Experiment 1 included 2 treatments, a control group in which calves always received unaltered milk (“control”; n = 14) and an experimental group in which calves received various concentrations of a bitter substance in milk (“bitter”; n = 12 calves). Treatments were assigned with pseudo-randomization; the first treatment assignment was randomly assigned by coin flip, and the remaining calves were assigned to treatments in an alternating pattern. Treatments were balanced for sex and breed (Table 1).

On study days, calves were fed different concentrations of the bitter additive (Quinine hydrochloride dihydrate, MilliporeSigma, Darmstadt, Germany) in milk during the afternoon meal (1500 h). The calves were fed on a kilogram basis and offered 4.0 kg milk replacer (3.8 L). Control calves received 0.0 g/L of quinine each study day. Bitter calves were given 0.0 g/L, 0.2 g/L, 0.3 g/L, and 0.4 g/L quinine in milk on different study days. Calves in the bitter treatment received all the quinine concentrations of different days of the study using a Latin Square design. Because calf enrollment in the study was continuous, quinine concentration was blocked by day (i.e., all calves received the same quinine concentration on a single testing day). To ensure all calves received all quinine concentrations in a random order, the two weeks of study were blocked in 0.0 and 0.2 g/L quinine within a single week (alternating Monday and Thursday) and 0.3 and 0.4 g/L quinine within a single week (alternating Monday and Thursday). Thus, there were 8 possible random orders that calves received quinine concentration including 1) 0.0, 0.2, 0.3, 0.4 g/L, 2) 0.2, 0.0, 0.3, 0.4 g/L, 3) 0.0, 0.2, 0.4, 0.3 g/L, 4) 0.2, 0.0, 0.4, 0.3 g/L, 5) 0.3, 0.4, 0.0, 0.2 g/L, 6) 0.3, 0.4, 0.2, 0.0 g/L, 7) 0.4, 0.3, 0.2, 0.0 g/L, and 8) 0.4, 0.3, 0.0, 0.2 g/L.

To ensure an even distribution of quinine in milk, the quinine powder was reconstituted in 50 mL water until fully dissolved and then added to the bucket containing milk. Another 50 mL of water was added to the quinine vial and shaken, then added to the milk replacer to ensure any quinine residue was added to the milk. The reconstituted quinine-water and milk replacer were thoroughly mixed using a handheld whisk. Control milk was supplemented with an equal volume of water (100 mL) to reach an equivalent weight to bitter milk. Milk was weighed after the water and quinine-water solution had been added to the control and bitter milk, respectively.

After Experiment 1, calves advanced to Experiment 2 of the study if they met the inclusion criteria. Experiment 2 included control (n = 12) and bitter (n = 11) treatments (Table 1). Calves were offered 4.0 kg of unaltered (“control”) or bitter milk (“bitter”; 0.35 g/L quinine) during the test period. This test period was defined as the second meal of a day, where the time between meals was manipulated: 6, 12, and 16 h following their morning meal. Calves experienced all three milk-withhold times regardless of treatment. The concentration of quinine was set based on Experiment 1 results, where calves reduced intake at 0.4 g/L compared to the other concentrations. For Experiment 2, this was reduced to 0.35 g/L quinine due to the pronounced decline of milk consumed observed at 0.4 g/L quinine in milk. Quinine and water were added to milk using the same protocol as Experiment 1. Treatments were balanced for sex (Table 1), and to control for treatment order from Experiment 1 to 2 (control-control = 5, control-bitter = 5, bitter-bitter = 5, bitter-control = 6); 4 and 2 calves were enrolled in only Experiment 1 or 2, respectively, due to not meeting the inclusion criteria. These times were selected based on the common feed intervals seen in the United States (USDA, 2016). This experiment was performed using a Latin-Square design; all calves received all milk withhold times in a random order. Because study enrollment was continuous, feed withholding time was blocked to individual days (i.e., all calves received the same feed withholding time on the same day). There were 6 orders that calves received different feed withholding times, including: 1) 6, 12, 16 h, 2) 6, 16, 12 h, 3) 12, 6, 16 h, 4) 12, 16, 6 h, 5) 16, 6, 12 h, and 6) 16, 12, and 6 h.

All data were recorded in Microsoft Excel (Microsoft Corp., Redmond, WA). Behavioral output was exported from Boris software (Friard and Gamba, 2016) in Microsoft Excel for analysis. Observations for each calf were summed into single values for each behavior type per calf including the total amount of time spent drinking (s) and the number of bouts of drinking, aversive behaviors, and non-drinking oral behaviors. Statistical analysis was performed in SAS (Version 9.4, Cary, NC). Data were visually inspected for distribution in SAS (PROC UNIVARIATE) before statistical analysis by each experiment to determine the appropriate model distributions to be utilized. For all models, univariable analysis (linear regression; PROC GLIMMIX) was performed between the outcomes and predictor variables. Fixed effects were included in multivariable models if P ≤ 0.20. Manual backward stepwise elimination was then used to remove fixed effects with the greatest P-value from the multivariable models; fixed effects were retained in the final model if P ≤ 0.20. Residuals were visually assessed with quantile-quantile and histogram plots for all models and models with the lowest Akaike Information Criterion (AIC) were selected. Significance and tendencies were declared at P < 0.05 and P < 0.10, respectively.

In Experiment 1, calves consumed less milk when the greatest concentration of quinine was present compared to all other concentrations (0.4 g/L vs. ≤ 0.3 g/L). Lower milk consumption was not observed when milk was altered with 0.2 g/L or 0.3 g/L quinine compared to milk without quinine. Calves also consumed more calf starter when the highest concentration of quinine was provided, compared to other lower concentrations, corresponding to the decreased milk consumption also observed at the highest concentration of quinine. An increase in calf starter intake was likely an effort to compensate for the decreased energy and nutrients typically received from milk when refused by calves. The depression of milk intake at the greatest quinine concentration is consistent with previous literature that evaluated cattle responses to bitter additives. In general, domestic ruminants are capable of tasting bitter flavors, though they may be less sensitive to bitter additives than other mammals (Ginane et al., 2011). When taste reception was evaluated in calves, quinine sulphate (bitter flavor) produced a relatively small electrophysiological response when passed over the tongue of anesthetized calves in comparison to other substances, including sucrose (sweet flavor), sodium chloride (salty flavor), acetic acid (sour flavor), and others (Bernard, 1964). Calf reactions to quinine hydrochloride in water at different concentrations showed that calves were first indifferent to quinine hydrochloride at lower concentrations (0.63 and 1.25 mg/100 ml) and then transitioned immediately to rejection at higher concentrations (2.5 mg/100 ml) (Goatcher and Church, 1970). Notably, all quinine concentrations utilized in Experiment 1 were greater than those in Goatcher and Church (1970). Therefore, this study builds upon the findings that show cattle have low sensitivity to bitter flavors but still find them unpalatable.

Calves started and stopped drinking milk more frequently (drinking bouts) and performed more aversive behaviors (nostril or muzzle lick, head bob or swing, lip smack, and attempt to drink milk replacer) at all quinine concentrations compared to unaltered milk, which may suggest that calves were able to taste quinine at all concentrations. Similarly, calves in another study also had more fragmented feeding behaviors (more pauses per suckling bout) when they were provided an acidified versus an unacidified milk replacer (Todd et al., 2018). Despite reduced intake of milk only occurring at the highest quinine concentration, the increased feeding bouts and performance of aversive behaviors at all quinine concentrations may indicate that calves can taste quinine and that it was aversive at all concentrations.

The expression of non-drinking oral behaviors only increased at a quinine concentration of 0.4 g/L compared to all other concentrations. Non-drinking oral behaviors in this study consisted of attempts to consume feed from a source other than the altered milk, including the consumption of calf starter between drinking bouts or attempting to drink from their neighbors’ milk bucket, and sucking or chewing on another calf or non-nutritive object. This increase in non-drinking oral behaviors occurred concurrently with depression in milk intake at quinine concentrations of 0.4 g/L. An increase in non-drinking oral behaviors may be an indicator of frustration because calves were unwilling to consume bitter milk but were still motivated to eat. Frustration in response to restricted access to expected feed has been documented in other livestock species. For example, when hens were trained to receive a feed reward in response to pressing a key, an increase in pecking of an alternative substrate (feathers) and repeated pecking of the trigger for the reward were observed when the feed reward was withheld (Rodenburg et al., 2002). In this study, the non-drinking oral behaviors may show frustration from palatability restriction from milk rather than physical restriction from milk. It is also possible that calves were hungry but sought an alternate feed source because they chose not to drink from the bitter milk available to them. Further research is needed to link these non-drinking oral behaviors with frustration from restriction of an expected feed reward and/or hunger in pre-weaned calves.

The results from Experiment 1 provided a basis that quinine has the potential to be used as a method to test calf motivation to consume milk through taste aversion. Application of a taste aversion test would be useful in field research where options for behavioral assessment are limited and to increase the number of experimental units enrolled in a study, which is frequently limited by the human effort required to review animal behavior for research. However, the low response of calves to quinine at smaller concentrations with an increase to high aversion at 0.4 g/L may present a challenge to the applicability of this test. Future research is encouraged test calf preference between multiple quinine concentrations in feed or the preference between different aversive flavor profiles.

As in Experiment 1, calves consumed less milk when a bitter additive was present than when the milk was unaltered. However, calves did not consume significantly different amounts of bitter or unaltered milk at different lengths of time between meals. This suggests that calves were not more motivated to drink bitter milk offered after milk was withheld for up to 16 h. However, a lack of motivation to consume milk would be contrary to previous literature indicating that calves are highly motivated to consume milk when it is restricted (Morita et al., 1999; de Passillé and Rushen, 2006; Rosenberger et al., 2017). Given the novelty of using taste aversion to assess calf motivation, these results should be interpreted cautiously. For instance, individual variation in willingness to consume a novel feed may have influenced the outcomes. There was a large amount of individual variation in the willingness to consume bitter milk after milk was withheld; calf intake of bitter milk after a 12 h inter-meal interval ranged from 0.1% to 99.7% of the offered milk. Previous work in dairy calves has shown that personality is related to calf response stressors such as diarrhea, dehorning, and weaning (Woodrum Setser et al., 2024a), as well as mixing into new groups (Occhiuto et al., 2023). For example, calves with a low score for the trait ‘Active’ had more unrewarded visits to an automatic milk feeder during weaning and, thus, were considered hungrier (De Paula Vieira et al., 2008; Woodrum Setser et al., 2024c). Some individual calves in the current study may have been more sensitive to hunger (consuming all the bitter milk) or have been more sensitive to a bitter additive (refused all of it). The large amount of variation in the amount of bitter milk consumed may also suggest that a larger sample size would be required to detect a difference in motivation using this specific motivation test design, or increasing the intensity of the bitter additive may decrease the amount of variation observed.

It is also possible that calf hunger was affected by the extra 1.9 L of milk fed in the morning on test days before application of the taste aversion motivation test. Additional milk was provided to ensure adequate milk was provided if milk was not consumed during test feedings, as an ethical consideration. Access to a resource outside of the test period has been shown to influence the elasticity of demand for the resource within a motivation test. For instance, one study explored rat motivation for water after water was manipulated before the motivation test period; rats that received water 1 h, 2 h, or 4 h before the test were less willing to push a weighted lever to receive water than rats who did not receive water before the motivation test (Ladewig et al., 2002). While the additional milk provided in this study occurs with the morning meal which was well before motivation test period following feed withholding, it is unknown how long after receiving additional milk in the morning feeding calves experience increased satiety. There is some evidence that at 6 h calves were less hungry, as there is greater variation in consumption of the unaltered milk than is observed after feed was withheld for 12 and 16 h; however, this difference was numerical and not statistically significant. Further research should evaluate length of satiety based on meal size in calves.

Animals show high motivation for food across species, and food is consistently classified as a resource with inelastic demand, meaning animals will proportionately increase their effort to achieve the resource even as the cost to obtain it increases (Dawkins, 1990; Jensen and Pedersen, 2008). Kirkden et al. (2003) suggest that measuring other behaviors during a motivation test, including time and effort to obtain a resource, can give a more wholistic view of animal motivation rather than simply measuring the amount of the resource consumed. Indeed, in studies evaluating feeding behaviors in calves and cows, the assessment of other behaviors during the test, rather than just feed intake was essential to understanding individual differences in willingness to consume a novel feed (Costa et al., 2020; Woodrum Setser et al., 2024b). Therefore, it is important to consider other measures, such as calf behaviors during the test, to gain a holistic picture of calf experience when milk is withheld.

The first additional behavior to consider is calf starter intake. Calves increased their intake of calf starter when milk was withheld for longer period; calves consumed more after 12 h than at 6, and tended to consume more at 16 h than after 6h. Increased solid feed consumption in response to restricted milk allowance is a well-accepted indicator of hunger (Khan et al., 2011). For example, calves consume more concentrate when smaller milk volumes are provided before weaning (4 L vs 12 L; Borderas et al., 2009; Steele et al., 2017). Therefore, greater calf starter intake during longer inter-meal intervals may have been observed in response to calf hunger. It is also possible that calf starter intake may have influenced the degree of hunger experienced during the motivation test, as calves always had access to solid feed. However, the increase in calf starter intake was seen regardless of treatment. This suggests that calves were likely hungrier when fed 12 and 16 h after their morning meal than 6 h, especially because calves fed at the 6 h interval were also fed their normal afternoon meal. Further, calves fed bitter milk tended to consume more calf starter than calves fed unaltered milk. Calves in the bitter treatment group were likely hungrier than those in the control group because they consumed less milk, thus compensating for the limited feed intake. Withholding calf starter on test days could have prohibited calves from compensating for a lack of feed and elucidated more treatment differences. However, we felt as though removing all feed sources from calves would be an ethical concern due to negative affective states caused by hunger. Calf starter intake also increased with calf age, where older calves consumed more calf starter than younger calves, which is consistent with expectations (Neave et al., 2019). Since both feed withhold time and calf age were associated with calf starter intake, one should use caution generalizing these finding across different calf ages, where younger calves with less developed rumens may not be consuming calf starter at significant levels yet (Eckert et al., 2015). However, within this study calves were the same age at the start of Experiment 2 and the ages at which calves received different feed withholding times were balanced by the Latin Square design.

Milk withhold time and treatment also impacted the behaviors expressed during the test period. Regardless of treatment, calves increased the number of drinking bouts when milk was withheld for 12 h compared to 6 h. An interaction was also seen between treatment and feed withhold time; calves fed unaltered milk started and stopped drinking less frequently after milk was withheld for 6 h compared to 12 and 16 h and had more drinking bouts at 12 than 16 h. However, this relationship was not observed in calves fed bitter milk. Additionally, we did not observe that milk treatment affected the number of drinking bouts. These results are contrary to our findings in Experiment 1, which found calves performed more drinking bouts when a bitter additive was in their milk, and previous research, which also saw an increase in drinking bouts with acidified milk (Todd et al., 2018). This difference in results is likely due to the combined effect with withholding feed. Because this increase in drinking bouts is only observed within the control group, it can be assumed that this increase in drinking bouts is associated with the amount of time milk is withheld. A study found calves fed 5% of their body weight twice daily performed more suckling bouts than when fed 10% of their body weight (Jongman et al., 2020); further supporting the relationship between hunger and a greater number of drinking bouts at meal time. Notably, drinking bouts were greater after 12 h than after 16 h regardless of treatment, perhaps indicating there may be a ceiling effect from hunger effect on the number of drinking bouts.

The amount of time milk was withheld was associated with the number of non-drinking oral behavior bouts. Calves performed more non-drinking oral behaviors when feed was withheld for 12 and 16 h compared to 6 h. Previous research reported that non-nutritive sucking, a behavior grouped into non-drinking oral behaviors in this study, also increased with calf hunger (Herskin et al., 2010). As previously discussed, the non-drinking oral behaviors around feeding may indicate frustration, as seen in chickens (Rodenburg et al., 2002). Therefore, the increased non-drinking oral behaviors suggest that calves were hungrier and more frustrated when feed was withheld for longer periods. It is important to note that ‘longer’ periods include the 12 h interval, the time between meals on farms that feed twice daily. Most farms in the U.S. (94.6%) feed calves twice daily (USDA, 2016). Comparative hunger behaviors at 12 and 16 h without feed may suggest that calves should be fed more frequently than the current standard practice. We also observed that the dairy-beef crossbreeds expressed more non-drinking oral behaviors than Holstein calves, which may indicate that they were more susceptible to hunger than their purebred counterparts. This may be because dairy-beef crossbreeds have been shown to have greater growth rates and thus may have greater energy demands (Vestergaard et al., 2019; Berry, 2021).

As in Experiment 1, calves fed bitter milk expressed a greater number of aversive behaviors than calves fed unaltered milk. This further supports the idea that these behaviors are an aversive reaction to a bitter additive, reminiscent of gaping and lip smacking observed in rodents, primates, and sheep (Grill and Norgren, 1978; Steiner et al., 2001; Henslee et al., 2019). Data from this study is also the first to show that dairy calves perform behaviors which indicate feed aversion in response to a bitter flavor. However, milk withhold time was also associated with the number of aversive behaviors. Regardless of treatment, calves expressed fewer aversive behaviors after milk was withheld for 16 h compared to both 6 h and 12 h. This may be a result of increased hunger reducing the expression of these behaviors, but more research is needed to understand the relationship between these behaviors and hunger.

To the authors knowledge, this is the first test that utilized taste aversion to measure motivation to access feed in calves intended for field research. This test was intended to replace more labor-intensive behavioral observations that limit research on commercial farms. However, this study highlights the importance of additional behavioral observations to gain an accurate picture of calf hunger as the test is currently designed. The following are considerations and recommendations for future application of taste aversion to measure calf motivation for milk and areas where this motivation test may need further refinement for field application.

The concentration of the bitter additive or flavor profile utilized in this taste aversion motivation test may require further refinement. The quinine concentration used in Experiment 2 (0.35 g/L) may have been too high a stimulus to accurately measure changes in motivation to consume milk after it was withheld. This concentration was utilized due to the divergent consumption of milk at higher (0.4 g/L) vs. lower (≤ 0.3 g/L) quinine concentrations in Experiment 1. As with other behavioral assessments in animals, such as judgment bias, the sensitivity of animals to a reward or punishment needs to be carefully considered, as an incorrect stimulus may lead to false negative results (Kremer et al., 2022). Within motivation tests, animals are appraising the cost required versus their desire to obtain a resource (Jensen and Pedersen, 2008). In Experiment 2, the quinine concentration may have set the ‘cost’ too high to distinguish a change in motivation to consume milk after a 6-hour interval between meals compared to 16 hours. We suggest that this methodology should be applied to other flavors to test the efficacy of taste aversive as a model of calf motivation to drink milk. Acidic flavors may be worthwhile to explore, as cattle are able to taste these flavors and find them aversive (Ginane et al., 2011; Todd et al., 2018). Future applications of this taste aversion motivation test should consider calf sensitivity to the aversive flavor (i.e., bitter or acidic flavors) utilized and subsequently the concentration utilized.