Body ovoidal with broad posterior half (Fig. 1A), covered with scale-like tegumental spines (Fig. 2A) from anterior to posterior ends. Oral sucker terminal, and 32 (34 in 1 of 10 specimens) circumoral spines arranged in 2 alternative rows (Figs. 1C, 2B). Esophagus short. Ceca large, bifurcated about midway between ventral and oral suckers, terminated slightly in front of ovary. Ventral sucker smaller than oral sucker and located at median 2 / 5 of body. Round non-ciliated and uni-ciliated papillae and tegumental spines located on half of lower part of ventral sucker but upper part of ventral sucker, between ventral sucker and genital pore, without tegumental spines (Fig. 2C). In mid-dorsal tegument, spines pectinate, with long uniciliated sensory papillae (Fig. 2D). Seminal vesicle large, bipartite, with anterior part transversely elongated and posterior one rounded. Genital pore anterior to ventral sucker. Ovary oval, located on right side of posterior half of body. Seminal receptacle lies in middle of posterior half of body, dorsal to ovary, between seminal vesicle and anterior border of excretory bladder. Laurer's canal prominent, opening dorsally. Mehlis' gland and ootype opposite to ovary. Uterus short, coiled between ovary and seminal vesicle, filled with 4-16 (av. 10.4) eggs. Vitelline follicles large, scattered laterally from posterior end to posterior border of pharynx. Eggs oval, shell surface with lattice design, yellowish brown, operculum distinct (Fig. 1D). Excretory bladder X-shaped, with short posterior tube. Excretory pore terminal, with low density of tegumental spines (Fig. 2E).

In the present study, the adult flukes of C. formosanus have been first identified in Lao PDR. The worm size was slightly smaller than those reported by previous authors [1,3], but it is considered to have been due to a shorter developmental period in our specimens. The important morphological features included the presence of 32 (rarely 34) circumoral spines arranged in two alternative rows, a large bipartite seminal vesicle, an oval-shaped ovary, egg shell surface with lattice design, and an X-shaped excretory bladder. The freshwater fish, Puntius brevis, is suggested to be a source of animal or, possibly human, infections with C. formosanus.

In the speciation of flukes belonging to the genus Centrocestus, the number and size of circumoral spines, the margin forms of the ovary and testes, the size and number of eggs in the uterus, the ridge patterns on the egg shell surface, the level of ceca ends, the shape of excretory bladder, and the presence of prepharynx were adopted as useful keys [1,4]. Among these characters, the number of circumoral spines was accepted as the most reliable one, although the number varied showing a range in some species, at times overlapping one another [1,4]. Counting the exact number is difficult in some specimens. Waikagul et al. [4] classified the reported species of Centrocestus into 3 big groups, i.e., those with 26-36 spines (C. yokogawai, C. caninus, C. longus, C. formosanus, C. cuspidatus, and C. asadai), 38-48 spines (C. nycticoracis, C. kurokawai, and C. armatus), and 50-60 spines (C. polyspinosus). Among the first group, they finally recognized only 2 species, C. cuspidatus, which has a V-shaped excretory bladder, and C. caninus, which has an X-shaped excretory bladder [4]. Species of the first group other than the two, i.e., C. yokogawai, C. longus, C. formosanus, and C. asadai, were regarded similar and synonymized with C. caninus [4].

However, the synonymy of C. formosanus with C. caninus is put to a question because of several reasons. First, the number of circumoral spines is 26-30 in C. caninus [4,14], whereas that of C. formosanus is 30-36 [1,3,15], not overlapping significantly. The latter has in most cases 32 spines [1,3,15] and rarely 30, 34, or 36 [1]. Therefore, the number of circumoral spines may be a unique feature for both C. formosanus and C. caninus. Second, C. caninus was originally described based on only 1 specimen obtained from a dog in Taiwan [14]. The number of circumoral spines was counted 28, but, as the preservation status of the worm was not good, it was suggested that some of the spines originally present may have been lost [14,16,17]. The worm was tentatively designated as a variant of C. cuspidatus, i.e., C. cuspidatus caninus, until material for further study would be available [14]. Separately from this, C. formosanus was found and described as a new species in Taiwan [3], and Faust and Nishigori [17] considered that Leiper's specimen of C. cuspidatus caninus was in fact C. formosanus. Yamaguti [18] raised C. cuspidatus caninus up to a specific level, i.e., C. caninus, and synonymized C. formosanus, C. longus, and C. yokogawai with C. caninus without appropriate reasons. Later, he retained C. formosanus as a valid species, whereas no more describing C. caninus [19]. A third point is that C. caninus was not reported for a long time after it was first reported based on only 1 specimen in 1913 [14], until 1997 when 6 specimens recovered from 2 human infections in Thailand were assigned as C. caninus [4]. We believe that treating C. caninus as a specific level [18] is problematic, and even if C. caninus (as C. cuspidatus caninus) had in fact more than 28 circumoral spines, for example, 32, C. formosanus should have the taxonomic priority.

Among the Centrocestus species, those reported from human infections are C. armatus (experimental and natural) [20,21], C. kurokawai (natural) [22], C. cuspidatus (natural) [23], C. caninus (natural) [4,5], and C. formosanus (experimental) [3]. In the report of human infections from Lao PDR [5], the worm was tentatively designated, because the number of specimens was only 4 and detailed morphological studies were not possible. With regard to C. formosanus, the possible presence of natural human infections has been suggested [15,24]. In this regard, epidemiological studies are required to verify this suggestion.

The second intermediate hosts of C. formosanus were reported to be a variety of freshwater fishes, Carassius auratus, Pseudorasbora parva, Channa formosana, Gambusia affinis, Ophicephalus tadianus, Misgurnus anguillicaudatus, Zacco platypus, Cyprinus carpio, Glossogobius giurus, Anabas testudineus, and Puntius semifasciolatus, frogs, Rana limnocharis, and toads, Bufo melanostictus [1]. The present study adds P. brevis as a new second intermediate host of C. formosanus. The metacercariae are encysted mostly in the gills of fish [1,7], but in heavily infected fish a small number of metacercariae may also be present under the scale of muscle [1]. As the gills are generally discarded during cooking fishes, infected people may be few in number. However, caution should be paid as the number of uterine eggs in Centrocestus spp. adults is small, less than 30-40 eggs per worm [1,4,25], and the chance of detecting eggs in fecal examination is low particularly in light infection cases. Moreover, the eggs of C. formosanus closely resemble those of O. viverrini, having lattice patterns on the egg shell; hence, researchers may find difficulty in differentiating them in fecal examinations.