POSSIBILITIES OF RECORDING REPRODUCTIVE TRAITS FROM SOW MANAGEMENT PROGRAM DATA IN COMMERCIAL PIG EVALUATION

Field data from the reproduction monitoring programme of the Al-Statio n in Ascheberg were used to consider the possibility to organise a com mercial product evaluation for reproductive traits from sow management programme data, Comparisons between 7 breeds were made using hierarch ically structured data (farms nested within breed) and non-hierarchica lly structured data (several breeds per farm or farms with a reference breed). For 3 breeds only a representative sample especially under co nsideration of number of farms was found, The traits mating success, n umber of piglets born alive and farrowing interval were analysed. As a dditional fixed effects, beside breed and farm parity (4 classes), sea son (6 months) and mating type (artificial insemination or natural mat ing) wore included in the model. Only the non-hierarchically structure d data showed for number of piglets boil alive significant breed diffe rences. The hierarchically structured data showed especially for breed s with higher number of farms for number of piglets born alive very si milar results compared to the non-hierarchically structured data. The lower number of degrees of freedom did not prove these differences to be significant. The analysed data clearly show that for the 7 breeds t here are not enough farms with at least to breeds to get non-hierarchi cally structured data. But the increasing amount of farms using sow ma nagement programmes will in future lend to larger samples for more tha n 3 breeds. These representative samples of hierarchically structured data will then prove the observed differences of 0.2 to 0.4 piglets bo rn alive to be significant.