Thesis

This study aimed to use radio frequency identification (RFID) technology combined with camera image data to establish individual oviposition records of breeding geese and to investigate their laying performance and individual behavioral differences. First-parity White Roman geese were used in this study and were reared in an environmentally controlled goose house under a floor-rearing system with natural mating. During the experimental period, the daily numbers of nest eggs and floor eggs were recorded manually, while the RFID system and image data were used to identify individual nest-laid oviposition events. A total of 14,138 eggs were recorded from manual laying records, with an average of 29.3 eggs per female goose. The proportion of eggs laid in nest was 86.4%, whereas floor eggs accounted for 13.6%. The distribution of floor eggs showed that floor eggs were not randomly distributed within the goose house but were clearly concentrated around the feeding troughs. The distribution of nest eggs also showed uneven spatial use, with higher proportions of eggs laid in the right and middle nest areas and a lower proportion in the left nest area. This indicates that even when all nest had the same structure, breeding geese may still show preferences for different nest locations. After RFID identification, image verification, and manual interpretation, the effective identification rate of oviposition events was 44.6%. This result indicates that the RFID system used in this study can provide individual-level laying records, although some nest eggs could not be completely assigned to the laying individuals. Based on the confirmed individual oviposition records, individual nest-box use behavior was further analyzed. Three behavioral indicators were quantified: mean distance between nests used in consecutive ovipositions (MDN), percentage of nests used for laying (PNL), and the degree of concentration in oviposition distribution, expressed as the Gini coefficient. Among individuals with 10 or more nest eggs, K-means clustering analysis and principal component analysis classified the breeding geese into three groups according to their nesting behavior: wide-range preference type, wide-range dispersed type, and fixed preference type. These results indicate that different nest use patterns existed among individual breeding geese. Correlation analysis showed that the number of nest eggs (NE) was significantly positively correlated with PNL and the Gini coefficient, whereas NE was not significantly correlated with MDN. This suggests that individuals with higher numbers of nest eggs not only used more nest, but their oviposition distribution may also have been concentrated in specific nest.
Keywords: white roman goose, radio frequency identification, nest.