No manipulation which harms the well-being of cows surveyed was carried out during our study, only observations have been made.

The present study was carried out in 100 dairy farms in the province of Algiers (Mitidja), located on the north-central coast of the country, with 1200 dairy cows (12.0 ± 7.9 cows/farm), with a minimum of 6 cows by farm and a maximum of 53 having an average daily milk production of 16 L. These cows belonged to different breeds: Holstein (44.6%), Montbeliarde (34.3%), Fleckvieh (9.7%) and Brown Swiss (11.4%), with an average of 2 breeds per farm. The farms visited are conducted in permanent (53.0) or partial (47.0%) tied stall with access to outdoor loafing area (28.0%) or a pasture (19.0%) from spring to summer The choice of the study sample was made from the list of cattle farmers in the province of Algiers. This list contains only the ones selected according to production type (dairy cattle), farms joining the national milk rehabilitation program (which requires the possession of health approval for enabling them to deliver their milk directly to a government processing unit or milk through a milk collector), the number of dairy cows (≥ 6) (minimum to have a health approval), the availability and cooperation of farmers to collect information. The visits to 100 selected farms were organized according to a well established schedule with the owners of these farms.

The choice was based on 6 indicators included in the European Welfare Quality® (2009) protocol for dairy cattle: lying time, lying outside the lying area, collision and cleanliness assessed on three body areas: udder, hind limbs and hind train. These indicators concerned a very important aspect of animal welfare, housing and farmer management and its impact on farm animal comfort. This choice was motivated by the fact that these indicators can be easily spotted on the animal, which allows early detection of the problem in order to act quickly to relieve the animal's pain. In addition, this does not require the animal to be handled in an aversive and stressful manner, or the application of a cumbersome and time-consuming detection method or technique. These indicators are defined in terms of alert and alarm thresholds (Welfare Quality^®, 2009), which allow the observer (whether a researcher or a farmer) to be alerted to the extent and severity of the situation, e.g. the degree (severe, mild, minimal) and duration of the impairment (acute or chronic), in order to act quickly to improve the situation.

The observations on cows started just after the morning milking and lasted one day per farm. To facilitate the recording and analysis of the observations, observation sheets were designed and used for each farm, based on the WQ^® (2009) Protocole, containing data on the animal condition; the degree of freedom (access to a pasture or loafing area); their comfort (cleanliness, injuries, difficulty in lying down and getting up, etc.), which made it possible to characterize the farms surveyed while situating their welfare level. Once all the measurements had been completed, the comfort scores were calculated.

The calculation of comfort scores around resting involves the calculation of a weighted sum for each measure of comfort according to alert and alarm thresholds defined by the WQ^® (2009).

Data processing was performed using Genstat Version 15.0 software (VSN International Ltd., UK). Descriptive statistics (mean, standard error, min and max) were used to describe the well-being level of surveyed farms on the comfort aspect. Also, identify the existence of co-variability between the different measurements of well-being studied by Pearson correlation tests and Chi-square test. The threshold of significance was set at 0.05.

Lying time (average duration and frequency) is a very important and determining criterion to measure the animal's comfort and can serve as an indicator of the ease with which the cow can perform it. Indeed, when they lie down or stand up, cows go through a kneeling position. At this point, the knees carry more than half the body weight and the softness of the floor will inevitably affect the comfort of the animals. Hard floors may cause them to lie down less often, and this may be partially compensated by an increase in the average lying time (Rushen, Haley, & de Passille, 2007; Cazin et al., 2014).

In our study, the average time taken by cows to lie down was (5.9s± 0.89) with moderate variability (min: 4.0s and max 8.1s) (Table.1), between farms visited. This mean time was similar to that of De Boyer des Roches et al. (2014) in France (5.9 ± 0.1 s). However, it is in contrast to several studies in other countries: Brorkens et al. (2009) (4.15 ± 1.0s); Silva (2013) (4.93 ± 1.15) and Ostjic-Andric (2011) (6.53 ± 0.45s).

Table 1
Scores of criterion “comfort around resting of farms surveyed.

M= mean, SE= standard error, Min= minimum, Max= maximum.

Therefore, 41.0% of farms exceeded the alert threshold (> 6.3s) and 39.0% the alarm threshold (> 5.2s). These thresholds were higher than those found by Silva (2013) in Portugal (8.3% of farms exceeded the alert threshold (> 6.30s) and 25.0% the alarm thresholds (> 5.20s). In contrast, our alarm thresholds were lower compared to De Boyer des Roches et al. (2014) in France (72.0% > 5.20s). These thresholds signaled difficulties to lying down. As a result, serious problems of well-being (WQ^®, 2009; Popescu et al., 2013).

The longer lying time is mainly related to the quality of floor, especially hard or concrete floors, which cause cows to lie down less often, although this reduction can be partly compensated by a longer average lying time (Rushen et al., 2007). This finding is consistent with several studies that have shown that cows prefer soft floors to hard or concrete floors (Lidfors, 1989; Haley, de Passille´, & Rushen, 2001, Tucker & Weary 2001). Norring, Manninen, de Passille, Rushen, and Saloniem (2010) also showed that 80% of lying positions were observed on carpet and only 20% on concrete. Thus, the longer lying sequence, the more difficult it is for cows to lie down, which explains the alarm and warning thresholds found in the present study as well as the significant correlation detected between comfort around resting and time spent lying down (r = -0.70; p < 0.0001) (Figure 1).

The presence of cows lying outside the lying area is also a resting comfort criterion, which generally reflects a mismatch between the lying area and the size of the cow. This problem exists with varying frequencies not only in Algerian farms (3.00±9.31) (Table.1) but also in other countries: Popescu et al. (2014) (3.19%), De Boyer des Roches et al. (2014) in France (0, 34%); Silva (2013) in Portugal (20%).

In our study, a small proportion (1%) of cows exceeded the alert threshold. This proportion was lower than that of De Boyer des Roches et al. (2014) in France (19.1%). This low frequency of cows lying outside the lying area was highly dependent on the level of maintenance of the lying area (Sunderland, 2002; Bowell et al., 2003), its size and density which forced some cows (the dominated) to lie outside the lying area (Phillips, 2002; Mattiello et al., 2009; Krug, Haskell, Nunes, & StilWell, 2015). This explains the significant correlation found between this measure and comfort (r = -0.33; p = 0.0008) (Figure 2).

Figure 1
Correlation between time need to lie down and comfort around resting.

Figure 2
Correlation between lying outside lying down area and comfort around resting.

Collision is a key issue for animal comfort and welfare (Brorkens et al., 2009; Ostojic-Andric et al., 2011; De Boyer des Roches et al., 2014; Popescu et al., 2014). It indicates a painful condition due to repeated movements with the rearing equipment. Consequently, the occurrence of injuries and diseases (lameness) is increased. In our study, a low average frequency of collisions with rearing equipment was observed in the farms visited (3 ± 9.56) (Table 1). This frequency was close to that found by Brorkens et al. (2009) (1.78%), but much lower than that reported by De Boyer des Roches et al. (2014) in France (27.1). Therefore, a small proportion of surveyed farms cows exceeded the alert threshold (2%). This threshold was much lower than that reported by De Boyer des Roches et al. (2014) in France (47.3%).

This state of discomfort and uneasiness is closely linked to the farmer's management: inadequacy of the lying area with cow movements, short and poorly placed tethering chains, slippery and unsuitable floor. Hence the significant correlation found in our study between the frequency of collisions and comfort (r = -0.43; p < 0.0001) (Figure.3).

Figure 3
Correlation between collision to housing equipment and comfort around resting.

Cow cleanliness influences cow comfort, health and hygiene of different parts of the animal's body, including the udder, milk quality and skin integrity. This has been demonstrated by several studies in different countries (Whay et al., 2008; Norring et al., 2008; Andreasen & Folkman, 2012; De Boyer des Roches et al., 2014; Radeski, Janevski, & Ilieski, 2015).

Regarding the cleanliness of dairy cows in our study, we found a high average prevalence of cows with dirty udders (62.6% ± 21.9), dirty hindquarters (60.6% ± 21.5) and dirty hind limbs (60.6 ± 21.4) (Table.1). As a result, a large number of farms exceeded the alert thresholds for these measures: 100.0% for udder cleanliness, 86.0% for hindquarters cleanliness and 63.0% for hind limbs. This high level of uncleanliness indicates serious problems of welfare. Our results are in agreement with those of De Boyer des Roches et al. ( 2014) in France, who noted a high proportion of farms exceeding the alert thresholds: Hind legs (95.4%), hind quarters (93.9%) and udders (70.7%), which explains the correlations found between cows with dirty limbs and those lying outside the lying area (r = 0.21 ; p = 0.03), dirty udders and dirty limbs (r = 0.21; P = 0.03), as well as cows with dirty hindquarters and those lying outside the lying area (r = 0.21; P = 0.003) (Figure 4).

Figure 4
Correlation between the different indicators of cleanliness of farms surveyed.

These frequencies of dirtiness in the surveyed farms were much higher than those of Whay et al. (2008) and De Boyer des Roches et al. (2014) for the udder (22.2; 26.5%) and hindquarters (10.7; 51.5%) respectively, but close to those of Radeski et al. (2015) (65.2%) for the udder. In contrast, our results were lower than those reported by Radeski et al. (2015) (85.5%) and Whay et al. (2008) regarding the frequency of hindquarter soiling and De Boyer des Roches (2014) and Radeski et al. (2015) (100.0; 80.4; 86.5) regarding hind limbs soiling.

This level of dirtiness reflects the poor rearing conditions that constitute a degraded animal environment and depend on a low frequency of bedding replacement (Norring et al., 2008), hygiene of lying area (Bowell et al., 2003; Hulsen, 2007; Leach, Knierim, & Whay, 2009, Ruud, Boe, & Osteras, 2010, Andreasen & Folkman, 2012) and dirty skin and hair can cause itching, decrease thermoregulatory and antimicrobial defense properties and may cause dermatitis (Winckler et al., 2003).

This level of dirtiness reflects poor husbandry conditions that constitute a degraded environment for animal and depended on low frequency of bedding replacement (Norring et al., 2008), hygiene of lying area (Bowell et al., 2003; Hulsen, 2007; Leach et al., 2009; Ruud et al., 2010; Andreasen & Folkman, 2012), and coat cows which can cause itching, decreases thermoregulatory and antimicrobial defense properties and can lead to dermatitis (Winckler et al., 2003).

Several studies have shown an association between dirtiness and mastitis (Sant'Anna & da Costa, 2011). Therefore, udder cleanliness was closely related to milk quality (Ellis et al., 2007). Indeed, Reneau et al. (2005) and Delaval (2006) have shown that the consequences of udder and skin infections increase rapidly as cows get dirtier and that one extra point can increase the somatic cell count in the milk tank by 50,000 milliliters. Therefore, it is essential to keep the cows clean to ensure good milk hygiene and avoid possible contamination. On the other hand, the high prevalence of dirt in the limbs and hindquarters would reflect the cleanliness of the lying area in combination with a high cow density. In addition, when stall sizes are inadequate, cows may lie partially or completely outside the lying area. As a result, they become dirtier and allow skin lesions and limb infections to develop (Mattiello et al., 2009; Von Keyserlingk, Barrientos, Ito, Galo, & Weary, 2012). Thus, efforts are needed to improve the cleanliness of the animals and their environment to overcome several problems related to cow health, behavior and welfare.

* Author for correspondense. E-mail: a.benatallah@ensv.dz