The impact of unrecorded culling (URC) of first lactation daughters, prior to receiving an official test day record, on the integrity of sire genetic evaluations for production traits was investigated. Simulation was used to generate a complete set of realistic production records with full pedigree. Ten or twenty percent of the sires in the population were randomly chosen to be affected by 2%, 5%, 10% or 20% URC of the phenotypically worst daughters based on mean test day milk yield. Bias in EBV for milk yield, and correlated traits fat yield and protein yield, increased proportionately as culling intensity increased (5-48%). Sires re-ranked significantly for all traits at all culling percentages. The prominent re-ranking of elite sires under high URC levels could result in the loss of genetic improvement, while re-ranking under low to moderate URC levels were negligible. Results showed that low to moderate levels of unrecorded culling were not a practical concern for genetic evaluations in this study.