Livestock Research for Rural Development 22 (2) 2010 Guide for preparation of papers LRRD News

Citation of this paper

Genetic parameters for milk yield, age at first calving and interval between first and second calving in milk Murrah buffaloes

L O Seno, V L Cardoso**, L El Faro**, R C Sesana*, R R Aspilcueta-Borquis*, G M F de Camargo* and H Tonhati*

Federal University of Grande Dourados, Faculty of Agricultural Sciences - FCA/UFGD. Rodovia Dourados à Itahum, km 12. P. O. Box 533, CEP 79.804-970, Dourados – MS, Brazil
* São Paulo State University, Faculty of Agriculture and Veterinary Medicine, Departament of Animal Sciences - FCAV/UNESP. Via de Acesso Prof. Paulo Donato Castellane s/n. CEP 14884-900, Jaboticabal – SP, Brazil
** Agência Paulista de Tecnologia dos Agronegócios - APTA, Pólo Regional Centro Leste. Avenida Bandeirantes, 2419. CEP 14.030-670, Ribeirão Preto – SP, Brazil
[This study was supported by Fundação de Amparo à Pesquisa do Estado de São Paulo (Fapesp) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)]


In the present study, data of 1,578 first lactation females, calving from 1985 to 2006 were analysed with the purpose of estimating genetic parameters for milk yield (MY), age at first calving (AFC) and interval between first and second calving (IBFSC) in dairy buffaloes of the Murrah breed in Brazil.


Heritability estimates for MY, AFC and IBFSC traits were 0.20, 0.07 and 0.14, respectively. Genetic correlations between MY and AFC and IBFSC were -0.12 and 0.07, respectively, while the corresponding phenotypic correlations were -0.15 and 0.30, respectively. Genetic and phenotypic correlations between AFC and IBFSC were 0.35 and 0.37, respectively. Genetic correlation between MY and AFC showed desirable negative association, suggesting that daughters of the bulls with high breeding values for MY could reach physiological mature at a precocious age. Genetic correlation between MY and IBFSC, showed that the selection for milk production could result in the increase of calving intervals.

Keywords: dairy buffalo, genetic and phenotypic correlation, productive and reproductive traits


The Brazil south-east region has an increasing demand of high quality dairy products such as the buffalo cheese. So, the buffalo production in this region is designated to the milk production and its industrialization. The increase of dairy products demand along the last decades promoted great valuation of the bubaline specie in the scene of dairy cattle. The producers realized the great potential of this market and started to invest in this activity. Basically, the investments were applied in improvements of conditions and/or management practices.


According to Bagnato and Oltenacu (1993) milk yield and fertility are the main factors that affect the profitability of milk herds. As the milk yield is related to the variations in the reproductive activity, then the shorter calving intervals can be associated to bigger milk production during the animal’s productive life, besides the possible increase in the number of calves per year. Thus, the genetic importance of the fertility in these herds must be evaluated according to the reproductive performance of the buffalo and its relations to the milk yield. However, there is an antagonism between the milk production and the fertility of an animal. Pryce et al (2002) showed that there is a genetic correlation between milk yield and calving intervals that vary from 0.22 and 0.67. It indicates that cows with high milk yield merit have a bad reproductive performance. However, a good management may improve the reproduction performances and changes this situation (Roxström et al 2001).


Cassiano et al (2004) estimated variance components and genetic parameters of the traits: interval between first and second calving, age at first calving and interval between calvings. The buffaloes used were from Carabao, Jafarabadi, Mediterranean and Murrah breeds in the Brazilian Amazonia region. The heritability estimates for interval between first and second calving varied from 0.04 to 0.05, for interval between calvings varied from 0.0 to 0.26. These results indicate a big environmental influence. The heritability estimates for age at first calving varied from 0.12 to 0.38.


Ramos et al (2006) studied the traits of milk yield and interval between calvings in buffaloes. The heritabilities estimates were 0.21 and 0.22 to the milk yield and interval between calvings, respectively. The genetic, phenotypic and environmental correlations between the traits were -0.22, 0.01 and 0.03 respectively. The authors concluded that the negative correlation suggests an existence of an antagonism between milk yield and interval between calvings.


The aim of this present study was to obtain the estimation of the components of variance, heritability and genetic and phenotypic associations between the milk yield, age at first calving and interval between first and second calving in dairy buffaloes and also the calculation of correlated answers and genetic tendency of traits during the years.


Material and methods 

The information used was originated from the Dairy Bubaline Test Program developed by the Animal Science Department of São Paulo State University in Jaboticabal - SP, Brazil. The database includes 13 herds of dairy Murrah buffaloes. These herds are distributed in the state of São Paulo. The production system was described for Seno et al (2006). The animals were raised in Brachiaria decumbens pastures and supplemented with 1 kg of concentrate feed (with 20% crude protein and 72.3% of TDN) for each 3 kg of milk produced. In the dry season (April to October), they were also supplemented with forage specially sugar cane, urea and mineral salt ad libitum. The buffaloes were milked twice a day with a milking machine. The milkings were done with the calves closer in order to stimulate the buffaloes’ milk production. The measures of milk yield occurred monthly.


Lactations which have less than 90 days and more than 400 days were eliminated, as well as, milk yields under 300 kg and higher than 3,300 kg. The contemporary groups that have less than 5 observations were also eliminated. After these preliminary restrictions and eliminations pertinent to each trait, it remained in the database information of 1,578 buffaloes at first lactation that given calves from 1985 to 2005. The traits analyzed were milk yield (MY), age at first calving (AFC) and interval between first and second calving (IBFSC).


The pedigree archive used in the analyses had identification of the animal, father and mother which totalized 11,749 animals in the relative matrix. This archive was revised using the farms’ archives for the animals that were born from 1974 to 2006. The calving seasons were divided as: April to September (season 1) and October to March (season 2).


Single and two-trait analyses were realized to MY, AFC and IBFSC traits using the animal model. The variance components were obtained by restricted maximum likelihood method (MTDFREML), developed for Boldman et al (1995). For MY and IBFSC traits the model included herd-year- calving season (contemporary group of calving - CGC) as fixed effects, and age of buffalo as covariate (linear and quadratic), animal and error as random effects. The model for AFC included the same random effects and the fixed effects of herd-year-season of birth (contemporary group of birth - CGB). The convergence criterion was 10-9 and to each convergence, the program was restarted with estimates of previous apparent convergence as initial values.


The animal genetic values were predicted to each trait and were estimated the simple correlations (Pearson) and rank correlations (Spearman) (SAS 1999). The genetic gain and the correlated answer were also calculated using the heritability estimates the phenotypic standard deviations obtained for the population analyzed. A selection of the 5% best sires was adopted, it corresponds to a selection intensity of 2.06 (Lush 1964). The females were randomly substituted in the herds and the selection intensity was equal to zero. The intensity factor means to the studied traits were 1.03. Genetic trends to the traits were estimated by linear regression of average breeding values on birth year of the animals.


Result and discussion 

The number of observations (N), the means, the standard deviations (SD), and the minimum and maximum values of the contemporary groups for the traits: MY, AFC and IBFSC are in table 1. The mean observed to the MY trait was 1,594.4 kg in a mean of 271.6 days of lactation. It was superior to the mean obtained by Tonhati et al (2000a) which was 1,259.47 kg and similar to the ones obtained by Ramos et al (2006) and Tonhati et al (2000b) which were 1,650.00 kg and 1,496.00 kg, respectively. All these studies were done in Brazil with Murrah buffaloes. Shabade et al (1993) studying Murrah buffaloes in India and Rosati and van Vleck (2002) studying Mediterranean buffaloes in Italy obtained bigger values for the same trait which were 1,892.21 kg and 2,286.80 kg, respectively.


The mean obtained to the trait AFC (Table 1) was 1,093.6 days, approximately 36 months. It was lower than the means obtained by Tonhati et al (2000b) and Mohamed et al (1993) which were 39 and 38 months, respectively. No data of IBFSC trait in buffaloes was found in the literature to compare with the ones obtained in the present work. In dairy cattle, Dong and van Vleck (1989) observed means of 382 and 386 days of IBFSC in two different data set of Holstein cows. Ramos et al (2006) and Tonhati et al (2000b) studying the trait interval between calvings by repeatability models in Murrah buffaloes obtained the values of 385.0 and 432.4 days, respectively.

Table 1.  Number of observations (N), overall means, standard deviations (SD), variation coefficients (CV%), minimum and maximum values and number of contemporary groups (NCG) for milk yield (MY), age at first calving (AFC) and interval between first and second calving (IBFSC).






Minimum, kg

Maximum, kg


MY, kg








AFC, days








IBFSC, days








Heritability estimate for MY trait (Table 2) was lower than estimate obtained by Tonhati et al (2000b) which was 0.38. The value obtained (0.21) has the same magnitude that those estimated by Ramos et al (2006) and higher than the verified by Rosati and van Vleck (2002) which was 0.14. According to the last authors, there are many aspects that influenced the values obtained. There is a big variability within and across herds, it makes the phenotypic and environmental variances become bigger. There are also mistakes in the paternal identification of the animals and it makes a genetic variability loss.

Table 2.  Estimates of variance components1, heritability (h2) and standard error (SE) for milk yield (MY), age at first calving (AFC) and interval between first and second calving (IBFSC) traits, obtained from one-trait analyses





h2 ± SE

MY, kg




0.20 ± 0.01

AFC, days




0.07 ± 0.05

IBFSC, days




0.14 ± 0.07

1 σ2a =additive variance; σ2e = residual variance; σ2p =phenotypic variance.

The estimates of heritability for AFC and IBFSC traits (Table 2) were very different from the values observed in literature. Cassiano et al (2004) had obtained, to the breeds Carabao, Murrah, Mediterranean and Jafarabadi, in the Brazilian Amazonia, heritability estimates from 0.12 to 0.38 and 0.04 to 0.05 for the same traits, respectively. Ramos et al (2006) used information of several calves and obtained a heritability estimate of 0.02 for calving interval (CI). The greater estimative obtained in the present study is due to the high variability at first lactation, since in repeatability models only those animals with great potential for milk yield remain longer in herds.


The differences between variance components estimated between single and two traits analysis or MY, AFC and IBFSC were not very big. Genetic and phenotypic correlations between MY, AFC and IBFSC were low (Table 3). These results indicate that selection for MY hardly affect the others traits genetically. It indicates that may be necessary adjustments in feeding and management practices to provide changes to AFC and IBFSC traits. These practices can result in lower CI and it improves the economic efficiency of the production system. Similarly, Tonhati et al (2000b) observed lower genetic correlation between MY and CI which was 0.04. Ramos et al (2006) had different results for the genetic correlation between CI and MY which was -0.22. It indicates that it is possible to have indirect gain, if a selection is done for MY, it may result in lower a CI.

Table 3.  Heritability estimates (diagonal), genetic (above diagonal) and phenotypic (below diagonal) correlations for milk yield (MY), age at first calving (AFC) and interval between first and second calving (IBFSC) traits in two trait animal models

















Differently from results obtained for MY and IBFSC traits, desirable genetic correlations were observed for MY and AFC and for AFC and IBFSC. In the first correlation, results suggest that daughters of bulls with high genetic value for MY will show physiologically maturity earlier. This result corresponded to that one reported by Jahageerdar et al (1997) which was -0.32 for the same traits in Surti buffaloes in India. However, a different genetic correlation (0.63) was estimated by Tonhati et al (2000b) for a herd with Murrah and crossbred animal in Brazil. Similarly, the result observed for the genetic correlation between AFC and IBFSC indicates that the selection for age at first calving, will reduce the IBFSC.


The average breeding values to MY were 13.81 kg. The regression of the average breeding values on the birth year resulted in a regression coefficient with a positive trend (P<0.01), equivalent to 0.207% of annual genetic change in the average observed to MY (Figure 1).

Figure 1.  Genetic trend of average breeding values of the animals according to birth year for milk yield (MY)

This value corresponds to, approximately, 9.03 g/day/year or 3.30 kg of annual gain (y = 3.30x -42.27), or 108.86 kg during the 33 years studied. The results suggest that despite the genetic trend was positive, the selection for MY is not efficient. It could be explained by the lack of breeding programs for buffalo milk production systems in Brazil.


A regression coefficient that isn’t significant of -0.02 days/year (y = -0.02x +2.46) to AFC (Figure 2). In spite of the regression coefficient isn’t significant, the result indicates that how the selection is made according to the productions observed, and the AFC wasn’t altered.

Figure 2.  Genetic trend of average breeding values of the animals according to birth year for age at first calving (AFC)

The average of the breeding values to the IBFSC is 3.99 days. As a result of the regression of the breeding values on the birth year, was observed a regression coefficient which corresponds to a gain of 0.3358 day/year (y = 0.34x -1.72) which is significant (P<0.01), and suggests a gain in the IBFSC during the years (Figure 3).

Figure 3.  Genetic trend of average breeding values of the animals according to birth year for interval between first and second calving (IBFSC)

This annual genetic change corresponds to 0.074% in the average of the trait observed or a gain of 11.05 days in 33 years. The results obtained by Ramos et al (2006) were lower in comparison of the ones obtained in this study, the authors observed a regression coefficient of 0.08day/year or 0.019% of annual genetic change in average observed for CI, which represents a gain in 1.8 day during the 21 years of the study, it means that, despite of a small genetic gain to MY has occurred, this change didn’t promote undesirable genetic change in IBFSC. 


The selection intensity was considered the same to all of the traits. The answers to the selection were 84.97 kg, -8.56 days and -11.47 days by generation to the traits MY, AFC and IBFSC. The correlated answers, assuming a direct selection to MY, were -1.74 days and 0.96 days to AFC and IBFSC respectively. When the selection is done emphasizing the reduction of AFC, the correlated answers to MY and IBFSC were 6.03 kg and 2.84 days respectively by generation.


The low genetic correlation between MY, AFC and IBFSC indicate that if a selection is done to MY, it will not promote big changes in the other traits. Undesired bigger correlated values to the CI may be changed by better nutritional and management practices. A good strategy would be the use of economic selection index, combining the genetic gain and the economic value of the traits, resulting in a bigger selection to the traits with higher economic values.


The estimates for the simple correlation between the genetic values (Pearson) and for the rank correlation (Spearman) among classifications by genetic values were significant (P<0.01). These correlations are shown in table 4. Both of the correlations (Pearson and Spearman) between MY and AFC and IBFSC were low. However, they were positive, indicating that the selection of the bulls with a higher MY genetic value may result in daughters with a higher IBFSC and AFC. Similar estimates were observed between AFC and IBFSC, but the genetic value correlation were favorable.

Table 4.  Simple correlations (Pearson’s1) and rank correlation (Spearman’s2) between milk yield (MY) breeding values and age at first calving (AFC) and interval between first and second calving (IBFSC).

















1 above of diagonal; 2 below of diagonal




Bagnato A and Oltenacu P A 1993 Genetic study of fertility traits and production in different parities in Italian Frisian cattle. Journal of Animal Breeding Genetics 110:126-134.


Boldman K G, Kriese L A, Van Vleck L D et al 1995 A manual for use of MTDFREML. Lincoln. USA. 1993. 120 p


Cassiano L A P, Mariante A da S, Mcmanus C et al 2004 Parâmetros genéticos das características produtivas e reprodutivas de búfalos na Amazônia brasileira; Pesquisa Agropecuária Brasileira 39 (5):451-457


Dong M C and van Vleck L D 1989 Correlations among first and second lactation milk yield and calving interval. Journal of Dairy Science 72:1933-1936


Jahageerdar S, Govindaiah M G and Jayashankar M R 1997 Genetic studies of first lactation performance of Surti buffaloes. Buffalo Bulletin 16 (3):51-55.


Lush J L 1964. Como a seleção muda uma população – expressão dos resultados. Melhoramento Genético dos Animais Domésticos. Rio de Janeiro. Brasil. p. 186-217.


Mohamed M, Kawther A M, Khattab A S 1993 Genetic relationship of age and weight at first calving with first lactation milk yield in Egyptian bufalloes. Buffalo Bulletin 12 (2):38-46.


Pryce J E, Coffey M P, Brotherstone S H et al 2002 Genetic relationships between calving interval and body condition score conditional on milk yield. Journal of Dairy Science 85:1590–1595


Ramos A A, Malhado C H M, Carneiro P L S et al 2006 Caracterização fenotípica e genética da produção de leite e do intervalo entre partos em bubalinos da Raça Murrah. Pesquisa Agropecuária Brasileira 41 (8):1261-1267


Rosati A and Van Vleck L D 2002 Estimation of genetic parameters for milk, fat, protein and mozzarella cheese production for the Italian river buffalo Bubalus bubalis population. Livestock Production Science 74:185–190.


Roxström A, Strandberg E, Berglund B et al 2001 Genetic and environmental correlations among female fertility traits and milk production in different parities of Swedish Red and White Dairy cattle. Acta Agriculturae Scandinavica 51 (1):7-14.


Seno L O, Cardoso V L and Tonhati H 2006 Responses to selection for milk traits in dairy buffaloes; Genetics and Molecular Research 5 (4):790-796


Shabade N S, Jagtap D Z and Behle N D 1993 Factors affecting production and production efficiency traits of first lactation in Murrah buffaloes; Indian Journal of Animal Science 63 (11):1212-1213.


Statistical Analyses System - SAS. User’s guide. version 8.0. Cary: 1999.


Tonhati H, Cerón-Muñoz M F, Oliveira J A et al 2000a  Parâmetros genéticos para produção de leite, gordura e proteína em bubalinos. Revista Brasileira de Zootecnia (suplemento 1) 29 (6): 2051-2056


Tonhati H, Vasconcellos F B and Albuquerque L G 2000b Genetic aspects of productive and reproductive traits in a Murrah buffalo herd in São Paulo, Brazil. Journal of Animal Breeding Genetics 117:331-336.

Received 19 February 2009; Accepted 3 January 2010; Published 7 February 2010

Go to top