UTILITY VALUE OF HALF-BRED HORSES TESTED FOR RACING PERFORMANCE ON THE WROCŁAW HORSE-RACING TRACK

Henryk Geringer de Oedenberg, Maciej Dobrowolski, Anna Górecka, Ewa Guzik, Justyna Domańska
Institute of Animal Breeding, Department of Horse Breeding and Equestrianism, Wrocław University of Environmental and Life Sciences, Poland

ABSTRACT

The study involved 3-year-old Half-Bred horses running at the Wrocław Partynice Horse Racing Track during 1999–2005. In all, pedigrees and racing careers of 527 horses have been analysed. Racing performance was analysed basing on three criteria: individual success index (SI), career earnings (CE), and mean earnings per start (ME). Significant effects of gender, trainer, and number of starts on performance of the horses were found. Effects of breed, season, and breeder were non-significant. The performance appraisal criteria, i.e. career earnings (h² = 0.61) and mean earnings per race (h²= 0.58), should be considered as adequate due to their heritability.

Key words: Half-Bred horses, genetic parameters, racing performance.

INTRODUCTION

The racing-track performance test of Half-Bred horses allows selection of sound, mentally strong animals, fully suited for disciplines of the equestrianism. The test represents an alternative to  obtaining a license or the right to enter the studbook.

Poland is among the few European countries organizing Half-Bred horse races. A number of examples illustrate how reliable and objective the tests are for evaluation of horses in terms of their prospective racing or jumping career. Hence, studies on performance have fundamental importance for selection and further breeding work.

Racing performance, a measurable phenotypic trait, is formed by a number of environmental factors, besides horse's origin [6]. Determination of the weight of each particular effect on the racing performance of the horse guarantee reliable evaluation of its performance and, in perspective, its breeding value.

Domestic literature brings many reports on Half-Bred horses used in racing [3, 4]. Studies on racing performance of Half-Bred horses were started by Izabela Bek-Kaczkowska [1], who analysed populations present in 1988–1998.

The aim of the study was to evaluate racing performance of Half-Bred horses participating in races at the Wrocław Horse-Racing Track during 1999–2005.

MATERIAL AND METHODS

The analysed dataset contained information on racing performance of three-year-old Half-Bred horses running during 1999–2005. The analysis included horses that participated in flat races over a distance of 1600 to 2400 meters. The data were collected from the Racing Newsletter published by the Track, records of horses born during 1996–2002, as well as from the Racing Programmes. The analysis included 527 horses (3076 starts) bred in 12 state stud-farms as well as by individual breeders. The horse's performance was evaluated basing on three criteria: individual success index and (SI), career earnings (CE), mean earnings per race (ME).

The analysed population was ranged by effects of the breed (1. Wielkopolski, 2. Anglo-Arabian, 3. Małopolski, 4. Noble Half-Bred), gender (1. mare, 2. stallion, 3. gelding), breeder (1. State Agricultural Property Agency (ANR) units such as stud-farms and stallion farms, 2. private farms with more than 6 horses, 3. private farms below 6 horses), trainer (1, ..., 5), number of starts (A. up to 6 starts, B. from 7 to 11 starts, C. from 12 to 16 starts).

The resulting data were log-transformed and analysed statistically using the GLM procedure and the SAS statistical package (SAS, 2000). ANOVA was carried out in order to test the significance of particular effects in relation to the level of the studied traits. Significance of differences between means was tested using the Duncan test.
Individual success index (SI) was calculated on the number of starts and career earnings, according to the following formula:

where:

season's earnings by the given horse,
x_a mean for the age group (the sum of earnings of all the horses of the same age divided by the number of horses at this age). The variance components were estimated with REML (SAS, 2000), according to the following model:

where:
y_ijklmno – empirical value,
µ – population mean,
R_i – fixed effect of the i-th breed (i = 1, ..., 4),
P_j – fixed effect of the j-th gender (j = 1, 2, 3),
S_k – fixed effect of the k-th season of start (k = 1, ..., 7),
H_l – fixed effect of the l-th breeder (l = l, 2, 3),
T_m – fixed effect of the m-th trainer (m = l, ..., 5),
LS_n – fixed effect of the n-th group of starts number A, B, C (n = 1, 2, 3),
o_o – random effect of o-th sire,
e_ijklmno – residual effect.

The following formula was used to calculate the heritability coefficient:

where:
– sire's component,
– within-class component of phenotypic variance.

Fitness of the model to the structure of the data was tested by means of GLM of the SAS package (SAS, 2000).

The NESTED procedure of the SAS package (SAS, 2000) was used to determine the relationships between the phenotypic and genetic values of the racing performance indices (SI, CE, ME) attained by horses during trials.

RESULTS AND DISCUSSION

Basing on the analysed data, a significant effects of gender, trainer, and number of starts on the racing performance of horses were found. Effects of breed, season, and breeder were non-significant.

The highest success index and maen earnings per start were achieved by geldings, followed by mares (Table 1). This may have resulted from the organisation of performance appraisal methods of Half-Bred horses in Poland, where best stallions usually reach training centres and, therefore, do not enter racing performance trials. The high value of geldings could have resulted from a low number of racing horses taking part (14 animals), thus this fact is of informative value only. Similar relationships were observed in a population of Arabian horses [9], whereas Thoroughbred stallions exhibited the best racing performance, although geldings were not included in the cited analysis [5].

Table 1. Means of racing performance measures for horses by different sex

Trait	Mares			Stallions			Geldings
	n		sd	n		sd	n		sd
SI	294	1.03	1.27	219	0.908 a	1.16	14	1.06 a	0.61
CE	294	5006.49	6645.36	219	4280.84	5433.91	14	4955.21	2941.28
ME	294	850.92 a	1196.26	219	778.07 b	1014.88	14	1036.4 ab	529.58

Within the particular lines, the means marked with the same letter differ significantly at: small letters (P ≤ 0.05), big letters (P ≤ 0.01), SI – success coefficient, CE – earnings, ME – mean of earnings.

The highest mean racing performance indices were attained by trainer 3, statistically significantly better compared to trainer 4 (Table 2). The significance of this effect was found at the level of  P ≤ 0.02 for SI and career earnings and P ≤ 0.03 for mean earnings per start. The significance of trainer's effect was reported also by Sobczyńska [7] as well as Chrzanowski and Koebcke [2], who studied a population of Thoroughbreds. The significance of the effect of number of starts differed depending on the analysed criterion of performance (Table 3). It should be stressed that horses of group B, i.e. those starting 7 to 11 times in a season, exhibited better parameters of racing performance. Horses of group A, which entered less than 6 races, were usually sent back to their stud-farms after a series of unsuccessful initial starts. The medium-level achievements of group C horses may have resulted from the fact that such middle-group horses were often entered into the race to fill the gaps. A small size of this group (14 animals) suggests a low importance of such horses. The highest performance by all criteria was achieved by the horses belonging to the state-owned stud farms, although no significant differences were found between the stud farms as such (Table 4).

Table 2. Mean of racing performance measures of horses by different trainers

Trainer	I			II			III			IV			V
Trait	n		sd	n		sd	n		sd	n		sd	n		sd
SI	91	0.96	0.97	91	1.02	1.16	99	1.30 a	1.74	86	0.82 a	0.85	160	0.85	1.11
CE	91	4618.65	4628.65	91	4880.97	5707.82	99	6045.51 a	8086.88	86	3866.02 a	3964.43	160	4270.86	6480.88
ME	91	824.01	950.51	91	801.67	976.61	99	1104.54 a	1415.73	86	678.60 a	837.20	160	746.45	1162.87

Explanation as in Table 1.

Table 3. Means of racing performance measures for horses by numbers of starts (A – to 6 starts, B – from 7 to 11 starts, C – from 12 to 16 starts)

Trait	A			B			C
	n		sd	n		sd	n		sd
SI	231	0.722 a	1.186	282	1.208 ab	1.220	14	0.796 b	0.801
CE	231	3383.3 a	5564.79	282	5820.11 a	6392.79	14	3997.50 b	3807.21
ME	231	849.43 A	1219.23	282	829.84 B	103.700	14	346.23 AB	321.80

Explanation as in Table 1.

Table 4. Means of racing performance measures for horses by origin

Trait	State breeding			Private breeding			Others to 5 horses
	n		sd	n		sd	n		sd
SI	385	1.024	1.244	32	0.763	0.857	110	0.907	1.215
CE	385	4892.65	6332.25	32	3659.06	3933.77	110	4371.86	5756.06
ME	385	856.88	1154.67	32	515.61	572.06	110	806.17	1063.0

Explanation as in Table 1.

Małopolski horses represented the largest group among the studied population (335), followed by Noble Half-bred (134), Wielkopolski (49), and pure-bred Anglo-Arabians (9 horses, Table 5).

Anglo-Arabians were the best horses in terms of racing performance, followed by Małopolski, without a statistically significant superiority, however. Apart from the season of 1999, career earnings and mean earnings  remained on a constant level (Table 6).

Table 5. Mean of racing performance measures for horses by different breed

Trait	wlkp			xxoo			m			sp
	n		sd	n		sd	n		sd	n		sd
SI	49	0.79	0.67	9	1.51	2.057	335	1.05	1.32	134	0.73	0.87
CE	49	3565.9	3084.8	9	6389.89	8461.53	335	5221.76	7189.03	134	3530.4	4283.13
ME	49	601.38	574.56	9	922.29	1190.31	335	923.97	1270.91	134	660.21	855.91

Explanation as in Table 1.

Table 6. Means of racing performance measures for horses by season of starts

Trait	1999			2000			2001			2002			2003			2004			2005
	n		sd	n		sd	n		sd	n		Sd	n		sd	n		sd	n		sd
SI	95	1.00	1.25	70	1.01	1.14	83	1.01	1.27	72	0.98	1.14	73	1.02	1.36	70	0.85	0.85	64	1.000	1.456
CE	95	4111.59	5341.01	70	4699.43	5341.01	83	4996.69	6282.32	72	5099.72	5879.85	73	4852.60	6507.18	70	4664.71	7061.11	64	4633.59	6746.49
ME	95	662.18	904.61	70	855.27	1054.18	83	827.95	1067.12	72	857.11	1111.12	73	876.99	1257.31	70	895.17	1295.47	64	862.32	1135.55

Explanation as in Table 1.

Racing performance measures represent traits that are of low to middle heritability and, according to Kaproń [6], remain within the range 0.01 to 0.60. These values depend on the criterion of performance, size and structure of the studied population, and, last but not least, on the applied statistical method. Analysing genetic parameters of the studied population, one should stress a high heritability of career earnings and mean earnings per start, as well as a low SI heritability. Heritability based on the sire component was h² = 0.16 (V(h²) = 0.02) for SI, h² = 0.61 (V(h²) = 0.09) for career earnings, and h² = 0.58 (V(h²) = 0.07) for the mean earnings per start (Table 7). Heritability of SI (h² = 0.16) was similar to that reported by Chrzanowski and Koebcke [2] for Thoroughbreds.

Table 7. Genetic parameters of racing performance of half breed horses

Trait	SI	CE	ME	h2
SI		0.21	0.17	0.16
CE	0.45		0.26	0.61
ME	0.28	0.56		0.58

Explanation as in Table 1. h2 heritability coefficient. Genetic correlation coefficients (rG ) − above the diagonal. Phenotipic correlation cooefficients (rP) − under the diagonal.

Low values of this index could imply a little linkage between performance (a phenotypic value) and breeding value for this trait. On the other hand, the value of heritability index calculated for career earnings and mean earnings per start was found much higher compared to that reported by Sobczyńska and Łukaszewicz [8] for Thoroughbred horses (h² = 0.14). Low correlations, both phenotypic and genetic, between SI and the remaining racing performance, along with a low heritability of SI, means that this parameter is of little use in selection. A high heritability of career earnings and mean earnings per start, on the other hand, are the characteristics that make these parameters fully suitable in the breeding work.

CONCLUSIONS

1. Career earnings and mean earnings per start were found good measures of racing performance for the studied population.

2. Effects of gender, trainer, number of starts were significant in terms of performance of the horses.

3. No significant effects were found for breed, season, and breeder in relation to racing performance of the analysed group of horses.

REFERENCES

1. Bek-Kaczkowska I., Chudoba K., 2000. Wartość hodowlana ogierów półkrwi szacowana metodą BLUP Animal Model [Breeding value of Half-Bred stallions estimated using BLUP Animal Model]. Zesz. Nauk. Prz. Hod. 50, 43-50 [in Polish].

2. Chrzanowski S., Koebcke K., 1993. The estimation of breeding value of Thoroughbred horses on the basis of racing performance of their progeny at the age 2 and 3 years. Ann. Wars. Agric. Univ. Anim. Sci. 29, 35-39.

3. Dobrowolski M., Geringer H., 2003. Porównanie wartości hodowlanej ogierów ras półkrwi, określonej metodą BLUP AM, na podstawie wyników prób dzielności przeprowadzonych w zakładach treningowych w latach 1997-2000 [Comparison of breeding values of Half-Bred horses estimated using BLUP AM, based on racing performance trials results carried out in training centres during 1997-2000]. Zesz. Nauk. Prz. Hod. 68(5), 205-212 [in Polish].

4. Geringer H., Olczyk A., Kiełbasiewicz A., 2003. Wartość hodowlana ogierów ocenianych na podstawie potomstwa startującego w dyscyplinie skoków przez przeszkody w latach 1990-2000 [Breeding value of stallions based on the progeny taking part in show jumping during 1990-2000]. Zesz. Nauk. Prz. Hod. 68(5), 101-107 [in Polish].

5. Grzesiak W., Pikuła R., 2003. Analiza zmienności mierników wartości wyścigowej koni pełnej krwi angielskiej [Analysis of variability of racing performance parameters in Thoroughbred horses]. Zesz. Nauk. Prz. Hod. 68(5), 234-240 [in Polish].

6. Kaproń M., 1999. Metody doskonalenia kon. [Methods of horses improvement]. Wyd. AR Lublin [in Polish].

7. Sobczyńska M., 2004. Wpływ "efektu trenera" na parametry genetyczne cech wyścigowych koni pełnej krwi angielskiej i czystej krwi arabskiej [Effect of trainer on genetic parameters of racing traits of Thoroughbred and Arabian horses]. Zesz. Nauk. Prz. Hod. 72(5), 27-32 [in Polish].

8. Sobczyńska M., Łukaszewicz M., 2003. Heritability of racing merit of Arab horses. Anim. Sci. Pap. Rep. 21(4), 233-239.

9. Pikuła R., Grzesiak W., 2003. Analiza zmienności mierników wartości wyścigowej koni czystej krwi arabskiej [Analysis of variability of racing performance parameters in Arabian horses]. Zesz. Nauk. Prz. Hod. 68(5), 229-234 [in Polish].

Accepted for print: 27.11.2008

---

Henryk Geringer de Oedenberg
Institute of Animal Breeding,
Department of Horse Breeding and Equestrianism,
Wrocław University of Environmental and Life Sciences, Poland
Kożuchowska 5a, 51-631 Wrocław, Poland
email: hengen@ozi.ar.wroc.pl

Maciej Dobrowolski
Institute of Animal Breeding,
Department of Horse Breeding and Equestrianism,
Wrocław University of Environmental and Life Sciences, Poland
Kożuchowska 5a, 51-631 Wrocław, Poland

Anna Górecka
Institute of Animal Breeding,
Department of Horse Breeding and Equestrianism,
Wrocław University of Environmental and Life Sciences, Poland
Kożuchowska 5a, 51-631 Wrocław, Poland

Ewa Guzik
Institute of Animal Breeding,
Department of Horse Breeding and Equestrianism,
Wrocław University of Environmental and Life Sciences, Poland
Kożuchowska 5a, 51-631 Wrocław, Poland

Justyna Domańska
Institute of Animal Breeding,
Department of Horse Breeding and Equestrianism,
Wrocław University of Environmental and Life Sciences, Poland
Kożuchowska 5a, 51-631 Wrocław, Poland

---

Responses to this article, comments are invited and should be submitted within three months of the publication of the article. If accepted for publication, they will be published in the chapter headed 'Discussions' and hyperlinked to the article.

---

Main  -  Issues  -  How to Submit  -  From the Publisher  -  Search  -  Subscription