Livestock Research for Rural Development 26 (10) 2014 Guide for preparation of papers LRRD Newsletter

Citation of this paper

Lactation curves of Friesian-Sanga and Sanga cows in Ghana

K Darfour-Oduro, B A Hagan and A Asafu-Adjaye

CSIR-Animal Research Institute, P. O. Box AH 20, Achimota, Accra, Ghana


Average daily milk yield records of 81 Friesian-Sanga crossbred cows obtained from 1999 to 2008 and for 46 Sanga cows obtained from 2003 to 2008 were used to abstract lactation curve and their parameters for the incomplete gamma function:  yn=anbe-cn; where yn is the predicted average daily milk yield in week n; a is a constant representing the level of initial yield of the cow, b is representing the rate of increase to peak milk production and c is the rate of decline after peak milk production.  


Season of calving significantly influenced only parameter b for Friesian-Sanga cows and all three parameters of Sanga cows.  Lactation number was important as a source of variation for only Friesian-Sanga cows with their parameters a and c being significantly affected.  The year an animal calved had a significant effect on parameter c for Friesian-Sanga cows and parameters a and b for their Sanga counterparts.  Significant negative correlations were observed between parameters a and b and parameters b and c for both breeds.  Residual mean squares and correlation analysis between actual and predicted values indicated that second parity Friesian-Sanga cows and their primiparous Sanga counterparts were the best and worst respectively in terms of how good the incomplete gamma function fitted the data.  Lactation curves derived by the incomplete gamma function in this study were similar to those observed for other cows in the tropics.  The incomplete gamma function can be relied on in terms of fitting lactation curves for these two cattle breeds in Ghana.    

Keywords: Accra plains, incomplete gamma function, parity, season of calving, year of calving


Lactation curve described as a graphical representation of the relationship between milk yield and lactation length (Fadlelmoula et al 2007) and whose shape has been used to indicate depressed performance as a result of environmental stresses (Nassuna-Mosoke et al 2007). It is further useful for health monitoring, genetic evaluation, feeding and economic management decisions of dairy cows (Scherchand et al 1995; Bouallegue et al 2013).  Effective management of animals for optimum milk production under the stressful environmental conditions of Ghana is lacking.


Parameters of lactation curves of indigenous cattle in Ghana in smallholder farms (Okantah 1992) with limited records and Holstein-Friesians raised on-station (Ahunu and Kabuga 1994; Ahunu et al 1999) cannot be applied to other breeds of cattle as these parameters differ depending on the breed in question.  Furthermore, for the same breed, parameters may differ depending on the prevailing environment in which the breed finds itself.


Abstraction of lactation curves for cattle has been possible by employing several algebraic equations. However the incomplete gamma function (Wood 1969) is the best known for describing the lactation pattern.  The parameters a, b and c associated with the lactation curve represent milk production at the beginning of lactation, slope before peak milk production and slope after peak milk production (Wood 1969).  These parameters are influenced by year of calving, season of calving and parity of cow (Ahunu and Kabuga 1994; Mario et al 2005; Aziz et al 2006; Fadlelmoula et al 2007; Kopec et al 2013).


The Friesian-Sanga and Sanga are two breed groups of cattle raised by the Animal Research Institute of the Council for Scientific and Industrial Research (CSIR-ARI), Ghana with the objective of evolving a dual-purpose cattle breed for use on the Accra Plains.  Knowledge of the lactation potential of these breed groupss will inform the decision as to the necessary interventions needed for optimal performance.


The objectives of this study is to describe the lactation curve of the Friesian-Sanga and Sanga cows using the incomplete gamma function and determine the effects of some environmental factors on the curve’s parameters. 

Materials and methods

Location and management of cattle herds


The study was based on data from milk records of  Sanga (Photo 1) and Friesian-Sanga crossbred (Photo 2) cows kept at the Animal Research Institute on the Accra Plains of Ghana. The area has a bimodal rainfall pattern with a major wet season occurring from April to July and a minor season from September to November. The remaining months constitute the dry period. Annual rainfall and temperature ranges between 600-1000 mm and 15-34C respectively (Okantah et al 2005). The management of the animals was described by Sottie et al (2009). Purposely, the animals were raised under agropastoral system to acclimatize them to the management practices on the Accra plains.

Photo 1. Sanga cows Photo 2. Friesian x Sanga crossbred

Partial milking was practised. Under this system of milking, calves were separated from their dams in the evening and were brought to suckle for a few minutes to stimulate milk let down before milking (Karikari et al 2008). The daily milk was collected in a bowl and transferred into a measuring cylinder after which the volumetric values were converted into weight using a milk specific gravity of 1.03. Daily milk yield from the cows were recorded in a small notebook and later entered in Excel spreadsheet.


Statistical analysis


A total of 31,709 daily milk records collected from 1999 to 2008 involving 81 Friesian-Sanga crossbred cows in 136 lactations and 13,836 daily milk records involving 46 Sanga cows in 67 lactations obtained from 2003 to 2008 were used in this study. Data were analysed separately for each breed group. Lactation curve was described by the incomplete gamma function, as suggested by Wood (1967):    






  yn is the average daily milk yield in week n

  a, b, c are parameters describing the curve’s shape

  e is the base of the natural logarithm

  n is weeks for which the cow has been lactating   


Parameter a is a constant representing the level of initial yield of the cow, b is representing the rate of increase to peak milk production and c is the rate of decline after peak milk production.  The parameters a, b and c were estimated by fitting the incomplete gamma function to individual cows using the nonlinear regression procedures (SAS 1999). The effects of season of calving, parity and year of calving on each of the estimated parameters were examined using the GLM procedure of SAS (SAS 1999). Seasons were classified as major rains (April-July), minor rains (September-November) and the dry period (December-March, August). Friesian-Sanga cows had their parities classified as 1st, 2nd, 3rd and 4th.  Sanga cows’ parities were classified as 1st, 2nd and 3rd. Year was classified as 1999-2000, 2001-2002, 2003-2004, 2005-2006 and 2007-2008.

Differences among means of a parameter for different factors were analysed by the PDIFF/SAS. The statistical model for the parameters was as follows: 


Yijkl= µ+Si+Pj+Tk+eijkl




Yijkl is the lth coefficient of the Wood curve for the ith season of calving, jth parity of dam and kth year of calving  

µ is the overall mean

Si is the fixed effect of the ith season of calving

Pj is the fixed effect of the jth parity of dam

Tk is the fixed effect of the kth year of calving

eijkl is the random error term associated with each observation 


The correlation procedure (SAS 1999) was used to analyze the correlation among parameters a, b and c.  Mean lactation curves were also fitted for all parities using the incomplete gamma function. Residual mean squares and correlation coefficients among observed and predicted average daily milk yields were used to determine how well the incomplete gamma function fitted the data for each parity.  


The means and standard errors of the lactation curve parameters for Friesian-Sanga and Sanga cows are presented in Tables 1 and 2, respectively. Friesian-Sanga cows, on the average, had higher parameters a and c but a lower parameter b as compared with their Sanga counterparts.

Table 1. Least square means and standard error of lactation curve parameters

of Friesian-Sanga cows











Season of calving




0.309 ±0.046a












Lactation number





















Year of calving


























abc  Means in a column with different subscripts differ significantly (P<0.05)

Season of calving  


Parameter b for Friesian-Sanga cows calving in the major rainy season differed (P<0.05) from that of cows calving in the minor rainy season (Table 1). Season of calving affected (P<0.05) all parameters of Sanga cows with major rainy season calvers surpassing dry season calvers (Table 2). 


Lactation number influenced (P<0.05) parameters a and c of Friesian-Sanga cow. Parameter a reached a maximum at third lactation while parameter c was highest at the first and third lactations. A decline in parameters a and c was evident at the fourth lactation. Lactation number however had no effect (P>0.05) on any parameter of Sanga cows.


Differences between parameter c for Friesian-Sanga cows and parameters a and b for Sanga cows were observed for years of calving. In general, a trend of increasing parameter c for Friesian-Sanga cows and parameter a for Sanga cows was observed with advancing years from 2003-2004.

Table 2. Least square means and standard error of lactation curve parameters of Sanga cows











Season of calving
















Lactation number
















Year of calving
















abc  Means in a column with different subscripts differ significantly (P<0.05)    

Correlation among parameters


With the exception of low correlations between parameters a and c for both breeds, all other correlations were high (P<0.01, Table 3). Correlations between parameters for Friesian-Sanga cows were negative.  Similarly, correlation between parameters for Sanga cows were negative except for a positive correlation between parameters a and c.

Table 3. Correlations among lactation curve parameters for Friesian-Sanga and Sanga cows         






















*Significant at (P<0.01), other correlations are not significant

The goodness-of-fit statistics for the gamma function for each parity are presented in Table 4.  Among the four lactation curves for Friesian-Sanga cows, the gamma function fitted average daily milk records best for second parity as indicated by the lowest residual mean square error of 0.005 and a correlation of 0.949 between actual and predicted values. All correlations among actual and predicted values for all parities were high and positive except for primiparous Sanga cows that had a low correlation of 0.154.

Table 4. Residual mean squares and simple correlations between actual and predicted values of the various lactation numbers as estimated by the Wood model for Friesian-Sanga and Sanga cows

Lactation number

Residual mean squares





























*Significant at (P<0.01)

Figures 1 and 2 show the lactation curves by parities of Friesian-Sanga and Sanga cows respectively. Lactation curves based on the gamma function of Friesian-Sanga cows were similar for parities 1 and 2 (Figures 1, top left and top right). Third parity Friesian-Sanga cows had a curve that was concave in shape (Figure 1, bottom left), and therefore did not follow normal lactation curve patterns. In contrast, fourth parity lactation curve of the same breed typified a normal lactation curve (Figure 1, bottom right).

Figure 1a-d: Observed and predicted lactation curves for first (top left), second (top right), third (bottom left) and fourth (bottom right) parities Friesian-Sanga cows

For primiparous Sanga cows, the lactation curve as estimated by the gamma function was more or less flat (Figure 2, top left), underestimating milk yield in the periods preceding the 5th week of lactation, overestimating milk yield between the 5th and 20th week of lactation, underestimating it from the 20th week to the 35th week and finally, overestimating milk yield after the 35th week.  Subsequent parities had curves (Figures 2, top right and bottom left), which to a large extent typified a normal lactation curve.

Figure 2a-c: Observed and predicted lactation curves for first (top left), second (top right) and third (bottom left) parities Sanga cows


Our results indicate that the incomplete gamma function can describe the lactation curve of both Friesian-Sanga and Sanga cows. 

In this study, parameter a, representing average initial milk yield for both Friesian-Sanga and Sanga cows is low and is indicative of the stressful environment the animals are raised. As the two breeds were raised under similar conditions especially between 2003 and 2008, differences among overall lactation curve parameter values could be attributed to genetic effects. The values for parameter a, reported in this study for both breeds were lower than the 4.095 and 1.869 for Bonsmara and Nguni breeds respectively using the weigh-suckle-weigh technique (Maiwashe et al 2013). The differences in initial milk yield could be attributed to both breed and milking technique.


Seasonal difference observed in performance of animals raised in the tropics is expected to be primarily a manifestation of variation in feed availability (Javed et al 2000). Generally, major rainy season calvers in this study had higher values for their lactation curve parameters than their dry season counterparts, an observation that could be explained by the fact that forage in the major rainy season are high in quantity and quality. Thus, animals calving in the major rainy season are better conditioned for milk production as compared with dry season calvers.


Increasing parameter a with increasing parities for Friesian-Sanga cows in this study corroborates observations by Okantah 1992; Ahunu and Kabuga 1994; Osorio-Arce et al 2005 Aziz et al 2006 and Bouallegue et al 2013 and indicate that generally older cows have higher initial milk production. Possible reasons for this observation are the fact that first parity cows normally are still developing and therefore need to meet their maintenance and growth requirement than produce milk  (Osorio-Arce et al 2005) and a tendency for greater feed intake in older cows than young ones (Singh and Gopal 1982). Stanton et al (1992) also explained that the body and mammary gland of young cows are still in the developmental stage hence the low initial milk yield of first parity cows. The relatively low initial milk production and rate of decline after peak production exhibited by Friesian-Sanga cows in their fourth parity tends to confirm the observation that cattle in the tropics attain peak production in their third or fourth lactation (Okantah 1992) and performance decline thereafter. For Sanga cows, the three parity levels may not have been enough to elicit significant differences in parameters a, b and c. 


As management of animals in this study have not changed much over the years, observed significant differences among parameters for year of calving could be due to annual climatic changes. Significant effect of year of calving on lactation curve parameters has been reported (Kaygisiz 1999; Rekik et al 2003; Yuksel and Yanar 2009). Increasing parameter c with advancing years for Friesian-Sanga cows gives an indication that persistency, which is the ability to maintain milk production after peak yield, decreases with progression in years. 


The negative correlations between parameters a and b observed for both Friesian-Sanga and Sanga cows is consistent with other studies (Ahunu and Kabuga, 1994; Terkeli et al 2000; Atashi et al 2009) and indicate that higher initial yield is associated with a lower rate of increase until peak yield. The negative correlations between parameters b and c for both breeds imply that cows that reach their peak faster have a slower decline after peak.


A flat lactation curve, a characteristic of primiparous Sanga cows in this study is evident of low nutrient intake (Licitra et al 1998) and high variability in milk performance of first parity cows. The weak correlation of 0.154 between predicted and observed values for primiparous cows may show an inability of the incomplete gamma function to accurately describe the lactation trajectory of these cows and calls for investigation of the suitability of other models for a more precise fit to the data. 


The relatively large variations at the end of the lactation curves could be partly because late lactation data is more variables leading to the difficulties in prediction by the incomplete gamma function (Kawonga et al 2012). 


Cows in other parities for both breeds except third parity Friesian-Sanga cows displayed standard curves, with goodness-of-fit statistics of mean square error and correlations indicating a good fit of model to the data. These standard curves, generally had transient peaks and a slow rate of decline after peak yield. Such curves reflect a low level of protein intake (Chamberlain 1993) and a stressful environment. 


A distinctive feature of third parity Friesian-Sanga cows was their continuously decreasing lactation curve, lacking a lactation peak, normally described as atypical (Olori 1999) and arises as a result of a negative b parameter, the parameter associated with the increasing phase of the lactation curve (Shanks et al 1981).



We are grateful to all the members of the Ruminant and Natural Resource Management Division of Animal Research Institute, Ghana and particularly Messrs A. Mamah, A. Alhassan and S.A. Quaye for their assistance in collecting data for this study.


Ahunu B K, Kabuga J D and Karikari P K 1999 A comparison of different models to characterize lactation curves of Friesians in a humid forest zone.  Ghana Journal of Agricultural Science 32: 31-38. 

Ahunu B K and Kabuga J D 1994 A study of the gamma function for describing lactation records in Ghana. Bulletin of Animal Health and Production in Africa 42: 147-152. 

Atashi H, Shahrbabak M M, and Shahrbabak H M 2009 Environmental factors affecting the shape components of the lactation curves in Holstein dairy cattle of Iran.  Livestock Research for Rural Development 21 (5), Article #60.  Retrieved June 7, 2010 from  

Aziz M A, Shalaby N A, El-Shafie O M, Mahdy A T and Nishida A 2006 Comparison between the shapes of lactation curve of Egyptian buffalo milk yield estimated by the incomplete gamma function and a new model.  Livestock Research for Rural D 18 (5), Article #59.  Retrieved June 7, 2010, from

Bouallegue M, Haddad B, Aschi M S and Hamouda M B 2013 Effect of environmental factors on lactation curves of milk production traits in Holstein-Friesian cows reared under North African conditions. Livestock Research for Rural Development. Volume 25, Article #75. Retrieved August 28, 2014, from

Chamberlain A 1993 Milk production in the tropics.  Longman, Essex. 

Fadlelmoula A A, Yousif I A and Abu Nikhaila A M 2007 Lactation curve and persistency of crossbred dairy cows in the Sudan. Journal of Applied Sciences Research 3(10): 1127-1133.  

Javed K, Mohiuddin G and Abdullah M 2000 Environmental factors affecting various productive traits in Sahiwal Cattle. Pakistan Veterinary Journal 20 (4): 187-192.  

Kawonga B S, Chagunda M G G, Gondwe T N, Gondwe S R and Banda J W 2012 Use of lactation models to develop a cow performance monitoring tool in smallholder dairy farms. Archiv Tierzucht 55(5): 427-437. 

Kaygisiz A 1999 Lactation curve traits of Simmental Cattle. Turkish Journal of Veterinary and Animal Sciences 23: 15-24.

Kopec T, Chladek G, Kucera J, Falta D, Hanus O and Roubal P 2013 The effect of the calving season on the Wood’s model parameters and characteristics of the lactation curve in Czech Fleckvieh cows. Archiv Tierzucht (Archives Animal Breeding 56:   

Licitra G, Lauria F, Carpino S, Schadt I, Sniffen C J and Van Soest P 1998 Improvement of the Streptomyces griseus method for degradable protein in ruminant feeds. Animal Feed Science and Technology 72: 1-10. 

Maiwashe A, Nengovhela N B, Nephawe K A, Sebei J, Netshilema T, Mashaba H D, Nesengani L and Norris D 2013 Estimates of lactation curve parameters for Bonsmara and Nguni cattle using the weigh-suckle-weigh technique. South African Journal of Animal Science 43(5): S12-S16. 

Nassuna-Musoke G M, Kabasa J D and John King M 2007 Lactation Curve of Temperate Friesians on Smallholder Farms in Tropical Central Uganda. Journal of Animal and Veterinary Advances 6(8): 950-954. 

Okantah S A 1992 Partial milking of cattle in smallholder herds on the Accra Plains: some factors affecting daily partial milk yield and milk composition. Animal Production 54: 15-21. 

Olori V E, Brotherstone S, Hill W G and McGuirk B J 1999 Fit of standard models of the lactation curve to weekly records of milk production of cows in a single herd. Livestock Production Science 58: 55-63. 

Osorio-Alce M M and Segura-Correa J C 2005 Factors affecting the lactation curve of Bos taurus x Bos indicus cows in a dual purpose system in the humid tropics of Tabasco, Mexico. Tecnica Pecuaria en México 43(1): 127-137. 

Rekik B, Ben Gara A, Ben Hamouda M and Hammami H 2003 Fitting lactation curves of dairy cattle in different types of herds in Tunisia. Livestock Science 83(2): 309-315. 

Ruiz R, Oregui L M and Herrero M 2000 Comparison of models for describing the lactation curve of Latxa Sheep and an analysis of factors affecting milk yield. Journal of Dairy Science 83: 2709-2719. 

SAS, 1999 Statistical Analysis Systems user’s guide (6th ed.). SAS Institute Inc, Raleigh, North Carolina, USA. 

Shanks R D, Berger P J, Freeman A E and Dickinson F N 1981 Genetic aspects of lactation curves. Journal of Dairy Science 64: 1852-1860. 

Sherchand L, McNew R W, Kellog D W and Johnson Z B 1995 Selection of a mathematical model to generate lactation curves using daily milk yields of Holstein cows. Journal Dairy Science 78: 2507-2513. 

Singh R P and Gopal R 1982 Persistency and peak milk yield of cattle in a rural area. Indian Journal Animal Science 52(7): 487-489. 

Stanton T L, Jones L R, Everett R W and Kachman S D 1992 Estimating milk, fat and protein lactation curves with a test day model. Journal of Dairy Science 75: 1691-1700.

Tekerli M, Akinci Z, Dogan I and Akcan A 2000 Factors Affecting the Shape of Lactation Curves of Holstein Cows from the Balikesir Province of Turkey. Journal Dairy Science 83: 1381-1386. 

Wood P D P 1969 Factors affecting the shape of the lactation curve in cattle. Animal Production 11: 307-316. 

Yuksel S and Yanar M 2009 Determination with different mathematical models of lactation curve parameters of Brown Swiss cows and the effects of some environmental factors on these parameters.  Journal of Lalahan Livestock Research Institute 49(1): 17-26.

Received 6 September 2014; Accepted 17 September 2014; Published 3 October 2014

Go to top