Monitoring of Legionella pneumophila populations in environmental water is very important for public health, because the bacterium is the primary cause of Legionnaires' disease. The conventional plate culture method, similar to ISO 11731 (2017) , is generally used to enumerate Legionella spp. in water samples. The method, however, is highly complicated and requires sample handling and colony identification skills due to the growth of non-target bacteria and fungi on the agar plates. Thus, the development of a simpler method that could deliver accurate enumeration results is urgently needed.The Legiolert ® / Quanti-Tray ® most probable number (MPN) method (IDEXX Laboratories, USA) is a novel method to enumerate L. pneumophila in potable (e.g., tap water) and non-potable (e.g., cooling tower water) water samples. According to the manufacturer's instructions, the principle of this method is based on a bacterial enzyme detection technology that signals the presence of L. pneumophila through utilization of a substrate present in the Legiolert reagent. L. pneumophila cells grow rapidly and reproduce using the rich supply of nutrients present in the Legiolert reagent. Actively growing strains of L. pneumophila use the added substrate to produce a brown color indicator. There are four different protocols in the manufacturer's instructions, two for potable water and the other two for non-potable water. Recent papers have reported method comparison studies for the enumeration of L. pneumophila between Legiolert and the conventional plate culture method from potable water, such as hot water taps, showers, and ice machines (Sartory et al., 2017;Petrisek and Hall, 2017;Spies et al., 2018) and non-potable water, such as cooling tower water (Petrisek and Hall, 2017;Rech et al., 2018) . In this paper, we report the first comparison of the enumeration of L. pneumophila from bath water samples using the two Legiolert potable water methods and the conventional plate culture method.All water samples were collected from natural hot springs (53 samples) and public baths using tap water (79 samples) between May 2018 and March 2019. These samples were collected in sterile 500-mL polypropylene bottles with sodium thiosulfate, and were examined within 3 d after sample collection.The Legiolert/Quanti-Tray MPN method was carried out according to manufacturer's instructions. We compared the two potable water protocols (10 mL and 100 mL protocols) described in the instructions. In the 10 mL protocol (Legiolert-10 mL) , 10 mL of each water sample was transferred into a sterile 120 mL vessel and diluted with 90 mL of sterile deionized water. In the 100