Organic farming promotes a combination of providing good-quality feedstuffs, no use of chemical products, appropriate livestock husbandry systems and correct management practices to deal with the principles of health, ecology, fairness and care ( IFOAM, 2012 ). Organic systems are highly dependent of the environment ( Falconer & Mackay, 1996 ), and to deal with these principles the European Union organic regulation ( EC, 2007 ) specifically states that the choice of breeds should take into account their capacity to adapt to local conditions, their vitality and their resistance to disease or health problems.

During the last half of the 20 ^th century, dairy genetic selection was nearly exclusively focused on the Holstein-Friesian breed to greatly improve milk production. The resulting cow could be considered a high maintenance animal for highly standardised intensive systems, an environment that clearly differs from that of organic production, mainly regarding feeding regimens and medical treatments ( Ahlman, 2010 ). In parallel, this selection process has resulted in reduced reproductive efficiency, extended calving intervals, increased health problems, increased culling rates, and decreased productive life ( González-Recio et al ., 2006 ; Bluhm, 2009 ) especially when these highly selected cows are out of intensive nutritional and environmental management systems. The most notable exception to the high-producer Holstein-Friesian genetic selection was carried out in Australia and New Zealand, countries with seasonal, site adapted, and pasture-based milk production systems ( Macdonald et al ., 2008 ; Baudracco et al ., 2010 , 2011 ) which are closer to the organic principles.

In this context, as organic dairy farming was growing in Europe, farmers realised that the available cows were not adapted to the organic requirements. Cows used in organic farming were similar to those in conventional systems ( Endendijk & Baars, 2001 ; Nauta, 2001 ) and their suitability for organic productive conditions was questionable in that their potential performance under organic systems was not optimal ( Diepen et al ., 2007 ). With the aim of solving this problem, experts over the world propose to use Holstein-Friesian cows adapted to organic systems ( Baars & Endendijk, 1990 ; Endendijk & Baars, 2001 ; Brotherstone & Goddard, 2005 ), or alternatively other purebred cows ( Sundrum, 2001 ) or their crosses ( Nauta et al ., 2005 ; Rozzi et al ., 2007 ; Begley et al ., 2009 ).

In the Netherlands, a survey by Nauta et al . (2006a) showed that different production and marketing strategies demand different breeds: some local or native breeds adapt well to the organic systems because they utilize lower quality feed, are more resilient to climatic stress, and are more resistant to local parasites and diseases. Interest in crossbreeding has grown over the past decade among both producers and researchers ( Heins et al ., 2006b ) and crossbreeding of Holstein-Friesian and local breeds could help to minimize the decline of health, longevity and fertility by introducing favourable genes, removing inbreeding depression, and taking advantage of heterosis ( Lopez-Villalobos et al ., 2000 ; Rozzi et al ., 2007 ); however this option needs to be further investigated before used on a large scale ( Ahlman, 2010 ).

Organic milk production in Spain is small (16,413 tons/year; MAGRAMA, 2014 ) but has been growing rapidly in recent years. Organic dairy farmers are not satisfied with the productive performance of their cows, nearly 60% of them would like to have a better adapted breed ( Rodríguez-Bermúdez et al ., 2015 ). However, due to the difficulty of getting technical advice for acquiring suitable animals, they kept maintaining the same cows as when they were in conventional farming; only a minority (18%) started to incorporate new breeds or use crossbreeding.

The objective of this paper was to analyse the milk productive performance (milk yield, nutritional composition, and somatic cell count) of Holstein-Friesian cows in organic farming in North Spain compared with conventional farms in the region. When breed diversity was present in the same organic farm, milk productive performance of Holstein-Friesian was also compared to the other breed and/or crosses.

Animal productive performance is the result of the interaction between genotype and environment (G×E), this interaction being specially relevant when animals are reared under particular environmental conditions ( Diepen et al ., 2007 ). When animals are genetically adapted to specific (such as organic production) or extreme conditions they will be more productive and production costs will be lower ( Simm et al ., 2004 ). In this sense, it has been found that G×E caused a re-ranking of Holstein-Friesian animals between organic and conventional systems for fertility traits in Sweden and production traits in the Netherlands ( Nauta et al ., 2006b ), indicating that some Holstein-Friesian cows are better suited for organic production than others ( Ahlman, 2010 ). The results of our study indicate that at the circumstances of dairy farming in North Spain, organic farms produce less milk compared to conventional farms. Overall, organic Holstein-Friesian cows of various LG produced 52 to 59% less milk compared to their counterparts in zero-grazing conventional farms, but only about 10% less than pasture-based conventional farms. Differences in milk yield of Holstein-Friesian cows in organic and pasture-based conventional systems respect to those producing in zero-grazing conventional systems are likely greatly related to nutrition management ( Walker et al ., 2004 ). Horan et al . (2005) observed that Holstein-Friesian cows selected for high production have better milk yield under systems of high concentrate supplementation. As indicated in Table 1 , in our study cows in zero-grazing conventional systems are supplemented with higher concentrate compared with the organic and grazing-based conventional groups. Our results are in accordance with a survey conducted in organic farms across 12 European countries that showed a lower milk yield per cow (5 to 25%) in organic than in conventional farms ( Offermann & Nieberg, 2000 ). In addition, Ahlman (2010) showed similar results (lower production in organic) when considering Swedish Holstein and Swedish Red cows in organic and conventional systems in Sweden.

However, in our study in the organic farms where breed diversity exits, Holstein-Friesian cows produce numerically (but not statistically significant) more milk litres than other breeds and crosses. It is well assumed that in zero-grazing conventional systems Holstein-Friesian cows have higher milk yields compared to other breeds ( Heins et al ., 2006a ; Bjelland et al ., 2011 ), because cows with high genetic potential for milk yield have better responses to supplements (Baudraco et al ., 2010). In organic systems, information on the effect of breed on milk yield is sparse, although seemingly lower than in conventional systems: organic systems are based on pasture utilisation with low concentrate supplementation, a diet that could make interbreed differences less evident. In fact, when comparing different type of diets in conventional systems, Oldenbroek (1988) found that differences between Holstein-Friesian and other breeds, although always present, were more evident with diets with higher amounts of concentrate compared with those based on forages. Moreover, the evolution of the milk yield along the productive life of our organic Holstein-Friesian cows was very similar to that observed in their conventional counterparts. Milk yield production along the productive life has been well studied ( Ray et al ., 1992 ; Mellado et al ., 2011 ; Ríos-Utrera et al ., 2013 ): the increase of milk yield with increased parity is due to differences in the control of tissue mobilization between primiparous and multiparous cows ( Wathes et al ., 2007 ); lower consumption of feed per day of primiparous cows than multiparous cows ( Dado & Allen, 1994 ), and increasing body size of older cows over that of first-lactation animals ( Mellado et al ., 2011 ). In general it is assumed that milk yield per lactation increased significantly until 3 ^rd lactation and is maximum between 4 ^th and 5 ^th lactation ( Ray et al ., 1992 ; Gader et al ., 2007 ; Ríos-Utrera et al ., 2013 ). The fact that older Holstein-Friesian cows (LG-3) on organic farms were able to better maintain the milk yield could be interpreted as a positive fact. Longevity has decreased in conventional dairy populations due to selection mainly on production traits ( Essl, 1998 ) and milk production has been shown to be unfavourably correlated with functional traits (animal health, longevity, and reproduction) ( Ahlman, 2010 ); this situation leads to cows being culled because of health and fertility reasons, so productive life of modern cows is rapidly declining ( Oltenacu & Algers, 2005 ).

When analysing milk composition (namely % fat and % protein) no statistically significant differences were found between Holstein-Friesian cows in organic systems compared with conventional systems, even though both organic and zero-grazing conventional Holstein-Friesian cows produce milk with numerically high fat content (3.88%) compared with cows in pasture-based conventional systems (3.58%). On the contrary, protein concentrations were less variable and differences between production systems were slight. It is well assumed that fat and protein production is highly related to nutrition ( Gordon & McMurray, 1979 ; King et al ., 1990 ; Walker et al ., 2001 ), depending not only on the quantity of forage and supplements but also in the composition of them ( Palmquist, 1993 ); Walker et al ., 2004 ). Consequently, differences in fat and protein production depend not only on the production system, but also along the year and the region. Ahlman (2010) , in a study in Sweden, found lower fat and protein yield in organic farms compared with conventional farms, although the opposite was found in Canada by Rozzi et al . (2007) .

However, when comparing Holstein-Friesian cows with other breeds/crosses in the same organic farm, all breeds and crosses showed statistically significant higher milk protein concentrations than purebred Holstein-Friesian. On the contrary, differences in the fat content were lower, with the only exception of Holstein × Brown Swiss cows in farm#2 that produced milk with a high fat content compared to pure Holstein-Friesian. It is well assumed that in conventional systems Holstein-Friesian cows produce higher milk volumes but with a lower fat and protein content compared to other breeds such as Brown Swiss, Jersey or Swedish Blond ( Oldenbroek, 1988 ; Heins et al ., 2006a ; Vance et al ., 2009 ; Prendiville et al ., 2011 ). Crossbred cows may be beneficial under a milk payment scheme that rewards higher components in contrast to systems where the payment is based on milk volume; in those systems care would have to be taken to select breeds, and bulls within breeds that could maintain milk volume ( Bluhm, 2009 ). In organic farming information of the effect of breed on milk composition is sparse. Similarly to milk yield, organic (pasture-based) diets would make the interbreed influence on milk composition less evident compared with those based on large amounts of concentrate ( Oldenbroek, 1988 ).

The evolution of milk composition along the productive life of Holstein-Friesian cows reared on organic farms is very similar to that observed in their conventional counterparties; in general fat and protein concentrations increase from LG-1 to LG-2 and significantly decreases in the LG-3. The only exceptions were the slight (1%) although significant progressive decrease of % protein from LG-1 to LG-3 for the organic Holstein-Friesian cows and the increase of % fat also from LG-1 to LG-3 for the zero-grazing conventional. It is generally assumed that fat and protein percentages decreased when increasing lactation number ( Schutz et al ., 1990 ; Yoon et al ., 2004 ), and we are not aware of the reasons behind this different result apparently related to the production system. We are conscious that the number of farms in our study is small and these results may be due to chance.

When considering the sanitary status of the farms, no significant differences on the SCC were observed between Holstein-Friesian cows from different farm systems or between Holstein-Friesian cows and other breeds or crosses within the same organic farm. As expected, SCC significantly increased with number of lactation in all production systems and breeds ( Reneau, 1986 ), particularly in cows with a high number of lactations ( Tancin, 2013 ). Numerous studies have compared the SCC in organic and conventional farms worldwide, however no consensus has been found. Even though some studies indicate a lower SCC in organic dairy farms ( Busato et al ., 2000 ; Hamilton et al ., 2006 ), most of them found the opposite ( Hovi & Roderick, 2000 ; Zwald et al ., 2004 ; Roesch et al ., 2007 ; Rozzi et al ., 2007 ; Sundberg et al ., 2009 ) or very little difference between both production systems ( Hardeng & Edge, 2001 ; Vaarst et al ., 2006 ; Fall et al ., 2008 ; Haskell et al ., 2009 ). In addition to clinical and subclinical udder disease, factors like husbandry, management, genetics, and nutrition and associated metabolic and endocrine changes have some influence on the SCC. In this sense, at a hypothetical identical sanitary condition the lower milk yield and the higher mean age of lactating cows in organic farms compared with conventional farms could leave to a higher SCC in the former ( Villar & López-Alonso, 2015 ). Differences in the udder health between organic and conventional herds are not necessarily only due to a direct effect of the organic management, and many other differences in management routines may exist between organic and conventional farms besides the “purely” organic and such differences might vary across studies ( Fall et al ., 2008 ). In fact, in studies conducted using identical experimental conditions, no differences were observed between organic and conventional herds ( López Villalobos & Scott, 2003 ; Fall et al ., 2008 ; Fall & Emanuelson, 2009 ). In our study, comparisons between SCC were always made at the same conditions, particularly number of parities and mastitis management (antibiotic based); however, if the SCC count was made at the bulk tank level it would be likely that SCC would be higher in organic herds because the mean age of the herd is higher, and the milk yield lower in organics ( Table 1 ). Our results indicate that at the conditions of organic production of Spain the udder sanitary status is similar to that of conventional systems along all the productive life, and consequently it does not like that Holstein-Friesian cows could have a higher risk to suffer udder disorders related to a lack of adaptation to the productive system.

In conclusion, the detailed evaluation of milk productive performance of organic dairy farms in North Spain, by comparing Holstein-Friesian cows (the main dairy breed in the region) from different production systems and, within the same organic farm, with other breeds and their crosses, did not show a lack of adaptation of Holstein-Friesian cows to the organic conditions. In organic farms, Holstein-Friesian cows produce slightly less milk (10%, but with a lower ingestion of concentrate) that in pasture-based conventional farms, but milk was similar in nutritional composition and for somatic cell counts. The limited available data from organic farms where breed diversity exists indicate that Holstein-Friesian cows produce numerically more milk than other breeds and crosses but with statistically lower protein content. Considering that in Spain organic milk production is mostly applied to liquid milk consumption and that the payment system is only based on milk volume, Holstein-Friesian cows would better fit the farmer interests than other breeds or crosses; in fact only 13% of organic farmers would consider the milk solid concentration as a relevant trait when selecting their cows. Other advantages of using Holstein-Friesian in organic dairy farming are more available information and fewer difficulties to find replacement heifers compared with other breeds. However, in addition to the productive performance, other parameters related to reproductive efficiency, animal health and consumer’s preferences should be fully considered when choosing Holstein-Friesian or other breeds or crosses for organic dairy. If finally, Holstein-Friesian was the selected breed, efforts should be made to find the cow best adapted to the organic conditions. New productive, reproductive, nutritional and economic studies would be needed to develop a genetic merit index for organic systems. This genetic merit index could be useful not only for organic dairy systems but would also be relevant for pasture-based conventional systems.

The authors thank all the farmers who participated in the study and the staff of regional dairy laboratories for analyzing the milk samples.