SciELO - Scientific Electronic Library Online

vol.50 issue3Anaerobic digestion technologies for the treatment of pig wastesMeat yield of grazing river buffalos (Bubalus bubalis) author indexsubject indexarticles search
Home Pagealphabetic serial listing  


Services on Demand




  • Have no cited articlesCited by SciELO

Related links

  • Have no similar articlesSimilars in SciELO


Cuban Journal of Agricultural Science

Print version ISSN 0864-0408On-line version ISSN 2079-3480

Cuban J. Agric. Sci. vol.50 no.3 Mayabeque Jul.-Sept. 2016


Cuban Journal of Agricultural Science, 50(3): 355-364, 2016, ISSN: 2079-3480




Effect of milk replacers on growth curves up to 90 days old in developing Siboney de Cuba females


Efecto de reemplazantes lecheros en las curvas de crecimiento hasta noventa días de edad de hembras Siboney de Cuba en desarrollo



Á.C. Alonso,I L. D. Soto,I Bertha Chongo,II Verena Torres,II A. Zamora,II

IDirección Técnica de Desarrollo. Empresa Pecuaria Genética “Camilo Cienfuegos”. Consolación del Sur, Pinar del Río, Cuba.
IIInstituto de Ciencia Animal. Apartado Postal 24. San José de Las Lajas, Mayabeque, Cuba.




Under production conditions, the effect of two technologies of milk replacers on growth curves (up to 90 days) of developing Siboney de Cuba females was evaluated. For that purpose, 30 females were used, randomly divided, at their arrival to the rearing unit, into two groups (15 animals each). For growth analysis, linear (linear, square linear and square logarithmic models) and non-linear (logistic, Gompertz and Richards) regression models were tested. For linear models, gains were estimated by the least square method. For non-linear ones, the interactive method of Marquardt was applied. The best adjustment was determined from different statistical criteria. ANOVA was applied for weights and daily mean gains, reached in different moments. The square logarithmic model had the best fit of growth, with R2 => 0.93 and P < 0.001 in both treatments, from which the performance of liveweight increase in other group of animals may be predicted, with the use of similar combinations per treatment. Daily mean gain and liveweight, up to 90 d, differed in both treatments after a month of birth. For females of the second treatment, daily mean gains were near 500 g/animal/d, and liveweight was around 69.2 kg up to weaning.

Key words: milk replacers, growth curves, Siboney de Cuba females.


En condiciones de producción se evaluó el efecto de dos tecnologías de reemplazantes lecheros en las curvas de crecimiento (hasta 90 d de edad) de hembras Siboney de Cuba en desarrollo. Se emplearon para ello 30 hembras, divididas aleatoriamente al llegar a la recría en dos grupos (15 animales cada uno). Para el análisis del crecimiento, se probaron modelos de regresión lineal (lineal, lineal cuadrático y cuadrático logarítmico) y no lineal (logístico, Gompertz y Richards). Para los modelos lineales, se estimaron las ganancias por el método de los mínimos cuadrados. Para los no lineales, se aplicó el método iterativo de Marquardt. El mejor ajuste se determinó a partir de diferentes criterios estadísticos. Se aplicó ANOVA a los pesos y ganancias medias diarias alcanzadas en diferentes momentos. El modelo que mejor ajustó el crecimiento fue el cuadrático logarítmico, con R2 => 0.93 y P < 0.001 en ambos tratamientos, a partir del cual se puede predecir el comportamiento del incremento de peso vivo en otro grupo de animales, si se utilizan combinaciones similares por tratamiento. La ganancia media diaria y el peso vivo, hasta los
90 d, difirieron en ambos tratamientos (P < 0.001) después de un mes del nacimiento. Para las hembras del segundo tratamiento, las ganancias medias diarias estuvieron próximas a los 500 g/ animal/d, y el peso vivo muy cercano a 69.2 kg hasta el destete.

Palabras clave: remplazantes lecheros, curvas de crecimiento, hembras Siboney de Cuba.




Siboney de Cuba genotype appears as part of a genetic improvement program in Cuba, directed to achieve milk production levels according to the needs of the country. This genotype is characterized by its adaptation ability in tropical areas, expressed in heat tolerance, acceptable rates of growth and development, and an early incorporation with good levels of prolificacy, as described by López (1983) and Gregorich (1992).

At international level, researches conducted in calves during the first stages of life coincide in pointing out the marked physiological changes, occurring during the transition from pre-ruminant (provided with milk or milk replacers) to ruminant (intake of concentrate, hay and grasses) during the first weeks of life. Nutritionists and producers show more and more interest every day for the weaning stage due to the physiological changes expressed during this period and its impact on animal performance (Ybalmea 2011, Chongo et al. 2013).

Describing and predicting growth curves during this stage contributes to the characterization of these used production systems, and provides, at the same time, useful information for the development of selection strategies in correspondence to the used technology.

The objective of this study was to describe the growth curve up to 90 d old of Siboney de Cuba females, fed with two technologies of milk replacers in systems of artificial rearing unit.



This study was conducted at the artificial rearing unit 246,  with 50.39 ha of surface, belonging to the Unidad Empresarial de Base (UEB) “Loma de Candelaria” from Empresa Pecuaria Genética “Camilo Cienfuegos”, located at southeast of Pinar del Río province, Cuba.

An amount of 30 Siboney de Cuba female calves were used. These animals were born in December, 2013, and were between 7 and 9 d old at their arrival to the rearing unit. The number of their tattoo and ear tag was individually controlled and they were weighed in a DI-162 Model (Digital) scale, according to the method of weighing technique established by the Cuban regulations (NC 74 - 49). Later, these animals were randomly divided into two groups (15 animals each) and were located in previously cleaned and disinfected individual cribs, with free access to water and long hay at will, starting from 30 d.

During all the studied stage, hay was produced in the same unit, from the hand-made reaping of pangola (Digitaria decumbens) and star grass (Cynodon nlemfuensis). For the experimental development, four types of milk replacers were used, and soybean meal for completing one of the used formulations. Combinations of milk replacers, as well as doses provided to females per treatment are described in tables 1 and 2.

Milk replacer was prepared in aluminum containers of 20 L, which were poured hot water between 45 and 50 ºC, from a solar heater installed at the unit. The water was mixed with the amounts of milk replacer designed for each stage. Later, liquid feed was offered to the calves at a temperature between 37 and 39 ºC.

Doses of starting concentrate were also included and the rules established for each stage up to 90 d were used (table 3)

After 30 d, female calves from each treatment were individually weighed during the first hours of the morning with their stomach empty. The result of liveweight (LW) was individually controlled. Weighing procedure was carried out the same way every month, up to 90 d.

For processing all the obtained data, three non-linear and three linear regression models, which are most used ones to estimate animal performance, according to Ratkowsky (1983):


y = LW or daily mean gain (DMG)

age =animal age

alpha (a), beta (b) and gamma (c) = parameters

Parameter estimations were determined by the method of least squares, in the case of linear regression, and by the iterative method of Marquardt (1963), for the non-linear models (Logistic, Gompertz and Richards).

Statistical criteria for selecting the model of the best fit were performed according to indicators described by Guerra et al. (2003) and Fernández (2004):

1.       Determination coefficient (R2)

2.       Parameters of models (a, b, c)

3.       Mean square of the error (MSe)

4.       Standard error of model parameters SE(a), SE(b), SE(c)

5.       Model significance

6.       Residual Analysis

Gathered information was processed by the statistical package SPSS, version 11.5.1 (Visauta 1998, IBM Corporation 2003). For evaluating liveweights and daily mean gains, the multiple range test of Duncan (1955) was also used, which allowed the comparison between treatments by Infostat package, version 1.0 (Di Rienzo et al. 2001).



Growth modeling with application of linear and non-linear models to developing females (up to 90 d), fed different combinations of milk replacers, was based on the descriptive analysis of what happened at this growth stage in two feeding systems used in artificial rearing.

Evaluating growth (up to weaning age) of Siboney de Cuba females, at this important stage, allows to predict the performance of LW developments at early ages. In addition, it also allows to evaluate biologically important parameters in a stage considered as slow growth stage during the life of bovines, in which significant physiological changes within their digestive system occur. These changes depend, in a large extent, on the used feeding systems.

Among the evaluated models (linear and non-linear), Richards model was discarded because it showed no significant fit in LW.

By comparing the fit of models that evaluate LW performance in Siboney de Cuba females, up to 90 d old for both treatments (table 4), it was demonstrated that all showed a very good fit, because of the flexibility they have (Arias et al. 2010), with high and similar determination coefficients (R2), indicating good precision in estimated liveweights.

Although all models show high R2 fit and differed among them (P <0.001) for the linear model (treatment 1) and Gompertz (treatment 2) (Table 2), these presented the lowest coefficients with 0.91. However, for the other models, R2 values, in both treatments, had the best fit, with 0.93 and differ among them (P <0.001). Similar results were obtained by Robert-Granié et al. (2002), Mejías (2008) and Alonso et al. (2014) in R2 fit, describing growth curves in cattle. These authors agree to inform, as in this study, linear growth in certain phases of the curve.

For selecting the best fit model, MSe is another statistical criterion to consider. In both treatments, the square logarithmic model showed lower values than the rest of the tested models, with values of 14.88 and 12.86, respectively. This allowed to estimate, with more precision, the effects of studied factors and, in turn, explain the existence of a lower variance in LW analysis, which is similar to that described by de Freitas (2005) and Alonso (2015). This result expresses, to a large extent, the variability of liveweights of females until the age at weaning at each weighing. Likewise, it also allows the selection of this model as the one with the best goodness of fit for both treatments.

After fitting the model, residuals were calculated (figure 1), which, according to Torres et al. (2012), is an essential procedure for knowing their general performance and determining whether there is a underestimation or overestimation at any part of growth models (Fabens 1965).

Dispersion range of residuals in both treatments was between 0.1 and -0.1, with higher concentration between 0.05 and -0.05, indicating good accuracy in the observed liveweights. For females of treatment 2, fitted weights showed lower dispersion, especially at 60 and 90 d, which is related to the little difference found between the weights observed in both weighings and liveweights calculated by the model fitted for a similar stage. In general, for both treatments, performance of residuals for the model of best fit was random, which is similar to results reported by Alonso (2015) for the same stage.

In the analysis of model parameters, parameter a of the square logarithmic model showed that the value reached for treatment 1 was 1.49, while for treatment 2, it showed similar results with 1.5 (P <0.001), and least standard errors for both cases up to weaning age. These results are indicators of development to reach the asymptotic weight of the female under study. However, it cannot be stated that these are the ones that predict mean weight with which females reach maturity, because she still has a long period of growth and development to evaluate.

For parameter b of the selected model, results for both treatments presented the same values (0.003) of high significance and low standard errors, which is related to the LW at birth, being very similar, because females of treatment 1 had a mean LW of 31.8 kg, while those from the second treatment had 31.5 kg.

Finally, the parameter c of the selected model indicates values of -0.00007 and SE 0.00005 of high significance for females of treatment 1, while females of treatment 2, with -0.00001 and SE 0.00004, show similar and highly significant values in this parameter. These results refer to an adequate growth speed for females in both treatments, although much better for those of treatment 2, in a phase in which bovines grows more slowly, in correspondence with reports of Torres et al. (2012) and Alonso (2015),who state that animals with high growth values represent an early maturity, compared to those of similar initial weight.

In figure 2 a and b, the performance between observed LW values and estimated LW of the square logarithmic model, for both treatments, was similar, for each studied cases, which confirms the correct model selection. This result coincides with studies of Guerra et al. (2003), Mejías (2008), Alonso et al. (2014), and Alonso (2015) regarding the goodness of fit of linear models to express the standard growth curve in the early stages of any living organism.

Although females from treatment 1 showed a good fit in observed and estimated LW values, treatment 2, from the second month of life, marked a favorable difference of observed weights with higher variability. The same happened in this treatment at 90 d of age.

This may be related to 2.5 kg of concentrate, provided to females from treatment 2 after week 8, which, joined to intake of certain level of fiber (hay) at will, may contribute to a better assimilation of ration and the best use of nutrient content of solid feeds. Ruminants, as poly-gastric animals, start to ingest these feeds after the sixth week of age (Plaza and Ybalmea 2008), if it is compared with 1.7 kg of concentrate supplied to females for the same stage, together with the offering of certain level of fiber (hay) at will and the liquid diet was maintained up to week 12, including 230 g of soybean meal, which is difficult to assimilate by the calf at this stage.

The previous explanation, together with the supply of RaltecMilk-17-1 and RaltecStar T-1 to females from treatment 2, from the second to the seventh week, allowed the earliest weaning in this group, and daily mean gains of 482.22 g/d in the second month of life and 497.61 g/d in the third (table 5). These values were superior to the gains obtained in treatment 1. As a consequence, next to the third month of age (84 d), calves could double the LW they were born with, which corresponds to the stated Johnson (1996) and Ybalmea et al. (2013).

Table 5 shows LW and gains of females with intake of both combinations of milk replacers at different ages during the experimental period. LW of females, from 7 to 30 d of the experimental period, had no significant differences among treatments. After this stage, females from treatment 1 are disadvantaged regarding those from treatment 2, with inferior weights in each weighing, which differed (P <0.001) between them at 60 and 90 d. These results are related to characteristics and presentation of combined feeds offered in the second treatment. According Stobo et al. (1966, cit. by Guerreiro 1991) and Lesmeister and Heinrichs (2004), the inclusion of amounts of beginning concentrate, together with fiber intake (long hay, offered at will), improves ruminal and reticular capacity, may increase the weight of tissues, and exerts a physical stimulation in the early development of digestive tract in a complex stage for the improvement of calf gastrointestinal system, especially near 60 d old, corresponding to results of Plaza and Ybalmea (2008).

The DMG up to 30 d differ among treatments (P <0.001) and were better in the females from treatment 1 regarding those from 2, because of the presence of a calf with gastro-enteric disorders in this treatment.

In the second and third month, the DMG obtained in treatment 2 were superior to 450 g/d, confirming the efficiency of feed conversion, which was offered through the used combinations. Although calves, due to their natural and physiological condition, do not like to consume dry feeds at early ages (Ybalmea 2011). Animals will accept better this feed if it is palatable and has granulated appearance. This contributes to increased ruminal fermentation from very early ages (Plaza 1983, Chongo et al. 2013), with good efficiency, reflected on high LW increase indexes.

Responses of calves from treatment 1 could be associate to the presence of soybean meal in the feed offered during the first stage of life. It is known that meal is difficult to assimilate by the calf after the first month of birth, due to its possible antitrypsic and antigenic effect, and to the limited availability of intestinal amylase, vegetable oils and fats, which are the worst used by calves, according to Toullec et al. (1975, 1994) and Ybalmea (2011).

It may be concluded that, among linear models, the square logarithmic had the best fit to express growth curves up to 90 d of age in Siboney of Cuba females, regardless of the use of different milk replacers, and of the combination of RaltecMilk 17-1 with RaltecStar T-1 obtained more favorable DMG from 30 to 90 d after birth. These values were very close to the 500 g/animal/d, and 69.2 kg of mean LW up to weaning, with respect to the use of the combination of Raltec-Milk 17-2 and Nuclimix (Campi) plus soybean meal.

Thanks to the Empresa Pecuaria Genética “Camilo Cienfuegos”, to workers of the artificial rearing unit 246 to allow the completion of the research process and to the Department of Biomathematics from the Institute of Animal Science for their help in data processing.



Alonso, A. C. 2015. Factores que afectan los indicadores de eficiencia en la producción de hembras en desarrollo Siboney de Cuba hasta primera lactancia,  en la U.E.B "Loma de Candelaria". Ph.D. Thesis, Instituto de Ciencia Animal, Mayabeque, Cuba, 110 p.

Alonso, Á. C., Chongo, B., Zamora, A., Torres, V., Noda, A. C. & Acosta, F. 2014. “Growth curves of developing Siboney de Cuba females, up to 18 months old”. Cuban Journal of Agricultural Science, 48(3): 227–234.

Arias, Y., Gonzales, P. D., Montes, I. & Domínguez, A. 2010. “Estimación de la curva de crecimiento en futuros sementales de la raza buffalypso”. Ciencia y Tecnología Ganadera, 4(3): 136–142, ISSN: 1999-4494, 1998-3050.

Chongo, B., Zamora, A., Martínez, Y. & Alonso, A. 2013. “Actualidad y perspectivas de la crianza de terneros. Retos para el trópico”. In: IV Congreso Internacional de Producción Animal, Cuba: Instituto de Ciencia Animal, p. 2102, ISBN: 978-959-7171-49-2.

de Freitas, A. R. 2005. “Curvas de crescimento na produção animal”. Revista Brasileira de Zootecnia, 34(3), ISSN: 1516-3598, DOI: 10.1590/S1516-35982005000300010.

Di Rienzo, J. A., Casanoves, F., Balzarini, M. G., González, L., Tablada, M. & Robledo, C. W. 2001. InfoStat. version 2001, [Windows], Universidad Nacional de Córdoba, Argentina: Grupo InfoStat, Available: <> .

Duncan, D. B. 1955. “Multiple Range and Multiple F Tests”. Biometrics, 11(1): 1–42, ISSN: 0006-341X, DOI: 10.2307/3001478.

Fabens, A. J. 1965. “Properties and fitting of the Von Bertalanffy growth curve”. Growth, 29(3): 265–289, ISSN: 0017-4793.

Fernández, L. 2004. Modelos Estadísticos–Matemáticos en el análisis de la curva de lactancia y factores que afectan en el genotipo Siboney de Cuba. Ph.D. Thesis, Universidad Agraria de La Habana, La Habana, Cuba, 113 p.

Gregorich, J. 1992. Comportamiento de la actividad estral y la fertilidad en rebaños Siboney. Graduated Thesis, Universidad de Granma, Granma, Cuba, 72 p.

Guerra, C. W., Cabrera, A. & Fernández, L. 2003. “Criteria for the selection of statistical models in scientific research”. Cuban Journal of Agricultural Science, 37(1): 3–10, ISSN: 2079-3480.

Guerreiro, R. 1991. Factores que influyen en el comportamiento en la recría Anacaona, de la Empresa Pecuaria Genética Bacuranao. Graduated Thesis, Universidad Agraria de La Habana, La Habana, Cuba, 84 p.

IBM Corporation 2003. IBM SPSS Statistics. version 11.5, [Windows], U.S: IBM Corporation, Available: <>.

Johnson, E. R. 1996. “Neonatal growth and development of the bovine stomach”. Animal Production in Australia, 21: 211–214, ISSN: 0728-5965.

Lesmeister, K. E. & Heinrichs, A. J. 2004. “Effects of Corn Processing on Growth Characteristics, Rumen Development, and Rumen Parameters in Neonatal Dairy Calves”. Journal of Dairy Science, 87(10): 3439–3450, ISSN: 0022-0302, DOI: 10.3168/jds.S0022-0302 (04)73479-7.

López, D. 1983. Características de la población base de un nuevo genotipo lechero tropical: El Siboney de Cuba. Ph.D. Thesis, Instituto Superior de Ciencias Agropecuarias de La Habana  «Fructuoso Rodríguez Pérez», La Habana, Cuba, 221 p.

Marquardt, D. W. 1963. “An Algorithm for Least-Squares Estimation of Nonlinear Parameters”. Journal of the Society for Industrial and Applied Mathematics, 11(2): 431–441, ISSN: 0368-4245, 2168-3484, DOI: 10.1137/0111030.

Mejías, R. A. 2008. Sistema para la producción de hembras bovinas de reposición con asociación de gramíneas con leguminosas. Ph.D. Thesis, Instituto de Ciencia Animal, La Habana, Cuba, 108 p.

Plaza, J. 1983. Efecto del alimento fibroso en el comportamiento y desarrollo ruminal de los terneros. Ph.D. Thesis, Instituto de Ciencia Animal, La Habana, Cuba, 201 p.

Plaza, J. & Ybalmea, R. 2008. “Effect of whole milk and milk replacers on the behavior of replacement female calves”. Cuban Journal of Agricultural Science, 42(4): 351–354, ISSN: 2079-3480.

Ratkowsky, D. A. 1983. Nonlinear regression modeling: a unified practical approach. (ser. Statistics, textbooks and monographs, no. ser. 48), New York: M. Dekker, 276 p., ISBN: 978-0-8247-1907-4.

Robert-Granié, C., Heude, B. & Foulley, J. L. 2002. “Modelling the growth curve of Maine-Anjou beef cattle using heteroskedastic random coefficients models”. Genetics Selection Evolution, 34(4): 423, ISSN: 1297-9686, DOI: 10.1186/1297-9686-34-4-423.

Stobo, I. J., Roy, J. H. & Gaston, H. J. 1966. “Rumen development in the calf. 1. The effect of diets containing different proportions of concentrates to hay on rumen development”. The British Journal of Nutrition, 20(2): 171–188, ISSN: 0007-1145.

Torres, V., Barbosa, I., Meyer, R., Noda, A. & Sarduy, L. 2012. “Criteria of goodness of fit test in the selection of non-linear models for the description of biological performances”. Cuban Journal of Agricultural Science, 46(4): 345–350, ISSN: 2079-3480.

Toullec, R., Lallès, J. P. & Bouchez, P. 1994. “Replacement of skim milk with soya bean protein concentrates and whey in milk replacers for veal calves”. Animal Feed Science and Technology, 50(1–2): 101–112, ISSN: 0377-8401, DOI: 10.1016/0377-8401(94)90012-4.

Toullec, R., Patureau-Mirand, P., Thivend, P. & Vermorel, M. 1975. “Bases physiologiques de la realisation des aliments d’allaitement pour veaux”. Alimentation et la Vie, 63: 70, ISSN: 0065-6267.

Visauta, V. B. 1998. Análisis estadístico con SPSS para Windows: estadística multivariante. Madrid, España: McGraw-Hill, 358 p., ISBN: 978-84-481-2074-0.

Ybalmea, R. 2011. Contribución al uso de reemplazantes lecheros de producción nacional en la alimentación de terneros. Ph.D. Thesis, Instituto de Ciencia Animal, La Habana, Cuba, 143 p.

Ybalmea, R., Alonso, A., Marrero, D. & Fernández, L. 2013. “Uso de reemplazantes lecheros (RL), en la crianza artificial de terneros, en condiciones de producción”. In: IV Congreso Internacional de Producción Animal, Cuba: Instituto de Ciencia Animal, pp. 2108–2113, ISBN: 978-959-7171-49-2.



Received: 18/7/2016
Accepted: 28/10/2016



Á.C. Alonso, Dirección Técnica de Desarrollo. Empresa Pecuaria Genética “Camilo Cienfuegos”. Consolación del Sur, Pinar del Río, Cuba. Email:

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License