Milk Leakage—An Increased Risk in Automatic Milking Systems

J. Dairy Sci. 86:3488–3497 © American Dairy Science Association, 2003.

Milk Leakage—An Increased Risk in Automatic Milking Systems K. Persson Waller,*,† T. Westermark,* T. Ekman,† and K. Svennersten-Sjaunja‡ *Department of Ruminant and Porcine Diseases, National Veterinary Institute, SE-751 89 Uppsala, Sweden, †Department of Obstetrics and Gynaecology, and ‡Department of Animal Nutrition and Management, Swedish University of Agricultural Sciences, SE-750 07 Uppsala, Sweden

ABSTRACT Milk leakage (ML), or milk observed dripping or flowing from one or more teats between milkings, has been associated with increased risk of udder infections and mastitis in dairy cows. Preliminary observations indicate that ML might occur more often in automatic milking systems (AMS) than in conventional milking systems (CMS), but comparative data on the incidence of ML in AMS or in CMS are not available. Therefore, the occurrence of ML at various observation periods was studied in one AMS with cows housed in a free-stall barn in comparison to CMS with cows housed either in a free-stall barn or a tie-stall barn and milked at regular intervals in a herringbone milking parlor. Relationships between ML and other cow and management factors were also examined. In each of 2 yr, all cows (n = 230 total; 46 cows present both years) were observed at 2-h intervals during six 24-h periods. At least one ML occurred in 39.0 (AMS) vs. 11.2% (CMS) of individual cows and in 16.2 (AMS) vs. 2.9% (CMS) of 24-h cow days studied. Milk leakage was not related to milk production, parity, stage of lactation, or estrous status. However, in the AMS, 62% of primiparous and 28% of multiparous cows leaked milk at least once. Milk leakage occurred more often in rear than in forequarters. Cows were usually lying down when ML was observed, but intervals from previous milking varied, especially in AMS. In AMS, about one-fifth of the ML observations occurred ≤4 h after milking, and half of those were associated with disturbances at the previous milking. Milk flow rate was higher in quarters leaking milk than in other quarters. Strategies to reduce milk leakage in AMS may be important to minimize potential risks of udder disease. (Key words: automatic milking systems, dairy cow, milk leakage, udder health) Abbreviation key: AMU = automatic milking unit, AMS = automatic milking system, CMS = conventional

Received February 7, 2003. Accepted July 24, 2003. Corresponding author: K. Persson Waller; e-mail: Karin. [email protected].

milking systems, FSB = free-stall barn with cows milked in a milking parlor, ML = milk leakage, TSB = tie-stall barn with cows milked in a milking parlor, UDS = udder disease score. INTRODUCTION Milk leakage (ML), or milk dripping or flowing from one or several of the teats, between milkings is a phenomenon that occurs occasionally in dairy cows. Preliminary observations (data not shown) have indicated that this might be more common in automatic milking systems (AMS) than in conventional milking systems (CMS). The problem should not be neglected, as ML is an important risk factor for mastitis (e.g., Jørstad et al., 1989; Schukken et al., 1990; Elbers et al., 1998; Waage et al., 1998). Milk leakage can occur if the closing mechanism of the teat canal is compromised, such as when the teat end is damaged (Jørstad et al., 1989). However, the patency of the teat canal can also be affected by its anatomy because wide (McDonald, 1975; Jørstad et al., 1989) and short (Lacy-Hulbert and Hillerton, 1995) teat canals have been associated with a higher incidence of mastitis. Milk leakage can also occur if the internal milk pressure overcomes the closing forces of the teat canal (e.g., when the milk let-down reflex is activated before milking). Moreover, milking routines and conditions (e.g., vacuum and pulsation levels, liner design, and weight of cluster) are important for successful emptying of the udder at milking (Hamann and Dodd, 1992; Mein, 1992). An AMS relies on well-functioning cow traffic between the automatic milking unit (AMU) and the feeding and resting areas. However, unlike a CMS, where milking frequency and interval are similar for all cows, there is more variation between and within cows in milking time and frequency, and in intervals between milkings in an AMS. Uneven milking intervals can influence both milk yield (Ipema et al., 1997) and milk SCC (Hamann and Gyodi, 2000), but the effect on the incidence of ML is not known. In AMS, quarters are often milked individually, which minimizes overmilking and is positive for udder health (Seeman 1997; Berglund et al., 2002). However, early studies showed that

3488

MILK LEAKAGE IN AUTOMATIC MILKING SYSTEMS

failure of milking cluster attachment during milking could occur in up to 15% of the milkings (Mottram et al., 1995; Ipema et al., 1997). This can result in no milking or incomplete milking of one or more quarters, which might result in increased risk of ML. However, no differences in strip yield (Svennersten-Sjaunja et al., 2000) or residual milk (Hopster et al., 2002) were observed when comparing conventional and automatic milking. Use of AMS has increased in many countries. Studies including AMS farms in different countries indicate that herd udder health, as measured by SCC, may decline after introduction of the new system (Everitt et al., 2002; Kruip et al., 2002; van der Vorst and de Koning, 2002). However, it has also been demonstrated that the udder health was not negatively affected in some individual herds, which had good udder health at the introduction of AMS (Svennersten-Sjaunja et al., 2000; Hamann and Reinecke, 2002). Reasons behind observed increases in SCC in several farms are not fully clear, but an increased occurrence of ML in such systems could be one factor of importance. Data on the incidence of ML in AMS in comparison with the incidence in CMS are, to our knowledge, not available. Therefore, the primary aim of this 2-yr study was to compare the occurrence of ML in an AMS in comparison with 2 other groups of cows managed using a CMS. A second objective was to identify cow and management factors associated with higher risks of ML. MATERIALS AND METHODS Animals and Housing Systems Lactating cows from the Kungsa¨ngen Research Centre, Swedish University of Agricultural Sciences, Uppsala, were used. The cows were of the Swedish Red and White breed and were housed in 3 barns with different housing and management systems; an AMS (Voluntary Milking System, DeLaval, Tumba, Sweden) with cubicles or free stalls, a free-stall barn (FSB), and a tiestall barn (TSB). Cows from both the FSB and TSB groups were milked in the same double-8 herringbone milking parlor. There were 120 cows present during the study periods in the first year and 110 cows the second year, with 46 individual cows present in both years. The number of individual cows and cows per observation occasion varied among the barns and years studied (Table 1). Overall, no statistically significant differences in parity, lactation month, or daily milk production were observed among barn systems. The overall proportions of cows in parity 1, 2, 3, and ≥4 were 45, 23, 16, and 16%, respectively. However, the proportion of first-lactation cows in the barns was 31, 79, and 39%

3489

in AMS, FSB, and TSB, respectively, when combining data from both years. The high proportion of first-lactation cows in the FSB was due to management decisions. Due to this discrepancy, caution must be taken when comparing ML occurrence between the FSB and the other systems. Milking In the AMS, cows were milked, on average, 2.7 times per day. The time between milkings varied due to milk production but was never less than 6 h; cows that had not been milked for 14 h were fetched manually. This was only done once per day, which means that the actual time between milkings could be longer than 14 h. The AMU was equipped with quarter milking units. The vacuum level was 42 kPa, the pulsation rate was 60 cycles/min, the pulsator ratio was 65:35, and 999 007-83 liners (DeLaval, Tumba, Sweden) were used. The teat cups were automatically removed, with a 6-s delay, at a flow rate of 210 g/min per quarter. Cows in the FSB and TSB barns were milked twice daily (at 0530 and 1600 h) in a conventional double-8 herringbone parlor with Harmony milking units (DeLaval), 999 007-83 liners (DeLaval), and using Duovac (DeLaval). During the high vacuum phase, the vacuum level was 44.5 kPa, the pulsation rate was 60 cycles/min, and the pulsator ratio was 65:35. During the low vacuum phase, the corresponding values were 33 kPa, 50 cycles/min and 30:70 for the pulsator ratio. The vacuum was switched from a low to high level at a milk flow of 300 g/min, and the teat cups were automatically removed when the milk flow decreased to 300 g/min and continued below that level for 24 s. In the AMU, the cows were fed 0.5 kg of concentrates at each milking. The FSB cows were fed up to 2 kg of concentrates at milking depending on milk production, whereas the TSB cows got 0.2 kg/milking. Feeding In the AMS, the feeding areas were reached either by passing through the AMU or by passing through selection gates. Silage only (yr 1) or a mix of silage, concentrate, and hay (yr 2) was fed ad libitum, whereas concentrates were fed individually in automatic feeding stations according to milk production. Feed troughs were filled at 4- to 5-h intervals between 0530 and 1830 h. In the FSB, cows were fed a mix of silage, concentrate, and hay four times per day (at 0430, 0900, 1400, and 1700 h). In the TSB, cows were fed silage and concentrates four times per day (at similar times as FSB), plus 1 kg of hay in the mornings. Daily supplementation of vitamins and minerals was given ad libitum from Journal of Dairy Science Vol. 86, No. 11, 2003

3490

PERSSON WALLER ET AL. Table 1. The total number of dairy cows per observation occasion and number of individual cows, and their mean (SD) parity, lactation month, and daily milk production in 3 management systems (automatic milking system [AMS], free-stall barn [FSB], and tie-stall barn [TSB] with milking parlor), studied during six 24h periods during 3 to 4 wk in each of 2 yr. Variables

Year

AMS

FSB

TSB

Cows per observation occasion

1 2 1 2 1 2 1 2 1 2

46 to 49 47 to 52 53 52 2.0 (0.6) 2.3 (1.1) 4.9 (2.0) 4.6 (2.4) 31.2 (6.9) 28.7 (6.7)

16 to 18 31 to 32 19 34 1.7 (0.8) 1.1 (0.1) 5.6 (2.6) 4.4 (1.9) 30.2 (5.2) 27.2 (3.5)

39 to 43 16 to 24 48 24 2.6 (1.5) 2.8 (1.6) 4.6 (2.1) 7.0 (2.9) 30.3 (7.0) 23.8 (5.6)

Individual cows Parity1 Lactation month1 Daily milk production (kg/cow)1

Overall (yr 1+2) values did not differ significantly between barns (P > 0.05).

1

separate feed troughs in AMS and FSB, and individually in TSB. Description of the Study The study was performed in January and February in 2001 and 2002. The time of year was selected as most cows were high producing and the farm routines were stable and not affected by factors like calving or grazing seasons. All cows were observed by the same person at 2-h intervals during six 24-h periods spread over 3 to 4 wk (i.e., 72 observation periods per year). If the cow was lying down, the udder was gently lifted to enable inspection of all teats. At the start of the study periods, none of the cows had any visible traumatic teat injuries. Milk leakage was defined as milk observed dripping or flowing from one or more teats. If ML were observed, the following variables were recorded for the affected cow: cow number, teat(s) with leakage, lying/standing, activity, and place in barn. In the second year, the position (lying/standing) of cows not leaking milk was also registered at all occasions. For each occasion of ML, the time from the previous milking (precise estimate in AMS and approximate estimates in FSB and TSB) was calculated. Additional data for all cows were collected, or calculated, using monthly milk recordings and other information in the herd database (Table 2). In the Swedish milk-recording service run by the Swedish Dairy Association, individual cow milk samples are taken monthly for measurement of the SCC. In this service, an udder disease score (UDS) is also calculated based on the geometric mean of the last 2 or 3 monthly cell counts (Brolund, 1990). The aim of this score is to give a better estimate of the occurrence of subclinical mastitis in individual cows; the higher the score, the higher the probability for mastitis. On a separate occasion at the end of the study period each year, one person examined the shape and condition Journal of Dairy Science Vol. 86, No. 11, 2003

of the teat ends of all cows. The shape of teat ends was classified as round, flat, or pointed, and the condition (no, smooth, or rough callosity rings) of teat ends was classified according to Neijenhuis et al. (2000). Statistics Data were evaluated for udder quarter, cow, observation, and/or cow-day (24-h period/cow), and are presented as means and standard deviations (SD). Each cow’s SCC was log10-transformed before analyses to ensure normal distribution of the data. Differences between cows with or without ML in parity, lactation month, daily milk production, SCC, and UDS were evaluated statistically using one-way ANOVA (Statistica, StatSoft, Inc., Tulsa, OK). Differences between barns in cow-days and individual cows with ML, effect of increasing UDS on risk for ML, differences between cows lying or standing, fore- and hind-udder quarters, and the relationship between ML and milk flow rate in AMS were analyzed using the GLM procedures (SAS Institute, Cary, NC) after correction for cow when necessary. RESULTS Occurrence of ML The total numbers and the proportions of cow observations, cow-days, and individual cows with ML on at least one occasion during the study periods and years in the different systems are given in Table 3. Including both years, proportions of cow-days (P < 0.001) and individual cows (P < 0.001) with ML on at least one occasion during the study period were significantly higher in AMS than in either TSB or FSB. Differences between the TSB and FSB were not significant. However, because mainly primiparous cows were present in the FSB, the data must be interpreted with caution. Three of 46 cows present both years leaked milk both years.

3491

MILK LEAKAGE IN AUTOMATIC MILKING SYSTEMS Table 2. Variables collected, or calculated from the herd database for all cows with or without milk leakage included in the study. Variables

Explanation/comment

Parity Stage of lactation Daily milk production

1, 2, 3 or ≥4 Month after calving Milk yield (kg/d) at the monthly milk recording closest to the study period The cow SCC measured at the monthly milk recording closest to the study period A score (0 to 9) calculated from the 2 or 3 previous monthly milk cow SCC indicating the likelihood that the cow has mastitis (e.g., 0 = 0 to 9%; 9 = 90 to 100% probability of mastitis) Observation of cows in or not in estrus Diseases treated by a veterinarian during the study periods; classification is performed according to the Swedish national health recording scheme Flow rate (g/min) at 2 milkings/cow in association with the monthly recordings (only in the AMS during yr 2) Problems registered in association with the milking previous to (AMS) or ongoing at (FSB, TSB) the observation occasion

Cow SCC Udder disease score (UDS)1

Estrus Health Average and peak milk flow rate per quarter Problems associated with individual milkings 1

A more detailed explanation is given in Materials and Methods.

Among the cows observed to leak milk during the study period, the number of observations of ML per cow varied between 1 and 14. Combining both years, the mean (SD) number per cow was 3.8 (3.0), 1.7 (1.0), and 1.4 (0.6) in AMS, FSB, and TSB, respectively. Within cows that leaked milk, the proportion of cows leaking milk ≥3 times during the study periods was 44% in AMS, but only 14% in TSB and FSB (Table 4). In TSB and FSB, the highest number of ML per time point was observed just before morning milking, which started at 0530 h (Figure 1). Afternoon milking started at 1600 h in those groups. Distribution of ML was more variable over time in AMS (Figure 1). More details about the time of ML in relation to the previous milking are given below (Milking-Related Factors).

Among cows with many observations of ML, it was common that 3 or 4 of the teats were affected (data not shown). However, it was significantly (P < 0.05) more common for rear quarters (13.6%) to leak milk than fore quarters (9.3%). This was observed both when cows were lying and when they were standing at the observation of ML (data not shown). The teat shape and condition of 198 cows (792 teats) were inspected. Eighty-two percent of the teats had a round teat end and 18% had a flat teat end. No teats were classified as pointed. The proportion of round and flat teats with at least one ML was 11.8 and 9.9%, respectively. In most cases (487 teats, 61%), no callosity ring was observed on the teat ends. Of these teats, 12.3% had ML at some occasion. A slight thickening of the teat end classified as a slight

Table 3. Total number (%) of dairy cow observations, cow-days, and individual cows with milk leakage (ML) on at least one occasion in 3 management systems (automatic milking system [AMS], free-stall barn [FSB], and tie-stall barn [TSB] with milking parlor) studied during six 24-h periods within 3 to 4 wk in each of 2 yr. AMS

FSB

TSB

Variables

Year

Total no.

No. (%) with ML

Total no.

No. (%) with ML

Total no.

No. (%) with ML

Cow observations

1 2 1+2 1 2 1+2 1 2 1+2

3426 3552 6978 286 296 582 53 52 105

53 (1.5) 103 (2.9) 156 (2.2)1 32 (11.2) 62 (20.9) 94 (16.2)a 13 (24.5) 28 (53.8) 41 (39.0)a

1220 2256 3476 101 188 289 19 34 53

1 (0.1) 11 (0.5) 12 (0.3)1 1 (1.0) 10 (5.3) 11 (3.8)b 1 (5.3) 6 (17.6) 7 (13.2)b

2914 1464 4378 244 120 364 48 24 72

6 (0.2) 4 (0.3) 10 (0.2)1 5 (2.0) 3 (2.5) 8 (2.2)b 5 (10.4) 2 (8.3) 7 (9.7)b

Cow-days Individual cows

Values within the same row with different superscripts differ significantly at P < 0.05. Not analyzed.

a,b 1

Journal of Dairy Science Vol. 86, No. 11, 2003

3492

PERSSON WALLER ET AL. Table 4. Number (%) of dairy cows with milk leakage (ML) at 1, 2, or ≥ 3 occasions during six 24-h periods during 3 to 4 wk in each of 2 yr in 3 management systems (automatic milking system [AMS], free-stall barn [FSB], and tie-stall barn [TSB] with milking parlor). Barn

Year

1 ML

2 ML

≥3 ML

Total

AMS

1 2 1+2 1 2 1+2 1 2 1+2 1+2

6 10 16 (39) 1 4 5 (71) 4 1 5 (71) 26 (47)

1 6 7 (17) 0 1 1 (14) 1 0 1 (14) 9 (16)

6 12 18 (44) 0 1 1 (14) 0 1 1 (14) 20 (36)

13 28 41 (100) 1 6 7 (100) 5 2 7 (100) 55 (100)

FSB TSB Total

callosity ring was observed in 295 teats (37%), whereas the ring was considered moderate to thick in 10 teats (1%). Milk leakage was observed in 10.5 and 0 % of those teats, respectively. A slight roughness of the callosity ring was observed in 39 teats, and a moderate to thick rough ring was detected in two teats. None of those teats leaked milk. Cow Factors Including all cows from both years, the lactation number, lactation month, and daily milk production did not differ significantly between cows with or without ML (Table 5). However, the proportion of primiparous cows with at least one ML was 62, 15, and 3% in AMS (n = 34), FSB (n = 41), and TSB (n = 29), respectively. Corresponding numbers for multiparous cows were 28, 9, and 14% for AMS (n = 71), FSB (n = 11), and TSB (n = 42), respectively. In AMS, among cows with at least one ML, 48% of the primiparous cows and 45% of the multiparous cows leaked milk ≥3 times.

Figure 1. Numbers of observations of dairy cow milk leakage (ML) at each tine point when observations were made every second hour during six 24-h observation periods during 2 yr (1, 2) in three management systems (automatic milking system [AMS], free-stall barn with milking parlor [FSB], and tie-stall barn with milking parlor [TSB]). Journal of Dairy Science Vol. 86, No. 11, 2003

No significant differences were observed in cow-days with or without ML between cows in estrus or not in estrus. In total, 68 and 1271 cow-days were recorded for cows in estrus and not in estrus, respectively, and the proportions with ML were 8.1 and 9.0% in the 2 groups. It was significantly (P < 0.001) more common that cows lying down (104 of 3558 observations) leaked milk than cows standing up (14 of 3690 observations), according to observations in yr 2. There was no indication that the cows’ lying/standing behavior differed between the barns. The proportion of observations with cows lying down varied between 47 and 51% in the three barns. The cows were lying down in 85% (yr 1) and 92% (yr 2) of the ML observations. For 57% (yr 1) and 39% (yr 2) of the observations of ML when lying down, the cows were ruminating. For the remaining occasions, they were resting or sleeping. Milking-Related Factors Among the observations of ML, the time in relationship to the previous milking varied, especially in the AMS (Table 6). In yr 1, the mean (SD) time from previous milking was 7.6 (4.2) h, 13 h (n = 1), and 7.3 (1.5) h for AMS, FSB, and TSB, respectively. For yr 2, the corresponding values were 8.3 (4.1) h, 9.8 (1.7) h, and 9.5 (3.0) h, respectively. In the AMS, 22 and 15% of the total number of observations were made ≤4 h and >12h after milking, respectively (Table 6). When combining both years, 57% of the ML observations were made in primiparous cows and 43% in multiparous cows. In primiparous cows, 19, 27, 39, 13, and 1% of the ML observations (n = 89) were made ≤4 h, >4 to 8, >8 to 12, >12 to 16, and >6 h after milking, respectively. The corresponding proportions for the ML observations (n = 67) in multiparous cows were 25, 33, 25, 13, and 3%, respectively. In the AMS, automatic registrations were made if problems occurred at milking, such as when the AMU

3493

MILK LEAKAGE IN AUTOMATIC MILKING SYSTEMS Table 5. Means (SD) for parity, lactation month, daily milk production, milk log10 SCC, and udder disease score (UDS) of dairy cows without milk leakage (ML) or with ML on at least 1 occasion during six 24-h periods per year spread over 3 to 4 wk in each of 2 yr. Variables

Cows with ≥1 ML (n = 55)

Cows without ML (n = 175)

P

Parity Lactation month Milk production (kg/d) Log10 SCC UDS

2.1 (1.3) 4.7 (2.8) 27.7 (7.1) 4.74 (0.61) 2.1 (2.4)

2.2 (1.3) 5.1 (3.2) 29.4 (8.4) 4.63 (0.58) 1.5 (2.6)

0.53 0.15 0.16 0.22 0.11

was unable to attach the teat cup to one or several teats (missed milking), or when one or several teat cups were kicked off during milking (incomplete milking = <60% of expected milk volume). During the study periods, the overall average proportions of missed and incomplete milkings in the AMS were 2.2 and 3.6%, respectively. When combining the data from both years, such problems was registered at the previous milking in 32% of all observed occasions of ML in AMS, and more specifically in 41, 35, 31, 19, and 0% of the ML observations registered at ≤4, >4 to 8, >8 to 12, >12 to 16, and >6 h after milking, respectively. Missed or incomplete milking was registered at least once in 4 of 22 (18%) cows with one or two ML, and in 11 of 19 (58%) cows with ≥3 ML. In a majority of the latter cows, such milkingrelated problems had occurred at most ML occasions. The relationship between the number of ML and these milking problems were similar in primiparous and multiparous cows (data not shown). Other problems registered in association with ML were cleaning of the AMU (n = 1), cow chased off the waiting area (n = 1), and abnormal milk quality (as defined by the Voluntary Milking System) (n = 3). The first two problems would have resulted in a longer waiting time for the cow and thereby in a longer milking interval. The abnormal milk quality indicates mastitis, which can lead to a reduced milk let down. At 2 occasions of ML that occurred 9 to 11 h after the previous milking in 1 cow, that cow had been moved to the AMS the same morning, and on 3 occasions (occurring >8 h postmilking) in 2 cows, the

animals had passed the milking unit just before the time for a new milking permission. During the second year, AMS data on the mean and peak milk flow rate were available for each cow and for individual quarter within cow. No such data were available in the FSB and TSB. No significant differences were observed between cows with at least 1 ML and cows without ML in their mean average and maximal flows. However, when comparing quarters with or without ML, the mean (P < 0.001) and peak (P < 0.01) flow rate per quarter was significantly higher in quarters with ML. The average (SD) mean and peak flow rates for udder quarters not leaking milk were 629 (199) g/min and 986 (210) g/min, respectively. The corresponding values for udder quarters with at least one ML were 776 (278) and 1138 (316) g/min, respectively. The effect of ML frequency observed per quarter on the milk flow rate was also analyzed after dividing the ML quarters in 3 groups, 1 ML, 2 ML, and ≥3 ML. Both the mean (P < 0.001) and the peak (P < 0.001) flow rate increased significantly with increasing numbers of ML. The average (SD) mean flows in the 3 groups were 725 (256), 819 (245), and 833 (323) g/min, respectively, whereas the corresponding values for the peak flow were 1071 (308), 1158 (253), and 1243 (346) g/min, respectively. Effect on Udder Health The average monthly bulk SCC of the whole herd was 144,000 and 154,000 cells/ml during the first and

Table 6. Numbers (%; only for AMS) of observations of milk leakage in relation to time after previous milking in dairy cows in three management systems (automatic milking system [AMS], free-stall barn [FSB], and tie-stall barn [TSB] with milking parlor). AMS

FSB

TSB

Hours

Yr 1

Yr 2

Yr 1

Yr 2

Yr 1

Yr 2

≤4 >4 to 8 >8 to 12 >12 to 16 >16 Total

14 (26) 17 (32) 13 (25) 8 (15) 1 (2) 53 (100)

20 (19) 29 (28) 39 (38) 13 (13) 2 (2) 103 (100)

0 0 0 1 0 1

0 2 7 0 0 9

0 5 1 0 0 6

0 2 2 0 0 4

Journal of Dairy Science Vol. 86, No. 11, 2003

3494

PERSSON WALLER ET AL. Table 7. Total number of cow-days, and the number (%) of cow-days with milk leakage (ML) observed on at least 1 occasion during six 24-h periods during 3 to 4 wk in each of 2 yr in cows with different udder disease score (UDS).1 Cow-days, yr 1

Cow-days, yr 2

Cow-days, yr 1+ 2

UDS2

Total no.

No. with ML

Total no.

No. with ML

Total no.

No. (%) with ML

0 1 2 3 4 5 6 7 8 9 —

339 46 54 48 16 18 45 15 12 13 25

10 6 3 3 1 0 1 5 0 3 1

360 38 33 18 35 12 27 18 10 10 40

33 8 5 4 13 1 8 1 0 0 2

699 84 87 66 51 30 72 33 22 23 65

43 (6.2) 14 (16.7) 8 (9.2) 7 (10.6) 14 (27.5) 1 (3.3) 9 (125) 6 (18.2) 0 (0) 3 (13.0) 3 (4.6)

The increase in the percentage of ML between UDS 0 to UDS 4 was significant (P < 0.001). See Materials and Methods for explanation. UDS 0 = 0 to 9%, 1 = 10 to 19%, and 9 = 90 to 100% risk for mastitis; — = UDS not available because the time point was too soon after calving. 1 2

second years, respectively. Mean (SD) UDS during yr 1 were 2.0 (2.1), 1.6 (2.1), and 1.6 (2.0) for AMS, FSB, and TSB, respectively. The corresponding means for yr 2 were 1.4 (1.7), 1.4 (2.1), and 1.7 (2.2), respectively. The cow SCC and UDS were numerically, but not significantly, higher in cows with ML compared with cows without ML (Table 5). The number of cow-days with ML increased significantly (P < 0.001) when the UDS increased from score 0 to 4, but the incidence of ML varied markedly when scores were above 4 (Table 7). The number of cows treated for clinical mastitis was too small to evaluate the effect of ML on the risk of clinical mastitis. DISCUSSION Milk leakage was observed significantly more often and in a larger proportion of the cows in the AMS than in the CMS studied. The SCC and UDS were numerically, but not significantly, higher in cows with ML compared with in cows without ML. Moreover, a higher quarter milk flow rate was associated with an increased risk of ML. When including all cows, daily milk production, parity, and month of lactation did not differ significantly between cows with or without ML. However, the number of primiparous cows with ML was more than 6 times higher in AMS than in CMS, whereas the proportion of multiparous cows with ML in AMS was just over double that found in CMS. More than half of the observations of ML made in AMS occurred in primiparous cows. The reason for the high proportion of primiparous cows with ML in AMS is not clear, but it was not due to a larger proportion of missed or incomplete milkings than in multiparous cows. However, social dominance Journal of Dairy Science Vol. 86, No. 11, 2003

might be a factor of importance. Primiparous cows are often ranked lower than older cows (Dickson et al., 1970), and low-ranking cows have to wait longer for their turn to get access to the milking unit in AMS (Ketelaar-de Lauwere et al., 1996). Stefanowska et al. (1995) observed ML in some cows that had to wait to get access to the AMU. Preliminary data (K. SvennerstenSjaunja, personal communication) from a parallel study indicate that low-ranking cows in the AMS had higher oxytocin concentrations and spent more time close to the milking unit. Moreover, those cows had more observations of ML than high-ranking cows (K. SvennerstenSjaunja, unpublished data). Milk leakage was mainly observed when cows were lying down, but it is not known if social dominance can influence lying behavior. However, a long milking interval can result in a shorter time lying during the hours before milking, probably ¨ sterdue to the discomfort when lying on a full udder (O man and Redbo, 2001). It was 3 times more common that cows in AMS leaked milk ≥3 times, indicating a more persistent problem in these cows. Parity was not an explanatory factor for this finding because the proportions of primiparous and multiparous cows with multiple ML were similar. However, problems related to the previous milking were probably of importance. Missed or incomplete milking was registered in association with a large number of ML observations, and in almost 60% of the affected cows. However, in several cases, ML did not occur in the same quarter registered to have a problem at milking. Milk leakage was also a common finding when missed milking was simulated experimentally (Stefanowska et al., 2000). Milk leakage occurred most often in the rear quarters and when cows were lying. The proportions of cows

MILK LEAKAGE IN AUTOMATIC MILKING SYSTEMS

ruminating or resting/sleeping when lying were similar, indicating that this did not influence the occurrence of leakage, even though an increased release of oxytocin has been observed during stimulation of the oral cavity (Svennersten et al., 1990). The increased pressure from the hind legs on the rear part of the udder when lying could explain these findings. This is also consistent with the higher risk of udder infections and mastitis in rear quarters (Adkinson et al., 1993; Lancelot et al., 1997). Proportions of cows lying and standing at the observations did not differ among the barns, indicating that this was not an explanatory factor for the higher ML number in AMS. No influence of estrus on ML was observed. Such relationships might have been expected because inhibition of the oxytocin release and milk let down can occur at estrus, increasing the risk for incomplete milking (Wellnitz and Bruckmaier, 2000). In CMS, ML was mainly observed at a few of the observation points, whereas the ML in AMS occurred at all times. The time between a specific ML observation and the previous milking varied in AMS. Many (26 and 19% for yr 1 and 2, respectively) of the ML observations occurred ≤4 h after milking. Almost half of those occasions were explained by missed or incomplete milking of one or several quarters at the previous milking. Such problems were also associated with one-third of the ML observations occurring 4 to 8 h after milking. In 15% of the ML observations, the time from the previous milking was >12 h, indicating that cow traffic was suboptimal. Long milking intervals can be a result of social dominance because low-ranked cows might have to wait longer to the next milking than high-ranked cows for their turn at the AMU (Ketelaar-de Lauwere et al., 1996). In this study, time from milking to ML observation did not differ between primiparous and multiparous cows. A cow’s motivation to be milked and/or fed is another important factor, which is of importance for the length of the milking interval in AMS. The milk flow rate of a quarter is influenced by teat canal anatomy (Hamann and Dodd, 1992) and can be considered an indirect indicator of the patency of the teat canal. Both the mean and the peak milk flow rate were significantly higher in quarters with ML than in quarters without ML, and increased with increasing numbers of ML per quarter. The research farm at Kungsa¨ngen is run as one breeding unit, and the cows are randomly selected to the different barns. It is therefore not likely that the cows’ milk flow rates were different in the AMS than in the other systems. Thus, cow differences among the barns in milk flow rate could not explain the higher incidence of ML in the AMS. Quarters with high milk flow rates have been associated with an increased risk of infections and mastitis (Dodd and Neave, 1951; Slettbakk et al., 1990; Grindal

3495

and Hillerton, 1991; Grindal et al., 1991; Lacy-Hulbert and Hillerton 1995), and ML has been reported to be an important risk factor for clinical and subclinical mastitis (e.g., Jørstad et al., 1989; Schukken et al., 1990; Elbers et al., 1998; Waage et al., 1998). In the present study, the SCC and UDS were numerically, but not significantly, higher in cows with ML. Thus, there was no clear relationship between ML and udder health. Although the observed positive association between the occurrence of ML and an increase in UDS from 0 to 4 could indicate that ML may increase the risk for new infections, no such relationship was observed between ML and UDS higher than 4. The risk of udder infections in association with ML increases when the hygiene in the cows’ environment, especially in the bedding, is poor (Schukken et al., 1990, 1991). In the research farm studied, udder health and hygiene was very good, which probably explained the small effect on SCC and the low number of clinical mastitis cases. The high occurrence of ML in AMS can to some extent be explained by problems related to disturbances at milking or overly long milking intervals, indicating technical or management problems. However, as mentioned previously, these problems can only explain a part of the ML observations. Milk leakage can occur when the milk let down reflex is activated due to oxytocin secretion. The reflex is normally activated by stimulation of the teats and udder before milking by the milker. However, visual or auditory stimuli of the calf or of the milking process (Peeter et al., 1973; Pollock and Hurnik 1978) can also activate the reflex. In AMS, the cows get more or less constant visual and/or auditory stimuli from the AMU, which could stimulate oxytocin release and milk let down. This, in combination with a higher milk flow in certain quarters may be one explanation to the increased risk of ML in the AMS. However, more studies are needed to investigate the relevance of this hypothesis. Moreover, the effect of social dominance, especially in primiparous cows, and of quarter milking, milking frequency, and the level at cluster take-off on the occurrence of ML in AMS are other factors that need to be studied. The present study was performed during a limited period during the indoor season on a research farm in which the degree of activity was different from that of commercial farms. It is not known if season and activity can influence the occurrence of ML. Preliminary data (our unpublished results) indicate that less ML occurs during weekends compared with weekdays, when more activity is ongoing in the barns. Thus, further studies are also needed on the effect of season and on the occurrence of ML in commercial AMS dairy farms. Journal of Dairy Science Vol. 86, No. 11, 2003

3496

PERSSON WALLER ET AL.

CONCLUSIONS In conclusion, the risk for milk leakage between milkings was higher in an AMS than in either a free-stall or tie-stall housing system using a conventional milking parlor with fixed milking intervals. Overall, ML was not related to milk production, parity, stage of lactation, or estrous status. However, a larger proportion of the primiparous cows leaked milk in AMS than in the other systems. Milk leakage occurred more often in rear quarters. At most observations of ML, the cows were lying down, and times from previous milking varied considerably, especially in AMS. In AMS, about one-fifth of the observations occurred ≤4 h after milking, and half of those were associated with missed or incomplete milkings. Moreover, milk flow rates were higher in quarters leaking milk. More studies are needed on the etiology of ML and on the occurrence and possible effects on udder health of ML in commercial farms. ACKNOWLEDGMENTS Financial support from the Swedish Farmers’ Foundation for Agricultural Research is gratefully acknowledged, as is the help with statistical analyses by P. ¨ hagen, Swedish University of Agricultural Sciences. O REFERENCES Adkinson, R. W., K. H. Ingawa, D. C. Blouin, and S. C. Nickerson. 1993. Distribution of clinical mastitis among quarters of the bovine udder. J. Dairy Sci. 76:3452–3459. Berglund, I., G. Pettersson, and K. Svennersten-Sjaunja. 2002. Automatic milking: Effects on somatic cell count and teat end-quality. Livestock Prod. Sci. 78:115–124. Brolund, L. 1990. Technical utilisation of cell count in the milk recording service. Commun. No. 161, Swedish Dairy Assoc., Eskilstuna, Sweden. Dickson, D. P., G. R. Barr, L. P. Johnson, and D. A. Wieckert. 1970. Social dominance and temperament of Holstein cows. J. Dairy Sci. 53:904–907. Dodd, F. H., and F. K. Neave. 1951. Machine milking rate and mastitis. J. Dairy Res. 18:240–245. Elbers, A. R. W., J. D. Miltenburg, D. De Lange, A. P. P. Crauwels, H. W. Barkema, and Y. K. Schukken. 1998. Risk factors for clinical mastitis in a random sample of dairy herds from the southern part of Netherlands. J. Dairy Sci. 81:420–426. Everitt, B., T. Ekman, and M. Gyllenswa¨rd. 2002. Monitoring milk quality and udder health in Swedish AMS herds. Pages V72–V75 in Proc. First N. Am. Conf. on Robotic Milking, Toronto, Canada. Grindal, R. J., and J. E. Hillerton. 1991. Influence of milk flow rate on new intramammary infection in dairy cows. J. Dairy Res. 58:263–268. Grindal, R. J., A. W. Walton, and J. E. Hillerton. 1991. Influence of milk flow rate and streak canal length on new intramammary infection in dairy cows. J. Dairy Res. 58:383–388. Hamann, J., and F. H. Dodd. 1992. Milking routines. Pages 69–96 in Machine Milking and Lactation. A. J. Bramley, F. H. Dodd, G. A. Mein and J. A. Bramley, ed. Insight Books, Huntington, VT. Hamann, J., and P. Gyodi. 2000. Somatic cells and electrical conductivity in relation to milking frequency. Milchwissenschaft 55:303–307. Hamann, J., and F. Reinecke. 2002. Machine milking effects on udder health—Comparison of a conventional with a robotic milking sysJournal of Dairy Science Vol. 86, No. 11, 2003

tem. Pages 17–27 in Proc. of the First N. Amer. Conf. on Robotic Milking, Toronto, Canada. Hopster, H., R. M. Bruckmaier, J. T. N. Van der Werf, S. M. Korte, J. Macuhova, G. Korte-Bouws, and C. G. van Reenen. 2002. Stress responses during milking; comparing conventional and automatic milking in primiparous dairy cows. J. Dairy Sci. 85:3206–3216. Ipema, A. H., C. C. Ketelaar-de Lauwere, C. J. A. M. de Koning, A. C. Smits and J. Stefanowska. 1997. Robotic milking of dairy cows. Pages 290–297 in Proc. 3rd Int. Tagung: Bau, Technik und Umwelt in der Landwirtschaftlichen Nutztierhaltung, Kiel, Germany. Jørstad, A., T. B. Farver, and H. Riemann. 1989. Teat canal diameter and other cow factors with possible influence on somatic cell counts in cow milk. Acta Vet. Scand. 30:239–245. Ketelaar-de Lauwere, C. C., S. Devir, and J. H. M. Metz. 1996. The influence of social hierarchy on the time budget of cows and their visits to an automatic milking system. Appl. Anim. Behav. Sci. 49:199–211. Kruip, T. A. M., H. Morice, M. Robert, and W. Ouweltjes. 2002. Robotic milking and its effect on fertility and cell counts. J. Dairy Sci. 85:2576–2581. Lancelot, R., B. Faye, and F. Lescourret. 1997. Factors affecting the distribution of clinical mastitis among udder quarters in French dairy cows. Vet. Res. 28:45–53. Lacy-Hulbert, S. J., and J. E. Hillerton. 1995. Physical characteristics of the bovine teat canal and their influence on susceptibility to streptococcal infection. J. Dairy Res. 62:395–404. McDonald, J. S. 1975. Radiographic method for anatomic study of the teat canal: Characteristics related to resistance to new intramammary infection during lactation and the early dry period. Cornell Vet. 65:492–499. Mein, G. A. 1992. Action of the cluster during milking. Pages 97– 140 in Machine Milking and Lactation. A. J. Bramley, F. H. Dodd, G. A. Mein, and J. A. Bramley, ed. Insight Books, Huntington, VT. Mottram, T. T., R. C. Hall, D. S. Spencer, and C. J. Allen. 1995. The role of the cow in automatic teat attachment. J. Dairy Sci. 78:1873–1880. Neijenhuis, F., F. H. Barkema, H. Hogeveen, and J. P. T. M. Noordhuizen. 2000. Classification and longitudinal examination of callused teat ends in dairy cows. J. Dairy Sci. 83:2795–2804. ¨ sterman, S., and I. Redbo. 2001. Effects of milking frequency on O lying down and getting up behaviour in dairy cows. Appl. Anim. Behav. 70:167–176. Peeters, G., E. De Buysscher, and G. Vandevelde. 1973. Milk ejection in primiparous heifers in the presence of their calves. Zentralbl. Veterinarmed. A. 20:531. Pollock, W. E., and J. F. Hurnik. 1978. Effect of calf calls on rate of milk release of dairy cows. J. Dairy Sci. 61:1624. Schukken, Y. H., F. J. Grommers, D. Van de Geer, H. N. Erb, and A. Brand. 1990. Risk factor for clinical mastitis in herds with a low bulk milk somatic cell count. 1. Data and risk factors for all cases. J. Dairy Sci. 73:3463–3471. Schukken, Y. H., F. J. Grommers, D. Van de Geer, H. N. Erb, and A. Brand. 1991. Risk factor for clinical mastitis in herds with a low bulk milk somatic cell count. 2. Risk factors for Escherichia coli and Staphylococcus aureus. J. Dairy Sci. 74:826–832. Seeman, A. 1997. Ja¨mfo¨rande studie mellan fja¨rdedelsmjo¨lkning och heljuvermjo¨lkning med avseende pa˚ mjo¨lkavkastning, mjo¨lkflo¨den, mjo¨lkningstider och spenbehandling. (Comparative study between quarter milking and conventional milking according to milk production, milk flow, machine on time and teat treatment.) MSc Thesis, Swedish Univ. of Agricultural Sciences, Uppsala. Slettbakk, T. A., A. Jørstad, T. B. Farver, and D. W. Hird. 1990. Impact of milking characteristics and teat morphology on somatic cell counts in first-lactation Norwegian cattle. Pre. Vet. Med. 8:253–267. Stefanowska, J., M. Plavsic, A. H. Ipema, and M. M. Hendriks. 2000. The effect of omitted milking on the behaviour of cows in the context of cluster attachment failure during automatic milking. Appl. Anim. Behav. Sci. 67:277–291. Stefanowska, J., B. Sonck, and A. Keen. 1995. Behaviour of cows milked in fixed milking periods with automatic milking system

MILK LEAKAGE IN AUTOMATIC MILKING SYSTEMS on practical farms (in Dutch). IMAG-DLO nota 95-07, Wageningen, Netherlands. Svennersten, K., L. Nelson, and K. Uvna¨s-Moberg. 1990. Feeding induced oxytocin release in dairy cows. Acta Physiol. Scand. 140:295–296. Svennersten-Sjaunja, K., I. Berglund, and G. Pettersson. 2000. The milking process in an automatic milking system, evaluation of milk yield, teat condition and udder health. Pages 277–288 Proc. Int. Symp. on Robotic Milking, Lelystad, the Netherlands.

3497

Van der Vorst, Y., and K. De Koning. 2002. Automatic milking systems and milk quality in three European countries. Pages V1– V12 in Proc. of the First N. Amer. Conf. on Robotic Milking, Toronto, Canada. Waage, S., S. Sviland, and S. A. Odegaard. 1998. Identification of risk factors for clinical mastitis in dairy heifers. J. Dairy Sci. 81:1275–1284. Wellnitz, O., and R. M. Bruckmaier. 2000. Central and peripheral inhibition of milk ejection. Livest. Prod. Sci. 70:135–140.

Journal of Dairy Science Vol. 86, No. 11, 2003