Japanese eel (Anguilla japonica) populations have been rapidly decreasing due to overfishing and habitat loss since the 1970s in Japan. U nderstanding the detailed ecological aspects of eels is essential to establish effective conservation measures . To augment depleted stocks, cultured eels have been widely released by fisheries cooperatives . However, studies on the ecology and behavior of cultured eels after release are limited , and the effectiveness of stocking remains unclear. The aim of this study is to examine behavioral characteristics of wild and cultured eels released in the wild to consider appropriate management policies for eel populations in Japan.
Studies were conducted in Matsukawa-ura lagoon in Fukushima (northern Japan) and the Tonda River in Wakayama (central Japan) using ultrasonic telemetry, which estimates the position of each eel by ultrasonic signals. Twenty-eight and 13 receivers were set in the Matsukawa-ura and Tonda River, respectively. Twenty wild and 12 cultured eels were implanted with acoustic transmitters and released in Matsuka wa-ura in September 2016, and 20 wild and five cultured eels were similarly outfitted and released in the Tonda River in June 2018 . Temporal and spatial changes of received signal data of each eel were analyzed.
Some eels migrated to the sea, and some moved between the lagoon/ river and the sea several times after release . Migration from the Matsukawa-ura lagoon to the open sea occurred at night during ebb tide in autumn (October to November ; Table 1 ). Out of the eight eels that migrated to the sea, seven individuals were cultured. However, it is unlikely that the seven cultured eels (2 years old) were mature enough for a spawning migration. This migration behavior is thought to be abnormal for cultured eels that were raised under constant high temperature. In the Tonda River, three cultured and one wild eel migrated to the open sea in September and October. Seasonal patterns of behavioral activity were found: high in summer and autumn, and low in winter for both eel groups. Both eel groups had nocturnal feeding activity ; in particular wild eels were most active around dusk. Diel rhythms of cultured eels were less pronounced compared to wild ones. The results of the present study indicate that the behavior of cultured eels differs from that of wild eels and that this difference most likely is caused by the artificial rearing conditions implemented in the aquaculture ponds 0.5 - 2 years before release. Although some cultured eels migrated to the sea, it is difficult to ass ess the contribution of these eels to overall reproduction in the wild. Previous studies indicated that the release of cultured eels may negatively impact ecosystems. Therefore, the release of cultured eels may disrupt the ecosystem; and unknown effects on reproduction should be reduced as much as possible. Instead, establishing rules to prevent overfishing of glass eels and restoring habitats in rivers and coastal areas to foster the establishment of wild yellow eels as future spawning stock are important.