Skip to main content

Implications of the introduction of automatic milking on dairy farms


This work package reports on the socio-economic aspects the introduction of automatic milking systems on dairy farms. A survey among 107 farmers who have invested in an automatic milking (AM) system in Belgium, Denmark, Germany and the Netherlands was carried out. A questionnaire was designed to capture the characteristics and motivations of AM farmers, farm characteristics and the implications of the introduction of an AM-system. The results summarize the motivation and characteristics of farmers investing in an AM-system and the implications for labour use and the farmers’ quality of life. The average AM farmer is male, married, 44 years old and has about 3 children. On average, AM farms have about 87 dairy cows. Two thirds of the interviewed AM farmers state social reasons for investing in an AM-system, such as increased labour flexibility, improving social life and health concerns. AM farmers reported an average labour saving of 19.8%, which increases to 21.3% when only farms that have kept their herd size more or less constant are considered. Only few farmers disagree that AM has improved their health, while about one third has experienced no changes on their health. Two thirds of all farmers state that their quality of life has increased. Noteworthy is that the improvements in quality of life are relatively unaffected by the chosen growth strategy.
Prevention of antibiotic residues in milk is necessary to ensure health protection of consumers and to avoid failures during milk processing. In practice residues of antibiotics in milk are most often determined by microbial inhibitor tests. Treatment of mastitis with antibiotics is one of the most important causes associated with positive results of inhibitor tests on bulk tank milk. In conventional milking management factors like accidental milking of treated cows, not attending the withholding time and failures during milking are the main reasons for antibiotic residues in bulk milk together with insufficient cleaning of milking equipment. In automatic milking like in conventional milking management failures are reported as the main reasons for residues in milk. Little is known about the impact of milking intervals on the excretion of antibiotic residues in milk. Previous investigations were mainly focused on the effect of stripping on concentrations of antibiotics in milk after treatment of mastitis. The aim of the study reported here was to determine the excretion of antibiotics in milk in dependence on milking frequency. Four drugs containing six different antibiotics were studied in healthy cows milked with three different milking frequencies. The concentration of residues in milk at the end of the withholding period was compared to Maximum Residue Limits (MRL) set by Council Regulation 2377/90/EEC. Excretion trials: Excretion trials were performed under experimental conditions simulating milking frequencies as observed in AM systems. The healthy cows were treated with one injector for each quarter (worst case) in intervals according to manufacturers instructions. Cows were milked 2 times per day (interval 14/10 hours (reference), 3times per day (interval 8h) or 1.5 times per day (interval 16h). One drug was selected for further analysis in cows with clinical mastitis. Only the diseased udder quarter was treated. The cows with mastitis were milked twice a day or 1.5times per day. Results and Conclusions: The excretion time of antibiotic residues in milk of healthy cows varied with different milking frequencies after intramammary treatment. For three of the tested drugs significantly shorter excretion times were observed in cows milked 3 times per day compared to cows milked 1.5 times per day. Longest excretion times were observed for all antibiotics except for one drug in cows milked less frequently with 1.5milkings per day. Treatment trials in cows with clinical mastitis showed large variation in excretion times of individual cows and much shorter excretion times than in worst case studies in healthy cows. Although no significant differences were determined between groups milked 2 times or 1.5 times per day it should be considered that only one drug was tested in a low number of cows. A more pronounced influence of milking frequency may be observed for other drugs. Although results from studies in healthy cows cannot necessarily be transferred to cows with clinical mastitis, controlled milking of cows at least 2 times per day is recommended after treatment with antibiotics. If cows in AM systems are milked in shorter intervals after treatment higher concentrations of residues in milk have to be expected. This could increase the risk for carry over of antibiotics into milk of the next cow milked at the same place, especially if failures in cleaning procedures occur. According to the results presented, prolonged excretion of residues in milk of individual cows in connection with milking frequencies deviating from regular milking times twice per day, seem to be of minor importance for positive results of inhibitor tests in bulk tank milk, when recommendations for treatment are followed.
The quality of milk is a very important aspect of milk production. Milk payment systems and consumer acceptance are, to a great extent, based on it. In dairy processing, milk quality can be a decisive factor for the value of the product. Features in automatic milking like a 24 hour operation, more frequent milking at different intervals, different cleaning procedures and complicated visual control, may influence milk quality. Changes in milk quality after the introduction of automatic milking systems (AM-systems) on dairy farms in The Netherlands, Germany and Denmark were examined and the data were compared with milk quality results of farms with conventional milking technology. After introduction, a small, but significant increase in total bacterial count, somatic cell count, freezing point and free fatty acids was observed. The highest levels for total plate count and cell count are found in the first six months after introduction. After this period the milk quality slightly improves to a more stable level. Risk factors related with milk quality concern general farm characteristics, animal health, AM-system, cleaning and cooling, housing, management skills of the farmer and the hygiene on the farm. Total plate count was significantly related to milk yield of the herd, cleaning of the area around the AM-system and the overall hygiene on the farm. Bulk milk somatic cell count appeared to be significantly related to milk yield of the herd and the number of milkings before replacement of the liners. An increased milking frequency is not the only explanation of increased free fatty acid levels. Technical factors related to free fatty acids mainly concern the air inlet in the teat cups, bubbling (excessive air inlet) and a too long post run time of the milk pump. Precautionary measures like adjusted milking frequencies for cows in late lactation, and preventive maintenance will contribute to overcome free fatty acid problems. However, several questions regarding the causes and background of increased free fatty acid levels remain unclear.
One of the major concerns with automatic milking is whether it can be successfully combined with grazing. The main reason for that concern is that pasture is an important feed source for cows in many countries. During recent years the interest for animal welfare and health has also led to an increased interest in pasture and grazing. The role of grazing animals for the amenity of the landscape and for biodiversity must also be considered, as well as the need to find management systems that include grazing for organic farmers interested in automatic milking. On-farm studies and surveys performed on AM farms with grazing have shown that there is a large variation among farms in grazing time allowed. In a Danish on-farm study no significant relationship between the time cows spent on pasture (5.4-20.6 hrs) and the total number of milkings per 24 hours (2.1-2.8) was found. In a Dutch experiment with 24 lactating cows however, where zero grazing was compared with 12 and 24 hours of grazing, animals with 24-hour access to pasture had a significantly lower average milking frequency (2.3milkings/day) than the other treatments (2.5-2.8milkings/day). On farms where the cows had free admission to the pasture area most of them went out immediately after the door was opened in the morning, showing that they were eager to use the opportunity to graze. There were fairly large differences between the amount of time that was spent fetching cows late for milking, but farms normally used around 10-20 minutes daily. In experiments with about 64 cows, the effects of distance between pasture and barn and level of roughage supplements (maize silage) were studied during 2002 and 2003. No significant effects of distance between barn and pasture or the level of maize silage supplements on milk yield, milking interval or number of fetched animals were found. In another experiment that also focused on the effects of distance and level of roughage supplements 24 hour grazing was applied throughout the season. The group grazing near the barn had a significantly higher milk yield (kg), longer pasturing time and shorter milking interval than the group grazing at a longer distance. To encourage cows to return to the barn regularly, many farms that practice AM milking and grazing have chosen to offer the cows drinking water in the barn only. However, it has also been argued that this practice may lead to a lower water intake, which in turn could lead to a lower milk yield. There have also been concerns about animal welfare. The effect of offering cows drinking water in the barn only was compared with offering drinking water both in the barn and out on the pasture in two experiments. No significant differences in water intake, milk yield or milking frequency were observed. Two cow traffic systems were compared during the grazing season, free and controlled cow traffic. At the start of the grazing season the herd was divided into two groups. Free cow traffic gave a significantly lower number of milkings (per cow and day) and a significantly lower number of voluntary milkings compared with controlled cow traffic. The average number of animals to be fetched increased as the season progressed and was higher for the cows with free than with controlled cow traffic. A trend for decreasing visiting and milking frequencies was seen over time. Possibilities of using individual auditory signals to stimulate cows to go to the milking unit were explored. Preliminary results indicate that animals need a longer training period and perhaps a better reward system if auditory signals are to be used to entice the animals to go to the milking unit. General conclusions. It is possible to combine automatic milking and grazing. The utilization of the milking unit may be slightly lower during the grazing season, especially when animals are allowed to graze many hours per day due to a more uneven distribution of milkings over the 24-hour period. To maintain smooth cow traffic it is therefore advisable not to maximise the number of lactating animals in the barn during the grazing months. In the long run, it appeared beneficial in terms of labour not to be too eager to fetch the cows but to give them the chance to return voluntarily. Good production results can be obtained with longer distances between barn and pasture but distance may affect number of cows that need to be fetched. There are indications that feed supplementation strategy can be used as a management tool to achieve well-functioning cow traffic during the grazing season.
Within hygiene management in dairying effective teat cleaning before milking is a precondition to ensure high quality of raw milk. The teat cleaning efficiency of Automatic Milking (AM) systems was investigated applying different methods. All brands currently used in practice were included: DeLaval, Insentec, Lely Industries, Fullwood, Prolion/Gascoigne Melotte and Westfalia Landtechnik GmbH. Significant differences in teat cleaning efficiency of different brands were determined (p<0.05), showing the necessity to improve teat cleaning efficiency of at least two brands. Variation between individual farms was of significant influence on teat cleaning efficiency with differences between the 3 farms working with the same AM system (p<0.05). Also of influence was the initial contamination of teats before cleaning. These results indicate that farm management is important to ensure low teat contamination levels. High coliform counts in bulk tank milk exceeding 100cfu/ml suggested insufficiencies of teat cleaning on 8 farms, although in some cases these were accompanied by high counts of thermoduric bacteria indicating failures in system cleaning. On 18 farms with AM systems management aspects with regard to teat cleanliness were studied by means of a questionnaire-based interview with farmers and a check-list on the actual hygiene status of barns. The outcome was compared to teat contamination measured by determination of TBC (total bacterial counts) and ATP (adenosine-tri-phosphate) from teat swabs taken before teat cleaning. AM specific management factors associated with high teat contamination were: replacement of teat cleaning device less than twice per year, moderate/poor status of the teat cleaning device, average milking frequency per day < 2.5 and no selection of cows for robot acceptance (p<0.10). Factors not directly related to AM involved contamination of cubicles: less than one cubicle per cow, cows lying on alleys present in the herd, addition of fresh bedding material less than twice per day, no selection of cows for udder health, moderate/poor status of bedding material and moderate/poor status of claws were significantly related to high teat contamination. Additional factors like the general impression of the robot, cleaning frequency of the milking box, status of teat cups and the use of cow brushes in the stable were probably more closely related to the general attitude of the farmer towards hygiene than to teat cleanliness.
One of the main objectives within this work package was to investigate the state of public opinion with regard to automatic milking. In order to assess the central role of the mass media in this respect a descriptive content analysis of newspapers from six countries (Belgium, Denmark, Germany, The Netherlands, Sweden, and United Kingdom) was conducted in order to establish which events become news and to assess the ways in which newspapers cover agricultural and food issues. Content analysis is, “a research technique for the objective, systematic and quantitative description of the manifest content of communication”. The analysis was designed to examine news coverage in each country with regard to food production (especially milk and dairy products) and to identify specific national issues and concerns. Agricultural news items per country. In neither of the two Dutch language Belgian newspapers, with 313 articles covered in total, were any news items related to automatic milking found. In the two French language Belgian newspapers 105 articles were recorded. In the period studied neither featured a single news article on milk or dairy production. In the Two Danish newspapers no less than 427 items were recorded, easily the highest number in the six countries studied. Two large articles relating to automatic milking were found, both in Jyllands Posten and both positive. In the two German newspapers 95 articles were recorded, the lowest number in the six countries studied. No articles dealing with dairy production were found. In the two newspapers from the Netherlands 131 articles were recorded. Only one news item dealing with dairy issues was found - dealing with labelling of milk from free grazing cows. In the two newspapers from Sweden 186 articles were recorded. One large article about automatic milking was found in Svenska Dagbladet and it was positive. Moreover, one article dealt (positively) with the fact that new farm technology provides farmers with more free time. In the two British newspapers 128 articles were recorded. No articles dealing specifically with dairy production were found in either newspaper. Issues of general concern. The analysis in each country shows varying particular national concerns, reflecting the importance of local events. However, by aggregating the results across the six countries one obtains an indication of which issues are of more general concern. The list of five most frequently dealt with topics in the 14 newspapers is, interestingly, topped by the issue of GMO’s with 15% of all items, followed by BSE and akrylamid (accounted for by the large number of items in the Swedish newspapers), both with 12%. The economics of farming comes in fourth place (11%), followed by Nitrofen (11%). In the distribution of news items across more general categories one issue in particular dominated the news coverage of the sector: food scares/safety which accounted for just over a half of all news items and which was dominated by akrylamid, nitrofen, PCB and FMD. In second place comes animal welfare (9%) dominated by stories of animal mistreatment, especially in relation to transport of livestock. In third and fourth place came GMO’s and gene technology (both around 9%). Milk in the news. In the year studied automatic milking was virtually absent from the news contained in the 14 European newspapers studied. Of the 1,385 news items concerning agriculture and the food chain recorded only 3 dealt with this new technology. In terms of informing and sensitising the public to the advantages offered by automatic milking this result is not favourable. On the other hand, given the fact that much news is ‘bad’, in the sense of events being negatively evaluated by journalists (41% of all items were evaluated as negative, compared to 21% positive), it also implies that the development is not an issue and is perhaps best left alone. Moreover, milk and dairy products as a whole appear only infrequently in the news and seem largely to have escaped the bad press received by some other sectors. However, while milk as such escapes, issues surrounding livestock do receive significant (and negative) coverage - particularly BSE, foot and mouth disease, GMO’s and pesticides in fodder, and animal welfare. GMO’s and animal welfare in particular appear to be widespread concerns in the press being kept visible, at least in part, by powerful pressure groups. What is evident from these results is that the press is extremely sensitive to scares over food safety. Consequently, if for any reason one should ever break loose over milk quality, especially in relation to automatic milking, the damage to the imago of the sector in the short and medium term might be considerable and difficult to redress.
This work package and its results give an overview of the work done including definitions of normal and abnormal milk, suggestions for reference methods, and tests of AMS models according to these requirements. The general conditions of hygiene in milk production in the EU are defined by the Commission Directive 89/362/EEC (1989) but not all elements apply to automatic milking. The following text is proposed to be included in the coming EU Hygiene Directive: "Milking must be carried out hygienically ensuring in particular: - that milk from an animal is checked for abnormalities by the milker or a method achieving similar results and that only normal milk is used for human consumption and that abnormal, milk with a withholding period, and undesirable milk are excluded". The definition of abnormal milk caused by clinical mastitis is proposed to be based on the homogeneity of the milk and not on the colour since the colour of the milk changes with breed, stage of lactation, feedstuffs etc. The reference method is suggested as filtration of the milk through a filter with a pore size of 0.1 mm, and milk where clots are clearly visible in such a filter is then defined as being abnormal. Incidences of watery and yellowish milk may or may not be detected by this method. The current AMS models have systems to produce alarm lists of cows that should be checked for abnormalities in their milk and at present the systems are not intended for automatic diversion of milk. Five different AMS-models were tested in six herds. The sensitivities for detection of abnormal milk in the six herds varied from 13 to 50% when calculated for the actual milking, from 22 to 100% for the test days, and from 43 to 100% when calculated for the previous week. Specificities for the same time periods were found to be 87-100%, 85-100%, and 35-100%, respectively. The sensitivities and specificities are generally too low for automatic diversion of abnormal milk and it seems as if most of the models could benefit from application of more sophisticated algorithms or measurements more directly related to the definition of abnormal milk.
In 2000, the book Robotic Milking was published, reflecting the proceedings of an International Symposium, which was held in The Netherlands. At that time, commercial introduction of automatic milking systems was no longer obstructed by technological inadequacies. Particularly in a few west-European countries, systems were being installed at an increasing rate. However, it was recognised that the changeover from “traditional” to automatic milking affected the farming operation, herd management and control of milk quality profoundly and that many of the implications were still unknown. So, new challenges in various fields of dairy farming and new research areas emerged. Since this previous International Symposium, much has happened. In general automatic milking has been adopted as a realistic alternative for milking in the milking parlour. Systems have gradually been improved and, maybe even more important, farmers have become more familiar with their potential and limitations, both technically and in herd management. The level of scientific knowledge on automatic milking has increased largely as well because of research efforts all over the world. A significant share of these efforts has been made within the framework of a EU-granted project on the implications of the introduction of automatic milking on dairy farms with participation of seven research institutes and six industrial companies from six countries. The Proceedings reflect the knowledge on automatic milking generated all over the world in the last few years. Its contents can therefore be regarded as the present state of knowledge in the field of automatic milking, for a better understanding. Proceedings of the International Symposium Automatic Milking, a better understanding, organised in Lelystad, The Netherlands, and March 2004. Edited by A. Meijering, H. Hogeveen and C.J.A.M. de Koning. ISBN 9076998388, Wageningen Academic Publishers.
The cleaning of an automatic milking system is of great importance for the hygienic quality of the milk. From conventional milking several methods and procedures are known and applied to automatic milking systems. The diversity of the cleaning systems of automatic milking systems is given. Important differences do occur in procedure (circulation cleaning versus cleaning with boiling water), in time needed for cleaning and in water and energy consumption. Cleaning time differs between 37 and 130 minutes per day for standard cleaning frequencies, water consumption differs from 284 to 495l/day. The effect of the cleaning frequency was investigated by a comparison between 2 and 3 system cleanings per day during 9 weeks each. When three system cleanings were performed, bacterial quality of the milk was significantly better for TBC, coliform, thermoduric and psychrotrofic bacteria. The difference for TBC however was small: 13 versus 10x10{3}cfu/ml. So, with both frequencies the average TBC is far within the range for first quality milk. It is concluded that two cleanings per day could be sufficient for a good quality of milk, but the cleaning system has to be optimised, well maintained and controlled. Three times cleaning however reduces the risk of increased bacterial growth in the milking system. To prevent transfer of pathogens by the milking cluster, in AM-systems have a procedure for a cluster flush. The effect of the cluster flush on removal of pathogens is tested on one system, with cold water and with a disinfectant. 98,4 and 98,9% of the pathogens were removed, expressed as log-reduction 1,80 and 2,26. The extra removal of pathogens by using a disinfectant does not compensate for the extra risk for contamination of the milk produced. In a test on the effect of a cluster flush on the rate of new infections, 46 cows were milked with a deliberately infected cluster. Two cross positioned liners were flushed after infection, the other two liners were not flushed. In none of the quarters an inflammation occurred. The effect of a cluster flush on new infections could not be established. Because cluster flush does remove most of the bacteria from the liners, does not influence the milking capacity of the AMS and can be performed with small amounts of water, there are no reasons not to perform a cluster flush.
It is generally recognised by now that with automatic milking operational farm management is considerably changed compared to conventional machine milking, and is a key issue to fulfil requirements in practise. An overall picture of changes in management when milking with an automatic milking system however is not available. Each manufacturer has different solutions for management support and performance of automatic tasks, and users can adapt various parameter settings to their own preferences. This report describes differences between conventional and automatic milking with respect to operational farm management. From these differences a list of demands for operational management with automatic milking systems is generated and compared with the possibilities of currently available systems. Opportunities and shortcomings are indicated and discussed. A clear difference with conventional milking is that milking intervals have to be controlled for individual cows. Feeding strategy is a key element in this, especially when grazing is applied. Usually cows with too long intervals are fetched. With regard to health, in general much more indirect information is available, especially detection of mastitis partly depends on abnormalities detected with sensors. However, visual inspection of the animals remains an important method to control health. Because milkings are unattended, regularly abnormalities have to be checked. Because of the limited reliability of the alarms, and the fact that these are not yet integrated, this requires specific skills from the farmer. The automatic milking unit has to be maintained and its cleaning and functioning must be controlled more or less constantly. The cooling system has to be suited to automatic milking. In order to be able to comply with legislation some of the currently existing regulation must be adapted, avoiding double standards. Farmers must have affinity with automation. They have to work with secure schedules, for instance first enter treatment data in the computer and then treat sic cows and respond as should on alarms for system and animals. Udders and teats must be kept clean. Farmers in general are satisfied with the current possibilities of AM systems and do not have a clear view which further improvements are possible. Despite this, not all of the demands for automatic milking are yet fulfilled by the current systems but this is also due to insufficient legislation. Especially automatic separation of abnormal milk and secured teat cleaning should be realised. Furthermore monitoring of the equipment can be improved. With these improvements automatic milking potentially has advantages with respect to milk quality and food safety compared to conventional milking.
The welfare of dairy cows depends on several factors related to social interaction between cows, the management and the environmental conditions. In addition to milking, cow traffic systems and feeding are important issues. While free cow traffic gives most freedom to the cows, long and irregular milking intervals among individual cows counteract the success. Forced traffic results often in unsatisfactory queuing in front of the milking unit and too few feeding visits. Controlled traffic, which allows access to some basic feeds between milking without passing the milking unit, results in better overall welfare conditions. The cows´ interest to go to the waiting area in front of the milking unit depends on the time since last feeding and the number of cows already in the waiting area. Low-ranked cows have to wait longer than the high-ranked when queuing, and generally choose time periods when the high ranked are less active. Studies on stress-related hormones do however not reveal that they suffer. A concept for assessing animal welfare at herd level has been developed as a tool for the AM-herd, including information on the systems, systems application, animal behaviour and animal health status. The concept has been applied on 8 AM-herds during a year. Each herd received an annual report, which were discussed at the farm. The report included a short overview, a list of indicator values with reference to results from other farms, and a detailed documentation of raw data from the farm. Major herd differences were found for many indicators. The farmers in general appreciated the tool as a relevant decision support system.
In Denmark, the Netherlands, and the UK, 15 herds each were recruited for monitoring the impact of transition to automated milking on animal health. The herds recruited represented the types of AMS marketed in each country. Each farm was visited at least twice before installation of the AMS and a minimum of twice, but often up to six times, after installation. On these visits assessments were made of at least half of the cows or fifty animals on body condition and locomotion, and forty cows for teat condition (on some farms in the Netherlands and UK only). Farm data including milk production, milk quality, animal records on individual cow cell count, fertility, animal treatments, animal movements, veterinary purchases etc. Additionally data were collected on all other infections, diseases and conditions likely to affect milk quality, production, and attendance at the automated milking system and involve special staff time. The body conditions varied more between countries than in response to the introduction of AM. In Denmark and the UK there was no change in body condition between 3-6 months prior to AM installation and 6 months post installation. A slight but not significant drop occurred with the Dutch cows. On the Dutch farms the range of body condition narrowed significantly from 1.35 to 0.98 points score suggesting that the farms are managing body condition, probably feeding, better. However, most of the change was due to the farm with the thinnest cows increasing condition by three quarters of a point. In the UK the range increased mostly because the thinnest cows became thinner and the fattest cows became fatter suggesting an exacerbation of the lack of adequate feeding management on two of the poorer units. No change in locomotion had occurred by one month after AM installation. The scores had increased slightly, but not a significant difference, in both countries by 3months after AM installation. The ranges also increased and the average score increased on seven farms whilst unchanged on 6 farms in the UK. Scoring was continued on 12 of the UK farms. Twelve months after installation of AMS the lameness has increased significantly. Prior to installation eleven of fourteen UK herds were grazed but only six after installation. The poorer locomotion may reflect the increase in constant housing. The overall impact of conversion to AM may be assessed by comparing how each individual farm has coped in terms of the main indicators of animal health related to the changes in production methods. This has been attempted for the Dutch and the UK data. Comparing 12 Dutch farms only one improved in locomotion, body condition as well as cell counts. This farm had a significant improvement in locomotion score from a poor pre-installation standard. It was one of only two farms that reduced the cell count and the proportion of cows with a high cell count. However, the cows were already some of the thinnest in the study and became thinner so body condition may not truly have improved. Overall, little change was apparent. Locomotion improved in five herds and deteriorated in five herds. Body condition score decreased in eight herds but only by a small amount. It increased in two herds but not making the cows any fatter, just more typical. The only major deterioration was in average milk cell count and the proportion of cows with a cell count above a threshold, where only two of the herds produced better quality milk. Average milk yield in the Dutch herds decreased in continuation of a trend starting up to 12-months prior to installation of the AMS and the cows became thinner with only a small reduction in DIM. Data on herd fertility are only available, so far, from the UK and then only from 6 herds. Overall there is little evidence of major changes occurring in the common measures of fertility. None of the changes were statistically significant but all suggestive of poorer fertility, at least in the transition period from conventional milking to AM. Conclusions: No major problems in converting from conventional milking to AM have been identified but equally none of the 44 farms has been found to achieve a substantial improvement in any aspect of cow health. Given that most of the farms studied were confronted by many and varied problems in conversion, with no clear advice readily available, they have all achieved a remarkable success. The transition period to AMS comprises a period of higher risk to health that extends from weeks before installation when resources start to be diverted from cow management. The length of the transition will vary on individual farms related to many unique factors. Several potential problems may develop in the longer term and anticipation of these is necessary. Clearly AMS succeeds but its longer-term promises for animal welfare and milk quality are unfulfilled to date.