Selection of hens with intact beaks for high performance in multiple-bird cages has produced a stock that clearly exceeds its unselected control when hens have intact beaks but shows less advantage when comparisons involve effectively beak-trimmed birds. The selected stock performed at about the same level as a stock recently derived from a commercial source (but with two generations of relaxed selection) when birds had intact beaks but at a lower level when birds had their beaks trimmed. Relative incidences of deaths from cannibalistic pecking were partially responsible for the results obtained.