The count method allowed for the collection of uneaten pellets and the determination of FCR in fish, which was smaller than the FCR obtained by weighing. A partial loss of pellet fragments took place when the feces were separated from uneaten pellets when the weigh method was used, which accounted for the FCR difference.