Effect of pre- and postpartum supplementation of a pure glycerol product to dairy cows on feeding behavior, lying behavior, and reticulorumen pH.

The objective of this study was to quantify the effects of supplementing a low level of dry glycerol product pre- and postpartum on the feeding behavior, lying behavior, and reticulorumen pH of dairy cows. Multiparous Holstein dairy cows (n = 60) were enrolled in a 2 × 2 factorial design study. Twenty-one days before expected parturition, cows individually received a dry cow diet with (1) 250 g/d glycerol supplementation (GLY; 66% pure glycerol, United States Pharmacopeia grade), or (2) no supplementation (CON). Following parturition, cows were individually assigned to either (1) 250 g/d glycerol product (GLY; 66% pure glycerol), or (2) no supplementation (CON) to their partial mixed ration (PMR) for the first 21 d in milk (DIM). All cows were milked by an automated milking system and offered a target of 5.4 kg/d pellet (23% of target total dry matter intake [DMI]). For both treatment periods, cows were individually assigned to automated feed bins to measure PMR feeding behavior. Rumination time and lying behavior were monitored with electronic sensors for the whole study (-21 to 21 DIM). Reticulorumen pH boluses were administered to a subset of cows (n = 40) where pH was recorded every 10 min from 21 d prepartum to 21 d postpartum. Prepartum, cows fed GLY had fewer, larger meals and spent 20.2% more time feeding than CON while consuming feed at a similar rate. Cows on the CON diet prepartum spent more time lying down in more frequent bouts in the 21 d before calving. Following parturition, cows that received GLY prepartum continued to devote more time to eating, while tending to spend less time ruminating per kilogram of DMI. Cows receiving CON postpartum had larger meals with longer intervals between meals. In the first 21 DIM, cows receiving CON prepartum tended to have shorter, but significantly more frequent, lying bouts than cows fed GLY prepartum. Glycerol supplementation pre- and postpartum resulted in less time spent lying down following parturition. Minimal differences between treatments were observed for pre- and postpartum sorting behavior or reticulorumen pH. Overall, supplementation of glycerol pre- and postpartum altered cow time budgets, with cows spending more time eating pre- and postpartum, less time lying pre- and postpartum, and having fewer, larger meals prepartum when receiving glycerol prepartum, and with cows having slower feeding rates and smaller meals following parturition with postpartum glycerol supplementation.

Farm-level risk factors associated with increased milk ß-hydroxybutyrate and hyperketolactia prevalence on farms with automated milking systems.

The objectives of this study were to determine the farm-level hyperketolactia (HKL) prevalence, as diagnosed from milk ß-hydroxybutyrate (BHB) concentration, on dairy farms milking with an automatic milking system (AMS) and to describe the farm-level housing, management, and nutritional risk factors associated with increased farm-average milk BHB and the within-herd HKL prevalence in the first 45 DIM. Canadian AMS farms (n = 162; eastern Canada n = 8, Quebec n = 23, Ontario n = 75, western Canada n = 55) were visited once between April to September 2019 to record housing and herd management practices. The first test milk data for each cow under 45 DIM were collected, along with the final test of the previous lactations for all multiparous cows, from April 1, 2019 to September 30, 2020. The first test milk BHB was then used to classify each individual cow as having HKL (milk BHB ≥ 0.15 mmol/L) at the time of testing. Milk fat and protein content, milk BHB, and HKL prevalence were summarized by farm and lactation group (all, primiparous, and multiparous). During this same time period, formulated diets for dry and lactating cows, including ingredients and nutrient composition, and AMS milking data were collected. Data from the AMS were used to determine milking behaviors and milk production of each herd during the first 45 DIM. Multivariable regression models were used to associate herd-level housing, feeding management practices, and formulated nutrient composition with first test milk BHB concentrations and within-herd HKL levels separately for primiparous and multiparous cows. The within-herd HKL prevalence for all cows was 21.8%, with primiparous cows having a lower mean prevalence (12.2 ± 9.2%) than multiparous cows (26.6 ± 11.3%). Milk BHB concentration (0.095 ± 0.018 mmol/L) and HKL prevalence for primiparous cows were positively associated with formulated prepartum DMI and forage content of the dry cow diet while being negatively associated with formulated postpartum DMI, the major ingredient in the concentrate supplemented through the AMS, and the PMR-to-AMS concentrate ratio. However, multiparous cows' milk BHB concentration (0.12 ± 0.023 mmol/L) and HKL prevalence were positively associated with the length of the previous lactation, milk BHB at dry off, prepartum diet nonfiber carbohydrate content, and the major forage fed on farm, while tending to be negatively associated with feed bunk space during lactation. This is the first study to determine the farm-level risk factors associated with herd-level prevalence of HKL in AMS dairy herds, thus helping optimize management and guide diet formulation to promote the reduction of HKL prevalence.

Milking system and diet forage type effect on milk quality of Italian Holstein-Friesian.

Moving from conventional (CMS) to automatic (AMS) milking systems could affect milk quality. Moreover, the type and preservation methods of the forages used in the TMR, such as alfalfa hay (HTMR) or corn silage (STMR) have been demonstrated to modify milk composition. Thus, this study investigated the effect of implementing AMS and different diet forage types on the quality of Italian Holstein-Friesian bulk milk. Milk samples (n = 168) were collected monthly from 21 commercial farms in northern Italy during a period of 8 mo. Farms were categorized into 4 groups according to their milking system (CMS vs. AMS) and diet forage type (HTMR vs. STMR). Milk quality data were analyzed through the mixed procedure for repeated measurement of SAS with the milking system, diet forage type, and sampling day as fixed effects. Milking through the AMS led to lower milk fat, freezing point, and ß-LG A; longer coagulation time; and higher K content, pH, and ß-LG B than CMS. Cows fed STMR produced milk with greater fat, protein, casein, Mg content, titratable acidity, and ß-LG A, but with reduced curd firming time, freezing point, and ß-LG B than those fed HTMR. In conclusion, milk quality is not only altered by the diet's forage type and characteristics but also by the milking system.

Evaluating the performance of herd-specific Long Short-Term Memory models to identify automated health alerts associated with a ketosis diagnosis in early lactation cows.

The growing use of automated systems in the dairy industry generates a vast amount of cow-level data daily, creating opportunities for using these data to support real-time decision-making. Currently, various commercial systems offer built-in alert algorithms to identify cows requiring attention. To our knowledge, no work has been done to compare the use of models accounting for herd-level variability on their predictive ability against automated systems. Long Short-Term Memory (LSTM) models are machine learning models capable of learning temporal patterns and making predictions based on time series data. The objective of our study was to evaluate the ability of LSTM models to identify a health alert associated with a ketosis diagnosis (HAK) using deviations of daily milk yield, milk FPR, number of successful milkings, rumination time, and activity index from the herd median by parity and DIM, considering various time series lengths and numbers of d before HAK. Additionally, we aimed to use Explainable Artificial Intelligence method to understand the relationships between input variables and model outputs. Data on daily milk yield, milk fat-to-protein ratio (FPR), number of successful milkings, rumination time, activity, and health events during 0 to 21 d in milk (DIM) were retrospectively obtained from a commercial Holstein dairy farm in northern Indiana from February 2020 to January 2023. A total of 1,743 cows were included in the analysis (non-HAK = 1,550; HAK = 193). Variables were transformed based on deviations from the herd median by parity and DIM. Six LSTM models were developed to identify HAK 1, 2, and 3 d before farm diagnosis using historic cow-level data with varying time series lengths. Model performance was assessed using repeated stratified 10-fold cross-validation for 20 repeats. The Shapley additive explanations framework (SHAP) was used for model explanation. Model accuracy was 83, 74, and 70%, balanced error rate was 17 to 18, 26 to 28, and 34%, sensitivity was 81 to 83, 71 to 74, and 62%, specificity was 83, 74, and 71%, positive predictive value was 38, 25 to 27, and 21%, negative predictive value was 97 to 98, 95 to 96, and 94%, and area under the curve was 0.89 to 0.90, 0.80 to 0.81, and 0.72 for models identifying HAK 1, 2, and 3 d before diagnosis, respectively. Performance declined as the time interval between identification and farm diagnosis increased, and extending the time series length did not improve model performance. Model explanation revealed that cows with lower milk yield, number of successful milkings, rumination time, and activity, and higher milk FPR compared with herdmates of the same parity and DIM were more likely to be classified as HAK. Our results demonstrate the potential of LSTM models in identifying HAK using deviations of daily milk production variables, rumination time, and activity index from the herd median by parity and DIM. Future studies are needed to evaluate the performance of health alerts using LSTM models controlling for herd-specific metrics against commercial built-in algorithms in multiple farms and for other disorders.

Multivariable time series classification for clinical mastitis detection and prediction in automated milking systems.

In this study, we developed a machine learning framework to detect clinical mastitis (CM) at the current milking (i.e., the same milking) and predict CM at the next milking (i.e., one milking before CM occurrence) at the quarter level. Time series quarter-level milking data were extracted from an automated milking system (AMS). For both CM detection and prediction, the best classification performance was obtained from the decision tree-based ensemble models. Moreover, applying models on a data set containing data from the current milking and past 9 milkings before the current milking showed the best accuracy for detecting CM; modeling with a data set containing data from the current milking and past 7 milkings before the current milking yielded the best results for predicting CM. The models combined with oversampling methods resulted in specificity of 95 and 93% for CM detection and prediction, respectively, with the same sensitivity (82%) for both scenarios; when lowering specificity to 80 to 83%, undersampling techniques facilitated models to increase sensitivity to 95%. We propose a feasible machine learning framework to identify CM in a timely manner using imbalanced data from an AMS, which could provide useful information for farmers to manage the negative effects of CM.

Genetic, farm, and lactation effects on behavior and performance of US Holsteins in automated milking systems.

Selecting for favorable behavior and performance could enhance the efficiency of production in automated milking systems (AMS). The objectives of this study were to describe AMS behavior and performance in Holsteins, estimate genetic parameters among AMS traits, and determine genetic relationships of AMS traits with other routinely recorded traits. The edited data included 1,101,651 individual milking records and 394,636 daily records from 2,531 lactations and 1,714 cows that resided on 3 farms; data were obtained from the Dairy Data Warehouse (Assen, Netherlands) cloud. Traits considered were individual milking and daily totals for milk yield, milking time, milk harvest rate (the ratio of milk yield to milking time), milk flow rate, electrical conductivity, machine kickoffs, incomplete milkings, and blood in milk; the number of milkings per day and 305-d mature-equivalent milk yield (305ME) were also evaluated. Individual milkings were evaluated with mixed models that included fixed effects of week of lactation, lactation group (1, 2, ≥3), hour of day, and farm; random effects included cow within lactation, lactation group by week of lactation, and interactions of farm with date, hour, week of lactation, and year-season of calving. Daily records were evaluated with 3-trait animal models that included 305ME and 2 AMS traits with random additive genetic and permanent environment effects. Estimated breeding values were extracted and correlated with yield, conformation, and udder health genetic evaluations. Farm specific robot access policies had notable effects on week of lactation patterns for milk yield and number of milkings. Mature cows had higher milk harvest rates (2.05 kg/min) than first-lactation cows (1.73 kg/min) with larger differences in early lactation. First-lactation cows were more likely to kick off the machine (15.04%) than mature cows (8.62%), particularly in early lactation. Heritability estimates were generally lower for behavior traits (0.03 for incomplete milkings and 0.08 for kickoffs) than for milk harvest rate (0.30) and flow rate (0.55). Udder conformation traits did not have favorable genetic correlations with AMS traits, with the exception that longer teats were correlated with fewer kickoffs (-0.34) and incomplete milkings (-0.49); increased milk harvest rate and flow rate were unfavorably associated with genetic merit for udder health. There is genetic variation for milking efficiency and behavioral traits, suggesting genetic selection to enhance efficiency in AMS systems is possible. Genetic associations with udder conformation indicate that selection for udder morphology is unlikely to substantially improve milking efficiency. This calls for more direct selection of traits related to AMS efficiency.

Robotic milking: Feeding strategies and economic returns.

Cows in herds equipped with conventional milking parlors follow a structured, consistent, and social milking and feeding routine. Furthermore, in most cases cows in conventional herds receive all their nutrients from a total mixed ration, whereas in herds equipped with robotic or automatic milking systems (AMS) a fraction of their nutrients is provided during milking, mainly as a means to attract cows to the milking system. In this regards, AMS present both a challenge and an opportunity for feeding cows. The main challenge resides in maintaining a minimum and relatively constant milking frequency in AMS. However, milking frequency is dependent on many factors, including the social structure of the herd, the farm layout design, the type of traffic imposed to cows, the type of flooring, the health status of the cow (especially lameness, but also mastitis, metritis, among others), the stage of lactation, the parity, and the type of ration fed at the feed bunk and the concentrate offered in the AMS. Uneven milk frequency has been associated with milk losses and increased risk of mastitis, but most importantly it is a lost opportunity for milking the cow and generating profit. On the other hand, the opportunity from AMS resides in the possibility of milking more frequently and feeding cows more precisely or more closely to their nutrient needs on an individual basis, potentially resulting in a more profitable production system. But, feeding cows in the parlor or AMS has many challenges. On one side, feeding starchy, highly palatable ingredients in large amounts may upset rumen fermentation or alter feeding behavior after milking, whereas feeding high-fiber concentrates may compromise total energy intake and limit milking performance. Nevertheless, AMS (and some milking parlors, especially rotary ones) offer the possibility of feeding the cows to their estimated individual nutrient needs by combining different feeds on real time with the aim of maximizing profits rather than milk yield. This approach requires that not only the amount of feed offered to each cow but also the composition of the feed vary according to the different nutrient needs of the cows. This review discusses the opportunities and pitfalls of milking and feeding cows in an AMS and summarizes different feeding strategies to maximize profits by managing the nutrition of the cows individually.

Producer estimates of prevalence and perceived importance of lameness in dairy herds with tiestalls, freestalls, and automated milking systems.

Lameness is one of the most important welfare and productivity concerns in the dairy industry. Our objectives were to obtain producers' estimates of its prevalence and their perceptions of lameness, and to investigate how producers monitor lameness in tiestall (TS), freestall with milking parlor (FS), and automated milking system (AMS) herds. Forty focal cows per farm in 237 Canadian dairy herds were scored for lameness by trained researchers. On the same day, the producers completed a questionnaire. Mean herd-level prevalence of lameness estimated by producers was 9.0% (±0.9%; ±SE), whereas the researchers observed a mean prevalence of 22.2% (±0.9%). Correlation between producer- and researcher-estimated lameness prevalence was low (r = 0.19) and mean researcher prevalence was 1.6, 1.8, and 4.1 times higher in AMS, FS, and TS farms, respectively. A total of 48% of producers thought lameness was a moderate or major problem in their herds (TS = 34%; AMS =53%; FS = 59%). One third of producers considered lameness the highest ranked health problem they were trying to control, whereas two-thirds of producers (TS = 43%; AMS = 63%; FS = 71%) stated that they had made management changes to deal with lameness in the past 2 yr. Almost all producers (98%) stated they routinely check cows to identify new cases of lameness; however, 40% of producers did not keep records of lameness (AMS = 24%; FS = 23%; TS = 60%). A majority (69%) of producers treated lame cows themselves immediately after detection, whereas 13% relied on hoof-trimmer or veterinarians to plan treatment. Producers are aware of lameness as an issue in dairy herds and almost all monitor lameness as part of their daily routine. However, producers underestimate lameness prevalence, which highlights that lameness detection continues to be difficult in in all housing systems, especially in TS herds. Training to improve detection, record keeping, identification of farm-specific risk factors, and treatment planning for lame cows is likely to help decrease lameness prevalence.

Lying times of lactating cows on dairy farms with automatic milking systems and the relation to lameness, leg lesions, and body condition score.

Lying down and resting are important for optimal cow health, welfare, and production. In comparison with free stall farms with a milking parlor, farms with automated milking systems (AMS) may place less constraint on how long cows can lie down. However, few studies report lying times on AMS farms. The aims of this study were to describe the variation in lying times of dairy cows in AMS farms and to understand how much of the variation in individual lying times is related to cow-level factors, including lameness, the presence of hock and knee lesions, and body condition score (BCS). We visited 36 farms in Canada (Quebec: n = 10; Ontario: n = 10; British Columbia: n = 4; and Alberta: n = 5), and the United States (Michigan: n = 7). Gait scores, presence of hock and knee lesions, and BCS were recorded for 40 Holstein cows from each herd. Parity and days in milk were retrieved from farm records. Lying time was recorded across 4d using accelerometers (n = 1,377). Multivariable analysis was performed. Of scored cows, 15.1% were lame (i.e., obviously limping; 203 of 1,348 cows). Knee lesions were found in 27.1% (340 of 1,256 cows) and hock lesions were found in 30.8% (421 of 1,366 cows) of the animals. Daily lying time varied among cows. Cows spent a median duration of 11.4 h/d lying down (25th-75th percentile = 9.7-12.9 h), with a lying bout frequency of 9.5 bouts/d (25th-75th percentile = 7.5-12 bouts/d) and a median bout duration of 71 min (25th-75th percentile = 58-87 min/bout). Lameness was associated with cows lying down for 0.6 h/d longer in fewer, longer bouts. Increased lying time was also associated with increased parity, later stage of lactation and higher BCS. Older cows (parity ≥ 3) spent about 0.5 h/d more lying down compared with parity 1 cows, and cows with BCS ≥ 3.5 lay down on average 1 h/d longer than cows with BCS ≤ 2.25. Hock lesions were associated with shorter lying times in univariable models, but no associations were found in the multivariable models. We concluded that only a small proportion of the variation between cows in lying time is explained by lameness, leg lesions, and BCS.

Invited review: Changes in the dairy industry affecting dairy cattle health and welfare.

The dairy industry in the developed world has undergone profound changes over recent decades. In this paper, we present an overview of some of the most important recent changes in the dairy industry that affect health and welfare of dairy cows, as well as the science associated with these changes. Additionally, knowledge gaps are identified where research is needed to guide the dairy industry through changes that are occurring now or that we expect will occur in the future. The number of farms has decreased considerably, whereas herd size has increased. As a result, an increasing number of dairy farms depend on hired (nonfamily) labor. Regular professional communication and establishment of farm-specific protocols are essential to minimize human errors and ensure consistency of practices. Average milk production per cow has increased, partly because of improvements in nutrition and management but also because of genetic selection for milk production. Adoption of new technologies (e.g., automated calf feeders, cow activity monitors, and automated milking systems) is accelerating. However, utilization of the data and action lists that these systems generate for health and welfare of livestock is still largely unrealized, and more training of dairy farmers, their employees, and their advisors is necessary. Concurrently, to remain competitive and to preserve their social license to operate, farmers are increasingly required to adopt increased standards for food safety and biosecurity, become less reliant on the use of antimicrobials and hormones, and provide assurances regarding animal welfare. Partly because of increasing herd size but also in response to animal welfare regulations in some countries, the proportion of dairy herds housed in tiestalls has decreased considerably. Although in some countries access to pasture is regulated, in countries that traditionally practiced seasonal grazing, fewer farmers let their dairy cows graze in the summer. The proportion of organic dairy farms has increased globally and, given the pressure to decrease the use of antimicrobials and hormones, conventional farms may be able to learn from well-managed organic farms. The possibilities of using milk for disease diagnostics and monitoring are considerable, and dairy herd improvement associations will continue to expand the number of tests offered to diagnose diseases and pregnancy. Genetic and genomic selection for increased resistance to disease offers substantial potential but requires collection of additional phenotypic data. There is every expectation that changes in the dairy industry will be further accentuated and additional novel technologies and different management practices will be adopted in the future.

The costs of chronic mastitis: A simulation study of an automatic milking system farm.

Mastitis is a production disease in dairy farming that causes economic losses. Especially chronic mastitis (i.e., mastitis cases continuing longer than 28 days) can substantially affect the risk of transmission of intramammary infections (IMI) and total milk production losses. Insights into the impact of chronic mastitis on production and farm economics are needed to guide chronic mastitis decision-making. We aimed to estimate the costs of chronic mastitis with a Monte Carlo simulation model in which the costs of chronic mastitis were estimated as part of the total mastitis costs. The model simulated milk yields, IMI dynamics, somatic cell count (SCC), and pregnancy status on an average Dutch dairy farm with 100 cow places over 9 years. The model was parameterized using information from the literature and actual sensor data from automatic milking system (AMS) farms. The daily subclinical milk production losses were modeled using a generalized additive model and sensor data. Transmission of IMI was modeled as well. The model results indicated median total costs of mastitis of € 230 per generic IMI case (i.e., a weighted average of all pathogens). The most substantial cost factors were the extra mastitis cases due to transmission, culling, and milk production losses. Other significant costs originated from dry cow treatments and diverted milk. The model also indicated median total costs due to chronic mastitis of € 118 (51 % of the total mastitis costs). The share of chronic mastitis relative to the total mastitis costs was substantial. Transmission of contagious bacteria had the largest share among the chronic mastitis costs (51 % of the costs of chronic cases). The large share of chronic mastitis costs in the total mastitis costs indicates the economic importance of these mastitis cases. The results of the study point to the need for future research to focus on chronic mastitis and reducing its presence on the AMS dairy farm.

The Combined Use of Automated Milking System and Sensor Data to Improve Detection of Mild Lameness in Dairy Cattle.

This study aimed to develop a tool to detect mildly lame cows by combining already existing data from sensors, AMSs, and routinely recorded animal and farm data. For this purpose, ten dairy farms were visited every 30-42 days from January 2020 to May 2021. Locomotion scores (LCS, from one for nonlame to five for severely lame) and body condition scores (BCS) were assessed at each visit, resulting in a total of 594 recorded animals. A questionnaire about farm management and husbandry was completed for the inclusion of potential risk factors. A lameness incidence risk (LCS ≥ 2) was calculated and varied widely between farms with a range from 27.07 to 65.52%. Moreover, the impact of lameness on the derived sensor parameters was inspected and showed no significant impact of lameness on total rumination time. Behavioral patterns for eating, low activity, and medium activity differed significantly in lame cows compared to nonlame cows. Finally, random forest models for lameness detection were fit by including different combinations of influencing variables. The results of these models were compared according to accuracy, sensitivity, and specificity. The best performing model achieved an accuracy of 0.75 with a sensitivity of 0.72 and specificity of 0.78. These approaches with routinely available data and sensor data can deliver promising results for early lameness detection in dairy cattle. While experimental automated lameness detection systems have achieved improved predictive results, the benefit of this presented approach is that it uses results from existing, routinely recorded, and therefore widely available data.

Investigating off-grid systems for a mobile automated milking facility.

This study examined the potential of renewable energy sources to provide the necessary power for a mobile off-grid automated milking system (AMS) and associated facilities on pasture. This involved choosing the most cost-effective, environmentally friendly, and sustainable power supply for a mobile AMS in Sweden operating May-September and milking 20 cows per day. Weather data, input from the milking system manufacturer (DeLaval), and outputs from two mathematical models, Insight Maker and HOMER, were used to investigate the potential of different renewable energy sources (biodiesel-, ethanol-, or biogas-run generators, solar photovoltaic (PV) panels + batteries) to support the mobile system. Solar-based energy best fulfilled the key requirements of being environmentally friendly, cost-effective, and sustainable. It also gave the lowest net present cost (11,804 USD), levelized cost of energy (0.31 USD), and annual operating costs (178.26 USD) of all renewable energy options considered. Thus use of solar PV panels + batteries is recommended for the mobile AMS facility. Ways of addressing possible challenges that could arise during implementation, uncertainties in input parameters, and limitations in scaling-up and replicating the proposed set-up are discussed.

Milk losses linked to mastitis treatments at dairy farms with automatic milking systems.

Mastitis-associated milk losses in dairy cows have a massive impact on farm profitability and sustainability. In this study, we analyzed milk losses from 4 553 treated mastitis cases as recorded via treatment registers at 41 AMS dairy farms. Milk losses were estimated based on the difference between the expected and the actual production. To estimate the unperturbed lactation curve, we applied an iterative procedure using the Wood model and a variance-dependent threshold on the milk yield residuals. We calculated milk losses both in a fixed window around the first treatment day of each mastitis case and in the perturbations corresponding to this day, at the cow level as well as at the quarter level. In a fixed time window of day -5 to 30 around the first treatment, the absolute median milk losses per case were 101.5 kg, highly dependent on the parity and the lactation stage with absolute milk losses being highest in multiparous cows and at peak lactation. Relative milk losses expressed in percentage were highest on the first treatment day, and full recovery was often not reached within 30 days from treatment onset. In 62 % of the cases, we found a perturbation in milk yield at the cow level at the time of treatment. On average, perturbations started 8.7 days before the first treatment and median absolute milk losses increased to 128 kg of milk per perturbation. Mastitis is not expected to have equal effects on the four quarters so this study additionally investigated losses in the individual udder quarters. We used a data-based method leveraging milk yield and electrical conductivity to project the presumably inflamed quarter. Next, we compared losses with the average of presumably non-inflamed quarters. Median absolute losses in a fixed 36-day window around treatment varied between 50.2 kg for front and 59.3 kg for hind inflamed quarters compared to respectively 24.7 and 26.3 kg for the median losses in the non-inflamed quarters. Also here, these losses differed between lactation stages and parities. Expressed proportionally to expected yield, the relative median milk losses in inflamed quarters on the treatment day were 20 % higher in inflamed quarters with a higher variability and slower recovery. In 86 % of the treated mastitis cases, at least one perturbation was found at the quarter level. This analysis confirms the high impact of mastitis on milk production, and the large variation between quarter losses illustrates the potential of quarter analysis for on-farm monitoring at farms with an automated milking system.