Improving parlor efficiency in block calving pasture-based dairy systems through the application of a fixed milking time determined by daily milk yield and milking frequency

Adjusting end-of-milking criteria, in particular applying a maximum milking time determined by expected milk yield at an individual milking session, is one strategy to optimize parlor efficiency. However, this strategy can be difficult to apply practically on farm due to large differences in session milk yield, driven by milking interval, which affects milking routines and can be limited by in-parlor technology. The objective of this study was to test the hypothesis that a single fixed milking time (duration) could be applied at all milking sessions without compromising milk production or udder health for a range of milking intervals. To test the hypothesis, 4 experimental herds were established: (1) herd milked twice a day (TAD) using a 10- and 14-h interval, (2) herd milked TAD using an 8- and 16-h interval, (3) herd milked 3 times in 2 d using a 10–19–19-h interval, and (4) herd milked once a day (OAD). Herds consisted of 40 cows each, and were established for two 6-wk experimental periods, one in peak lactation and the other in mid-late lactation. Within each herd, half the cows had an end-of-milking criterion of 0.35 kg/min (Flow), and the other half had milking ended after a fixed period of time (FixedT) based on the average milking session yield, the daily milk yield divided by average number of milkings per day, irrespective of milking interval. We found no differences in daily milk yield between end-of-milking criteria due to residual milk from one milking likely increasing the proportion of milk in the udder cistern at the next milking session for the FixedT treatment. However, fat yield was compromised when the percentage of the herd with a truncated milking exceeded an estimated 33% at a milking session, which occurred in the TAD 8–16 herd due to the divergence from the average milking interval (in the case of TAD, 12–12 h). Applying a fixed milking time had no detrimental effects on udder health, except in the OAD herd in mid-late lactation, which had both a higher cell count and new intramammary infection rate. This warrants further investigation, although the majority of cultured bacteria were coagulase-negative staphylococci (CNS). Consequently, we conclude that, in general, with appropriate monitoring (e.g., weekly inspection) to ensure the proportion of the herd with truncated milkings does not exceed 33%, farmers in pasture-based dairy systems can use a fixed milking time to improve parlor efficiency.     

Influence of air intake on the concentration of free fatty acids and vacuum fluctuations during automatic milking

The main objective of the study was to determine whether the amount of air intake during quarter milking influences the concentration of free fatty acids (FFA) and vacuum fluctuations at the teat end when milking automatically. Air intake in the teat cup was restricted from the normal inlet of 4.5 to 7 L/min to 1.7 and 0 L/min on 2 farms and experiments were carried out as half-udder studies with 40 cows. Blockage of the air inlet reduced FFA from 1.02 to 0.77 mEq/100 g of fat in one herd and from 1.50 to 1.17 mEq/100 g of fat in the other herd. Milk yield per milking was the most significant factor influencing FFA. Air intake accounted for <20% of the variation in FFA concentration. Characteristics of the cow explained the most variation, which could mainly be assigned to the effects of milk yield, fat percentage, fat globule size, and fat globule size distribution. The interval between milkings was not significant when adjusting for milk yields. Blockage of the air inlet caused vacuum fluctuations at the teat end to increase from 15.4 to 21.5 kPa for one model of an automatic milking system (AMS), but from 12.8 to 53.6 kPa for another model. Measurements made with a flow simulator and water revealed that the AMS model and water flow were the most important factors influencing vacuum fluctuations, and that interactions existed between the diameter of the short milk tube and air intake. Free fatty acids in bulk milk from 5,980 herds averaged 0.75 mEq/L of milk for conventional herds and varied from 0.77 to 0.94 mEq/L of milk for the 5 AMS models on the Danish market. Fault detection in 55 herds pointed out that the most frequent faults in conventional herds were air leakages and intake of too much air in the cluster, whereas AMS herds had problems with the cooling and stirring of milk. Correction of the cooling faults caused FFA to decrease by 0.52 mEq/L in the AMS herds. We concluded that air intake during automatic milking is not the most important factor in reducing FFA, whereas milk yield per milking matters the most. More attention should be paid to the cooling and stirring of milk. Reducing the air intake causes vacuum fluctuations during milking to increase significantly.     

Genetic parameters of milking frequency and milk production traits in Canadian Holsteins milked by an automated milking system

Twice-a-day milking is currently the most frequently used milking schedule in Canadian dairy cattle. However, with an automated milking system (AMS), dairy cows can be milked more frequently. The objective of this study was to estimate genetic parameters for milking frequency and for production traits of cows milked within an AMS. Data were 141,927 daily records of 953 primiparous Holstein cows from 14 farms in Ontario and Quebec. Most cows visited the AMS 2 (46%) or 3 (37%) times a day. A 2-trait [daily (24-h) milking frequency and daily (24-h) milk yield] random regression daily animal model and a multiple-trait (milk, fat, protein yields, somatic cell score, and milking frequency) random regression test-day animal model were used for the estimation of (co)variance components. Both models included fixed effect of herd × test-date, fixed regressions on days in milk (DIM) nested within age at calving by season of calving, and random regressions for additive genetic and permanent environmental effects. Both fixed and random regressions were fitted with fourth-order Legendre polynomials on DIM. The number of cows in the multiple-trait test-day model was smaller compared with the daily animal model. Heritabilities from the daily model for daily (24-h) milking frequency and daily (24-h) milk yield ranged between 0.02 and 0.08 and 0.14 and 0.20, respectively. Genetic correlations between daily (24-h) milk yield and daily (24-h) milking frequency were largest at the end of lactation (0.80) and smallest in mid-lactation (0.27). Heritabilities from the test-day model for test-day milking frequency, milk, fat and protein yield, and somatic cell score were 0.14, 0.26, 0.20, 0.21, and 0.20, respectively. The genetic correlation was positive between test-day milking frequency and official test-day milk, fat, and protein yields, and negative between official test-day somatic cell score and test-day milking frequency.     

Simulation model of quarter milk flowrates to estimate quarter and cow milking duration and automated milking system's box duration

The aims of this research were (1) to develop a model to simulate a herd of cows and quarter milk flowrates for a milking and derive quarter and udder milking durations and box duration (i.e., the time a cow spends inside the robot) for a group of cows milked with an automatic milking system (AMS); (2) to validate the simulation by comparing the model outcomes with empirical data from a commercial AMS dairy farm; and (3) to apply teatcup removal settings to the simulation to predict their effect on quarter and cow milking duration and box duration in an AMS. For model development, a data set from an AMS farm with 32 robots milking over 1,500 cows was used to fit the parameters to the variables days in milk, parity, and milking interval, which were subsequently used to create a herd of cows. A second data set from 2019 from an AMS farm with 1 robot milking 60 cows that contained quarter milk flowrates (at 2-s intervals) was used to extract the parameters necessary to simulate quarter milk flowrates for a milking. We simulated a herd of cows, and each was assigned a parity, days in milk, milking interval, and milk production rate. We also simulated milk flowrates every 1 s for each quarter of each cow. We estimated quarter milking duration as the total time that flowrate was greater than 0.1 kg/min after a minimum of 1 min of milk flow. We incorporated a randomly sampled attachment time for each quarter and calculated cow milking duration as the time from the first quarter attached to the last quarter detached. We included a randomly sampled preparation time which, added to cow milking duration, represented box duration. For simulation application, we tested the effect of quarter teatcup removal settings on quarter and cow milking duration. The settings were based on absolute flowrate (0.2, 0.4, and 0.6 kg/min) or a percentage of the quarter's 30-s rolling average milk flowrate (20, 30, and 50%). We simulated over 84,000 quarter milkings and found that quarter milking duration (average 212 s) had a mean absolute percent error (MAPE) of 7.5% when compared with actual data. Simulated cow milking duration (average 415 s) had a MAPE of 8%, and box duration (average 510 s) had a MAPE of 12%. From simulation application, we determined that quarter milking duration and box duration were reduced by 19% (209 vs. 170 s) and 6.5% (512 vs. 479 s), respectively, when increasing the teatcup removal flowrate from 0.2 to 0.6 kg/min. Quarter milking duration and box duration were 7% (259 vs. 241 s) and 3% (590 vs. 573 s) longer respectively by using a teatcup removal setting of 20% of the quarter's rolling average milk flowrate, compared with 30%. Both results agree with previous research. This simulation model is useful for predicting quarter and cow milking and box duration in a group of cows and to analyze the effect of milking management practices on milking efficiency.     

Estimating efficiency in automatic milking systems

Milking data of 34 single automatic milking system (AMS) units on 29 Galician dairy farms were analyzed to determine the system capacity in each farm under actual working conditions. Number of cows, milk yield, milkings per cow per day, actual milking time, rejected milking time, cleaning time, and machine downtime were used to determine the number of cows milked per AMS unit to obtain the optimal values of milkings per cow and milk production. Multiple linear regression data analysis was used to model the linear relationship between the dependent variable, milk yield per AMS per year, and the predictor variables: number of cows per AMS, milkings per cow per day, milk flow rate, and rejections per AMS per year. An AMS unit milked 52.7±9.0 cows daily at 2.69±0.28 milkings per cow, with a total milking downtime of 1,947±978 h/yr and a milk yield of 549,734±126,432 kg/yr. The predictor variables cow and milk flow rate had a greater level of influence on the milk yield per AMS than milkings per cow and rejections, and explained the 87% of the variation. The AMS in Galician dairy farms could facilitate an increase of 16±8.5 cows per AMS without impairing milking performance; in this way, the quantity of milk obtained per robot annually could be increased (185,460±137,460 kg). This would make it possible to recoup the cost of the system earlier. In the present situation, the daily milking throughput could be maximized at 2.4 to 2.6 milkings per cow.     

Effects of automatic cluster remover settings on average milking duration,milk flow,and milk yield

A crossover study design was used in five commercial dairy herds to study the effect of altering the switch point settings for automatic cluster remover units on the average duration of unit attachment, milk flow, and milk yield. Automatic cluster remover switch point settings were alternated, for 1-wk periods, between 0.50 and 0.64 kg/min (1.1 and 1.4 lb/min) in one herd and between 0.73 and 0.82 kg/min (1.6 and 1.8 lb/min) in the four remaining herds. Parlor data were captured at 329 separate milking sessions (range 39 to 92 per herd), representing 239,393 individual cow milkings. While increasing the automatic cluster remover switch point setting was not associated with a change in average milking duration in one herd, it had the effect of significantly reducing the average milking duration by between 10.2 and 15.6 s per cow in the remaining four herds. Milk flow was significantly increased at higher switch point settings for all five herds. Higher automatic cluster remover switch point settings did not have a negative effect on milk yield in any of the herds studied and, in fact, were associated with increased milk yield in two of the five herds. Decreasing milking duration while either maintaining or increasing the volume of milk harvested should ultimately lead to improved milking efficiency and parlor performance. Modifying systems to increase automatic cluster remover switch point settings offers an important potential opportunity to increase parlor efficiency in commercial dairy herds.     

Estimating daily yields of cows milked three times a day

Factors to estimate daily yield from one or two sampled milkings of cows milked three times a day were derived with data for 21 mo from seven Pennsylvania herds and data for 1 yr from five Utah herds. Accuracy of estimation increased with number of milkings weighed; standard deviation of error was nearly halved when information was included from two milkings instead of one. Fat yield did not increase as much as milk yield with increased time between milkings (interval); therefore, fat percentage factors increased with interval. If fewer milkings were sampled than weighed, fat yield was estimated by adjusting sample-day fat percentage. Effect of interval on protein yield was nearly the same as on milk yield. Therefore, milk factors are recommended for protein, and no adjustment to protein percentage is necessary if fewer milkings are sampled than weighed. Factors derived for milk were tested on records from 53 California and Oregon herds. Standard deviations of errors were similar to those from Utah and Pennsylvania data.     

Increasing the revenues from automatic milking by using individual variation in milking characteristics

The objective of this study was to quantify individual variation in daily milk yield and milking duration in response to the length of the milking interval and to assess the economic potential of using this individual variation to optimize the use of an automated milking system. Random coefficient models were used to describe the individual effects of milking interval on daily milk yield and milking duration. The random coefficient models were fitted on a data set consisting of 4,915 records of normal uninterrupted milkings collected from 311 cows kept in 5 separate herds for 1 wk. The estimated random parameters showed considerable variation between individuals within herds in milk yield and milking duration in response to milking interval. In the actual situation, the herd consisted of 60 cows and the automatic milking system operated at an occupation rate (OR) of 64%. When maximizing daily milk revenues per automated milking system by optimizing individual milking intervals, the average milking interval was reduced from 0.421 d to 0.400 d, the daily milk yield at the herd level was increased from 1,883 to 1,909 kg/d, and milk revenues increased from €498 to €507/d. If an OR of 85% could be reached with the same herd size, the optimal milking interval would decrease to 0.238 d, milk yield would increase to 1,997 kg/d, and milk revenues would increase to €529/d. Consequently, more labor would be required for fetching the cows, and milking duration would increase. Alternatively, an OR of 85% could be achieved by increasing the herd size from 60 to 80 cows without decreasing the milking interval. Milk yield would then increase to 2,535 kg/d and milk revenues would increase to €673/d. For practical implementation on farms, a dynamic approach is recommended, by which the parameter estimates regarding the effect of interval length on milk yield and the effect of milk yield on milking duration are updated regularly and also the milk production response to concentrate intake is taken into account.     

Milk quality and automatic milking: fat globule size, natural creaming, and lipolysis

Thirty-eight Italian Friesian first-lactation cows were allocated to 2 groups to evaluate the effect of 1) an automatic milking system (AMS) vs. milking in a milking parlor (MP) on milk fat characteristics; and 2) milking interval (≤480, 481 to 600, 601 to 720, and >720 min) on the same variables. Milk fat was analyzed for content (% vol/vol), natural creaming (% of fat), and free fatty acids (FFA, mEq/100 g of fat). Distribution of milk fat globule size was evaluated to calculate average fat globule diameter (d₁), volume-surface average diameter (d₃₂), specific globule surface area, and mean interglobular distance. Milk yield was recorded to calculate hourly milk and milk fat yield. Milking system had no effect on milk yield, milk fat content, and hourly milk fat yield. Milk from AMS had less natural creaming and more FFA content than milk from MP. Fat globule size, globular surface area, and interglobular distance were not affected by milking system per se. Afternoon MP milkings had more fat content and hourly milk fat yield than AMS milkings when milking interval was >480 min. Milk fat FFA content was greater in AMS milkings when milking interval was ≤480 min than in milkings from MP and from AMS when milking interval was >600 min. Milking interval did not affect fat globule size, expressed as d₃₂. Results from this experiment indicate a limited effect of AMS per se on milk fat quality; a more important factor seems to be the increase in milking frequency, generally associated with AMS.     

Effect of milking interval on alveolar versus cisternal milk accumulation and milk production and composition in dairy ewes

Cisternal and alveolar milk fractions were measured in East Friesian crossbred dairy ewes (n = 32) after 4, 8, 12, 16, 20, or 24 h of milk accumulation in a 6 x 6 Latin square design by administration of an oxytocin receptor antagonist for recuperation of cisternal milk followed by injection of oxytocin to remove the alveolar fraction. Less than half (38 to 47%) of the total milk yield was stored within the cistern for the first 12 h of udder filling compared with up to 57% after 24 h of udder filling. Subsequent milk yield was significantly reduced following the 16-, 20-, and 24-h treatments. Cisternal milk fat percentage, but not milk protein percentage, was lower than in alveolar milk (4.49 vs. 7.92% milk fat, respectively), indicating that casein micelles pass more freely from the alveoli to the cistern between milkings compared with fat globules. Alveolar compared to cisternal somatic cell count was higher for the 16-, 20-, and 24-h treatments. Significant increases in cisternal milk yield and milk composition observed for the 24-h compared with the 20-h treatment demonstrated the importance of the cistern as a storage space when the alveoli and small intramammary ducts became full. The main difference between cisternal and alveolar milk fractions is the poor fat content of cisternal milk, which is an important reason for the milk ejection reflex to be present during machine milking of dairy ewes. In a second experiment, milking every 16 h compared with every 12 h during mid- to late-lactation did not effect milk yield, milk composition, and quality, or lactation length; however, a 25% savings in labor was achieved with the longer milking interval.     

Supplementary feeding at milking and minimum milking interval effects on cow traffic and milking performance in a pasture-based automatic milking system

In extensive pastoral dairy farming systems herds graze 12 months of the year with the majority fed a near-100% pasture or conserved pasture diet. The viability of automatic milking in these systems will depend partly upon the amount of supplementary feed necessary to encourage cows to walk from the pasture to the milking unit but also on the efficient use of the automatic milking system (AMS). This paper describes a study to determine the importance of offering concentrate in the milking unit and the effect of minimum milking interval on cow movement and milking performance in a pasture-based AMS. The effects of feeding rate (FR0=0 kg or FR1=1 kg crushed barley/d) and minimum milking interval (MM6=6 h or MM12=12 h) on cow movement and behaviour during milking were studied in a multi-factorial cross-over (feeding level only, 4 weeks per treatment) experiment involving 27 mixed-breed cows milked through a single AMS. Feeding 1 kg barley in the milking unit resulted in a higher visiting frequency to the pre-selection unit (FR0=4.6 visits/d, FR1=5.4 visits/d, sed=0.35, P<0.05) and a higher yield (FR0=22.5 kg/d, FR1=23.6 kg/d, sed=0.385, P<0.01) but had no effect on milking frequency (FR0=1.6 milkings/d, FR1=1.7 milkings/d, sed=0.04, NS). Minimum milking interval was the major factor influencing milking frequency (MM6=1.9, MM12=1.4 milkings/d, sed=0.15, P<0.01). The absence of feeding in the milking unit had no negative effect on behaviour during milking or the number of cows that had to be manually driven from the paddock. The results show that automatic milking can be combined with a near-100% pasture diet and that milking interval is an important determinant for maximizing milk harvested per AMS.     

Three times a day milking during the first half of lactation

Three milking frequency treatments were compared: twice daily milking; thrice daily milking until milk dropped below 24 kg; thrice daily milking until milk dropped below 31 kg. Three time milking was at least 45 days but no more than 150 days. Cows (12 to 14 per group) were managed alike except for milking frequency. In early lactation, increased milking had little effect. With time the superiority in yields increased such that cows on three times for 150 days were outproducing two time cows by 20%. Cumulative milk yields were greater for the thrice groups than for the twice group by 5% at 56 days, 11 and 8% at 154 days, 11 and 9% at 182 days, and 10% at 280 days. Fat percentage, adjusted for previous lactation fat percentage, averaged .2 to .3% lower for the cows milked three times until 24 kig but only .1% lower for cows milked thrice until 31 kg. Increased yield was primarily from prolonged peak yield and less subsequent decline. Switching from three to two milkings decreased yield 6 to 8% in the 1st wk. However, three time milking had a positive carryover, apparently due to higher starting yield at the point at which they were switched.     

Detecting and predicting changes in milk homogeneity using data from automatic milking systems

To ensure milk quality and detect cows with signs of mastitis, visual inspection of milk by prestripping quarters before milking is recommended in many countries. An objective method to find milk changed in homogeneity (i.e., with clots) is to use commercially available inline filters to inspect the milk. Due to the required manual labor, this method is not applicable in automatic milking systems (AMS). We investigated the possibility of detecting and predicting changes in milk homogeneity using data generated by AMS. In total, 21,335 quarter-level milk inspections were performed on 5,424 milkings of 624 unique cows on 4 farms by applying visual inspection of inline filters that assembled clots from the separate quarters during milking. Images of the filters with clots were scored for density, resulting in 892 observations with signs of clots for analysis (77% traces or mild cases, 15% moderate cases, and 8% heavy cases). The quarter density scores were combined into 1 score indicating the presence of clots during a single cow milking and into 2 scores summarizing the density scores in cow milkings during a 30-h sampling period. Data generated from the AMS, such as milk yield, milk flow, conductivity, and online somatic cell counts, were used as input to 4 multilayer perceptron models to detect or predict single milkings with clots and to detect milking periods with clots. All models resulted in high specificity (98–100%), showing that the models correctly classified cow milkings or cow milking periods with no clots observed. The ability to successfully classify cow milkings or cow periods with observed clots had a low sensitivity. The highest sensitivity (26%) was obtained by the model that detected clots in a single milking. The prevalence of clots in the data was low (2.4%), which was reflected in the results. The positive predictive value depends on the prevalence and was relatively high, with the highest positive predictive value (72%) reached in the model that detected clots during the 30-h sampling periods. The misclassification rate for cow milkings that included higher-density scores was lower, indicating that the models that detected or predicted clots in a single milking could better distinguish the heavier cases of clots. Using data from AMS to detect and predict changes in milk homogeneity seems to be possible, although the prediction performance for the definitions of clots used in this study was poor.     

Covariance among milking frequency, milk yield, and milk composition from automatically milked cows

Automatic milking systems allow cows voluntary access to milking and concentrates within set limits. This leads to large variation in milking intervals, both within and between cows, which further affects yield per milking and composition of milk. This study aimed to describe the degree to which differences in milking interval were attributable to individual cows, and how this correlated to individual differences in yield and composition of milk throughout lactation. Data from 288,366 milkings from 664 cow-lactations were used, of which 229,020 milkings had milk composition results. Cows were Holsteins, Red Danes, and Jerseys in parities 1, 2, and 3. Data were analyzed using a linear mixed model, with cow-lactation as a random effect and assuming heterogeneous residual variance over the lactation. Cow-lactation variance was fitted using linear spline functions with 5 knot-points. Residual variance was generally greatest in early lactation and declined thereafter. Accordingly, animal-related variance tended to increase with progression of lactation. Milking frequency (the reverse of milking interval) was found to be moderately repeatable throughout lactation. Daily milk yield expressed per milking was found to be highly repeatable in all breeds, with the highest values occurring by the end of lactation. Fat percentage had only moderate repeatability in early to mid lactation but increased toward the end of lactation. Individual level correlations showed that cows with higher milking frequency also had greater yields, but had lower fat percentage. Correlations were slightly weaker in very early lactation than in the remaining parts of lactation. We concluded that individual differences exist among cows milked automatically. Cows with higher yields are milked more often and have lower fat content in their milk.     

Effects on milk yield of milking interval regularity and teat cup attachment failures with robotic milking systems

A database consisting of 35291 milking records from 83 cows was built over a period of 10 months with the objectives of studying the effect of teat cup attachment failures and milking interval regularity on milk production with an automated milking system (AMS). The database collected records of lactation number, days in milk (DIM), milk production, interval between milkings (for both the entire udder and individual quarters in case of a teat cup attachment failure) and average and peak milk flows for each milking. The weekly coefficient of variation (CV) of milking intervals was used as a measure of milking regularity. DIM, milking intervals, and CV of milking intervals were divided into four categories coinciding with the four quartiles of their respective distributions. The data were analysed by analysis of variance with cow as a random effect and lactation number, DIM, the occurrence of a milking failure, and the intervals between milkings or the weekly CV of milking intervals as fixed effects. The incidence of attachment failures was 7.6% of total milkings. Milk production by quarters affected by a milking failure following the failure was numerically greater owing to the longer interval between milkings. When accounting for the effect of milking intervals, milk production by affected quarters following a milking failure was 26% lower than with regular milkings. However, the decrease in milk production by quarters affected by milking failures was more severe as DIM increased. Average and peak milk flows by quarters affected by a milking failure were lower than when milkings occurred normally. However, milk production recovered its former level within seven milkings following a milking failure. Uneven frequency (weekly CV of milking intervals >27%) decreased daily milk yield, and affected multiparous more negatively than primiparous cows.     

Factors for Estimating Daily Yield of Milk,Fat, and Protein from a Single Milking for Herds Milked Twice a Day

Existing factors for estimating daily yield from the yield of a single milking were analyzed by comparing estimated daily yields with milk sold for herds in the northeastern United States. Milk yield was overestimated consistently compared with bulk tank milk weights. Estimates of protein yield showed lack of fit by length of preceding interval. Milk factors were developed from testing plans with one measured milking (a.m.-p.m. testing plans) and from those with two measured milkings; resulting factors were essentially the same. New factors for components were developed from data from a.m.-p.m. testing plans because component percentages for individual milkings are determined only for a.m.-p.m. plans. Factors were reciprocals of portion of day's yield, which was modeled as a linear effect of preceding interval. Slopes for milk and protein were essentially the same. Slope for fat was about 70% of those for milk and protein, which indicated that fat was affected less by interval than were milk and protein. Therefore, fat percentage from a single milking should be adjusted to estimate day's fat percentage. Analysis of residuals from analysis for milk indicated an effect by stage of lactation. An additive adjustment is suggested. Data were collected from herds enrolled in a.m.-p.m. testing plans in the northeastern United States after the new factors had been implemented for more than 1 yr. They indicated that most lack of fit from old factors had been eliminated.     

Short communication: Cow- and quarter-level milking indicators and their associations with clinical mastitis in an automatic milking system

The aim of this study was to assess associations of cow-, udder-, and quarter-level factors with the risk of clinical mastitis (CM) in cows managed using an automatic milking system. The primary hypothesis was that quarter peak milk flow rate (QPMF) is associated with increased risk of CM. A retrospective, case-control study was undertaken using data from a 1,549 cow farm using 20 automatic milking system units. All data from cows milked during March to December 2015 was available for analysis. Cases (n = 82) were defined as cows diagnosed with their first case of CM between 24 and 300 d in milk in the current lactation. Healthy control cows (n = 6/case) were randomly matched based on identical parity, existence of milk records during the day in milk period corresponding to the 15-d window before case diagnosis, average conductivity of <5.5 mS/cm in that window, and no history of CM in the current lactation. Logistic regression was used to estimate effects of parity, quarter position, day in milk at diagnosis of CM, average of QPMF 15 d before CM diagnosis, udder milk yield, and milking interval on the probability of CM. Of the 6 predictor variables included in the model, only milking interval was significantly associated with the increased risk of quarter CM. We concluded that in a high-production, freestall-housed North American herd using automatic milking system, milking interval, but not QPMF, was associated with risk of CM.     

Accuracy of simplified sampling procedures for estimating milk composition in dairy ewes

Simplified designs of milk-composition recording, based on information from a single monthly milking, adjusted or not for interval between milkings and for milk yield, were simulated and evaluated for 2553 ewe-test-day records belonging to 155 lactations of Churra dairy ewes. Accuracy of simplified methods was evaluated by comparing estimated trait values (fat, protein, casein and total solid yields) with those observed both in a reference plan, where the two daily milkings were recorded at weekly intervals (A1), and in the official A4 milk recording (monthly records of the two daily milkings). Trait yields per lactation were estimated and adjusted to the only milking period (days in milk 30-120). Estimates of milk component traits were less precise when monthly designs, including the A4 design, were compared with a weekly sampling of both a.m. and p.m. milkings, with fat yield being the most difficult trait to estimate. All options with one daily milking every month were more accurate when the corresponding plan was based on, or began with, the a.m. milking. Adjustment for the preceding interval between milkings or milk yield did not improve sampling accuracy. The design alternating a.m. and p.m. milkings every month, without adjustment, is suggested for recoding milk component traits in dairy ewes.     

The effect of pre-milking teat-brushing on milk processing time in an automated milking system

Cow throughput in an automatic milking system (AMS) is limited by system parameters such as the time required for pre-milking udder preparation and cup attachment, physiological responses of the cow (such as milk let-down and milking-out rate), milking machine features and cow behaviour. A single-factor cross-over design was used to investigate the effect of pre-milking teat brushing on milk processing time in an AMS operating in an extensive grazing farming system. Teat brushing consisted of two roller brushes tracking up each teat three times (total brushing time of up to 45 s/cow). Cows were allocated to one of two treatment groups with either no brushing (NB) or brushing (B) for a 4-week period before being changed to the other treatment. Teat brushing resulted in shorter average cups-on-time (B = 506.1 s, NB = 541.0 s, P = 0.0001), longer average milk processing time (B = 10.30 min, NB = 9.76 min, P = 0.001) and no difference in daily milk yield (B = 14.67, NB = 14.71 kg/cow, P = 0.826). There was no difference between the two treatments in the success of cup attachment (B = 3.76%, NB = 5.10% unsuccessful milking attempts, P = 0.285). The estimated time cost of pre-milking teat brushing was 53 min for every 100 milkings, equivalent to an additional 5-6 milkings for every 100 milkings by an AMS. The importance of these potential time savings is discussed in relation to automatic milking in farming systems that aim for a lower per cow milking frequency and high ratio of cows to AMS.     

Effects of an automatic milking system on milk yield and quality of Mediterranean buffaloes

A study was carried out to evaluate the effects of an automatic milking system (AMS) on milk yield and composition of buffalo (Mediterranean-type Bubalus bubalis) cows. Performed from January 2015 to December 2015 in an organic buffalo dairy farm equipped with both a traditional tandem milking parlor and an AMS, the study involved 90 primiparous buffaloes randomly allotted to a tandem or AMS group from 5 to 10 d of lactation onward. Number of milkings per day and daily milk yield of each cow were recorded, and individual milk sampling was carried out twice a month. Compared with the tandem, the AMS group showed significantly higher daily milk yield and persistence of lactation. Use of the AMS resulted in higher protein and casein contents, and lower somatic cell and total bacterial counts, whereas fat, freezing point, and pH were unaffected by the system. We conclude that, in terms of milk yield and quality, automatic milking may be a suitable alternative to conventional milking for buffaloes.