GENOTYPE-ASTERISK-ENVIRONMENTAL INTERACTIONS IN A GERMAN PIG BREEDINGHERDBOOK SOCIETY USING CROSSBRED PROGENY INFORMATION

At nucleus level of the German 'Pig Breeding Society North-West', sele ction of growth and carcass traits is based upon performance records o f purebred, station-tested pigs. In a next step Pietrain AI-boars are selected on field-and/or station-tested crossbred progenies. It is sup posed that the genetic correlations (r(g)) between same traits measure d in different environments do not deviate from one. In order to proof this hypothesis, genetic parameters were estimated using REML, applie d to a multivariate animal model including the significant random inte ractions sireherd and sire*sex. Data set 1 (DS1) covered 896 and 2916 records of station (1 test-station) or field (6 commercial fattening herds) tested progenies of 46 AI-sires recorded during 1993-94. Data s et 2 (DS2) comprised 4572 progenies of 110 AI-boars recorded during 19 95-97 in 11 fattening herds. Traits measured in the test station were 'daily gain' (DG), 'estimated carcass lean content' (LP) and 'conducti vity' (C). Field recorded traits were 'carcass weight to age ratio' (N DG), 'Fat-O-Meater carcass lean content' (FOM) and 'FOM-reflection ind ex' (RI). Estimated h(2) in DS1 (DS2) were for FOM .31 (.37), NDG .30 (.02), RI .18 (.11), DG .69, LP .56 and C .36 (s(h2):.01-.11). The est imated siresex to phenotypic variance ratios for FOM ranged only betw een .02 and .04 whereas the sireherd variance ratios for NG were of p ractical impact with .09 +/- .02 (DSI) and .12 +/- .02 (DS2). The r(g) between station and field traits were .65 +/- .18 for LP to FOM and . 60 +/- .19 for C to RI. However, the r(g) between DG and NDG (-.29 +/- .19) was negative supporting the existence of genotypeherd interacti ons.