Servicios Personalizados
Revista
Articulo
Indicadores
- Citado por SciELO
- Accesos
Links relacionados
- Similares en SciELO
Compartir
Revista mexicana de ciencias pecuarias
versión On-line ISSN 2448-6698versión impresa ISSN 2007-1124
Rev. mex. de cienc. pecuarias vol.12 no.4 Mérida oct./dic. 2021 Epub 06-Jun-2022
https://doi.org/10.22319/rmcp.v12i4.5016
Technical notes
Comparison of equations to fit growth curves of Holstein, Jersey and Jersey x Holstein cows in a grazing system
aUniversidad Autónoma Chapingo. Departamento de Zootecnia, Posgrado en Producción Animal, km 38.5 Carretera México-Texcoco, 56230, Chapingo, Estado de México, México.
The objective of the study was to compare the goodness of fit of four nonlinear equations to describe the growth curves of Jersey, Holstein, and Jersey x Holstein cows in grazing. The Brody, Gompertz, von Bertalanffy, and Logistic equations were fitted to the data (n= 2,315) on weight and age of Jersey (n= 54), Holstein (n= 6) and Jersey x Holstein (n= 30) cows. For each animal, genotype and equation, the parameters A, b and k that generated the best-fit growth curves were estimated. In each of the four equations compared, parameter A corresponds to the upper asymptote of the curve and estimates the ‘mature weight’ of the animal, parameters b and k represent the integration constant and the maturation rate. For the growth curves of Jersey cows and Jersey x Holstein crosses, the Gompertz and von Bertalanffy equations produced the best fit. In contrast, the Logistic equation had the best fit for the growth curves of Holstein, followed closely by the Gompertz and von Bertalanffy equations. Under the management and feeding conditions of the animals in this study, the growth curves of the females of the three genotypes studied can be fitted with the von Bertalanffy equation.
Key words Dairy cattle; Nonlinear models; Growth pattern
El objetivo del estudio fue comparar la bondad de ajuste de cuatro ecuaciones no lineales para describir las curvas de crecimiento de vacas Jersey, Holstein, y Jersey x Holstein en pastoreo. Las ecuaciones de Brody, Gompertz, von Bertalanffy, y Logística se ajustaron a los datos (n= 2,315) de peso y edad de vacas Jersey (n= 54), Holstein (n= 6) y Jersey x Holstein (n= 30). Para cada animal, genotipo y ecuación, se estimaron los parámetros A, b y k que produjeron las curvas de crecimiento de mejor ajuste. En cada una de las cuatro ecuaciones comparadas, el parámetro A corresponde a la asíntota superior de la curva y estima el ‘peso maduro’ del animal, los parámetros b y k representan la constante de integración y la tasa de maduración. Para las curvas de crecimiento de las vacas Jersey y las cruzas de Jersey x Holstein, las ecuaciones de Gompertz y de von Bertalanffy produjeron el mejor ajuste. En contraste, la ecuación Logística tuvo el mejor ajuste para las curvas de crecimiento de Holstein, seguida muy de cerca por las ecuaciones de Gompertz y von Bertalanffy. En las condiciones de manejo y alimentación de los animales en este estudio, las curvas de crecimiento de las hembras de los tres genotipos estudiados pueden ajustarse con la ecuación de von Bertalanffy.
Palabras clave Ganado lechero; modelos no lineales; patrón de crecimiento
When the weight of the same individual at regular age intervals, from birth to maturity, is available, it is possible to fit a function that describes the weight curve according to age. This curve that allows condensing the growth path of an individual into a few parameters is called the growth curve1.
There are several equations that have been used to describe the growth of plants and animals2, including Brody’s3, Gompertz’s4, Logistic5 and von Bertalanffy’s5, all special cases of Richards’ equation5. Growth curves reflect the interrelationships between the individual’s inherent impulse to grow and mature, and the environment in which these impulses are expressed6. Some of these equations have been used to describe the growth of several species of zootechnical interest, such as Pantaneira bovines in Brazil7, and Holstein females to predict their growth from birth to first calving8. These equations have also been used to describe the growth of sheep genotypes9,10,11, dairy cattle12,13, and horses14.
Under grazing milk production systems with low use of supplemental feed and a mixture of bovine genotypes that differ in mature size, the average weight of the cow largely determines the maximum stocking rate supported by the system. Different bovine genotypes may have different mature weights and metabolizable energy requirements for maintenance. Fitting a growth curve from birth to maturity can be useful for calculating maintenance requirements over the life of the animal. The individual growth curves of animals can be estimated by fitting mixed nonlinear models, even with incomplete records of the growth curve of the animals of the herd. The objective of this study was to compare the goodness of fit of four nonlinear equations to describe the growth curves of Jersey, Holstein, and Jersey x Holstein cows grazing on mixed meadows of alfalfa-perennial ryegrass-orchard grass-white clover.
The study was conducted in the Module of Organic Milk Production in Grazing of the Experimental Farm of the Chapingo Autonomous University, Mexico. Díaz15 details the development of the Module. Briefly, the module started with 10 Holstein cows and 25 Jersey cows of first calving in May 2000. The animals that make up the herd are Jersey cows from the New Zealand genetic line, Holstein cows from the North American genetic line and cows originating from the cross of Holstein cows with Jersey semen.
The feeding of the animals is based on the grazing on meadows, hay that is preserved from these and corn silage that is produced in meadow areas that are renewed every year. The animals are managed in intensive rotational grazing in strips on mixed meadows of alfalfa (Medicago saltiva) variety Aragonese, orchard grass (Dactylis glomerata) variety Potomac, perennial ryegrass (Lolium perennial) variety Linn and white clover (Trifolium repens) variety Ladino.
The animals of the herd are managed in three groups: lactating, dry and growing animals. Depending on the growing conditions of the meadow, paddocks are assigned by group or grazing is carried out with the scheme of leaders and followers, the leaders are the cows in production and the followers the group of dry cows and growing cattle.
With this grazing scheme, the animals are provided with production conditions that guarantee sufficient freedom of movement, area for rest, fresh air, adequate light and temperature, clean water and sufficient food for the productive stage of the animal. In addition to the meadow forage, animals are provided with a freely accessible mineral mixture. This mixture was formulated based on a diagnosis of the mineral status of animals and meadow forage, with tests on blood, forage, feces, hair and milk16.
Cows give birth on the meadow, in the group of dry cows, providing them with assistance only when necessary. The calf remains with the mother in the meadow for the first two days of life consuming colostrum freely. On the third day, the mother joins the group of lactating cows, the calf is sent to a rearing room with natural ventilation, and it is assigned an individual cage with a raised floor. Only for females, a card is opened on which identification of the calf, date and weight at birth, identification of the mother and father are recorded. The identification of the calf consists of a tattoo on the right ear and earring on the left ear, using a four-digit system. During the first two weeks of life, the calves receive four liters of whole milk divided into two doses. Subsequently, in addition to milk, the calves have access to the meadow forage. Weaning is done at three months of age15.
The weight and age records of the animals of the genotypes studied were generated from the dates of birth, weighing dates and individual records of the live weight of animals at different stages of their life, from birth to maturity. The animals were periodically weighed at regular intervals of about 30 d. A Tru-Test digital scale (Tru-Test, Palmerston North, NZ), EC 2000 System with GP 600 Bars, with a capacity of 1,500 kg and an accuracy of 0.5 kg, was used. The live weight data used in this study were recorded from 2012 to 2016. The animals included in the study were weighed at birth, weaning, and monthly.
In total, there were 2,315 records of weight (kg) and age (d) generated by 54 Jersey, 6 Holstein, and 30 Jersey x Holstein cows, with ages ranging from 1 to 6,597 d.
The Brody3 model, a modified version of the Gompertz4 function, the Logistic5 and von Bertalanffy5 models were fitted to the weight and age data of the cows.
Brody:
Gompertz:
Logistic:
von Bertalanffy:
In all cases, Y t is the live weight (kg) recorded at age t (days); parameter A corresponds to the upper asymptote, which estimates the mature weight (kg) of the animal; parameter b is an integration constant related to the initial weight; parameter k is the maturation rate; and e is the base of natural logarithms12.
With the information of weight and age, descriptive statistics were calculated in different sections of the herd growth curve. Nine age intervals were generated: three that covered the interval from birth to one year, and six that covered the interval from two to twelve years, with two years apart. A similar partition was made to describe the growth curve of North American Angus cattle17. Descriptive statistics were obtained using the MEANS procedure of SAS18.
For each genotype, each of the equations evaluated was fitted by the NLMIXED
procedure of SAS18. Only the
parameter related to mature weight (
A
) was fitted as a random effect in the respective equation; the remaining
parameters of the growth curve (
b
and
k
) were considered fixed. A procedure similar to that of other studies, in
which stochastic models were fitted to describe individual variation in farm animal
growth, was followed19,20,21. The inclusion of a random term (
Brody:
Gompertz:
Logistic:
von Bertalanffy:
where Y
i,t
is the live weight of the animal i recorded on day
t of age, e is the base of the natural
logarithms (i.e. 2.718281),
A
is the predicted mature weight, a
i
is the random effect of the animal i for the parameter of
the growth curve related to the mature weight (
A
) of the animal
The fitting of these models generated, in addition to the estimators of the
parameters of the growth curve, estimators of variance for parameter
A
(
The fitting of each of the models to the three genotypes studied generated the coefficients of parameters A , b and k , to generate the individual growth curves of animals. In the SAS code used, an expression was also included to generate the coefficients of the fixed regression to describe the average growth curve of each genotype studied. These coefficients were used to generate, in each equation and genotype, the predicted values of live weight to generate the individual curves of the animals and the average growth curves per cow genotype, using the SGPLOT procedure of SAS18.
The descriptive statistics of the live weight for different age intervals of the growth curve of the genotypes evaluated are shown in Table 1. In general, it is observed that the weight increases markedly in the first three to four years, subsequently, it remains relatively constant. The standard deviation increases as the age of the animal advances; this is explained by a greater variation in adult weight in the different genotypes22.
Age (days) | Jersey | Holstein | Jersey x Holstein | ||||||
---|---|---|---|---|---|---|---|---|---|
n | Mean | SD | n | Mean | SD | n | Mean | SD | |
0- 30 | 10 | 26.2 | 6.6 | -- | -- | -- | 7 | 26.0 | 3.4 |
30- 120 | 19 | 56.1 | 11.9 | -- | -- | -- | 2 | 69.5 | 21.9 |
121- 200 | 44 | 81.6 | 14.2 | 2 | 110.5 | 2.1 | 27 | 104.2 | 34.4 |
201-365 | 92 | 128.4 | 29.5 | 5 | 136.2 | 24.9 | 72 | 151.4 | 35.1 |
366-730 | 201 | 245.9 | 46.3 | 12 | 276.0 | 56.6 | 157 | 255.3 | 58.9 |
731-1460 | 404 | 338.5 | 54.3 | 28 | 490.5 | 66.5 | 269 | 365.0 | 60.4 |
1461-2190 | 213 | 358.5 | 51.1 | 43 | 477.0 | 42.5 | 138 | 441.7 | 56.3 |
2191-2920 | 103 | 384.4 | 33.4 | 55 | 512.0 | 37.0 | 23 | 480.7 | 55.1 |
>2920 | 235 | 403.0 | 58.0 | 130 | 527.2 | 44.5 | 11 | 464.0 | 28.9 |
SD = standard deviation.
The estimators of the goodness of fit of the model, -2 log likelihood, the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC) are shown in Table 2. According to the magnitude of these estimators (smaller values indicate better fit), the von Bertalanffy and Gompertz models were the ones that best fitted the growth curves of females of the Jersey genotype. For the growth curves of the Jersey x Holstein crosses, the best fit was obtained with the von Bertalanffy equation, and for Holstein with the Logistic, followed very closely by the Gompertz and von Bertalanffy equations.
Genotype | Model | Indicator of the goodness of fit of the model1 | ||
---|---|---|---|---|
-2 log likelihood | AIC | BIC | ||
Jersey (J) | Brody | 12963 | 12973 | 12983 |
von Bertalanffy | 12892 | 12902 | 12912 | |
Gompertz | 12891 | 12901 | 12911 | |
Logistic | 12920 | 12930 | 12940 | |
Holstein (H) | Brody | 2945 | 2955 | 2954 |
von Bertalanffy | 2915 | 2925 | 2924 | |
Gompertz | 2909 | 2919 | 2918 | |
Logistic | 2895 | 2905 | 2904 | |
JH | Brody | 7143 | 7153 | 7260 |
von Bertalanffy | 7111 | 7121 | 7128 | |
Gompertz | 7122 | 7132 | 7139 | |
Logisticl | 7151 | 7161 | 7168 |
1AIC and BIC are the Akaike (AIC) and Bayesian (BIC) Information Criteria for assessing the goodness of fit of the respective model (lower values indicate better model fit).
In a study of bovine growth in Brazil, the Brody equation generated the best fit to describe the growth curve of Nellore bovines in confinement23. Likewise, the Brody equation best fitted the growth curves of Tropicarne bovines, which implies for this genotype a slow maturation rate, characteristic of the growth of bovines under grazing conditions in the tropics24. Brown et al25 reported that the Brody equation best fitted the growth curve in several bovine genotypes under different management and feeding conditions. In contrast, other researchers26 indicate that the Gompertz and von Bertalanffy equations best fitted to describe the growth of Hereford steers, which is consistent with what was found in this study for the Jersey and Jersey x Holstein genotypes.
The estimators of parameters A , b and k of the growth models evaluated, as well as the estimators of the variances of ‘mature weight’ and residual are presented in Table 3. Parameter A , which represents mature weight, was different in each genotype, depending on the equation used. The Brody equation fitted for the three genotypes the highest value of mature weight, which is consistent with what was obtained in other studies23, in which it is also mentioned that this equation overestimates mature weight. Herrera et al27 indicated that groups of animals with higher A value are less precocious. For Holstein, the von Bertalanffy and Logistic equations produced practically the same A value (512 kg). This value of mature weight for cows of the Holstein genotype is well below the estimators of 591, 566 and 543 kg obtained in Ireland under grazing conditions for two lines from Holstein-Friesian from North America-Europe and Holstein-Friesian from New Zealand, respectively13. The Brody equation yielded the lowest value of k . A small value of k represents a slower growth rate to reach asymptotic weight from initial weight, or slower maturation rates7,24,28.
Parameter | Jersey | Holstein | Jersey x Holstein | ||
---|---|---|---|---|---|
_________________________Brody equation________________________ | |||||
A | 400.0 ± 6.5 | 513.5 ± 13.6 | 455.8 ± 9.4 | ||
b | 1.04 ± 0.01 | 0.99 ± 0.03 | 1.02 ± 0.02 | ||
k | 1.71E-3 ± 5.6E-5 | 1.95E-3 ± 0.01 | 1.65E-3 ± 8.1E-5 | ||
|
1005.0 ± 156.8 | 800.0 ± 617.0 | 803.0 ± 121.0 | ||
|
1075.0 ± 43.7 | 1875.0 ± 154.2 | 1497.5 ± 100.0 | ||
_______________________Gompertz equation________________________ | |||||
A | 383.5 ± 5.6 | 510.6 ± 15.0 | 426.6 ± 7.9 | ||
b | 0.98 ± 0.03 | 1.30 ± 0.01 | 0.84 ± 0.04 | ||
k | 3.19E-3 ± 9.6E-5 | 3.86E-3 ± 3.42E-4 | 2.97E-3 ± 1.27E-4 | ||
|
918.7 ± 140.2 | 904.8 ± 2598.0 | 952.4 ± 143.2 | ||
|
996.4 ± 39.7 | 1731.2 ± 148.4 | 1238.0 ± 68.0 | ||
_________________________Logistic equation________________________ | |||||
A | 373.5 ± 5.3 | 512.3 ± 10.2 | 413.8 ± 7.5 | ||
b | 7.31 ± 0.41 | 13.2 ± 2.86 | 5.55 ± 0.39 | ||
k | 4.69E-3 ± 1.39E-4 | 5.43E-3 ± 4.02E-4 | 4.26E-3 ± 1.83E-4 | ||
|
900.8 ± 134.4 | 500.6 ± 437.7 | 958.7 ± 138.7 | ||
|
1000.4 ± 39.3 | 1661.5 ± 141.8 | 1282.7 ± 70.6 | ||
__________________von Bertalanffy equation________________________ | |||||
A | 388.4 ± 5.9 | 512.0 ± 10.6 | 432.3 ± 8.6 | ||
b | 0.64 ± 0.02 | 0.92 ± 0.16 | 0.59 ± 0.02 | ||
k | 2.70E-3 ± 8.3E-5 | 3.45E-3 ± 3.92E-4 | 2.57E-3 ± 1.07E-4 | ||
|
1013.7 ± 163.9 | 512.2 ± 411.7 | 1153.9 ± 191.7 | ||
|
998.1 ± 39.8 | 1778.0 ± 152.0 | 1135.3 ± 57.6 |
1
A
is the parameter of the growth curve that corresponds to the
upper asymptote, which estimates the mature weight (kg) of the
animal;
b
is an integration constant;
k
is the maturation rate; (
The fit of the Brody equation to the weight and age data of the three genotypes studied is shown in Figure 1. In the present work, the availability records from birth to weaning at the beginning of the project was low. For the Jersey and Jersey x Holstein genotypes, the Brody equation, in the case of the present work, predicted negative values for birth weight, while for Holstein, it estimated positive but low weights. This is consistent with what was estimated by Mgbere and Olutogun29, who argue that the Brody equation underestimates weight during the first days of life of N’Dama cattle, but that it fits well at ages greater than six months. Similarly, Berry et al13 found that the Brody equation predicts negative weights at early ages when used to fit the growth curves of Holstein cows from the North American, European Holstein and Holstein-Friesian genetic lines, managed under semi-grazing conditions. In Brazil30, they found that the Brody and Gompertz equations had the best fit of the growth curve from birth to adulthood of Caracu cows. This is consistent with the best fit of the Brody equation to describe the growth curve of Lagune beef cattle under grazing conditions31.
The fit of the Gompertz equation to the weight and age data of the three genotypes studied is shown in Figure 2. In the present study, birth weights for Holstein, predicted with the fit of the Gompertz equation, were very close to zero, while for Jersey and Jersey x Holstein, a better fit was observed. Several researchers32 used this equation to describe the growth of dairy cattle, as it was the one that best fitted their data. Berry et al13 also found that the Gompertz equation better fitted the growth curves of American Holstein, European Holstein and Holstein-Friesian cows, managed under grazing conditions, better than Brody’s.
The fit of the Logistic equation to the weight and age data of the three genotypes studied is shown in Figure 3. In previous studies describing the growth curve of dairy cattle, the Logistic equation underestimated the mature weight of the animals12. In this study, the Logistic equation yielded values identical to those of von Bertalanffy for this growth curve parameter.
The fit of the von Bertalanffy equation to the weight and age data of the three genotypes studied is shown in Figure 4. This equation fitted the live birth weights of Holstein with negative values and of Jersey and Jersey x Holstein, with weights close to zero, which differs from what was reported in other studies in which the von Bertalanffy equation overestimated the weight at early ages in N’Dama cattle29. Spanish researchers33 compared the Brody, Richards and von Bertalanffy equations to describe the growth curves of Retinta beef cows under extensive grazing conditions in Spain, finding that the von Bertalanffy equation had the best fit for the growth pattern of this genotype. In the present study, the von Bertalanffy equation generated the best fit in the growth curves of Jersey and Jersey x Holstein cows, and for Holstein cows, it produced parameters identical to those of the Logistic equation, which produced the best fit for the growth curves of this genotype.
In conclusion, the Gompertz and von Bertalanffy equations best describe the growth curve of Jersey cows; that of von Bertalanffy, that of Jersey x Holstein females; and the Logistics, that of Holstein females, followed by those of Gompertz and von Bertalanffy. For the management and feeding conditions of the animals in this study, and in order to adequately describe their growth pattern, the growth curves of the females of the three genotypes studied can be fitted with the von Bertalanffy equation.
Acknowledgements
To the National Council for Science and Technology (CONACYT) for the financial support granted to the first author to carry out Master of Science studies at Universidad Autónoma Chapingo.
REFERENCES
1. Val JE, Freitas MAR, Oliveira HN, Cardoso VL, Machado PF, Paneto JC. Indicadores de desempenho em rebanho da raça Holandesa: curvas de crescimento e altura, características reprodutivas, produtivas e parâmetros genéticos. Arq Bras Med Vet Zootec 2004;56(1):86-93. [ Links ]
2. Koya PR, Goshu AT. Solutions of rate-state equation describing biological growths. Am J Math and Stat 2004;3(6):305-311. [ Links ]
3. Thornley JHM, France J. Mathematical models in agriculture, 2nd ed. Wallingford: CABI Publishing; 2007. [ Links ]
4. Tjørve KMC, Tjørve E. The use of Gompertz models in growth analyses, and new Gompertz-model approach: an addition to the Unified-Richards family. PLoS One 2017;12(6) e0178691. https://doi.org/10.1371/journal.pone.0178691. [ Links ]
5. Tjørve E, Tjørve KMC. A unified approach to the Richards-model family for use in growth analyses: why we need only two model forms. J Theor Biol 2010;267:417-425. [ Links ]
6. Fitzhugh HA. Analysis of growth curves and strategies for altering their shape. J Anim Sci 1976;42(4):1036-1051. [ Links ]
7. Abreu UG, Cobuci JA, da Silva MVGB, Sereno JRB. Uso de modelos no lineales para el ajuste de la curva de crecimiento de bovinos Pantaneiros. Arch Zootec 2004;53(204):367-370. [ Links ]
8. Noor RR, Saefuddin A, Talib C. Comparison on accuracy of Logistic, Gompertz and von Bertalanffy models in predicting growth of new born calf until first mating of Holstein Friesian heifers. J Ind Trop Anim Agric 2012;37(3):151-160. [ Links ]
9. Gbangboche AB, Glele-Kakai R, Salifou S, Albuquerque LGD, Leroy PL. Comparison of non-linear growth models to describe the growth curve in West African Dwarf sheep. Animal 2008;2(7):1003-1012. [ Links ]
10. Malhado CHM, Carneiro PLS, Affonso PRAM, Souza Jr AAO, Sarmento JLR. Growth curves in Dorper sheep crossed with the local Brazilian breeds, Morada Nova, Rabo Largo, and Santa Inês. Small Ruminant Res 2009;84(1):16-21. [ Links ]
11. Behzadi MRB, Aslaminejad AA. A comparison of neural network and nonlinear regression predictions of sheep growth. J Anim Vet Adv 2010;9(16):2128-2131. [ Links ]
12. Perotto D, Cue RI, Lee AJ. Comparison of nonlinear functions for describing the growth curve of three genotypes of dairy cattle. Can J Anim Sci 1992;72(4):773-782. [ Links ]
13. Berry DP, Horan B, Dillon P. Comparison of growth curves of three strains of female dairy cattle. Anim Sci 2005;80(2):151-160. [ Links ]
14. McManus CM, Louvandini H, Campos VAL. Nonlinear growth curves for weight and height in four genetic groups of horses. Ciencia Anim Bras 2010;11(1):80-89. [ Links ]
15. Díaz HM. Indicaciones preliminares y el establecimiento de un módulo de producción de leche orgánica [tesis licenciatura]. Chapingo, Estado de México: Universidad Autónoma Chapingo; 2002. [ Links ]
16. Del Razo ROE. Características productivas y nivel de selenio y hormonas tiroideas en vacas lecheras suplementadas con bolos con selenio, bolos con selenio y yodo o bolos con yodo [tesis maestría]. Chapingo, Estado de México: Universidad Autónoma Chapingo; 2002. [ Links ]
17. Qing QBS. Comparison of four growth curve models in Angus cow: An application of Bayesian nonlinear mixed model. Report Presented to the Faculty of the Graduate School of the University of Texas at Austin; 2012. [ Links ]
18. SAS. SAS User’s Guide: Statistics (version 9.3 ed.). Cary NC, USA: SAS Inst. Inc. 2016. [ Links ]
19. Craig BA, Schinckel AP. Nonlinear mixed effects model for swine growth. Prof Anim Scient 2001;17(4):256-260. [ Links ]
20. Schinckel AP, Li N, Preckel PV, Einstein ME, Miller D. Development of a stochastic pig compositional growth model. Prof Anim Scient 2003;19(3):255-260. [ Links ]
21. Strathe AB, Danfaer A, Sorensen H, Kebreab E. A multilevel nonlinear mixed-effects approach to model growth in pigs. J Anim Sci 2010;88(2):638-649. [ Links ]
22. Echeverri ZJ, Salazar RV, Parra SJ. Análisis comparativo de los grupos genéticos Holstein, Jersey y algunos de sus cruces en un hato lechero del Norte de Antioquia en Colombia. Zoot Trop 2011;29(1):49-59. [ Links ]
23. Posada OS, Rosero NR, Rodríguez N, Costa CA. Estimación de parámetros de curvas de crecimiento de ganado Nellore criado en confinamiento. Rev MVZ Córdoba 2011;16(3):2701-2710. [ Links ]
24. Domínguez-Viveros J, Rodríguez-Almeida FA, Núñez-Domínguez R, Ramírez-Valverde R, Ortega-Gutiérrez A, Ruiz-Flores A. Ajuste de modelos no lineales y estimación de parámetros de crecimiento en bovinos Tropicarne. Agrociencia 2013;47(1):25-34. [ Links ]
25. Brown J, Fizugh H, Cartwright TC. A comparison of nonlinear models for describing weight-age relationships in cattle. J Anim Sci 1976;42(4):810-818. [ Links ]
26. Mazzini AR de A, Muniz JA, Silva FF, de Aquino LH. Curva de crescimento de novilhos Hereford: heterocedasticidade e resíduos autorregressivos. Cienc Rural 2005;35(2):422-427. [ Links ]
27. Herrera RAC, Vergara GOD, Cerón MMF, Agudelo-Gómez D, Arboleda ZEM. Curvas de crecimiento en bovinos cruzados utilizando el modelo Brody. Livest Res Rural Dev 2008, Livest Res Rural Dev 2008, http://www.lrrd.org/lrrd20/9/herr20140.htm . Consultado Dic 11, 2017. [ Links ]
28. Oliveira HN, Lôbo RB, Pereira CS. Comparação de modelos não-lineares para descrever o crescimento de fêmeas da raça Guzerá. Pesqui Agropecu Bras 2000;35(9):1843-1851. [ Links ]
29. Mgbere OO, Olutogun O. A Comparison of non-linear models for describing weight-age relationships in N'Dama cattle. J Appl Anim Res 2002;22(2):225-230. [ Links ]
30. Moreira RP, Mercadente MEZ, Pedrosa VB, Cyrillo JNSG, Henrique W. Growth curves of the Caracu breed. Ciências Agrárias 2016;37(4):2749-2758. [ Links ]
31. Gbangboche AB, Alkoiret TI, Toukourou Y, Kagbo A, Mensah GA. Growth curves for different body traits of Lagune Cattle. Res J Anim Sci 2011;5(2):17-24. [ Links ]
32. Vargas-Leitón B, Cuevas-Abrego M. Modelo estocástico para estimación de valores económicos de rasgos productivos y funcionales en bovinos lecheros. Agrociencia 2009;43(8):881-893. [ Links ]
33. López de Torre G, Candotti JJ, Reverter A, Bellido MM, Vasco P, García LJ, Brinks JS. Effects of growth curve parameters on cow efficiency. J Anim Sci 1992;70(9):2668-72. [ Links ]
Received: August 15, 2018; Accepted: February 17, 2021