Characteristics of egg numbers and mean egg weight were examined for their usefulness in the daily management of aviary systems for laying hens. 2. A number of 3238 brown Isabrown/Warren hens were housed in 1 compartment, a separated part of the house where the hens could move around freely, of a tiered-wired-floor aviary system (TWF-system). An automatic egg weighing and counting system (EWACS) was used to count and weigh eggs daily from 2 tiers of laying nests on 1 side of the compartment and the number of eggs for the whole compartment were counted daily by the farmer. Each tier was divided into 16 blocks of 5 individual laying nests. Two adjoining blocks were called a group. To prevent hens from walking along all the laying nests in a tier, partitions were placed on the perches in front of the laying nests, between nest groups 2-3, 4-5, and 6-7. 3. After the first 3 weeks of the laying period, the distribution of egg numbers over the nest groups within a tier became stable. If egg numbers were counted daily from only 1 nest group the coefficient of variation was 23.1%. If the eggs from the whole compartment were counted daily, the coefficient of variation for the number of eggs was 2.8%. The nest group, presence of a partition and tier level influenced the daily number of eggs. 4. The distribution of the mean egg weight over the different nest groups within a tier was stable for the whole laying period. The coefficient of variation of the daily mean egg weight for a nest group was 3.1%. The difference in mean egg weight between nest groups was small, between 0.1 and 0.6 g, and the level of tiers and the presence of partitions between nest groups had no effect on the mean egg weight. 5. It could be concluded that egg numbers could not be estimated reliably by taking samples from a group of laying nests or a tier, but that it was necessary to count all the eggs from a compartment. The daily mean egg weight, however, could be estimated reliably on the basis of a sample of eggs from a nest group or a tier. By using EWACS frequent samples could be taken, which diminished the coefficient of variation so that the reliability of the data increased.
