- Email: [email protected]
Use of the Fossomatic Method to Determine Somatic Cell Counts in Sheep Milk
Use of the Fossomatic Method to Determine Somatic Cell Counts in Sheep Milk CARLOS GONZALO, JESUS A. BARO, JUAN A. CARRIEDO, and FERMIN SAN PRIMITIVO Departamento de Produccion Animal Universidad de Leon 24071 Leon, Spain ABSTRACT
infection (13). The main sec methods (direct microscopic, Coulter counter, and Fossomatic) have been completely standardized for cow milk (3, 9, 11, 14). Recently, interest in milk cell counting in sheep research and mastitis control programs has increased. However, sec methods for sheep milk have not been widely evaluated or tested (10). Some counting methods, such as direct microscopic SCC (DMSCC) and Coulter counter, have been modified 0, 2) to adapt to sheep milk, which has a higher fat content than cow milk. The Fossomatic SCC method (FSCC) automatically processes the milk samples so that they need no prior adaptation. However, optimal conditions for use of this method have not been specified for sheep milk. Basically, the Fossomatic counter is a fluorescence microscope. The ethidium bromide dye penetrates the cell and forms a fluorescent complex with the nuclear DNA. Each cell produces an electrical pulse, which is amplified and recorded. The objectives of this study were to compare the FSCC with the DMSCC (reference method) in sheep milk and to evaluate the effect of the storage method (fresh, refrigerated, and frozen milk) and of the sample age on the sce of foremilk and strippings.
The Fossomatic method for SCC was compared with the direct microscopic method in 85 half-udder samples of sheep milk. The correlation coefficient was .986. The repeatability of the Fossomatic method showed average variation coefficients less than 5%. The carry-over effect between samples was less than .5%. The effect of the storage method (fresh milk, refrigerated at 4°C and frozen at -19°C) and the sample age were studied in 48 samples of foremilk and strippings. The storage method had a significant effect on the SCC variation. The average fresh, refrigerated, and frozen sample counts were 125,000, 110,000, and 82,000 cells/ml for foremilk and 201,000, 192,000, and 145,000 cells/ml for strippings, respectively. The effect of age on the refrigerated samples was also significant; counts were reduced by about 14% from d 1 to 7 in both types of milk. The effect of age on the frozen sample varied. These results suggest standardization of age and storage conditions of the milk samples to reduce variation of SCC. The milk must not be frozen. (Key words: sheep milk, somatic cell count, Fossomatic method) Abbreviation key: DMSCC microscopic SCC method, FSee matic SCC method.
= direct = Fosso-
INTRODUCTION
Somatic cell count in milk is a widely used method to evaluate the status of cow udder
Received June 3. 1992. Accepted August 14, 1992. 1993] Dairy Sci 76:115-119
MATERIALS AND METHODS Comparison of the FSCC and the DMSCC
Eighty-five samples of sheep milk preserved in .05% potassium dichromate were analyzed within 24 h after milking by FSCe and DMSCC. The samples were collected from half-udders and represent sec varying from 15,000 to 4 million cells/mt. For the DMSeC, the milk was heated to 40°C in a water bath and held at that tempera-
115
116
GONZALO ET AL.
ture for 15 min before being cooled to 20·C by careful stirring. The method used was that recommended by the International Dairy Federation (6), adapted to sheep milk by Gonzalo and Gaudioso (I). Two slides of each sample were prepared and counted. Slides were stained with May Grunwald-Giemsa (Vorqufmica, Vigo, Spain). The working factor was 1600. For the FSCC, the milk samples were heated to 40·C in a water bath and held at this temperature for 15 min. The samples were then double processed in a Fossomatic 90 (AlS N. Foss Electric, Hillen'ld, Denmark). The reagents were prepared following the manufacturer's instructions, which coincide with the method recommended by International Dairy Federation (6). Repeatability
The repeatability of the FSCC was evaluated by counting a number of subsamples from 20 milk samples with different numbers of cells (from 46,000 to 12.1 million cells/ml). The coefficients of variation for the data of each sample were calculated. In order to determine the possible carryover effect from one sample to another in the FSCC, a sample with low SCC (79,900 cells! ml) was counted, alternating with a sample with high SCC (7.88 million cells/ml). First, 15 counts were carried out with the low cell concentration sample. Then 15 counts were made alternating between high and low SCC milk. The low count sample was then counted 15 more times. Effect of the Storage Method and Sample Age
Thirty-five milliliters of foremilk and 25 ml of hand strippings (after machine removal) were obtained from 48 half-udders from a total of 24 primiparous, midlactation Churra sheep during morning milking. Potassium dichromate to a final concentration of .05% was used as a preservative for all samples. After collection, each sample of original milk (foremilk and stripping) was divided into eight aliquots (4 and 3 ml, respectively). One aliquot was used for SCC on the sampling day (fresh milk). Four aliquots were refrigerated at 4·C and Journal of Dairy Science Vol. 76. No.1. 1993
counted I, 3, 5, and 7 d after collection. The remaining three aliquots were frozen at -19·C and analyzed I, 7, and IS d after collection. All samples were counted in a Fossomatic 90 following the previously mentioned specifications. The frozen samples were thawed in a water bath at ambient temperature for 45 min before being heated at 40·C. Statistical Analysis
The statistical study of the storage method and sample age factors was made using analysis of variance according to the model Yijk
= JL
+ Mi + Dj(i) +
<& + fijk
where y
= variable SCC and log SCC (of foremilk and strippings), M· = effect of milk storage method, Di(j) = effect of day j in method i, <2k = effect of half udder k or block effect, and fijk = residual effect. 1
The Harvey methodology with the LSMLMW program (4, 5) was used to estimate the components of variance as well as the contrast of the differences between means. RESULTS AND DISCUSSION Correlation Between DMSCC and FSCC
The following regression equation was calculated: FSCC = .954 DMSCC + 14,393; the correlation coefficient was r = .986 ± .0183. The mean SCC were 428,000 cells!ml for DMSCC and 423,000 cells!ml for FSCC. These results show excellent agreement between the two SCC methods for sheep milk. For cow milk, the correlation coefficients between both methods were .974 and .988 when DMSCC working factors of 1()4 and 1()2, respectively, were used (14). The working factor that we used for DMSCC was 103.2 . Repeatability
Table I shows the repeatability of the FSCC in milk samples with a wide range of
117
FOSSOMATIC CELL COUNT IN SHEEP MILK TABLE I. Repeatability of the Fossomatic method in sheep milk samples with different SCC. Sample
SubArithmetic samples mean 00-3
- - (no.) - - I 32 46 2 30 79 3 30 130 4 30 167 5 30 186 6 30 220 7 30 310 8 30 391 9 30 427 10 30 512 11 28 527 12 32 654 13 25 779 14 26 904 15 25 1034 16 28 1123 17 28 1719 18 26 4655 19 29 5009 20 12,100 30
SD
cells/mt) 4.6 7.7 6.7 10.6 12.4 13.8 12.0 16.3 18.9 19.5 17.6 20.6 30.3 30.2 20.0 28.1 60.8 148.6 97.8 346.5
CV (%)
10.0 9.7 5.1 6.3 6.6 6.2 3.8 4.1 4.4 3.8 3.3 3.1 3.8 3.3 1.9 2.5 3.5 3.2 1.9 2.8
sec. The coefficients of variation decrease as the number of cells increases and, in most cases, are less than 5%. The International Dairy Federation (6) recommends coefficients of variation of the FSee lower than 7.5% for milk samples with cell counts between 400,000 and 600,000 cells/m1, which coincides with our results. The carry-over effect, when a milk sample with a low count was counted alternately with a sample with a much higher cell count, was
only .5%, which corresponds to an increase of 400 cells/ml when the cell content of the high count sample was 100 times greater (7.88 million cells/ml) than that of the low count sample (79,900 cells/ml). Thus, the carry-over effect between consecutive samples is not significant. Storage Method and Sample Age
Table 2 shows the analysis of variance results for foremilk and strippings, indicating the statistical significance and the percentage of variance for each variation factor. The block or half-udder variation factor was highly significant (P < .001), as expected. The storage method had a highly significant effect (P < .001); the proportion of explained variance was important (5.69 and 4.10% in foremilk and strippings, respectively). The effect of the sample age was also statistically significant, especially in the refrigeration method (P < .001). The analysis of variance for log see of foremilk and strippings (Table 3) decreased the percentage of variance explained by the block or half-udder variation factor and emphasized the variation percentages explained by other variation sources (storage method and sample age). Table 4 shows mean sec for each of the storage methods studied. The highest sec were for fresh milk. Refrigeration reduced the sec significantly, but this decrease was more marked in the freezing method, in which sec were much lower (P < .001) than those for refrigerated milk.
TABLE 2. Analysis of variance of the variation factors and proportion of variance explained by storage method, sample age in each method, and half-udder for the SCC variable of foremilk and stripping milk. SCC Stripping milk
SCC Foremilk Source of variation
df
Variance explained
F
Variance explained
F
(%)
Method Half-udder Days in refrigeration
2 47 3
212.5*** 237.8*** 11.6***
2 329
3.2*
(0/0)
5.69 90.79
199.0*** 316.9*** 12.4***
.46 Days in freezing Residual
4.10 93.19 .35
1.8NS• 1 3.06
2.36
lp> .05.
*p < .05. ***p < .001. Journal of Dairy Science Vol. 76, No. I, 1993
118
GONZALO ET AL.
TABLE 3. Analysis of variance of the variation factors and proportion of variance explained by storage method. sample age in each method. and half-udder for the logarithm see variable of foremilk and stripping milk. log see Foremilk Source of variation
df
log see Stripping milk
Variance explained
F
Variance explained
F
(%)
(%)
Method Half-udder Days in refrigeration
2 47 3
267.7·" 124.8··.. 13.0""
2 329
10.3··..
12.28 81.27
278.4"" 143.9...... 15.9·....
11.30 83.01 1.04
1.20 Days in freezing Residual
5.4" 4.64
5.25
"p < .01. ""p < .001.
The contrast of the difference between means, obtained from the ANOVA, was carried out for the see variable (Tables 4 and 5) and for its logarithmic transformation. In general, the contrasts carried out using logarithmic transformation gave results similar to those of the untransformed variable, although statistical significance of the differences was, in some cases, somewhat lower. Tables 4 and 5 show how the effect of refrigeration on the see was produced after 24 h of storage at 4°e for foremilk and 72 h for strippings. In stripping milk, the see for fresh milk (201,000 cells/ml) and for milk refrigerated for 24 h (204,000 cells/ml) did not differ, in contrast with results of Miller et al. (8) for cows, which showed a significant decline (P < .05) in stripping counts in the first 24 h in samples kept at ambient temperature. In refrigerated samples, see decreased by approximately 14% from the d 1 to 7 in both types of milk (Table 5). Kennedy et al. (7) recorded a much greater decrease (from 28 to
TABLE 4. Mean
see
36%) in cow milk samples preserved in potassium dichromate and kept at temperatures between 20 and 25°e for 8 d or longer. These results suggest that storage at ambient temperature may have a more negative effect on see than refrigeration. However, by using orthogonal contrasts, we have observed a very significant decline in the count (P < .001) in refrigeration. This decline was more marked in older samples (d 7), which also explains the results of Kennedy et al. (7). Freezing reduced see by 28 to 34% in both types of milk. This effect was comparable with that previously recorded by others (12, 14) in cow milk (22 to 57%) and may have been caused by cell nuclei that were less able to absorb the DNA-specific stain after freezing. A small auxiliary study was carried out to confirm this idea. Twenty samples of foremilk were counted using both methods (DMSee and FSeC) on d 0 (fresh milk) and d 7 (frozen milk). The average fresh milk counts were 103,000 cells/ml (DMSeC) and 110,000 cells/
of foremilk and stripping milk for each of the three storage methods studied. Storage method
see Foremilk x 1O-3/ml Stripping x 1O-3/ml
Fresh
Refrigerated
Frozen
X
X
X
125' 201'
SE 2.15 3.39
,.b,CMeans in a row with different superscripts differ (P < .05). Journal of Dairy Science Vol. 76. No.1. 1993
1l0b 192b
SE 1.07 1.69
82c 145C
SE 1.24 1.95
119
FOSSOMATIC CELL COUNT IN SHEEP MILK
TABLE 5. Mean SCC according to the sample age in refrigeration and freezing methods (foremilk and stripping milk). In freezing
In refrigeration SCC
I d
3 d
5 d
7 d
I d
7 d
15 d
Foremilk l x lO-3/ml Stripping2 x 1o-3/ml
116a 204a
115a 196ab
1lOa 192 b
l()()b 176c
84ab 148
78a 140
85 b 147
a,b,cMeans in a row with different superscripts differ (P < .05). lSE for foremilk = 2.15. 2SE for stripping milk = 3.39.
ml (FSCC). After 7 d of freezing, SCC were 46,000 and 63,000 cells/ml, respectively. The DMSCC allowed the verification of slight staining and a greater degree of nuclear degeneration after freezing. The effect of age in frozen samples varied (Table 5); counts were lower on d 7 than on d IS. Read et a1. (12) observed lower reduction percentages in the electronic and microscopic counts in older samples. Thus, the effect of age in frozen samples is not homogeneous. CONCLUSIONS
Our results suggest the need for standardization of age and storage conditions of milk samples to reduce variation of SCC as much as possible. Analysis of samples within the first 3 d after collection is particularly recommended. The milk must not be frozen. ACKNOWLEDGMENTS
This paper was developed within the project Gan90-0577, financed by the Comisi6n Interministerial de Ciencia y Tecnologia. REFERENCES 1 Gonzalo, C., and V. Gaudioso. 1983. Recuento celular en leche de oveja. Comparaci6n entre ordello mecamco y manual. Page 268 in Proc. 3rd Symp. Int. Ordello Mecanico Pequellos Rumiantes. Comite Espallol, Valladolid, Spain.
2 Green, T. J. 1984. Use of somatic cell counts for detection of subclinical mastitis in ewes. Vet. Rec. 114:43. 3 Greer, D.O., and J. L. Pearson. 1976. Reproducibility of electronic cell counts in milk; a study of 5 further factors, J. Dairy Res. 43:371. 4 Harvey, W, R. 1979, Least-Squares Analysis of Data with Unequal Subclass Numbers. ARS-USDA, ARSH4. Ohio State Univ" Columbus. 5 Harvey, W, R. 1990. User's guide for LSMLMW and MIXMDL. Mixed model least-squares and maximum likelihood computer program. Mimeo., Ohio State Univ., Columbus. 6lntemational Dairy Federation. 1984. Recommended Methods for Somatic Cell Counting in Milk. Doc. no. 168. Int. Dairy Fed. Brussels, Belgium. 7 Kennedy. B. W" M, S. Sethar, A. W. Tong, J. E. Moxley, and B. R. Downey. 1982. Environmental factors influencing test-day somatic cell counts in Holsteins. J. Dairy Sci. 65:275. 8 Miller, R. H., M. J. Paape, and J. C. Acton. 1986. Comparison of milk somatic cell counts by Coulter and Fossomatic counters. 1. Dairy Sci. 69: 1942. 9 National Mastitis Council. 1968. Subcommitee on screening test: direct microscopic somatic cell counts in milk. J. Milk Food Technol. 31:350. 10 Peris, c., P. Molina, N. Fernandez, M. Rodriguez, and A. Torres. 1991. Variation in somatic cell count, California mastitis test, and electrical conductivity among various fractions of ewe's milk. J. Dairy Sci. 74:1553. 11 Phipps, L. W. 1966. Determination of leucocyte concentrations in cow's milk with a Coulter counter. J. Dairy Res. 33:51. 12 Read, R. D., A. L. Reyes, and 1. G. Bradshaw. 1969. Effect of freezing milk samples on abnormal milk test results. J. Dairy Sci. 52:261. 13 Schalm, O. W., E. J. Carroll, and N. C. Jain. 1971. Bovine Mastitis. Lea and Febiger, Philadelphia, PA. 14 Schmidt, P. 1975. Fluoro-opto-electronic cell-counting on milk. J. Dairy Res. 42:227.
Journal of Dairy Science Vol. 76, No.1, 1993
infection (13). The main sec methods (direct microscopic, Coulter counter, and Fossomatic) have been completely standardized for cow milk (3, 9, 11, 14). Recently, interest in milk cell counting in sheep research and mastitis control programs has increased. However, sec methods for sheep milk have not been widely evaluated or tested (10). Some counting methods, such as direct microscopic SCC (DMSCC) and Coulter counter, have been modified 0, 2) to adapt to sheep milk, which has a higher fat content than cow milk. The Fossomatic SCC method (FSCC) automatically processes the milk samples so that they need no prior adaptation. However, optimal conditions for use of this method have not been specified for sheep milk. Basically, the Fossomatic counter is a fluorescence microscope. The ethidium bromide dye penetrates the cell and forms a fluorescent complex with the nuclear DNA. Each cell produces an electrical pulse, which is amplified and recorded. The objectives of this study were to compare the FSCC with the DMSCC (reference method) in sheep milk and to evaluate the effect of the storage method (fresh, refrigerated, and frozen milk) and of the sample age on the sce of foremilk and strippings.
The Fossomatic method for SCC was compared with the direct microscopic method in 85 half-udder samples of sheep milk. The correlation coefficient was .986. The repeatability of the Fossomatic method showed average variation coefficients less than 5%. The carry-over effect between samples was less than .5%. The effect of the storage method (fresh milk, refrigerated at 4°C and frozen at -19°C) and the sample age were studied in 48 samples of foremilk and strippings. The storage method had a significant effect on the SCC variation. The average fresh, refrigerated, and frozen sample counts were 125,000, 110,000, and 82,000 cells/ml for foremilk and 201,000, 192,000, and 145,000 cells/ml for strippings, respectively. The effect of age on the refrigerated samples was also significant; counts were reduced by about 14% from d 1 to 7 in both types of milk. The effect of age on the frozen sample varied. These results suggest standardization of age and storage conditions of the milk samples to reduce variation of SCC. The milk must not be frozen. (Key words: sheep milk, somatic cell count, Fossomatic method) Abbreviation key: DMSCC microscopic SCC method, FSee matic SCC method.
= direct = Fosso-
INTRODUCTION
Somatic cell count in milk is a widely used method to evaluate the status of cow udder
Received June 3. 1992. Accepted August 14, 1992. 1993] Dairy Sci 76:115-119
MATERIALS AND METHODS Comparison of the FSCC and the DMSCC
Eighty-five samples of sheep milk preserved in .05% potassium dichromate were analyzed within 24 h after milking by FSCe and DMSCC. The samples were collected from half-udders and represent sec varying from 15,000 to 4 million cells/mt. For the DMSeC, the milk was heated to 40°C in a water bath and held at that tempera-
115
116
GONZALO ET AL.
ture for 15 min before being cooled to 20·C by careful stirring. The method used was that recommended by the International Dairy Federation (6), adapted to sheep milk by Gonzalo and Gaudioso (I). Two slides of each sample were prepared and counted. Slides were stained with May Grunwald-Giemsa (Vorqufmica, Vigo, Spain). The working factor was 1600. For the FSCC, the milk samples were heated to 40·C in a water bath and held at this temperature for 15 min. The samples were then double processed in a Fossomatic 90 (AlS N. Foss Electric, Hillen'ld, Denmark). The reagents were prepared following the manufacturer's instructions, which coincide with the method recommended by International Dairy Federation (6). Repeatability
The repeatability of the FSCC was evaluated by counting a number of subsamples from 20 milk samples with different numbers of cells (from 46,000 to 12.1 million cells/ml). The coefficients of variation for the data of each sample were calculated. In order to determine the possible carryover effect from one sample to another in the FSCC, a sample with low SCC (79,900 cells! ml) was counted, alternating with a sample with high SCC (7.88 million cells/ml). First, 15 counts were carried out with the low cell concentration sample. Then 15 counts were made alternating between high and low SCC milk. The low count sample was then counted 15 more times. Effect of the Storage Method and Sample Age
Thirty-five milliliters of foremilk and 25 ml of hand strippings (after machine removal) were obtained from 48 half-udders from a total of 24 primiparous, midlactation Churra sheep during morning milking. Potassium dichromate to a final concentration of .05% was used as a preservative for all samples. After collection, each sample of original milk (foremilk and stripping) was divided into eight aliquots (4 and 3 ml, respectively). One aliquot was used for SCC on the sampling day (fresh milk). Four aliquots were refrigerated at 4·C and Journal of Dairy Science Vol. 76. No.1. 1993
counted I, 3, 5, and 7 d after collection. The remaining three aliquots were frozen at -19·C and analyzed I, 7, and IS d after collection. All samples were counted in a Fossomatic 90 following the previously mentioned specifications. The frozen samples were thawed in a water bath at ambient temperature for 45 min before being heated at 40·C. Statistical Analysis
The statistical study of the storage method and sample age factors was made using analysis of variance according to the model Yijk
= JL
+ Mi + Dj(i) +
<& + fijk
where y
= variable SCC and log SCC (of foremilk and strippings), M· = effect of milk storage method, Di(j) = effect of day j in method i, <2k = effect of half udder k or block effect, and fijk = residual effect. 1
The Harvey methodology with the LSMLMW program (4, 5) was used to estimate the components of variance as well as the contrast of the differences between means. RESULTS AND DISCUSSION Correlation Between DMSCC and FSCC
The following regression equation was calculated: FSCC = .954 DMSCC + 14,393; the correlation coefficient was r = .986 ± .0183. The mean SCC were 428,000 cells!ml for DMSCC and 423,000 cells!ml for FSCC. These results show excellent agreement between the two SCC methods for sheep milk. For cow milk, the correlation coefficients between both methods were .974 and .988 when DMSCC working factors of 1()4 and 1()2, respectively, were used (14). The working factor that we used for DMSCC was 103.2 . Repeatability
Table I shows the repeatability of the FSCC in milk samples with a wide range of
117
FOSSOMATIC CELL COUNT IN SHEEP MILK TABLE I. Repeatability of the Fossomatic method in sheep milk samples with different SCC. Sample
SubArithmetic samples mean 00-3
- - (no.) - - I 32 46 2 30 79 3 30 130 4 30 167 5 30 186 6 30 220 7 30 310 8 30 391 9 30 427 10 30 512 11 28 527 12 32 654 13 25 779 14 26 904 15 25 1034 16 28 1123 17 28 1719 18 26 4655 19 29 5009 20 12,100 30
SD
cells/mt) 4.6 7.7 6.7 10.6 12.4 13.8 12.0 16.3 18.9 19.5 17.6 20.6 30.3 30.2 20.0 28.1 60.8 148.6 97.8 346.5
CV (%)
10.0 9.7 5.1 6.3 6.6 6.2 3.8 4.1 4.4 3.8 3.3 3.1 3.8 3.3 1.9 2.5 3.5 3.2 1.9 2.8
sec. The coefficients of variation decrease as the number of cells increases and, in most cases, are less than 5%. The International Dairy Federation (6) recommends coefficients of variation of the FSee lower than 7.5% for milk samples with cell counts between 400,000 and 600,000 cells/m1, which coincides with our results. The carry-over effect, when a milk sample with a low count was counted alternately with a sample with a much higher cell count, was
only .5%, which corresponds to an increase of 400 cells/ml when the cell content of the high count sample was 100 times greater (7.88 million cells/ml) than that of the low count sample (79,900 cells/ml). Thus, the carry-over effect between consecutive samples is not significant. Storage Method and Sample Age
Table 2 shows the analysis of variance results for foremilk and strippings, indicating the statistical significance and the percentage of variance for each variation factor. The block or half-udder variation factor was highly significant (P < .001), as expected. The storage method had a highly significant effect (P < .001); the proportion of explained variance was important (5.69 and 4.10% in foremilk and strippings, respectively). The effect of the sample age was also statistically significant, especially in the refrigeration method (P < .001). The analysis of variance for log see of foremilk and strippings (Table 3) decreased the percentage of variance explained by the block or half-udder variation factor and emphasized the variation percentages explained by other variation sources (storage method and sample age). Table 4 shows mean sec for each of the storage methods studied. The highest sec were for fresh milk. Refrigeration reduced the sec significantly, but this decrease was more marked in the freezing method, in which sec were much lower (P < .001) than those for refrigerated milk.
TABLE 2. Analysis of variance of the variation factors and proportion of variance explained by storage method, sample age in each method, and half-udder for the SCC variable of foremilk and stripping milk. SCC Stripping milk
SCC Foremilk Source of variation
df
Variance explained
F
Variance explained
F
(%)
Method Half-udder Days in refrigeration
2 47 3
212.5*** 237.8*** 11.6***
2 329
3.2*
(0/0)
5.69 90.79
199.0*** 316.9*** 12.4***
.46 Days in freezing Residual
4.10 93.19 .35
1.8NS• 1 3.06
2.36
lp> .05.
*p < .05. ***p < .001. Journal of Dairy Science Vol. 76, No. I, 1993
118
GONZALO ET AL.
TABLE 3. Analysis of variance of the variation factors and proportion of variance explained by storage method. sample age in each method. and half-udder for the logarithm see variable of foremilk and stripping milk. log see Foremilk Source of variation
df
log see Stripping milk
Variance explained
F
Variance explained
F
(%)
(%)
Method Half-udder Days in refrigeration
2 47 3
267.7·" 124.8··.. 13.0""
2 329
10.3··..
12.28 81.27
278.4"" 143.9...... 15.9·....
11.30 83.01 1.04
1.20 Days in freezing Residual
5.4" 4.64
5.25
"p < .01. ""p < .001.
The contrast of the difference between means, obtained from the ANOVA, was carried out for the see variable (Tables 4 and 5) and for its logarithmic transformation. In general, the contrasts carried out using logarithmic transformation gave results similar to those of the untransformed variable, although statistical significance of the differences was, in some cases, somewhat lower. Tables 4 and 5 show how the effect of refrigeration on the see was produced after 24 h of storage at 4°e for foremilk and 72 h for strippings. In stripping milk, the see for fresh milk (201,000 cells/ml) and for milk refrigerated for 24 h (204,000 cells/ml) did not differ, in contrast with results of Miller et al. (8) for cows, which showed a significant decline (P < .05) in stripping counts in the first 24 h in samples kept at ambient temperature. In refrigerated samples, see decreased by approximately 14% from the d 1 to 7 in both types of milk (Table 5). Kennedy et al. (7) recorded a much greater decrease (from 28 to
TABLE 4. Mean
see
36%) in cow milk samples preserved in potassium dichromate and kept at temperatures between 20 and 25°e for 8 d or longer. These results suggest that storage at ambient temperature may have a more negative effect on see than refrigeration. However, by using orthogonal contrasts, we have observed a very significant decline in the count (P < .001) in refrigeration. This decline was more marked in older samples (d 7), which also explains the results of Kennedy et al. (7). Freezing reduced see by 28 to 34% in both types of milk. This effect was comparable with that previously recorded by others (12, 14) in cow milk (22 to 57%) and may have been caused by cell nuclei that were less able to absorb the DNA-specific stain after freezing. A small auxiliary study was carried out to confirm this idea. Twenty samples of foremilk were counted using both methods (DMSee and FSeC) on d 0 (fresh milk) and d 7 (frozen milk). The average fresh milk counts were 103,000 cells/ml (DMSeC) and 110,000 cells/
of foremilk and stripping milk for each of the three storage methods studied. Storage method
see Foremilk x 1O-3/ml Stripping x 1O-3/ml
Fresh
Refrigerated
Frozen
X
X
X
125' 201'
SE 2.15 3.39
,.b,CMeans in a row with different superscripts differ (P < .05). Journal of Dairy Science Vol. 76. No.1. 1993
1l0b 192b
SE 1.07 1.69
82c 145C
SE 1.24 1.95
119
FOSSOMATIC CELL COUNT IN SHEEP MILK
TABLE 5. Mean SCC according to the sample age in refrigeration and freezing methods (foremilk and stripping milk). In freezing
In refrigeration SCC
I d
3 d
5 d
7 d
I d
7 d
15 d
Foremilk l x lO-3/ml Stripping2 x 1o-3/ml
116a 204a
115a 196ab
1lOa 192 b
l()()b 176c
84ab 148
78a 140
85 b 147
a,b,cMeans in a row with different superscripts differ (P < .05). lSE for foremilk = 2.15. 2SE for stripping milk = 3.39.
ml (FSCC). After 7 d of freezing, SCC were 46,000 and 63,000 cells/ml, respectively. The DMSCC allowed the verification of slight staining and a greater degree of nuclear degeneration after freezing. The effect of age in frozen samples varied (Table 5); counts were lower on d 7 than on d IS. Read et a1. (12) observed lower reduction percentages in the electronic and microscopic counts in older samples. Thus, the effect of age in frozen samples is not homogeneous. CONCLUSIONS
Our results suggest the need for standardization of age and storage conditions of milk samples to reduce variation of SCC as much as possible. Analysis of samples within the first 3 d after collection is particularly recommended. The milk must not be frozen. ACKNOWLEDGMENTS
This paper was developed within the project Gan90-0577, financed by the Comisi6n Interministerial de Ciencia y Tecnologia. REFERENCES 1 Gonzalo, C., and V. Gaudioso. 1983. Recuento celular en leche de oveja. Comparaci6n entre ordello mecamco y manual. Page 268 in Proc. 3rd Symp. Int. Ordello Mecanico Pequellos Rumiantes. Comite Espallol, Valladolid, Spain.
2 Green, T. J. 1984. Use of somatic cell counts for detection of subclinical mastitis in ewes. Vet. Rec. 114:43. 3 Greer, D.O., and J. L. Pearson. 1976. Reproducibility of electronic cell counts in milk; a study of 5 further factors, J. Dairy Res. 43:371. 4 Harvey, W, R. 1979, Least-Squares Analysis of Data with Unequal Subclass Numbers. ARS-USDA, ARSH4. Ohio State Univ" Columbus. 5 Harvey, W, R. 1990. User's guide for LSMLMW and MIXMDL. Mixed model least-squares and maximum likelihood computer program. Mimeo., Ohio State Univ., Columbus. 6lntemational Dairy Federation. 1984. Recommended Methods for Somatic Cell Counting in Milk. Doc. no. 168. Int. Dairy Fed. Brussels, Belgium. 7 Kennedy. B. W" M, S. Sethar, A. W. Tong, J. E. Moxley, and B. R. Downey. 1982. Environmental factors influencing test-day somatic cell counts in Holsteins. J. Dairy Sci. 65:275. 8 Miller, R. H., M. J. Paape, and J. C. Acton. 1986. Comparison of milk somatic cell counts by Coulter and Fossomatic counters. 1. Dairy Sci. 69: 1942. 9 National Mastitis Council. 1968. Subcommitee on screening test: direct microscopic somatic cell counts in milk. J. Milk Food Technol. 31:350. 10 Peris, c., P. Molina, N. Fernandez, M. Rodriguez, and A. Torres. 1991. Variation in somatic cell count, California mastitis test, and electrical conductivity among various fractions of ewe's milk. J. Dairy Sci. 74:1553. 11 Phipps, L. W. 1966. Determination of leucocyte concentrations in cow's milk with a Coulter counter. J. Dairy Res. 33:51. 12 Read, R. D., A. L. Reyes, and 1. G. Bradshaw. 1969. Effect of freezing milk samples on abnormal milk test results. J. Dairy Sci. 52:261. 13 Schalm, O. W., E. J. Carroll, and N. C. Jain. 1971. Bovine Mastitis. Lea and Febiger, Philadelphia, PA. 14 Schmidt, P. 1975. Fluoro-opto-electronic cell-counting on milk. J. Dairy Res. 42:227.
Journal of Dairy Science Vol. 76, No.1, 1993