Milking of Brown Swiss and Austrian Simmental cows in a herringbone parlour or an automatic milking unit

Optimal Parameters to Milk Murciano-Granadina Goats in Mid and Low-Line Milking Parlours

Recent short-term studies on Murciano-Granadina goats have established that the optimal parameters to set up the milking machines are different according to the milk pipes height. Two groups of 52 fresh goats each were employed in 2 different experiments to confirm during an entire lactation period the best combinations of system vacuum pulsation rate and pulsator ratio in low-line and mid-line milking parlours. The experiment performed in the low-line milking parlour included one group milked with 40 KPa vacuum system level, 90 puls/min pulsation rate and 60/40 pulsation and a second group milked with 38 KPa vacuum system level, 90 puls/min pulsation rate and 60/40 pulsation ratio. The experiment carried out in mid-line included one group milked with 40 KPa vacuum system level, 90 puls/min pulsation rate and 60/40 pulsation ratio and a second group milked with 40 KPa vacuum system level, 120 puls/min pulsation rate and 60/40 pulsation ratio. Variables studied included milking efficiency, milk composition, cortisol, SCC and intramammary infections, teat-end oedema after milking and vacuum dynamics during milking. Considering the results of an entire lactation period, it was confirmed that when milking in mid-line, the combination of 40 KPa system vacuum, 90 cycles/min pulsation rate and 60/40 pulsator ratio showed optimal results of the above-mentioned variables. On the other hand, the use of 40 KPa in a low-line system increased the milk cortisol values (0.34 ± 0.1 vs. 0.44 ± 0.1 ng/mL) without any other advantage. Thus, the recommendation is to use a combination of 38 KPa system vacuum, 90 cycles/min pulsation rate and 60/40 pulsator ratio to enhance animal welfare.

Effect of Vacuum Level and Pulsation Parameters on Milking Efficiency and Animal Welfare of Murciano-Granadina Goats Milked in Mid-Line and Low-Line Milking Machines

Simple Summary

In recent years, several studies have been carried out to optimize milking efficiency in Murciano-Granadina goats, but to our knowledge there are no experiments combining different milking parameters (system vacuum, pulsation rate and pulsator ratio) in two different height level milking machines. In two short experiments, testing different combinations of the parameters mentioned, milk fractioning, milking efficiency, teat end status and animal welfare were analyzed. Results showed that the height of the milking machine pipes can have an influence on which parameters are most optimal for milking efficiency and animal welfare in the milking of Murciano-Granadina goats.

Abstract

The Murciano-Granadina goat breed has been described as a slow milking breed. As milking machine parameters can affect milk extraction in terms of yield and time employed, two experiments of one-month duration were performed with 88 goats in Latin square design to find the best combination of these parameters. One of them was carried out in a mid-line milking machine and one in a low-line milking machine. For each of them, two vacuum levels (36 and 40 kPa), two pulsation rates (90 and 120 cycles/min) and two pulsator ratios (50 and 60%) were used and milking efficiency, sanitary status of the mammary gland, milk cortisol, and teat end status were evaluated. Results showed that in milking machines installed in mid- and low-line, the use of 40 kPa system vacuum, 60% pulsator ratio and 90 or 120 cycles/min pulsation rate achieved optimum milking fractioning and efficiency. In the case of low-level milking machines, a similar combination with 36 kPa not only showed worse milking fractioning values, but also provided better values of teat end status and cortisol level.

Effects of small milking stalls on stress responses in dairy cows during milking in group milking parlors

Milking stall dimensions have not been adapted to the increase in cow body size caused by selection for better milking performance over the past decades. Improper milking stall dimensions might limit cow comfort, could lead to stress responses during milking and thus could negatively affect cow welfare. A crossover study was conducted in an experimental milking parlor that was converted from a herringbone (HB) to a side-by-side (SBS) parlor. The milking stall dimensions were modified in length and width and for HB also in depth (perpendicular distance between rump rail and breast rail). The stall dimensions applied during the experiments ranged from much smaller than common in European dairy farming to much larger. Treatments were applied for 2 wk per milking parlor type. In each milking parlor type, a total of 30 cows, kept in 2 groups were observed during milking for behavioral and physiological stress responses and for milking performance. In addition, milk cortisol levels and somatic cell counts were measured at the end of the 2-wk period. Outcome variables were selected based on a principal component analysis and analyzed using mixed effects models reflecting the experimental design. The results showed that the first cow per milking batch required more time (on average >40 s) to enter very small HB stalls than to enter small, large and very large stalls (<30 s). Also, cows yielded more milk per milking in very small and very large HB stalls than in the small and large milking stall dimensions [very small: 15.8 kg; 95% confidence interval (CI), 14.2-17.4 kg); small: 14.3 kg (95% CI, 12.8-15.9 kg); large: 14.6 kg (95% CI, 13.1-16.1 kg); very large: 16.1 kg (95% CI, 14.6-17.6 kg)]. The other behavioral, physiological and milk flow parameters as well as udder health were not affected by stall dimensions. For the SBS parlor, effects of milking stall dimensions were not detectable in any of the parameters. Despite the strong avoidance behavior to enter the milking parlor (measured as latency), no acute stress responses were found during milking. However, the study cannot exclude long-term effects of narrow stall dimensions on stress levels and possibly udder health, which should be investigated in future studies.

Effects of photoperiod and light intensity on milk production and milk composition of dairy cows in automatic milking system

The purpose of this study was to determine the effects of photoperiod and light intensity on milk production, milk composition, hormones levels and blood metabolites indices of Korean Holstein dairy cows in automatic milking system (AMS). A total of 24 Holstein dairy cows were selected and used to four subsequent treatments for the experimental periods of 60 days. The light programs consisted of (1) Control: the natural photoperiod with 14.2 h of the light period and 9.4 h of the dark period (below 10 Lux); (2) T1: 16 h of the long day photoperiod (LDPP) with 50 Lux of light; (3) T2: 16 h of LDPP with 100 Lux of light; and (4) T3: 16 h of LDPP with 200 Lux of light, respectively. Importantly, there was a significant difference in the thurl activity of dairy cows between the different light intensity programs (p < 0.05). Milk yield was higher in T1 and T2 (40.80 ± 1.71 and 39.90 ± 2.02 kg/d, respectively) than those of Control and T3 (32.18 ± 1.51 and 35.76 ± 2.80 kg/d, respectively) (p < 0.05), but DMI was lower in T1, T2, and T3 compared to Control (p < 0.05). Also, milk fat percentage, the contents of milk fat and total solids were higher in T2 than those in the others (p < 0.05). The average daily melatonin level in milk was high to T3 (28.20 ± 0.43 pg/mL), T2 (24.62 ± 0.32 pg/mL), T1 (19.78 ± 0.35 pg/mL), and Control (19.36 ± 0.45 pg/mL) in order (p < 0.05). Also, the cortisol levels in milk and blood were lower in treatment groups than in Control (p < 0.05). The results of this study showed that it will be effective to improve the milk yield and milk composition, and to reduce the stress of dairy cows when the light conditions regulate to extend the photoperiod to 16 h at a light emitting diode (LED) intensity of 100 Lux under the AMS in dairy farm.

Detection of cow hind-leg activity during milking by using a 3-dimensional accelerometer attached to the milking cluster

•

Hind-leg activity during milking in cows is a good welfare indicator

•

Cow hind-leg movements are transmitted to the milking claw

•

3D accelerometers attached to the milking claws accurately record cow hind-leg movements

Hind-leg activity in dairy cows during milking is considered an indicator of a stressful situation or discomfort. Automatic detection of cow hind-leg activity during milking could be used to monitor deficiencies at the cow, milking machine, or human level. We assumed that the milking cluster follows cow hind-leg movements during milking. Therefore, the aim of our study was to estimate the ability of a 3-dimensional (3D) accelerometer attached to the claw's hook of the milking cluster to indirectly measure cow hind-leg activity during milking. The behavior of 45 dairy cows during 1 morning and 1 evening milking in a 2 × 3 auto tandem milking parlor was monitored by using direct visual observations and a 3D accelerometer attached directly to the hook of each milking claw. We recorded when hind-leg foot-lifting or stepping (undifferentiated) and kicking behaviors occurred. Detected activities were obtained after applying a standard deviation filter (over a sliding window of 1 s) on the x (left/right lateral) acceleration axis only. We set the detection threshold at 0.13 g (i.e., 1.27 m·s⁻²), based on the visual evaluation of 10 “reference” graphs (from 5 randomly chosen cows). Periods of activity were created for each observed or detected activity (and for consecutive activities that were less than 3 s apart) to take into account the milking cluster's rocking motion resulting from a single movement. Observed and detected periods of activity were considered matching when they occurred simultaneously (with a 2-s delay acceptance in the visual observations). In total, 472 periods of hind-leg activity were observed. The detection of hind-leg activity by using a 3D accelerometer attached to the milking cluster was found to have 69% sensitivity, 99% specificity, 75% positive predictive value, 98% negative predictive value, and an overall accuracy of 97%. On average, cows were observed being active 0.94 times per minute and being detected active 0.86 times per minute. The observed and detected periods of activity were found to be correlated (r = 0.657). Therefore, the use of a 3D accelerometer attached to the milking cluster seems promising for reliably estimating cow hind-leg activity during milking. Nevertheless, further investigations are necessary to clarify whether other factors can cause variations in milking cluster movements, regardless of cow hind-leg activity, and to generalize our method to better identify deficiencies in cow–machine–human interactions.

Minimally Invasive Markers of Stress and Production Parameters in Dairy Cows before and after the Installation of a Voluntary Milking System

Automatic milking systems (AMS) are a low-labour alternative to conventional parlours, with previous studies demonstrating that cows vary in their ability to cope with the change to AMS. Cortisol expression can be combined with other measures to assess stress: saliva and hair have the advantage of requiring minimally invasive sampling. No work has investigated the long-term impact of introduction of AMS. The aims of the study were to assess short-term and chronic stress associated with a change in milking system by measuring salivary and hair cortisol levels and to assess the impact on health and production parameters. Cows from one farm changing their milking system were recruited to the study and sampled for saliva (n = 10) and hair (n = 12) before and after installation. Cortisol levels were measured using a salivary cortisol enzyme immunoassay kit. Body condition, lameness and milk parameters of the whole herd were regularly assessed. Salivary cortisol showed no diurnal pattern but was affected by lameness and gestation. Non-lame cows showed a reduction in salivary cortisol after AMS introduction (p < 0.001). Hair cortisol levels increased after AMS, but it was unclear if this change was seasonal. Milk yield increased by 13% and somatic cell count reduced by 28%. Body condition score was consistently good, but lameness remained high throughout the study. Production values alone do not represent high welfare. The high lameness and associated cortisol levels suggest that cow stress requires consideration when changing milking systems.

Transitioning from conventional to automatic milking: Effects on the human-animal relationship

There are several differences in how an automatic milking system (AMS; milking equipment not requiring human intervention for the milk harvesting process) and a conventional milking system (CMS) are managed, where the effect of milking system type on the human-animal relationship has remained unexplored. A survey and observations from 5 Australian dairy farms transitioning from CMS to AMS were taken twice, 1 yr apart, before and after transition to an AMS. The farmers completed a survey and had all farmer-cow interactions documented for 3 d. In addition, a random selection of lactating cows had their avoidance distance (the distance at which they move away from an approaching person) recorded and were involved in a handling test during both visits. The survey findings indicated that basic management practices remained mostly unchanged, whereas records of farmer-cow interactions showed 4 out of 5 farms had less interaction time after AMS transition. This was caused by a reduction in milk harvesting tasks, where a small portion of this time was re-invested into time that farmers spent around the cows without directly interacting with them and into tasks involving close cow contact. Overall, an approximate 27% decline was observed in avoidance distances of cows from an AMS compared with the CMS. A handling test was performed on 4 of the 5 farms before and after AMS transition, where the farmers were asked to move a selection of cows through a gate one at a time. In the AMS more vocal effort was required to move the cows, and the cows responded with a reduced occurrence of running past the farmer and reduced occurrence of slipping in an attempt to avoid the farmers compared with the CMS. Overall, results show that farmers spent less time interacting with cows in the AMS, and that cows were less fearful around people as seen by reduced avoidance distances and reduced stress responses to close handling.

Robotic milking: Technology, farm design, and effects on work flow

Robotic milking reduces labor demands on dairy farms of all sizes and offers a more flexible lifestyle for farm families milking up to 250 cows. Because milking is voluntary, barn layouts that encourage low-stress access by providing adequate open space near the milking stations and escape routes for waiting cows improve milking frequency and reduce fetching. Because lame cows attend less often, preventing lameness with comfortable stalls, clean alley floors, and effective foot bathing warrants special emphasis in robotic dairies. Variable milking intervals create challenges for foot bathing, sorting and handling, and dealing with special-needs cows. Appropriate cow routing and separation options at the milking stations are needed to address these challenges and ensure that the expected labor savings are realized. Protocols and layout and gating should make it possible for a herd worker to complete all handling tasks alone. Free traffic and guided traffic systems yield similar results when excellent management is applied or when the number of cows is well below capacity. In less ideal circumstances, guided traffic and the use of commitment pens result in longer standing times and stress, particularly for lower ranking cows, and poor management with free traffic results in more labor for fetching.

Step behaviour and autonomic nervous system activity in multiparous dairy cows during milking in a herringbone milking system

Behavioural and cardiac responses of multiparous dairy cows (n=24) during milking in a 2×4 stall herringbone milking system were evaluated in this study. Heart rate (HR), parasympathetic tone index (high frequency component, HF) of heart rate variability and sympathovagal balance indicator LF/HF ratio (the ratio of the low frequency (LF) and the HF component) were analysed. Measurement periods were established as follows: (1) standing calm (baseline), (2) udder preparation, (3) milking, (4) waiting after milking in the milking stall and (5) in the night (2 h after milking). Step behaviour was recorded and calculated per minute for the three phases of the milking process (udder preparation, milking and waiting after milking). HR was higher during udder preparation and milking compared with baseline (P=0.03, 0.027, respectively). HF was significantly lower than baseline levels during waiting in the milking stall after milking (P=0.009), however, during udder preparation, milking and 2 h after milking did not differ from baseline (P>0.05, in either case). LF/HF during the three phases of the milking process differed neither from baseline levels nor from each other. Steps occurred more often during waiting after milking than during udder preparation (P=0.042) or during milking (23; P=0.017). Our results suggest that the milking procedure itself was not stressful for these animals. After milking (following the removal of the last teat cup and before leaving the milking stall), both decreased parasympathetic tone (lower HF) and increased stepping rate indicated a sensitive period for animals during this phase.

Within-day variation and effect of acute stress on plasma and milk cortisol in lactating goats

An experiment was carried out to study the effect of acute stress on cortisol concentration in goat milk (milk cortisol; MC) and its suitability for stress assessment. Additional variables studied were plasma cortisol (PC), blood glucose (BG), and somatic cell count (SCC). Ten goats free of mastitis at 2 mo of lactation were divided into 2 groups. One group of animals was exposed to acute stress (visual and auditory, but not physical, contact with a barking dog for 3 min) immediately before the 1000 h sampling (STR), and the other group was considered as the control group, not exposed to acute stress (CON). The animals were gradually accustomed to the sampling schedule for 1 wk before the stress treatment (pre-experimental period). The variables PC, MC, BG, SCC, and milk composition (fat, lactose, whey protein, true protein, and casein) were recorded at 0600, 1000, 1130, 1400, 1800, 2200, and 0200 h. Milk variables were recorded in both glands separately. The right gland (n=10) was completely milked and the left gland (n=10) was milked to 30 mL at every sampling except at 1000 h, when complete milking was carried out on both glands. Variables were analyzed by using a linear mixed model. Both PC and MC concentrations varied throughout the day, with values being highest in the morning. Higher PC levels were recorded in the STR group immediately after treatment compared with CON (36.9 vs 16.3 ng/mL). For MC, differences between treatments were noted 1.5h after STR treatment (1.27 vs 0.25 ng/mL). Blood glucose increased immediately after treatment in STR animals, being different from CON until 1.5h after treatment sampling (101 vs. 58 mg/dL). The SCC was not increased by STR treatment, showing higher values in the morning and afternoon compared with evening and night (5.6±0.13 to 5.87±0.13 log10 SCC). The STR treatment had no significant effect on milk composition, and overall milk yield of a complete day after stress stimulus of both groups was similar (STR: 2,067±280 g/d; CON: 2,134±427 g/d). We conclude that plasma and milk cortisol concentrations in control dairy goats varied throughout the day; MC showed significant differences in the STR group if samples were taken 1.5h after the stressing stimulus, whereas PC showed significant differences in samples taken immediately after the stressing stimulus. Additionally, MC sampling has the advantage of being noninvasive compared with blood sampling.

Cardiac responses to palpation per rectum in lactating and nonlactating dairy cows

Interest in the monitoring of heart rate variability (HRV) has increased recently, as it gives more detailed and immediate information about the level of stress than traditional behavioral or hypothalamus-pituitary-adrenal measures. In this study, we evaluated heart rate (HR) and parasympathetic HRV parameters to monitor cardiac stress responses to palpation per rectum (PPR) in lactating (LACT; n = 11) and nonlactating (NLACT; n = 12) dairy cows. Heart rate and HRV were recorded from 40 min before PPR until 120 min after it was completed. Heart rate, the root mean square of successive differences (RMSSD), and the high-frequency component (HF) of HRV were analyzed by examining 5-min time windows. To compare cardiac responses to PPR between groups, changes in HR and HRV parameters were calculated as area under the curve (AUC) for LACT and NLACT cows. An immediate increase in HR was detected during PPR in both LACT (+21.4 ± 2.4 beats/min) and NLACT cows (+20.6 ± 2.3 beats/min); however, no differences were found between groups on the basis of parameters of AUC. The increase in HR in both groups along with a parallel decrease in RMSSD (LACT cows: -5.2 ± 0.4 ms; NLACT cows: -5.1 ± 0.4 ms) and HF [LACT cows: -10.1 ± 0.8 nu (where nu = normalized units); NLACT cows: -16.9 ± 1.2 nu] during PPR indicate an increase in the sympathetic, and a decrease in the parasympathetic tone of the autonomic nervous system. The increase in RMSSD (LACT cows: +7.3 ± 0.7 ms; NL cows: +17.8 ± 2.2 ms) and in HF (LACT cows: +24.3 ± 2.6 nu; NLACT cows: +32.7 ± 3.5 nu) immediately after PPR indicated a rapid increase in parasympathetic activity, which decreased under the baseline values 10 min following PPR. The amplitude and the maximum RMSSD and HF values were greater in NLACT cows than in LACT animals, suggesting a higher short-term cardiac responsiveness of NLACT cows. However, the magnitude and the duration of the stress response were greater in LACT cows, as indicated by the analysis of AUC parameters (area under the HRV response curve and time to return to baseline). Cow response to the PPR was more prominent in parasympathetic HRV measures than in HR. Based on our results, the effect of PPR on the cows' cardiac stress responses may have an impact on animal welfare on dairy farms, and investigating the effect of lactation on the cardiac stress reactions could prove useful in modeling bovine stress sensitivity. Further research is needed to find out whether the differences due to lactation are physiological or management related.

Welfare implication of measuring heart rate and heart rate variability in dairy cattle: literature review and conclusions for future research

Heart rate (HR) measurements have been used to determine stress in livestock species since the beginning of the 1970s. However, according to the latest studies in veterinary and behaviour-physiological sciences, heart rate variability (HRV) proved to be more precise for studying the activity of the autonomic nervous system. In dairy cattle, HR and HRV indices have been used to detect stress caused by routine management practices, pain or milking. This review provides the significance of HR and HRV measurements in dairy cattle by summarising current knowledge and research results in this area. First, the biological background and the interrelation of the autonomic regulation of cardiovascular function, stress, HR and HRV are discussed. Equipment and methodological approaches developed to measure interbeat intervals and estimate HRV in dairy cattle are described. The methods of HRV analysis in time, frequency and non-linear domains are also explained in detail emphasising their physiological background. Finally, the most important scientific results and potential possibilities for future research are presented.

Measuring zoo animal welfare: theory and practice

The assessment of animal welfare relates to investigations of how animals try to cope with their environment, and how easy or how difficult it is for them to do so. The use of rigorous scientific methods to assess this has grown over the past few decades, and so our understanding of the needs of animals has improved during this time. Much of the work in the field of animal welfare has been conducted on farm animals, but it is important to consider how the methods and approaches used in assessing farm animal welfare have been, and can be, adapted and applied to the measurement of welfare in animals in other domains, such as in zoos. This is beneficial to our understanding of both the theoretical knowledge, and the practicability of methods. In this article, some of the commonly-used methods for measuring animal welfare will be discussed, as well as some practical considerations in assessing the welfare of zoo animals.

Time budgets and adrenocortical activity of cows milked in a robot or a milking parlour: interrelationships and influence of social rank

Adrenocortical activity and time budgets in a robotic milking system in two variants of cow traffic (partially forced, free) were compared with a conventional milking system, focusing on the relationships between adrenocortical activity, time budgets, and social rank of the cows. Both groups were housed in identical conditions and consisted of 30 cows each. Within each experiment, direct observations of social behaviour and 24-hour video recordings were conducted during six two-day blocks. We calculated the dominance value and the time budget (‘lying’, ‘feeding’, ‘other activity’) of each cow. Faeces from each cow were collected to determine the concentration of cortisol metabolites (CCM) as an indicator of baseline adrenocortical activity and possible chronic stress. Analysis of CCM and time budgets did not indicate any adverse effects of the robotic milking system, with one exception: under partially-forced cow traffic, cows of low social rank had longer waiting times in front of the robot. Considering the fact that the number of cows milked by the robot was less than half of that recommended by the manufacturers, our results do not exclude the possibility that a higher stocking rate may lead to further adverse effects.

Comparison of Functional Aspects in Two Automatic Milking Systems and Auto-Tandem Milking Parlors

Milk yield, milking frequency, intermilking interval, teat-cup attachment success rate, and length of the milking procedure are important functional aspects of automatic milking systems (AMS). In this study, these variables were compared for 2 different models of AMS (AMS-1, with free cow traffic, and AMS-2, with selectively guided cow traffic) and auto-tandem milking parlors (ATM) on 4 farms each. Data on milking-stall visits and milkings of 20 cows were recorded on 3 successive days by means of video observations. Data were evaluated with mixed-effects models. Milk yield did not differ among the 3 milking systems. Milking frequency in the AMS was 2.47/d [95% confidence interval (CI) = (2.38, 2.56)], and was significantly higher than the 2 milkings/d in ATM. Milking frequency was lower for cows with a higher number of days in milk (DIM) in AMS-1 [change of -0.057/10 DIM, CI = (-0.070, -0.044)], but remained constant for cows with varying DIM in AMS-2 [change of -0.003/10 DIM, CI = (-0.034, 0.027)]. As a consequence, milking frequency was higher in early lactation [by 0.603, CI = (0.102, 1.103)] and lower in late lactation in AMS-1 than in AMS-2 [by -0.397, CI = (-0.785, -0.008)]. The intermilking interval showed the opposite pattern. Teat-cup attachment was more successful in AMS-1 than in AMS-2 (98.4 vs. 94.3% of the milkings), with some variation among farms (range: AMS-1 96.2 to 99.5%; AMS-2 91.5 to 96.1%). The length of the entire milking process did not differ among the milking systems [454 s, CI = (430, 478)], although the preparation phase was longer [changes in comparison with ATM: in AMS-1 by a factor of 2.90, CI = (2.30, 3.65), and in AMS-2 by 5.15, CI = (4.09, 6.48)] and the actual milking phase was shorter in both AMS-1 and AMS-2 than in ATM [changes in comparison with ATM: in AMS-1 by a factor of 0.76, CI = (0.62, 0.94), and in AMS-2 by 0.75, CI = (0.60, 0.93)]. The admission [changes in comparison with ATM: in AMS-1 by a factor of 2.56, CI = (1.55, 4.22), and in AMS-2 by 3.07, CI = (1.86, 5.08)] and preparation phases lasted longer in AMS-2 than in AMS-1, whereas the time required by the cows to leave the milking stall did not differ among the systems [changes in comparison with ATM: in AMS-1 by a factor of 0.89, CI = (0.55, 1.44), and in AMS-2 by 1.02, CI = (0.63, 1.66)]. In conclusion, different technical approaches to automatic milking led to differences in teat-cup attachment success rates, in the duration of several phases of the milking process, and in milking frequency. The capacity of an AMS could be further improved by shortening the preparation phase and reducing the proportion of failed milkings.

Milk Cortisol Concentration in Automatic Milking Systems Compared with Auto-Tandem Milking Parlors

Milk cortisol concentration was determined under routine management conditions on 4 farms with an auto-tandem milking parlor and 8 farms with 1 of 2 automatic milking systems (AMS). One of the AMS was a partially forced (AMSp) system, and the other was a free cow traffic (AMSf) system. Milk samples were collected for all the cows on a given farm (20 to 54 cows) for at least 1 d. Behavioral observations were made during the milking process for a subset of 16 to 20 cows per farm. Milk cortisol concentration was evaluated by milking system, time of day, behavior during milking, daily milk yield, and somatic cell count using linear mixed-effects models. Milk cortisol did not differ between systems (AMSp: 1.15 +/- 0.07; AMSf: 1.02 +/- 0.12; auto-tandem parlor: 1.01 +/- 0.16 nmol/L). Cortisol concentrations were lower in evening than in morning milkings (1.01 +/- 0.12 vs. 1.24 +/- 0.13 nmol/L). The daily periodicity of cortisol concentration was characterized by an early morning peak and a late afternoon elevation in AMSp. A bimodal pattern was not evident in AMSf. Finally, milk cortisol decreased by a factor of 0.915 in milking parlors, by 0.998 in AMSp, and increased by a factor of 1.161 in AMSf for each unit of ln(somatic cell count/1,000). We conclude that milking cows in milking parlors or AMS does not result in relevant stress differences as measured by milk cortisol concentrations. The biological relevance of the difference regarding the daily periodicity of milk cortisol concentrations observed between the AMSp and AMSf needs further investigation.