Abstract: The aim of the study was to evaluate the Hungarian Sporthorse population based on eventing competition performance. The database contained the results of 792 horses and 449 riders between 1996 and 2006. The eventing results were gathered from Hungary and other European countries. Blom transformed ranks were used to measure competition performance. Sporthorses competed in fourteen categories but only the easiest category (category ’A’) contained enough result to handle it as a single category (model I.). The other 13 categories were handled together based on professional reasons in a different model (model II.). In model III., all records were analysed together and the results were weighted according to the difficulty of the category. The competition results were classified into eight groups, each group had a weight (0-7) and it was multiplied by a constant 3, and the result of this formula was added to the original Blom score. The linear mixed models included fixed effects for age, sex, breeder, owner, location, year and random effects for animal and rider. The model II. contained one more fix effect for difficulty of the competition level. The distribution of number of horses and number of starts by sex were heterogeneous (P<0.05). For category ’A’ mares and geldings appeared in higher proportion, despite this fact ratio of stallions is greater in higher competition levels. Considering the goodness of fit in each model, model fitting to the weighted Blom scores was the best R2=0.82. In model III. every fix effects (age, sex, breeder, owner, location, year) were significant. Random effects for animal and rider were not significant, but form professional point of view they need to be included in the model. The variance components estimated for the weighted Blom scores were the highest also (0.52 rider effect, 0.09 animal effect), the animal x rider interaction effect was zero. The variance proportion of rider effect exceeded the variance proportion of animal effect in each model. Breeding values of eventing performance were predicted using model III. The reliability of the estimated breeding values was acceptable for only a few stallions. To improve the reliability of breeding values, more progenies should be used in eventing competitions (as a kind of progeny test) and more competition records needed (as a kind of own performance test).

Abstract: The objective of the study was to determine which cross-country obstacles are more difficult for eventing horses. Jumping scores were considered in terms of the horse's reaction to novelty and to the fearfulness of novel objects which are the fences situated in novel terrain. The data concerned 11 classes of One to Four Star level (stars showing the difficulty of the class) held at the Olympic Games and three international Three Day Events. A total of 400 entries, in which 259 horses jumped 372 obstacles were considered. Scores of 11,341 jumps at particular fences were categorized either as faulty jumps or non-faulty jumps. Factors describing the fences versus the jumping scores were studied with least square analysis of variance, with respect to the interaction between the star level and the fence traits. The overall frequency of faults at the cross-country amounted to 4.33 ± 0.57%. Among the effects analyzed, the difficulty of cross-country fences for the horses depends upon whether an obstacle is single or is an element of a combination, whether it is straight or requiring an effort in both height and spread, is broad or narrow, has a solid top or a brush, has an alternative or not and whether it is a water crossing or not. At One Star level, the less experienced horses react differently to certain fence traits compared to horses participating in Three or Four Star levels. It is concluded that the equine visionary system, being less developed towards identification of stationary objects than the human visionary system may be a key towards explaining the horse's behaviour while jumping the obstacle. The height-spread obstacles, those of the narrow front, with the brush and with the alternative seem to involve more faults because of the equine low-acuity vision. The same reason may justify the similar frequency of faults at the single obstacles and the first elements in combinations. Accurate methods of measuring equine vision would complement behavioural tests and should both be introduced into the selection of eventing horses.