gibb et al. (1998) suggested that the limited effect of bunk visit frequency on intake may result from reduced competition when feed is continuously available in the bunks. In the present study, goats had ad libitum access to feed, and at least 3 kg of feed was consistently maintained in the bunks to minimize empty periods, which may explain the weak association between bunk visit behavior and feed intake. The observed variation in feeding duration across the testing period indicates shifts in feeding behavior as goats progressed through the study, suggesting that goats’ feeding patterns may change with age and experience. Similar behavioral adjustments have been reported in ruminants as they become familiar with the diet, feeder design, and social environment (Nielsen 1999; Gipson et al. 2006; Forbes 2007; González et al. 2008). These findings indicate that feeding behavior is dynamic and influenced by age, experience, adaptation, and other factors.

Although Goetsch et al. (2010) reviewed numerous factors influencing goat feeding behavior, including breed, diet, production stage, and management, and Gipson et al. (2006) studied feeding behavior of Spanish and Boer crossed goats using an automated feeding system, few studies have simultaneously evaluated feeding behavior and feed efficiency in dairy and meat goat breeds using automated feeding systems. The present study provides novel insights into temporal changes in feeding behavior and feed efficiency traits in Alpine and Spanish goats, monitored using GrowSafe technology. These findings enhance our understanding of behavioral factors associated with feed utilization and may support selection strategies to improve feed efficiency in goats.

Conclusion

This study demonstrated the successful application of GrowSafe feeding technology to evaluate feed efficiency and feeding behavior traits in goats. Although most phenotypic traits did not differ significantly between breeds, Spanish goats tended to be more feed-efficient than Alpine goats, as indicated by lower feed intake and negative RFI values. Feeding behavior also changed with growth stage: goats exhibited more frequent feeder visits and lower eating rates early in development, followed by increased eating rates and reduced feeding time as they grew. These findings provide valuable insights into growth performance, feed efficiency, and feeding behavior in goats; however, further research involving additional breeds, larger populations, and diverse production systems is needed to confirm their broader applicability. In addition, GrowSafe technology, originally developed for larger livestock species, was successfully adapted for goats. This successful adaptation underscores the system’s potential as a practical tool for precision goat research and management, improving efficiency and productivity.

Glossary

Abbreviations

ADF

acid detergent fiber

ADG

average daily gain

BW

body weight

CP

crude protein

CV

coefficient of variance

DM

dry matter

DMI

dry matter intake

EE

ether extract

OM

organic matter

MBW

metabolic body weight

RFI

residual feed intake

TMR

total mixed ration

SD

standard deviation

TD

test duration

Artificial intelligence disclosure

The author(s) state that neither artificial intelligence nor artificial intelligence–assisted tools were used during the preparation of this work.

Funding

This work was supported by the U.S. Department of Agriculture, National Institute of Food and Agriculture (USDA-NIFA), through the Evans-Allen 1890 Research Formula Program under Section 1445. The contents are solely the responsibility of the authors and do not necessarily represent the official views of USDA-NIFA.

Conflicts of interest

None declared.