The Nomenclature of Trichopodus pectoralis REGAN , 1910 ; Trichopus cantoris SAUVAGE , 1884 and Osphronemus saigonensis BORODIN , 1930 ( Teleostei : Peciformes : Osphronemidae )

Only four species are so far known of the Southeast Asian genus Trichopodus LACEPÈDE, 1801: Trichopodus trichopterus (PALLAS, 1770) (not 1777, as often wrongly quoted); Trichopodus leerii (BLEEKER, 1852); Trichopodus microlepis (GÜNTHER, 1861) and Trichopodus pectoralis REGAN, 1910. With 150 up to 180 mm TL the last species is the biggest. In their home country all species are caught for everyday consumption and furthermore are considered well-known aquarium fi sh. In its original range (South Vietnam, Thailand, and Malayan Peninsula) Trichopodus pectoralis is also grown in ponds and paddy fi elds. There were also undertaken several artifi cial, partly successful introduction attempts in other countries (South China [Hong Kong], Sri Lanka, Indonesia, The Philippines, New Caledonia, Papua New Guinea, Pakistan, Japan und Columbia) (SMITH, 1945; WELCOME, 1988). In contrast to the economic interest in the SnakeSkin Gourami, its identifi cation as a discrete species and its nomenclature, however, seemed diffi cult. In the following the nomenclature of the Snake-Skin Gourami will be explained and commented and the status of two other taxa is discussed. The Nomenclature of Trichopodus pectoralis REGAN, 1910; Trichopus cantoris SAUVAGE, 1884 and Osphronemus saigonensis BORODIN, 1930 (Teleostei: Peciformes: Osphronemidae)


Introduction
Only four species are so far known of the Southeast Asian genus Trichopodus LACEPÈDE, 1801: Trichopodus trichopterus (PALLAS, 1770) (not 1777, as often wrongly quoted); Trichopodus leerii (BLEEKER, 1852); Trichopodus microlepis (GÜNTHER, 1861) and Trichopodus pectoralis REGAN, 1910.With 150 up to 180 mm TL the last species is the biggest.In their home country all species are caught for everyday consumption and furthermore are considered well-known aquarium fi sh.In its original range (South Vietnam, Thailand, and Malayan Peninsula) Trichopodus pectoralis is also grown in ponds and paddy fi elds.There were also undertaken several artifi cial, partly successful introduction attempts in other countries (South China [Hong Kong], Sri Lanka, Indonesia, The Philippines, New Caledonia, Papua New Guinea, Pakistan, Japan und Columbia) (SMITH, 1945;WELCOME, 1988).
In contrast to the economic interest in the Snake-Skin Gourami, its identifi cation as a discrete species and its nomenclature, however, seemed diffi cult.In the following the nomenclature of the Snake-Skin Gourami will be explained and commented and the status of two other taxa is discussed.

The most important features of the Snake-Skin Gourami in comparison to the other three Trichopodus species
There are two old basic pattern in the colouration of the Southeast Asian gourami, which very probably originated during the phylogenesis of the species of the genera Trichopodus and which can be assumed as plesiomorph.On the one hand, it is a blackish, longitudinal band starting from the mouth towards the eye down to the caudal peduncle.On the other hand there is a pattern of dark, oblique bars, often with silver shining gaps (best depicted in REGAN, [1910] Image LXXIX, Fig. 1), covering the whole body and parts of the median fi ns.These two features are not similarly distinct in all four species and can, when occurring in a species, superimpose each other more or less strongly and thus change the appearance of the respective individuals (Miller & Robison 1974).They also play a signifi cant role in the verbal descriptions of the Three-Spot and Snake-Skin Gourami, and they are an important reason for the double description of the latter by GÜNTHER (1861) and REGAN (1910).
Trichopodus microlepis (GÜNTHER, 1861), the species with most apomorphic features is nearly monochrome greenish silver-coloured apart from red portions on the ventral fi ns, the front part of the anal fi n and in the iris.It hardly ever, respectively temporarily, has the black longitudinal band, whereas the pattern of oblique bars does not occur at all.In contrast to the other species of the genus, its forehead often is saddle-shaped dented.Trichopodus leerii (BLEEKER, 1852) only has the black longitudinal band which can develop into a black spot on the caudal peduncle.Instead of the oblique bars, this species possesses a mosaic of light silvery spots which are dark-outlined and stretch over the whole body, the dorsal fi n, anal fi n and caudal fi n.In the males, each ray in the posterior part of the anal fi n terminating in a short silvery fi lament; breast, ventral fi ns and the front part of the anal fi n are of a strong brickred (all this being evidence indicating that SAUVAGE [1884] and partly also CANTOR [1850] dealt with this species).In Trichopodus trichopterus (PALLAS, 1770) the black longitudinal band is reduced to two round black spots in the middle of the body and on the caudal peduncle, the dark pattern of oblique bars is temporarily quite distinctive (except in some breeding forms).In Trichopodus pectoralis (REGAN, 1910) one fi nds both, the black longitudinal band as well as the pattern of oblique bars, more or less distinct according to the respective psychological condition (KÖHLER, 1905;MILLER & ROBISON, 1974).It is important for the history of the nomenclature of the Snake-Skin Gourami that GÜNTHER (1861) emphasized the existence of the black longitudinal band in his description of the variety ß cantoris, of which he thought as an occasionally existing connection between the two "trichopterus points".REGAN'S (1910) eye was more caught by the oblique bars of his specimens, which led him to emphasize these more.
Selected meristic data of the Trichopodus species after REGAN (1910):

Chronology of the Nomenclature of the Snake-Skin Gourami
(1) The fi rst scientifi c reference of the Snake-Skin Gourami can be traced to the English doctor and natural scientist THEODOR CANTOR (1850).In his comprehensive "Catalogue of the Malayan Fishes" CANTOR described a series of freshwater Gourami from Pinang, the Malayan Peninsula, the Moluccas, Madura and Java using the name Trichopodus trichopterus (PAL-LAS, 1770).He gave a list of synonyms all referring to the species fi rst scientifi cally valid described by PALLAS.Their signifi cant features as a whole do by no means agree with the appearance of a typical Trichopodus trichopterus.In fact, they also refer to (at least) two species that had not yet been described then: to the Pearl Gourami, described by BLEEKER (1852) as Trichopus leerii (CANTOR: "...all scales iridescent, edged with reddish brown forming an irregular network...dorsal spines and rays whitish grey, their membrane dark grey with numerous white rounded spots; caudal membrane and rays like the dorsal; anal spines carmine, their membrane and rays whitish, each ray terminating in a short silvery fi lament..."); and to the Snake-Skin Gourami, described by GÜNTHER in 1861 as Osphromenus trichopterus var.β cantoris (CANTOR: "...from the angle of the mouth through the iris, below the silvery lateral line to the root of the caudal a black zigzag band, widening at the termination into a large spot...").This zigzag band is a clear indication for the Snake-Skin Gourami because the Pearl Gourami's longitudinal band does not have a zigzag form.The meristic and other anatomical data of CANTOR ( 1850) are less helpful because of the interference of the respective features.
(2) ALBERT GÜNTHER (1861) was the fi rst ichthyologist who noticed the inconsistent character of Cantor's "Trichopodus trichopterus" collection.Consequently he grouped the specimens of his opinion polytypical species trichopterus into three varieties under the genus name Osphromenus: in Var.α koelreuteri (the actual Three-Spot Gourami, Trichopodus trichopterus, with the two prominent black spots in the middle of the sides of the body and the caudal peduncle; in Var.β cantoris (the Snake-Skin Gourami, later named Trichopodus pectoralis by REGAN [1910], with the temporarily appearing dark zigzag longitudinal band, etc. GÜNTHER's description based on a skin of an adult specimen from the CANTOR collection with the added remark "Cant.Catal.").According to the information provided by DR.JAMES MCLAINE (in lit.), this skin can unfortunately no longer be found in the ichthyological collection of the Natural History Museum of London.Finally, GÜNTHER described the Var.γ leerii -also based on specimen from the CANTOR collection (it meant the Pearl Gourami, Trichopodus leerii, in the meantime discovered by BLEEKER (1852) as new but not recognized as an independent species by GÜNTHER; also temporarily with a dark longitudinal band andespecially noted by GÜNTHER and also by CANTOR already -covered with many light silvery dark-outlined spots).Thus GÜNTHER ( 1861) was well able to differentiate among those species, which CANTOR (1850) then had "lumped together", even though GÜNTHER did not allow them species level.Furthermore GÜNTHER (1861) described the Moonlight Gourami as Osphromemus microlepis and -completely incomprehensible for us -a trichopterus population from Siam as Osphromemus siamensis.
(3) In the years 1881 and 1884 the Frenchman H.-E. SAUVAGE mentioned three Southeast Asian gouramis of which the fi rst two were already discovered as synonyms for the earlier described taxa: (4) In 1905 the Magdeburg schoolmaster and editor of the "Blätter für Aquarien-und Terrarienkunde" W. KÖHLER published an article titled "Osphromenus trichopterus (PALL.)var.cantoris GÜNTHER" in the above mentioned paper.However, to simplify matters KÖH-LER purposefully used the term Osphromenus cantoris in the text eleven times without being able to estimate the nomenclatorial consequences deducible from that.
(5) In 1910 the Englishman TATE REGAN published his revision "The Asiatic Fishes of the Family Anabantidae" which for a long time would remain trendsetting.In his list of synonyms on the Three-Spot Gourami, referred to as Trichopodus trichopterus by him, he already pointed out that only a part of the fi sh attributed to this taxon by CANTOR, really corresponds to what we emphasized at the beginning.Unfortunate-ly REGAN had nonetheless partly adopted the opinion of his predecessor GÜNTHER with regard to the alleged variability of the Three-Spot Gourami, Trichopodus trichopterus, when he assumes: "...sometimes a blackish lateral band through the spots from the eye to the caudal fi n." Thus he also included the GÜNTHER variety cantoris in the species characteristics of Trichopodus trichopterus, although it defi nitely represents an independent species.On the basis of six relatively large specimens from Siam and Singapore he described that subsequently as the new species Trichopodus pectoralis.REGAN's fi ndings show, the meristic data of T. trichopterus and T. pectoralis partly overlap, apart from the pectoral fi ns (head-long in T. trichopterus, longer than the head in T. pectoralis) and the upper lateral line series (40-52 in T. trichopterus, 55-63 in T. pectoralis), although this could not always be clearly verifi ed in larger series.There remains the pattern of coloura-tion of T. pectoralis, of which REGAN says: "Head and Body with oblique dark cross-bands; an interrupted lateral band from eye to caudal fi n, sometimes present on the head only...".Thus REGAN had well described and pictured the different patterns of colouration of the Snake-Skin Gourami, especially emphasizing the temporal dark pattern of oblique bars.This perhaps is the reason he did not recognize the identity of GÜNTHER's Osphromenus trichopterus var.cantoris and his Trichopodus pectoralis.
(6) For the time being, BORODIN (1930)   to the combination of characteristics described, the origin, the size of 162 mm (TL?) as well as on the basis of the pictures I received courtesy of DAMARIS ROD-RIGUEZ, it is doubtlessly a Snake-Skin Gourami.The oblique bars mentioned by BORODIN, which had faded in 1939 already, can not be recognized anymore, the body is brownish grey without any pattern elements.
On the basis of the pictures 55 resp.56 scales could be ascertained in the upper lateral line series above the LI; nearly 49 -50 scales in the LI, P clearly longer than the head, 3.2 times contained in the SI, head 3.8 times in the SI, relatively short, the profi le of the forehead slightly convex, not at all sunken; biggest body height before D begins, 2.  (1952), MUNRO (1955), TAKI (1974), KOTTELAT (1989), TALWAR & JHINGRAN (1992), KOTTELAT et al. (1993), KOTTELAT et al. (1995) and many others.The name Osphromenus saigonensis was -as far as known to the author -no longer used as a valid nomen after its coinage by BORODIN (1930).

Nomenclatorial Interpretation of the Results
The fi rst name of nomenclatorial relevance for the Snake-Skin Gourami is Osphromenus trichopterus var.β cantoris assigned by GÜNTHER (1861).According to article 45.6.4. of the ICZN it is subspecifi c, because before 1961 it was assigned "var." after the binomen and because GÜNTHER did not explicitly indicate that he wanted to create an infra subspecifi c nomen, The name Trichopus cantoris allocated by SAUVAGE (1884) can not automatically be taken as an uprating of the status fi rst assigned by GÜNTHER (1861), because it can not conclusively be verifi ed that it refers to the same species.It is much more likely that Trichopus cantoris SAUVAGE, 1884 constitutes a junior synonym to Trichopus leerii BLEEKER, 1852.At the same time Trichopus cantoris SAUVAGE, 1884 is -according to article 57.3.1.ICZN -a younger secondary homonym for Osphromenus trichopterus var.β cantoris, since today both taxa are in the same genus Trichopodus.In both cases the name is not available.
In his article KÖHLER (1905) used the name Osphromenus cantoris for the Snake-Skin Gourami eleven times.Furthermore he gave a description of the appearance (especially of the ability to change the colour) and behaviour, he named parts of its artifi cially extended area, and he published two distinct photographs of the fi sh.He has thus met the formal requirements for uprating the formerly subspecies Snake-Skin Gourami to species level.
If applying the principle of priority (Article 23 ICZN) and using today's common genus name Trichopodus for the Southeast Asian gourami, the Snake-Skin Gourami would have to be named Trichopodus cantoris (GÜNTHER, 1861).The species is dedicated to the ichthyologist CANTOR.The species name cantoris is the genitive of this noun and thus the original spelling remains (MAHNERT in lit., DUBOIS (2007), Ar- To keep the system stable and in consideration of the economic signifi cance of the Snake-Skin Gourami it would surely be reasonable retain the previous name pectoralis for that species.However, the attempt to protect it through a request to the ICZN, would probably fail in view of the evidences and a majority among the experts who in similar cases defi nitely favoured the principle of priority.Unless this attempt would be undertaken by exactly that majority!
tikel 31.1.1ICZN).The name Trichopodus pectoralis REGAN, 1910 is doubtlessly a junior synonym for Trichopodus cantoris (GÜNTHER, 1861), as was fi rst indicated by RIEHL & BAENSCH (1983/84)!According to the principle of priority the name pectoralis is invalid.Regardless of that, the status of the six syntypes of Trichopodus pectoralis (BMNH 1862.11.1.232-233)examined by Regan, deposited and partly available in the Natural History Museum of London, remains unaffected by the change of the name.They become even more important since the holotype for the nomen Osphromenus trichopterus var.β cantoris GÜNTHER, 1861 is untraceable in the same museum.
BLANC (1963)1)had added the reference "Cant.Catal.".The question is whether (a) SAUVAGE's material was identical with what GÜNTHER defi ned as the cantoris variety; or (b) being not conspecifi c with GÜNTHER's var.β cantoris?The case can no longer be examined with the object itself: probably SAUVAGE had -based on the measurements -only one specimen, but he did not denote it as a type because he referred to another author.Consequently one does not fi nd any reference to this taxon byBLANC (1963)in the type catalogue of the Anabantoids ROBERTS (1989))pinnis SAUVAGE 1881 according toREGAN (1910)is Trichopodus microlepis(GÜNTHER, 1861), and Trichopus siamensis SAUVAGE 1881 according toROBERTS (1989)is Trichopodus trichopterus (PALLAS, 1770).A specimen described by SAUVAGE as Trichopus cantoris CANT. in 1884 has not yet been identifi ed.It originates from the lower reaches of the river Pérak in the district of Kinta on the southern Malayan Peninsula.CANTOR never named a taxon cantoris after himself and thus SAUVAGE -possible unintentionally -became the author of the name.Perhaps SAUVAGE's abbreviation "Cant:" referred to GÜNTHER's variety β cantoris of Osphromenus trichopterus.and Snakeheads of the Museum National d´Histoire Naturelle in Paris.Eventually, PATRICE PRUVOST, the collection manager of the Paris fi sh collection confi rmed for the author that the collection documentation does not include material collected by SAUVAGE under the name Trichopus cantoris.In the collection documentation, let it be understood, he did not men-Trichopodus pectoralis and Trichopodus leerii.Both temporarily have a black longitudinal band SAUVAGE refers to.In the case of Trichopodus pectoralis it is often interrupted, respectively forms a zigzag, in the case of Trichopodus leerii it is rather a narrow longitudinal stripe.Some particulars speak for Trichopodus leerii and against the other species: thus the low TL of 65 mm, more especially several silver, dark-outlined spots or dots on the body and the base of the anal fi n, which are typical of the Pearl Gourami, but do not occur in the Snake-Skin Gourami, as well as the fact that the respective species is known from Sumatra.From there BLEEKER (1852) had described his Trichopodus leerii.This could support the idea that Trichopus cantoris SAUVAGE, 1884, is a junior synonym of Trichopodus leerii (BLEEKER, 1852), and thus can not be identical with the GÜNTHER variety cantoris.Perhaps someone someday discovers the Gourami mentioned by SAUVAGE in the Parisian collection and can correct this statement if necessary.

Usage of the species names cantoris and pectoralis in the late 19th and 20th Century
7 times contained in the SI; 8 D-spines, as BORODIN indicates, which is slightly high for T. pectoralis, but absolutely possible.By no means the fi sh is a Trichopodus trichopterus, as VI-ERKE (1978) assumed without knowing the holotype, because BORODIN emphasized repeatedly: "...absence of black spots".Apart from these two species there are no other autochthon gouramis near Saigon.