Revision of the South American freshwater fish genus Laemolyta Cope , 1872 ( Ostariophysi : Characiformes : Anostomidae )

The anostomid genus Laemolyta Cope, 1872, is redefined.Various morphological, especially osteological characters in addition to the commonly utilized features of dentition proved useful for its characterization. A taxonomic revision of all species was made using meristics, morphometrics and color pattern. Five species are recognized: Laemolyta fernandezi Myers, 1950, from the río Orinoco (Venezuela) and the sub-basins Tocantins/Araguaia and Xingu, L. orinocensis (Steindachner, 1879), restricted to the río Orinoco, L. garmani (Borodin, 1931) and L. proxima (Garman, 1890), from the Amazon basin with the latter also occurring in the Essequibo River (Guiana), and L. taeniata (Kner, 1859), from the Amazon and Orinoco basins. Laemolyta garmani macra is considered a synonym of L. garmani, L. petiti a synonym of L. fernandezi, and L. nitens and L. varia synonyms of L. proxima. Lectotypes are designated herein for L. orinocencis and L. taeniata.


Introduction
The family Anostomidae has been recognized as a monophyletic assemblage by Vari (1983) primarily based on synapomorphies related to modifications of the gill apparatus and suspensorium, but the monophyletic nature of the included genera has not been demonstrated.A better morphological characterization of the species of Laemolyta is critical since the species in the genus have never been adequately described.The original descriptions are imprecise and based on a limited number of sometimes young specimens.
In this work the species of the genus Laemolyta are systematically reviewed and their osteological characters are analyzed and compared to those of other anostomid genera in order to find morphological characters other than dentition and mouth position to better characterize the genus.A phylogenetic analysis of characters was not attempted because this would require the examination of all anostomid genera and outgroups, an effort that extends far beyond the scope of this paper.The relationships of the species and a phyloge-netic definition of Laemolyta will be the subject of a subsequent paper.

Material and Methods
Counts and measurements were taken on the left side of the specimens, except when the structure being measured or counted was recognizably abnormal or damaged on that side.All measurements were taken point to point with calipers and data recorded to tenths of a millimeter.Head length and other parts of body are presented as percents of standard length and subunits of the head are presented as percents of head length.Greatest body depth was taken at dorsal-fin origin.Counts and measurements are presented in tables and were taken according to Fink & Weitzman (1974), except for measurements involving the anterior delimiter of the specimen.In this study, they were always taken from the anterior end of the closed lower jaw rather than the premaxillary because of the upturned mouth in Laemolyta.Lateral-line scale count includes all perforated scales, even those posterior to the hypural joint.Predorsal scales were counted only in species where the scale arrangement was regular.Total number of vertebrae includes the four vertebrae of the Weberian apparatus and the fused PU1 + U1 of the caudal fin which was considered as a single element.Vertebral counts were taken from radiographs of specimens and from specimens cleared and stained following the method of Taylor and Van Dyke (1985).In the meristic data, asterisks indicate values for holotype.In counts of fin rays, unbranched rays are indicated by lower case Roman numerals, and branched rays by Arabic numerals.Number of gill-rakers includes those of the dorsal and ventral limb of first gill arch, but with such counts taken only from a subset of specimens of each species because of the damage to some head bones that is inevitable in making this count accurately.
Teeth are described for each premaxilla and dentary, from the medial to lateral regions of the mouth.Thus, the sequence is first tooth (premaxillary) or symphyseal (dentary), second, third, and fourth teeth.
The material examined is listed by country and, then within each country data for each lot includes state, province or city, department or district, more precise locality data, institutional abbreviation, catalog number, number of specimens measured and counted in each lot, followed by the range of standard lengths.For type material collecting date and collector are added.Sex of specimens was determined by a macroscopic examination of the gonads.No sexual dimorphism was found in meristic and morphometric data and color pattern.
Diagnosis.Laemolyta is distinguished from all other genera of the Anostomidae in having the symphyseal tooth of the dentary approximately equal in size to the second tooth with both teeth spatulated, very compressed, and with the upper edge straight (vs.symphyseal tooth sometimes much larger than the second tooth or slightly larger than the second tooth and sometimes cusped as the remaining dentary teeth, thick or recurved) (Fig. 1).Adult specimens have a supra-terminal mouth, a condition that is shared only with Anostomoides among anostomids.In spite of Winterbottom's interpretation of a supraterminal mouth in the Anostominae, in our interpretation all members of this subfamily have a superior mouth.Species of Anostomoides, however, have the symphyseal tooth longer than the second tooth on the dentary.Laemolyta further differs from all other examined genera of the Anostomidae in possessing a protuberance on the anteroventral region of the ectopterygoid (Fig. 2), in the elongated mesopterygoid (Fig. 2) with a small ventrally-directed process on its lateral surface and its ventral margin poorly developed, and in the relatively wide premaxilla with a fold on its median ventral margin that is prolonged as a rounded process medially, the upper premaxillary margin slightly recurved (Fig. 3).
Cope ( 1872), described the genus Laemolyta to include Schizodon taeniatus Kner, 1859, that according to him differed from the remaining species of Schizodon known at the time in which the branchiostegal membranes were entirely free from the isthmus, the dentary teeth truncate with their   upper edge entire and premaxillary teeth crenate.The typespecies of the genus, Laemolyta taeniata, had been first described in Schizodon by Kner (1859) but was later included in Anostomus by Günther (1864).
In his study of the genus Anostomus, Garman (1890) recognized three groups or sub-genera based primarily on the snout shape, position of the mouth opening, and form of dentition: Anostomus, Schizodontopsis, and Schizodon.Schizodontopsis (=Laemolyta) was primarily characterized by having its mouth aligned obliquely forward and upturned and by the lower jaw bearing large truncate teeth with their upper edges straight.He described two species that he assigned to this group: Anostomus (Schizodontopsis) proximus and Anostomus (Schizodontopsis) varius and considered Anostomus (Schizodontopsis) nitens to be a "variation" of Anostomus (Schizodontopsis) varius.
Eigenmann & Eigenmann (1891) listed in Laemolyta the species that Garman (1890) considered to belong in the subgenus Schizodontopsis and others including L. taeniata, L. proximus, L. varius, L. varius nitens, and L. orinocensis. Subsequently, however, Eigenmann (1910) included only the typespecies of the genus Schizodon taeniatus Kner, 1859 in Laemolyta; and allocated the other species previously included in this genus to Anostomus.In the same paper Eigenmann considered the sub-genus Schizodontopsis Garman, 1890 to be synonymous with Laemolyta. Myers (1950) retained Schizodontopsis Garman (1890) in the synonymy of Laemolyta and based on Garman (1890), Borodin (1931), andSteindachner (1879) prepared a list of eight species of Laemolyta: L. fasciata Pearson, 1924, L. garmani (Borodin, 1931), L. nitens (Garman, 1890), L. proxima (Garman, 1890), L. taeniata (Kner, 1859), and L. varia (Garman, 1890) all from the Amazon basin, L. orinocensis (Steindachner, 1879) from the río Orinoco basin and L. laticeps (Eigenmann, 1912) from Guiana.Additionally, he described Laemolyta fernandezi, based on a 72.5 mm SL juvenile collected in the upper río Orinoco.Myers noted that in Laemolyta the branchiostegal membranes are in actuality closely connected to the isthmus, contrary to what was reported by Cope (1872).Myers also distinguished Laemolyta from Schizodon, mainly by dentition, with Laemolyta having the dentary teeth truncate with their upper edges straight whereas Schizodon has cuspidate teeth on both jaws.Fowler (1950) included in Laemolyta all species recognized by Myers (1950), with the exception of L. laticeps and L. orinocensis, and added L. borellii (Boulenger, 1900) of the rio Paraguay basin to the genus.The form of dentition described by Cope (1872) and Myers (1950) for Laemolyta was observed in all of the species listed above, except for L. fasciata, L. borellii and L. laticeps.
Laemolyta fasciata was described by Pearson (1924) based on two immature specimens (20 mm SL) from Bolivia with the taxonomic position of the species becoming very confused when Ahl (1937) shifted the species to Anostomus.Ahl reported that Laemolyta fasciata possessed the characteristics of Anostomus (Schizodontopsis) Garman, 1890 as a con-sequence of the fact that he did not recognize Garman's subgenera.Since Anostomus fasciatus was already preoccupied in Anostomus (Spix & Agassiz, 1829), Ahl renamed the species described by Pearson as Anostomus pearsoni.Myers (1950) did not recognize Ahl's combination as valid, because according that author the name fasciatus was not occupied in Laemolyta.
The type material of Laemolyta fasciata was not examined during this study but all measurements and counts on that specimen were taken by Anostomoides Mônica Toledo-Piza (DZ-IBUSP) and the digitalized picture and radiograph of the specimen were provided by David Catania (CAS).This information was insufficient to permit a definitive identification because the specimen is a juvenile (15.7 mm SL) in very poor condition.According to H. A. Britski (MZUSP) (pers.commun.) the general characteristics of the specimen, in particular its color pattern suggest that this specimen might belong to a species of Leporinus.Böhlke (1958) discussed this question and if this is true, the valid combination would become Leporinus pearsoni (Ahl, 1937), which would be a homonym of Leporinus pearsoni Fowler, 1940.The problem about the correct name for the species in question can only be resolved after a more detailed study.
Laemolyta borellii (Boulenger, 1900) included by Fowler (1950) in that genus, has teeth with five or six cusps of about the same size on both jaws, a characteristic of the genus Schizodon according to Garavello & Britski (1990) and Garavello (1994).
Laemolyta laticeps cited by Borodin (1931) and Myers (1950) belongs to Anostomoides Pellegrin, 1908, as suggested by Géry (1974: 149, in footnote).Géry examined the type and mentioned that it was very similar to Anostomoides atrianalis.One of us (KCM) examined the holotype of Schizodontopsis laticeps Eigenmann, 1912 and confirmed that it is not a species of Laemolyta.The lower jaw teeth of the specimen have rounded upper margins and the symphyseal tooth is considerably more developed than the second tooth, both of which are characteristic of Anostomoides.
According to Géry (1977) the genus Anostomus is composed of two groups that could be considered distinct genera or subgenera: Anostomus (Laemolyta) and Anostomus (Anostomus).The Laemolyta group is distinct in having the mouth opening slightly less upturned (not quite aligned with the dorsal surface of the snout), and in having multicuspidate teeth on the premaxilla and truncate teeth on the dentary.Géry also emphasized the great similarity between the dentition of young specimens of Laemolyta and Anostomus and the manner in which they orient in the water column with the head downturned.
In young specimens of Laemolyta the premaxillary teeth are very similar to those of adult specimens of Anostomus, and it was probably for this reason that Géry considered Laemolyta to be a sub-genus of Anostomus.The supra-terminal mouth is not useful to distinguish young specimens of Laemolyta from some other anostomids.According to H. A. Britski (MZUSP) (pers.commun.)young specimens of some Leporinus species have such a form of the mouth.According to Winterbottom's (1980) phylogenetic analysis of the members of the subfamily Anostominae the inclusion of Laemolyta in Anostomus is not justifiable since Anostomus is a monophyletic group within the subfamily, with Laemolyta proposed as the possible sister-group to the Anostominae.
Description.Morphometric and meristic data presented in Table 1.Body moderately large (SL = 63.6-212.9mm).Dorsal profile of body straight or slightly concave from snout tip to Color in alcohol.Adult specimens usually with dark, but sometimes very inconspicuous, longitudinal stripe, covering 1.5 to almost 3 longitudinal scale rows.Stripe extends from posterior region of opercle to caudal peduncle where it usually ends in form of darker triangular or oval spot.Upper region of body usually dark, with dark coloration extending ventrally to below midlateral dark stripe.Ventral region of body unpigmented from ventral limit of dark midlateral stripe on body to midventral one.Body with 4 transverse dark but sometimes inconspicuous bars or blotches, sometimes inconspicuous.First series situated between opercle and dorsal fin, but closer to opercle; second below dorsal fin, and usually more conspicuous; third located between pelvic and anal fins and last, anterior to adipose fin.First and second series usually more conspicuous and last series usually diffuse (Fig. 6).
Young specimens (Fig. 7) with midlateral longitudinal stripe indistinct, extending from anterior part of snout to rear of caudal peduncle and usually finishing in form of triangular spot.Four transverse bars or blotches present in adults absent.Upper part of body sometimes with approximately 14 narrow transverse dark bars extending ventrally to near lateral line.Adult specimens sometimes retain faded transverse bars.Adult specimens from río Orinoco basin (Fig. 8) with bars faded and midlateral longitudinal stripe on body very conspicuous.Dorsal, pectoral, and ventral fins hyaline with scattered dark chromatophores.Dark chromatophores also scattered along fin margin and over median rays of caudal fin.Adipose fin usually dark.
Some specimens with 4 transverse dark bars quite evident, with dark blotch present on caudal peduncle and dark midlateral stripe on body faded (Fig. 9).
Distribution: Río Orinoco and Amazon basin (rios Xingu, Tocantins and Araguaia) (Fig. 10).Myers, 1950, was described based on a young specimen (73.1 mm SL), that was considered to be a new species in having a compressed body and more longitudinal scale rows (8/7) than L. garmani, L. proxima and L. taeniata (4-6/4-7).According to Myers the specimen was collected in the upper río Orinoco, but the precise locality was not provided.The description of L. petiti Géry, 1964, is also based on a young specimen (76.4 mm SL), collected in the rio Araguaia.When Géry (1964) described his new species he recognized its similarity with L. fernandezi, but noted that a comparison between the two nominal forms was impossible because of the succinct description and lack of illustration of L. fernandezi.Later on Géry (1974) redescribed L. fernandezi and compared the data he obtained with those of L. petiti and found that L. fernandezi has fewer longitudinal scale rows (7-7 ½ / 6-6 ½), more lateral line scales (55) and a deeper caudal-peduncle than did L. petiti.

Remarks. Laemolyta fernandezi
The comparison of meristic and morphometric data among specimens of L. fernandezi from the Orinoco, Tocantins/ Araguaia and rio Xingu basins revealed no significant differences between these populations.The number of vertebrae in a sample from the río Orinoco is 41 (10 specimens) and 42 (1 specimen -holotype) and from the rios Tocantins/Araguaia 39 (15 specimens), 40 (3 specimens) and 41 (1 specimen).The number of vertebrae in 4 specimens from the rio Xingu basin is 39 (1 specimen), 40 (2 specimens) and 41 (1 specimen).The comparison of body depth, that seemed to be visually different among the populations originating from the three basins mentioned, also did not show any significant difference (Fig. 11).
In spite of the slight difference in number of vertebrae we prefer to recognize just one widely distributed species represented by allopatric populations, especially because no other meristic or morphometric differences were found to distinguish the populations.
Laemolyta fernandezi was recorded from the rios Tocantins and Araguaia by Santos & Jégu (1989) and is the most common anostomid species in the lower rio Tocantins (Santos et al., 1984).It is to date the only species of Laemolyta found in the central portion of the rio Araguaia basin.In the lower rio Araguaia it occurs in sympatry with L. garmani.
Laemolyta fernandezi reaches 25 cm SL and eats small invertebrates (Santos & Jegú, 1989), vegetable material and periphiton (Santos et al., 1984).Sexual maturity seems to occur in specimens at approximately 130 mm SL judging from direct observation of the gonads.Reproduction seems to occur during the flood season (Santos et al., 1984).Remarks.Borodin (1931) described Laemolyta garmani and considered it to be very similar to L. taeniata, but distinguished his species by having larger and fewer scales and by the presence of a darker longitudinal stripe, fins, and scales.He also mentioned that the longitudinal dark stripe extends onto the middle caudal-fin rays and is clearly visible in preserved specimens even after 64 years.These characteristics and the presence of dark chromatophores on the fins and scales although less evident, are also present in L. taeniata, and, thus, are not useful to distinguish the two species.The two nominal species, however, can be separated primarily on the basis of scale size and consequently scale counts.
Laemolyta garmani was originally considered by Géry (1964) to be a subspecies of L. taeniata but later that author (1974) raised it to the species level and at the same time described a new subspecies, Laemolyta garmani macra based on a small sample of young specimens that had a shallower body.The holotype of this subspecies was not examined but M. Toledo-Piza (DZ-IBUSP) made counts and measurements on the specimen and a photo (Fig. 14) was provided by C. Weber (MNHG).
Figure 15 demonstrates that the body depth of the holotype of Laemolyta garmani macra does not differ significantly when compared to all the examined specimens of L. garmani.Although body depth varies to a certain extent among the specimens, there is no statistical difference to justify recognition of a subspecies.Distribution.Amazon basin (Fig. 13).In his original description of Laemolyta taeniata, Kner (1859) mentioned that two specimens from Barra do rio Negro had thinner bodies, and some other divergent unspecified characters relative to the other examined specimens.One of these specimens, considered to be the syntype of L. taeniata (NMW 56988:2) was examined and proved to be identical to all the other specimens of L. garmani studied.This species was collected in sympatry with Laemolyta proxima and L. taeniata in Peru.
Color in alcohol.Large adult specimens with dorsal part of body usually dark, extending ventrally to 1 or 2 longitudinal scale rows above lateral line.Conspicuous dark midlateral longitudinal stripe covering 1.5 to 2 longitudinal scale rows extending from tip of snout to rear of caudal peduncle.Ventral part of body lighter from lower limit of longitudinal dark stripe to midventral line.Some specimens with very inconspicuous dark blotch present posterior to opercular margin and another one under dorsal-fin base.Dorsal, pectoral, and pelvic fins hyaline.Scattered dark chromatophores on anal and caudal fins and along dorsal and ventral edges and median rays of caudal fin.
Young specimens (Fig. 18) with dark, midlateral, longitudinal stripe more conspicuous and narrower at opercle and caudal peduncle.Some specimens with narrow, transverse, dark bars on dorsal part of body and with anal fin dark.
Distribution.Rio Orinoco basin (Fig. 13).1879) to be very similar to L. taeniata (Kner), but to differ from it in scales counts.In the original description of L. orinocensis, Steindachner (1879) referred to the presence of 3.5 scales between the lateral line and the ventral margin of the body, but Garman (1890) observed that this count is probably incorrect since six scales can be easily counted below the lateral line in the figure of the species in Steindachner's paper.
Laemolyta orinocensis and L. taeniata are very similar in having the dark, longitudinal midlateral stripe on body more evident in young specimens; however, in L. orinocensis the dorsal and ventral margins of the stripe are not absolutely straight.
In the syntype specimens (NMW 62820:1 and 62595:1) the longitudinal dark stripe is no longer apparent.In these specimens, however, a vestigial longitudinal dark stripe is still visible, but most of the stripe is silvery due to the presence of guanine on the scales.The silvery color on the centers of the scales above the lateral line form a reticulate pattern.In the original description, Steindachner (1879) designated two syntypes (NMW 62820:1 and 62595:1).The one specimen in NMW 62820:1 is herein designated as the lectotype.
Remarks.Laemolyta proxima was described by Garman (1890) as Anostomus (Schizodontopsis) proximus together with A. (Schizodontopsis) varius and A. (Schizodontopsis) nitens.These three species were recognized as distinct by Borodin (1931) who, nonetheless, considered the first A. proximus and A. varius to be very similar on the basis of their body shape and color pattern, and with A. nitens very similar to A. varius but having the body more elongate and with an overall lighter coloration.Géry (1974) considered these three species as part of a complex and divided that into two units: the proxima-varia group, which was characterized by the presence of a dark, midlateral longitudinal stripe and nitens, which had dark spots over an inconspicuous dark longitudinal stripe, a larger orbital diameter, and the body more elongate.In the same paper, Géry cast doubts on the recognition of two different species within the first group.The same author (1977) included Laemolyta as a subgenus of Anostomus, considered L. nitens to be a valid species and provided an identification key which suggested that Anostomus (Laemolyta) proximus is a synonym of Anostomus (Laemolyta) varius.
According to Garman (1890) Laemolyta proxima would be closely related to L. taeniata from which it could be distinguished in having smaller and consequently more numerous scales.Indeed some of the examined specimens of L. proxima species have a color pattern identical to that of L. taeniata, but the differences in scale counts make their identification possible.
Notwithstanding the differences in color pattern and number of scales between Laemolyta varia and L. proxima re-   ported by Garman (1890), an examination of the syntypes of L. varia indicated that these characters fall within the range of variation found in L. proxima.The two species are, thus, considered synonymous, with L. proxima utilized because it appears first in Garman's publication.
The color pattern described by Garman (1890) to characterize Laemolyta nitens dark longitudinal stripe, dark spots along the lateral line, and the presence in some specimens of narrow, dark, vertical bars and faint dark spot on the caudal peduncle also occurs in some specimens of L. proxima (Fig. 22).This, however, falls within the color pattern variation in this species.The comparison of meristic and morphometric data between the two nominal species failed to reveal any significant difference, and L. nitens is, thus, placed as a junior synonym of L. proxima.Lasso (1992) reported Laemolyta proxima from the río Suapare (tributary of the río Orinoco, Venezuela).Two out of the 3 specimens in the lot that served as the basis of that record (MHNLS 5512) were examined.One specimen proved to be L. taeniata and the other L. orinocensis.The third specimen, probably the one illustrated in the paper, also looks very much like L. taeniata, because of the clearly visible dark longitudinal stripe, and the whitish areas on the scales above the lateral line that form forming light stripes and which are characteristic of this species.
Data from Santos & Jégu (1996) indicate that Laemolyta proxima is found in floodplain lakes in large shoals during certain parts of the year, sometimes occurring together with L. garmani.Those authors note that L. proxima is fished commercially.The species is omnivorous, but feeds mainly on insect larvae and vegetable material (Santos, 1982).It is widely distributed through the Amazon Basin including Peru.
Diagnosis.Laemolyta taeniata is the only species in the genus having 5 longitudinal scale rows from the lateral line to the origin of the dorsal fin (vs. 4 and 6-8 in the remaining species).Its color pattern is similar to that of L. garmani and L. orinocensis, but it can be distinguished from L. garmani by the presence of a dark, midlateral longitudinal stripe and the more conspicuous, narrow light stripes above the lateral line and from L. orinocensis in having fewer scales along the lateral line (42-46 vs. 47-52).Description.Meristic and morphometric data presented in Table 6.Body moderately large (SL = 54.9-288.8mm).Dorsal profile of body almost straight and posterodorsally inclined from anterior tip of snout to end of occipital spine, slightly convex from that point to origin of dorsal fin, straight and posteroventrally inclined along dorsal-fin base and straight or slightly convex from base of posterior-most end of dorsalfin ray to origin of adipose fin.Gill rakers on dorsal limb of first gill arch 9-12 and 11-15 on ventral limb.
Color in alcohol.Dorsal region of body generally dark, dark coloration extending ventrally to 1 longitudinal scale row above lateral line.Conspicuous, dark, longitudinal stripe extending from tip of snout to caudal peduncle.Stripe narrower on snout and caudal peduncle.Scales above lateral line with central posterior areas whitish and forming narrow light stripes or linear light dots.
Remarks.In the original description three specimens were designated syntypes of Schizodon taeniatus.Specimen NMW 81379:2 is herein designated the lectotype.
Laemolyta taeniata reaches the largest size among its congeners and it is the second most widely distributed member of the genus in the Amazon basin, being particularly common in the central Amazon and upper rio Madeira basin.It has also been recorded in the río Orinoco basin where it has been confused with L. orinocensis.
Data from Santos & Jégu (1996) indicate that L. taeniata is usually found in lakes and main river channels where it forms large migratory shoals in company of species of other genera such as Leporinus fasciatus and L. agassizi.
Although not included in the list of examined material, specimens from São Domingos, (MZUSP 3850, specimen damaged) and rio Tocantins, below the dam of Serra da Mesa, Goiás (SM 19-221 da Universidade Federal do Rio de Janeiro, specimen examined, but not included in analysis) were examined.

Fig. 5 .
Fig. 5. Scatter plot of body depth on standard length for all species of Laemolyta.

Fig. 11 .
Fig. 11.Scatter plot of body depth on standard length for populations of Laemolyta fernandezi.

Fig. 15 .
Fig. 15.Scatter plot of body depth on standard length for specimens and holotypes Laemolyta garmani.

Table 2 .
Frequency distribution of lateral line scales in Laemolyta.