Microscopic differential cell counts in milk for the evaluation of inflammatory reactions in clinically healthy and subclinically infected bovine mammary glands

Department of Animal Sciences, Institute of Veterinary Medicine, Division of Microbiology and Animal Hygiene, Faculty of Agricultural Sciences, Georg-August-University Göttingen, Burckhardtweg 2, D-37077 Göttingen, Germany Department of Animal Breeding, University of Kassel, Nordbahnhofstraße 1a, D-37213 Witzenhausen, Germany Landesbetrieb Hessisches Landeslabor, Schubertstraße 60, D-35392 Giessen, Germany Regierungspräsidium Giessen, Milk Control, Schanzenfeldstraße 8, D-35578 Wetzlar, Germany

For mastitis diagnosis, traditional and well-established tests including somatic cell count (SCC) and microbial culturebased methods are standard (Viguier et al. 2009).According to current definitions of udder health in Germany, SCC 4 100 × 10 3 cells/ml in quarter foremilk samples are in the physiological range (DVG, 2002).It is well known that the crossover of normal cellular defence in the mammary gland into an inflammatory reaction starts at a level of > 100 × 10 3 cells/ml (Harmon, 1994;DVG, 2002).However, SCC varies with the status of lactation, age, stress of the animals, time and frequency of milking, season, and status of udder infection (Dohoo & Meek, 1982;Harmon, 1994).SCC is a robust quantitative estimate, but it does not divide the cells present in milk into different cell types (Kehrli & Shuster, 1994;Rivas et al. 2001).
In the mammary gland, number and distribution of leucocytes are important for the success of udder defences against invading pathogens (Leitner et al. 2003).Lymphocytes, macrophages, and polymorphonuclear leucocytes (PMN) play an important role in immune reactions within the mammary gland (Paape et al. 1979;Sordillo & Nickerson, 1988).Induction and suppression of immune responses are regulated by lymphocytes (Nickerson, 1989).They recognize antigens through membrane receptors specific for invading pathogens (Sordillo et al. 1997).Macrophages are active phagocytic cells in the mammary gland and capable of ingesting bacteria, cellular debris and accumulated milk components (Sordillo & Nickerson, 1988).Milk or tissue macrophages recognize the invading pathogens and initiate an immune response by the release of chemo-attractants inducing the rapid recruitment of PMN into the mammary gland (Paape et al. 2002;Oviedo-Boyso et al. 2007).The main task of PMN is the defence of invading bacteria at the beginning of an acute inflammatory process (Paape et al. 1979;Oviedo-Boyso et al. 2007).Not only does the number of PMN increase enormously, but also their level of defence activity (Targowski, 1983;Paape et al. 2003).
The distribution of leucocyte types varies in normal milk without any symptoms of mastitis.Some previous studies found lymphocyte proportions between 14 and 80%, macrophage proportions between 12 and 46%, and those of PMN between 6 and 50% (Rivas et al. 2001;Merle et al. 2007;Koess & Hamann, 2008).In mastitis milk, PMN proportions of up to 95% have been reported (Paape et al. 1979;Kehrli & Shuster, 1994).During various phases of inflammation SCC differs in total numbers, whereas differential cell count (DCC) varies in composition of the cell populations involved (Nickerson, 1989).Therefore, in addition to SCC, determination of different types of immune cells present in milk is beneficial for describing udder health status (Pillai et al. 2001;Rivas et al. 2001).So far, however, there is little knowledge on DCC and the qualitative role of milk leucocytes in healthy udders because DCC in low-SCC milk is difficult to perform (Dosogne et al. 2003).
Data from a previous study (Schwarz et al. 2010) indicated a high standard of udder health in a representative part of the dairy cow population in the German federal state Hesse and confirmed the threshold of 100 × 10 3 cells/ml differentiating between normal and abnormal secretion of quarters.However, unexpectedly high numbers of mastitis pathogens in the SCC range 4 100 × 10 3 cells/ml were found.They could already be detected at a threshold of 1 × 10 3 cells/ml.Based on these data we suspected inflammatory processes even within the SCC range of mammary glands classified as healthy according to current definitions.Therefore, the objective of this study was the detailed evaluation of health status in udder quarters with SCC clearly < 100 × 10 3 cells/ml based on a statistical analysis of DCC.Leucocytes were isolated from quarter foremilk samples and differentiated into lymphocytes, macrophages, and PMN using microscopy.

Animals and farms
Twenty dairy cows in good condition and without previous history of mastitis were selected from four German dairy farms (A-D) for a detailed analysis of their udder health status based on DCC in quarter foremilk samples.The animals, Holstein-Frisian cows (n = 18) and German Simmental cows (n = 2), were in different lactations (1-6) and stages of the lactation.Six cows were in their first, seven in their second, three in their third, one in her fourth, two in their fifth and one in her sixth lactation.Seven animals were in an early stage of lactation (28-86 d), eight were in mid lactation (107-177 d) and five in a late stage of lactation (212-289 d).Foremilk samples from 72 udder quarters of the 20 cows classified as normal secreting (SCC 4 100 × 10 3 cells/ml and no pathogen) were selected for DCC analysis.A further 6 quarter foremilk samples with SCC of (100-1824) × 10 3 cells/ml were chosen from 6 cows as control group.Clinical mastitis symptoms such as flecks in milk, swelling or redness of the udder quarters could only be observed in quarters with SCC > 100 × 10 3 cells/ml.
In Farms A-D, 52-109 dairy cows were housed in pen barns and milked twice a day in milking parlours.Milking operations were similar in all farms.After forestripping into a foremilk cup, the milkers used damp cotton tissues for udder cleaning.Teats were dipped after milking with iodine solution.Feeding comprised a total mixed ration consisting of grass and maize silage, rape grist and cereals.Water was available ad libitum.Farm A produced high quality milk, while farms B-D were conventional milk producers.The average herd annual milk yields of the four farms ranged between 6500 and 9900 kg.

Milk sampling
Quarter foremilk samples were obtained according to DVG (2000) standards.Before milking, teat ends were scrubbed with 70% ethanol and the first two squirts of milk were discarded.Ten millilitres of milk per udder quarter was collected aseptically in a sterile 14-ml plastic sample tube (Greiner Bio-one, Frickenhausen, Germany).Four millilitres was used for SCC and bacteriological examinations, the remaining 6 ml was subjected to DCC analysis.
Quarter foremilk samples were taken on farms B-D during morning milking.Further processing occurred within 4 h.Samples on farm A were collected during evening milking and analysed within 15 h.

Somatic cell counts and bacteriological examinations
SCC was determined using a Fossomatic 5000 (Foss Electric, Hillerød, Denmark).Cytobacteriological analysis of all quarter foremilk samples was performed according to IDF (1981) standards.Promptly after collecting the quarter foremilk samples and cooled-transportation to the laboratory, 10 μl of milk was streaked onto a quadrant of a 7% bovine blood agar plate containing 0•05% aesculin (Merck, Darmstadt, Germany), incubated for 48 h at 37 °C, and examined for bacterial growth.

Differential cell counts
Six millilitres of each quarter foremilk sample was transferred into a sterile 14-ml plastic tube.Milk samples were then centrifuged at 200 g at 4 °C for 15 min.Cream layers and supernatants were discarded and cells were washed once in PBS by centrifugation at 200 g at 4 °C for 15 min.Cell pellets were finally resuspended in 20 μl PBS.To obtain as many cells as possible for DCC analysis on the microscope slide, the whole sediment of the tube was spread over an area of 2 cm 2 .Cell staining was performed according to the method of Pappenheim (1912).
Evaluation of the slides followed using light microscopy and oil immersion (100-fold magnification).One-hundred cells of each slide were counted meander-shaped and differentiated into lymphocytes, macrophages and PMN.Cell identification occurred according to standard methods (Coles, 1974;Lee et al. 1980).Lymphocytes were identified based on their circular form (5-10 μm) and the typical shape of the nucleus that almost fills the cell leaving a very thin rim of cytoplasm.Cells of 8-30 μm in size containing a little nucleus and pale staining were considered as macrophages.The group of PMN was characterized as cells of 10-14 μm in size and segmented nuclei.They were intensely coloured and contained granula in the cytoplasm.

Statistical analyses
Associations between values for individual cell populations and values for SCC were analysed by applying linear mixed models as implemented in the SAS program (version 9.1, SAS Institute, Cary NC, USA).The statistical model included fixed and random effects as well as a regression on SCC up to the third polynomial degree, in order to fit regression curves.The non-significant regression coefficients of different polynomial structures were removed from the model by using F-statistics sum of square type I tests at P < 0•05 instead of likelihood ratio tests.Based on type I sums of squares at P < 0•05, a sequential analysis approach is appropriate for polynomial formulated models (Littell et al. 1998).The applied statistical model [1] was defined as follows: For verification of results of model [1], and to test DCC data especially for differences in the SCC range 4 100 × 10 3 cells/ ml, analysis of variance was additionally done by including a fixed effect of the SCC group and removing the SCC covariates from the statistical model.Therefore, all 78 udder quarters were classified into SCC groups I-IV as defined in a previous study (Schwarz et al. 2010).Group IV represented the control quarters.Twenty-five (31%) of the 78 samples analysed belonged to group I according to their SCC values of 4 6•25 × 10 3 cells/ml.Forty-one samples (53%) showed SCC values of (> 6•25 to 4 25) × 10 3 cells/ml (group II).Six samples (8%) with SCC of (> 25 to 4 100) × 10 3 cells/ml were categorized into group III.A further six samples (8%) with SCC > 100 × 10 3 cells/ml were sorted into group IV.The applied statistical model [2] was defined as follows:

Somatic cell counts and bacteriological status of quarter foremilk samples
The 78 udder quarters selected showed a geometric mean value for SCC of 11•78 × 10 3 cells/ml and a median of 10 × 10 3 cells/ml.The quarter with the lowest SCC contained 1 × 10 3 cells/ml, the quarter with the highest SCC contained 1824 × 10 3 cells/ml.Udder pathogenic microorganisms were identified in only two of the 78 quarter foremilk samples.In two of the six control quarters (SCC of and Staphylococcus aureus were detected.

Differential cell counts of quarter foremilk samples depending on somatic cell counts
For a more detailed evaluation of the udder health status 100 cells per quarter foremilk sample were differentiated into lymphocytes, macrophages, and PMN.In view of DCC (n = 78) the proportions of lymphocytes lay between 2 and 92% with a mean of 44•56% and SD of 21•63% (Table 1).Proportions of macrophages ranged between 8 and 68% with a mean of 34•85% and SD of 15•20%.PMN proportions varied between 0 and 88% with a mean of 19•94% and SD of 21•40%.
Because of the wide variations found within the cell populations, particularly in case of lymphocytes and PMN, DCC data were tested for correlation with SCC using the statistical model [1].
Percentages of PMN (Fig. 1B) and lymphocytes (Fig. 1A) emerged in contrary directions at rising SCC.PMN increased constantly from 0% at SCC of 1 × 10 3 cells/ml to a maximum of 88% at 139 × 10 3 cells/ml.At SCC of 1824 × 10 3 cells/ml the proportion of PMN was 86%.Interestingly, PMN was already the predominant cell population at a SCC level of 43 × 10 3 cells/ml with a proportion of 62%.This event was observed in three udder quarters [SCC (43-45) × 10 3 cells/ ml] of three different cows housed in two different farms.2).

Statistical analysis (model [1]) indicated a significant
Within the SCC range of (3-100) × 10 3 cells/ml macrophage proportions lay between 8% and 68% (Fig. 1C).At SCC < 3 × 10 3 cells/ml and in samples with > 100 × 10 3 cells/ ml proportions of macrophages were < 30%.Statistical analysis (model [1]) revealed a significant (P < 0•01) negative correlation between macrophage percentage and SCC (Table 2).In addition, the position of the udder quarter, the lactation number and the stage of lactation had no significant impact on the individual cell populations (Table 2).However, percentages of lymphocytes and macrophages were significantly influenced by the farm (Table 2).
Statistical analysis (model [1]) revealed a significant negative correlation between the percentages of lymphocytes and SCC, a significant negative correlation between macrophages and SCC, as well as a significant positive correlation between PMN and SCC.To test DCC data especially for differences in the SCC range 4 100 × 10 3 cells/ ml, a second statistical analysis (model [2]) was performed (Fig. 2).

Discussion
Together with SCC, determination of DCC in milk is an important tool characterizing udder health (Pillai et al. 2001).There are clear SCC cut-offs to differentiate between normal and abnormal secretion of quarters.But even in healthy udders, inflammations can be suspected under special circumstances (Schwarz et al. 2010).The immunological status of mammary glands classified by DCC is poorly investigated.Reviewing the literature Medzhitov (2007) reported that there might be a lack of knowledge on host defence in asymptomatic infections because almost all studies performed so far concentrated on symptomatic infections.In the present study, we differentiated leucocytes purified from quarter foremilk samples to improve knowledge of the immunological status of clinically healthy and subclinically infected bovine mammary glands.While SCC of > 100 × 10 3 cells/ml is normally related to inflammatory processes inside the mammary gland, a SCC range of 4 100 × 10 3 cells/ml is in a physiological band (Harmon, 1994;DVG, 2002) but can also be related to latent mastitis in the presence of pathogens (Schwarz et al. 2010).Here, we predominantly analysed milk samples with SCC < 50 × 10 3 cells/ml to have a high informative value about DCC in low-SCC milk.Our control group (SCC > 100 × 10 3 cells/ml) included only six quarters.However, the SCC range > 100 × 10 3 cells/ml has been studied extensively before and PMN has been reported generally to be the dominant cell population in mastitic milk (e.g.Leitner et al. 2000;Merle et al. 2005;Koess & Hamann, 2008).
It is known that the milk fraction collected has an impact on both SCC and DCC (Sarikaya et al. 2005;Sarikaya & Bruckmaier, 2006;Olde Riekerink et al. 2007).SCC values in foremilk samples of quarters with a total quarter milk SCC > 100 × 10 3 cells/ml were significantly higher than in cisternal milk (Sarikaya & Bruckmaier, 2006).In samples with SCC < 100 × 10 3 cells/ml differences of SCC between foremilk and cisternal milk were only minor.While no changes of DCC during milking could be observed in milk with SCC < 200 × 10 3 cells/ml, proportions of PMN were higher and proportions of macrophages were lower in milk (SCC > 200 × 10 3 cells/ml) collected post milking compared with milk collected premilking or during the milking process (Olde Riekerink et al. 2007). Sarikaya et al. (2005) also reported that the proportion of macrophages decreased while that of PMN increased during milking.Since we concentrated predominantly on the analysis of low-SCC milk, we presume that the foremilk samples taken in our study are representative for the analysis of the udder health status.
Our results indicate that lymphocytes were the predominant cell population in healthy mammary glands.Milk samples with an extremely low SCC value of 4 6•25 × 10 3 cells/ml revealed high lymphocyte proportions of up to 92% (mean value: 55%).In a SCC range of (> 6•25 to 4 25) × 10 3 cells/ml a high mean proportion of lymphocytes (49%) was determined too.Information on DCC in milk samples with such low SCC from other field studies is rare.Only Koess & Hamann (2008) reported a mean value of 25% for the proportion of lymphocytes in the SCC range of (0-50) × 10 3 cells/ml.Merle et al. ( 2007) measured a mean proportion of 25% of lymphocytes in milk samples with SCC < 100 × 10 3 cells/ml.In an experimental study (Rivas et al. 2001) lymphocyte proportions between 54 and 80% were measured pre-inoculation in udder quarters with SCC < 200 × 10 3 cells/ml.Data from our study showed higher proportions of lymphocytes in milk with SCC < 100 × 10 3 cells/ml than reported before.This difference resulted from the analysis of milk with very low SCC, because lymphocytes were the dominant cell population in these samples.The proportions of 2-16% of lymphocytes in milk secreted by diseased udder quarters (SCC > 100 × 10 3 cells/ml) were clearly lower than those in healthy quarters.Similar observations were described before (Rivas et al. 2001;Merle et al. 2007;Koess & Hamann, 2008).
In udder quarters classified as normal secreting (SCC 4 100 × 10 3 cells/ml) PMN proportions ranged from 0 to 63%.Our data show that PMN, particularly in milk samples with SCC values 4 6•25 × 10 3 cells/ml, were rare (mean PMN proportion: 10%).At a SCC level of (> 6•25 to 4 25) × 10 3 cells/ml the mean proportion of PMN of 14% was also low.Comparable data for such low SCC values are not available from the literature.Only a mean PMN proportion of 30% in udder quarters with SCC of (0-50) × 10 3 cells/ml was reported previously (Koess & Hamann, 2008).However, because of an increased transfer of PMN from blood into the mammary gland at the beginning of an inflammation (Kehrli & Shuster, 1994;Paape et al. 2002;Paape et al. 2003), a high percentage of PMN in milk is an important indicator of inflammatory reactions (Pillai et al. 2001;Paape et al. 2002).
PMN have been reported previously as the predominant cell population in secretions of diseased mammary glands (Paape et al. 1979;Kehrli & Shuster, 1994).We made the unexpected observation that in milk of udder quarters classified as normal secreting PMN dominated already at SCC 5 43 × 10 3 cells/ml.This finding suggested that inflammatory processes appear already in a SCC range that is clearly below the cut-off value of 100 × 10 3 cells/ml.Factors that might have triggered the elevated proportion of PMN might be manifold.A dairy cow is under constant pressure from udder pathogenic microorganisms in the environment.The elevated PMN proportion could be evidence for the initial phase of an inflammation.In this regard it is also possible that PMN are able to defend against pathogens successfully and prevent mastitis.However, although we could not isolate any pathogens in such quarters they might be not healthy anyhow.Negative bacteriological results could depend on intermittent pathogen shedding (Sears et al. 1990), presence of antimicrobials or other inhibitors in milk (Reiter, 1978).At the time of examination pathogens could also be ingested by phagocytes or survive intracellularly in the host (Newbould & Neave, 1965;Hill et al. 1978).Shedding of too low amounts of pathogens or ceased growth may be further reasons for negative bacteriological results (Sears et al. 1990).
The interdependence of infections, inflammatory processes, and immune responses in individual udder quarters is discussed controversially in the literature.Some authors suggested that individual udder quarters within a cow can be influenced by infections of neighbouring quarters (Merle et al. 2007) whereas others did not find any evidence for an interdependence of udder quarters (Wever & Emanuelson, 1989) because they did not find DCC to be affected by the bacteriological status of adjacent quarters.Our data indicated no immunological interdepencence between the four udder quarters at low and high SCC levels.In the three udder quarters of three different cows with SCC of (43-45) × 10 3 cells/ml in foremilk samples, elevated PMN proportions between 54 and 63% were determined.In the remaining nine quarters of these cows clearly lower SCC Cell differentiation in milk values of (4-19) × 10 3 cells/ml, lower PMN proportions of 6-18%, and no bacterial infection were detected.Furthermore, no interactions between the quarters were observed in the six cows of the control group with high SCC values > 100 × 10 3 cells/ml and PMN proportions of 65-92% in one udder quarter.In these animals SCC < 100 × 10 3 cells/ ml and PMN proportions of 4-39%, respectively, were detected in the other three quarters.
Another factor that might have triggered the elevated percentage of PMN is stress (Davis et al. 2008).Although we did not measure parameters related to stress, such as corticosterone in plasma, the influence of stress in our study might be minimal because the animals analysed were kept under optimal conditions and according to national guidelines.No obvious symptoms of stress (i.e.kicking during pre-milking preparation of the udder or during taking the quarter foremilk samples) was observed.
The antidromic trend of lymphocyte and PMN percentages at increasing SCC is caused by the composition of milk leucocytes.Since they consist primarily of lymphocytes, macrophages and PMN (Sordillo & Nickerson, 1988), the increase of the percentage of one cell population implies the decrease of at least one of the other cell populations.However, our statistical analysis indicated a significant impact of the farms on percentages of both macrophages and lymphocytes.This impact might be due to a nonrandomized selection of cows within the farms and different numbers of cows selected per farm.While cows with healthy mammary glands were predominantly selected from farms A, B and D, samples from cows with diseased quarters were predominantly collected from farm C.
Besides PMN, macrophages also possess phagocytic functions.Milk samples with extremely low SCC values of 4 6•25 × 10 3 cells/ml showed a mean proportion of 32% of macrophages.In the SCC range of (> 6•25 to 4 25) × 10 3 cells/ml the mean macrophage proportion was 38%.In the literature a mean value of 43% of macrophages was reported for the SCC range of (0-50) × 10 3 cells/ml (Koess & Hamann, 2008).In the case of diseased udder quarters (SCC > 100 × 10 3 cells/ml) we measured macrophage proportions of 9-28%.These results lay also within a wide range of 4-48% as mentioned in other studies (Rivas et al. 2001;Merle et al. 2007;Koess & Hamann, 2008).While we found that macrophages were the second dominant cell population in almost all samples tested in relationship to lymphocytes and PMN, they had been reported to be the predominant cell population in milk of healthy mammary glands (Lee et al. 1980).This difference might be explainable by different definitions of healthy mammary glands.In our study we focused on the analysis of immune cells in milk with very low SCC values.However, Lee et al. (1980) defined mammary glands as healthy based on negative bacteriological examinations and did not present any SCC values.
Besides leucocytes, epithelial cells can also be found in milk.In the literature (Lee et al. 1980;Koess & Hamann, 2008) low proportions of epithelial cells of 1-3%, which were similar to our examinations (data not shown), were described.Other researchers reported epithelial cell proportions of 10-19% (Miller et al. 1991) or even 44% (Leitner et al. 2000).However, proportions of 5 10% should be discussed critically.Miller et al. (1991) analysed milk of primiparous cows during the first 75 d of lactation, whereas Leitner et al. (2000) measured epithelial cells by flow cytometry based on a non-specific identification procedure.
Variations in the distribution of leucocytes in milk from non-infected mammary glands as shown in other studies, were probably dependent on differences in methods, sampling, investigators (Schröder & Hamann, 2005), breed (Leitner et al. 2003), stage of lactation (Vangroenweghe et al. 2001;Dosogne et al. 2003) and variable SCC.Contrary to other authors (Schröder & Hamann, 2005), we did not observe influences of the composition of the sample tubes (glass or plastic) on differences of phagocytic cell percentages (data not shown).
In our study bacteriological examinations revealed udder pathogenic microorganisms in only 2 of the 78 udder quarters analysed (2 of the 6 control quarters).Esch.coli, however, was found in a quarter with high SCC of 967 × 10 3 cells/ml (DCC: lymphocytes 6%, PMN 85%, macrophages 9%), whereas Staph.aureus was detected in a quarter with SCC of 1824 × 10 3 cells/ml (DCC: lymphocytes 5%, PMN 86%, macrophages 9%).These few bacteriological findings did not allow any assessment.However, other studies (Piccinini et al. 1999) also found PMN to be the dominant cell population in milk of udder quarters infected with major pathogens.

Conclusion
SCC is an undisputed and well-established criterion for the evaluation of udder health and milk quality.However, in addition to SCC, DCC can be used for a more detailed analysis of udder health status.Analysing DCC of mammary glands classified as normal secreting by SCC < 100 × 10 3 cells/ml, inflammatory reactions were already detectable at a SCC level of 5 43 × 10 3 cells/ml due to predominating PMN proportions in foremilk samples of the corresponding udder quarters.This is the first study indicating inflammatory reactions in udder quarters with SCC that were clearly below the current threshold of 100 × 10 3 cells/ml.
observation for the individual cell population (lymphocytes, macrophages, PMN) of the individual udder quarter of cow l μ = overall mean effect herd i = fixed effect of the i-th herd of cow l parity j = fixed effect of the j-th lactation number of cow l [1, 2, 5 3] DIM k = fixed effect of the k-th stage of lactation of cow l [early, 28-86 d; mid, 107-177 d; late, 212-289 d] cow l = random effect of cow l SCC ijkl = value for SCC of the individual udder quarter of cow where: y ijklm = observation for the individual cell population (lymphocytes, macrophages, PMN) of the individual udder quarter of cow l μ = overall mean effect herd i = fixed effect of the i-th herd of cow l parity j = fixed effect of the j-th lactation number of cow l [1, 2, 5 3] DIM k = fixed effect of the k-th stage of lactation of cow l [early, 28-86 d; mid, 107-177 d; late, 212-289 d] cow l = random effect of cow l group m = fixed effect of the m-th SCC-group [I, 4 6•25 × 10 3 cells/ml; II, (> 6•25 to 4 25) × 10 3 cells/ml; III, (> 25 to 4 100) × 10 3 cells/ml; IV, > 100 × 10 3 cells/ ml] e ijklm = random residual effect

Fig. 1 .†
Fig. 1.A-C.Differential cell counts (DCC) depending on somatic cell counts (SCC): A, Proportions of lymphocytes (○ LYM) pictured in combination with a calculated potential trendline; B, Proportions of polymorphonuclear leucocytes (.PMN) pictured in combination with a calculated logarithmic trendline.C, Proportions of macrophages ( ▵ MAC); each symbol represents the result of one udder quarter analysed, but overlapping is possible.

Table 1 .
General overview of differential cell counts (DCC) of 78 quarter foremilk samples analysed by microscopy