Redescription of Leporinus jamesi ( Characiformes : Anostomidae ) , a poorly known species of Leporinus from the lowlands of the central Amazon , Brazil

The poorly known Leporinus jamesi is redescribed. The species was originally described based on a single specimen collected in the rio Solimões at Manacapuru, in the central Amazon, Brazil. The holotype went missing before the species description was finished and published, and remained lost for more than a hundred years. Leporinus jamesi is distinguished from its congeners by having pectoral and pelvic fins dark, 42 to 45 scales on the lateral line, 16 scale series around the caudal peduncle, a body with two conspicuous dark midlateral blotches (the blotch on the caudal peduncle absent or inconspicuous), and four teeth on the premaxilla and dentary, including a bicuspid symphyseal tooth on the premaxilla. A principal component analysis on morphometric traits between combined samples of L. jamesi and L. amazonicus was performed showing significant morphometric differences between these species. In addition, inaccuracies in Borodin’s descriptions of various species of the genus Leporinus are discussed.


Introduction
Despite its great diversity, relatively few species of Leporinus Agassiz, 1829 have been reported to occur in the lowlands of the central Amazon.Beginning in 1865, some species were described on the basis of specimens collected during the Thayer Expedition in central Amazon.Steindachner (1876) described Leporinus agassizi, L. muelleri (= L. moralesi), L. nattereri, and L. trifasciatus, and Borodin (1929) described Leporinus fasciatus altipinnis, L. cylindriformis and L. jamesi.From the mid-twenty century to the present, only seven species of Leporinus were described from the central Amazon, including: L. brunneus Myers, 1950; L. klausewitzi Géry, 1960;L. pitingai Santos & Jégu, 1996; L. uatumensis Santos & Jégu, 1996; L. falcipinnis Mahnert, Géry & Muller, 1997;L. vanzoi Britski & Garavello, 2005; and L. amazonicus Santos & Zuanon, 2008.Leporinus cylindriformis Borodin, 1929 andL. jamesi Borodin, 1929 were both described based on single type specimens, which were illustrated in various positions (Borodin, 1929: plates 9 and 10).Until now, both species have been known only from their holotypes (Sidlauskas et al., 2011).Assignation of additional specimens to L. jamesi has been hindered because its holotype remained missing for approximately one hundred years.It was lost sometime after the drawings were prepared and before the description was published (Borodin, 1929), and rediscovered in Harvard's Museum of Comparative Zoology less than three decades ago (Horacio Higuchi & Karsten Hartel, pers. comm.).
The aim of this paper is to redescribe Leporinus jamesi and diagnose it from congeners.In order to do that, we examined the holotype and additional material, especially specimens collected during the "Expedição Permanente da Amazônia", during which the Amazon river channel and several tributaries were sampled from 1965 to 1970, and specimens recently collected by researchers of the "Instituto Nacional de Pesquisas da Amazônia".

Material and Methods
Measurements and counts were obtained according to Garavello & Santos (2009), with the addition of counts of middorsal scales taken from the tip of the supraoccipital to the dorsal-fin origin; middorsal scales from the dorsal-fin terminus and the caudal-fin origin; middorsal scales between the adipose-fin terminus and the caudal-fin origin; scales along the ventral midline between the isthmus and the pelvic fin; scales between the pelvic fin and anus; scales between the anus and anal-fin origin; and scales from the anal-fin terminus to caudal-fin origin.Examined specimens belong to the following institutions: Academy of Natural Sciences of Drexel University, Philadelphia (ANSP); Natural History Museum, London (BMNH); California Academy of Sciences, San Francisco (CAS); Museum of Comparative Zoology, Cambridge (MCZ); Instituto de Ciencias Naturales, Museo de Historia Natural, Facultad de Ciencias, Universidad Nacional de Colombia, Bogotá (ICNMHN); Museo de Historia Natural "Javier Prado" de la Universidad Nacional Mayor de San Marcos, Lima (MUSM); Museu de Zoologia da Universidade de São Paulo, São Paulo (MZUSP); Instituto Nacional de Pesquisas da Amazônia, Manaus (INPA); and Laboratório de Ictiologia Sistemática do Departamento de Ecologia e Biologia Evolutiva da Universidade Federal de São Carlos, São Carlos (LISDEBE).The principal component analysis follows the procedure of Bookstein et al. (1985) and was performed in Statistics Analysis System software (SAS Institute Inc.).

Diagnosis
Dorsal-fin rays posteriorly reaching five or six dorsal scales from dorsal-fin terminus, when adpressed.Pectoral-fin tip reaching up to five or six lateral body scales when adpressed; pelvic fin reaching fourth abdominal scale anterior to anus.Anal fin short, tip of longest rays separate from caudal-fin origin by two or three scales, when adpressed; margin of anal fin truncate or convex, when extended.Dorsal-fin rays i-ii,9 (holotype with i,9); pectoral-fin rays i,15; pelvic-fin rays i,8; anal-fin rays ii,7; caudal fin rays i,17,i.Caudal fin forked; upper lobe longer than lower.
Color in alcohol.Color pattern on body and fins somewhat variable; specimens usually with two small and elongate darkbrown blotches on lateral line, first spanning two scale rows vertically and up to five scales longitudinally, and positioned ventral to middle of dorsal-fin base; second spanning one scale row vertically and up to six scales longitudinally, positioned ventral to adipose fin.An inconspicuous and smaller third blotch on caudal peduncle rarely present.In juveniles (up to 60 mm SL), small blotches sometimes appear as interrupted dark-brown stripes from vertical through dorsal-fin origin to caudal peduncle.Dorsum in juvenile specimens with 12 or 13 short transversal dark-brown bars sometimes meeting small dark-brown blotches laterally; three of these bars situated anterior to dorsal-fin origin, two ventral to middle of dorsalfin base, five between dorsal and adipose fins, and two or three on caudal peduncle.Head dark brown on interorbital space and supraoccipital bone; snout, anterior region of eyes, and infraorbital and subopercular regions light brown; isthmus and abdomen yellowish.Fin membranes with brown chromatophores concentrated on base or covering base of pectoral, pelvic and anal fins; dorsal fin with dark blotches occasionally; adipose fin dark brown.Caudal fin with a column of scales forming a dark arrow-shaped stripe facing backwards at end of caudal peduncle; caudal-fin rays dark-brown, with hyaline interradial membranes.
Distribution.Leporinus jamesi is known from the rio Amazonas, rio Solimões, rio Madeira, río Ucayali and lower portion of their tributaries at the central Amazonian region in Brazil and Peru (Fig. 3).Scales relatively small; 42 to 45 (holotype with 43) perforated scales on lateral line; five series of scales between dorsal-fin origin and lateral line; five series of scales between lateral line and pelvic-fin origin; 16 series of scales around caudal peduncle; median dorsal line of trunk with 11 scales between supraoccipital and origin of dorsal fin, 12 or 13 between dorsal and adipose fins, and 8 or 9 between adipose fin and upper caudal-fin base; median ventral line with 18 to 21 scales between isthmus and origin of pelvic fins, 9 to 11 between pelvic fins and anal opening; 2, 3, or 4 scales between anus and anal fin, and 6 to 8 between anal fin and lower caudal-fin base.views, and also details of the head, snout, mouth and teeth.They are also generally precise in terms of body shape, proportions, color pattern, and especially number, shape and arrangement of teeth.However, his drawings are sometimes imprecise with respect to the number of scales on the lateral line, and also regarding the number of transverse series and circumpeduncular series of scales, especially if compared with the actual specimens on which Borodin based his descriptions.
In The coloration of the holotype of Leporinus jamesi also does not agree with the pattern represented in the plate 9, fig. 1, of Borodin (1929), especially with respect to the dark blotches on the fins.However, the lack of fin markings in the holotype might be attributed to fading after more than a century of storage in alcohol.Except for the blotches on the fins and the scale counts, the drawing of L. jamesi in plate 9 agrees with that of the holotype (MCZ 20439) in all other aspects.Roetter's illustration of L. jamesi also shows conspicuous dark blotches on the dorsal

Discussion
The holotype of Leporinus jamesi was missing by the time that Borodin prepared the description of the species (Borodin, 1929: 270) and, consequently, his description was based in part on the drawing prepared by Paulus Roetter, and in part on the notes made by Garman.However, there are certain inconsistencies between the drawings, Borodin's description and the holotype.Inaccuracies are also observed in the descriptions of other species and subspecies described in Borodin (1929).Borodin's (1929) descriptions were based on specimens collected in Brazil by the Thayer Expedition, conducted by Louis Agassiz during 1865 and 1866, and deposited in the Museum of Comparative Zoology at Harvard University (MCZ).Borodin was later invited by Thomas Barbour, director of the museum at that time, to study the collection of Leporinus in connection with drawings of the species and notes made by Garman found in the MCZ after the decease of the latter in 1927 (Borodin, 1929: 270).Borodin stated that Garman's notes are "fragmentary rough drafts", but he credited the authorship of L. jamesi "completely to Mr. Garman, because the specimen, which served for the drawing and the fairly good description, has not yet been found in the museum's collections".Myers (1950: 193), in his study of the anostomid genera, also recognized Borodin's (1929) inaccuracies, stating that "Borodin has reviewed the species in the Harvard Museum; this is an important paper, not for its author's work, which is bad, but for the excellent plates made under Louis Agassiz's and Garman's direction, and for its rather complete coverage of Brazilian Amazon forms.However, even the Agassiz plates do not sufficiently and accurately illustrate the dentition".Borodin (1929) did not identify his illustrator, but the drawings of the species of Leporinus he described were made by Paulus Roetter, an artist working under Agassiz's supervision in 1867, just after the decease of Jaques Burkhardt, who was Agassiz's long time illustrator and field companion during the Thayer Expedition (Britski, pers. comm.).Roetter's drawings are accurate and depict several different positions of the fishes, including unusual frontal and ventral and pectoral fins (Fig. 4a).Those conspicuous dark blotches might represent an "embellishment by the artist", perhaps based on the observation of a specimen with only inconspicuous marks on the fins (as the specimen of L. jamesi in Fig. 4b).However, it is also possible that the holotype may once have had those conspicuous marks, which are now completely faded (Fig. 1a), since fresh specimens of some species of Leporinus have similar blotches on the dorsal and pectoral fins (Fig. 4c).In either case, the dark blotches on the dorsal and pectoral fins probably represent intraspecific variation.
As a result of inconsistencies between the descriptions, the drawings, and the reality, it is sometimes difficult to identify specimens based exclusively on Borodin (1929).This might also help to explain why specimens of Leporinus jamesi were traits (Bookstein et al., 1985) and the morphometric study of the genus Leporinus conducted by Garavello et al. (1991) (Table 2; Fig. 6).The first principal component (PC1) accounts for 93.55% of total among group variation while the second principal component (PC2) accounts for the remaining 6.45%.The ordination of the two species along principal components in Fig. 6 indicates that they are indeed distinct.
The logarithm transformed values of the morphometric traits: interorbital distance (0.41365); body depth (0.38592); caudal peduncle depth (0.35085) and head length (0.34007), are the fastest growing allometric vectors indexing sizerelated shape (Table 2; Fig. 6).Differences between Leporinus jamesi and L. amazonicus on the principal component 1 (PC1) are due to different average sizes of the samples, with L. amazonicus having a larger average size than L. jamesi.However, interorbital distance (-0.46615), trunk length (0.43702) and orbital diameter (0.43447) were confirmed as differential morphometric characters along the principal component 2 (PC2) that discriminate the two species (Fig. 6).Summing up, based on the morphometric analysis, L. jamesi differ from L. amazonicus by the proportionally larger orbital diameter and shorter interorbital distance, in addition to the slightly different body and caudal peduncle depth, and head length.
ascribed to new species when Garavello (1979) started to study the species of Leporinus.Johannes (1999), for instance, followed Garavello (1979)'s unpublished Ph.D. Dissertation in regards to the concept of Leporinus "amazonensis".However, during Garavello's (1979) study, the holotype of L. jamesi was still missing, and as a consequence, the species was recognized as valid based only on its original description (Borodin, 1929).Garavello (1979) also recognized some specimens of Leporinus from the central Amazon as belonging to an undescribed species, and provisionally named it "Leporinus amazonensis".Nevertheless, this name represents a nomem nudum since it was used only in an unpublished study by Garavello, and also because Johhannes (1999) did not present a formal description of the species (see article 16 of the ICZN).
Leporinus jamesi is similar in terms of external morphology and co-occurs with L. amazonicus in the central Amazon.The differences between L. jamesi and L. amazonicus are more obvious when comparing young individuals of both species (Figs.1c and 5).Leporinus jamesi is distinguished from L. amazonicus by having a body relatively deeper, with the dorsal profile straight from snout tip to the dorsal-fin origin (vs.body relatively slender, with dorsal profile convex from snout tip to dorsal-fin origin), three dark transverse bars on dorsum between the supraoccipital and dorsal-fin origin in specimens of approximately 60 mm SL (vs.five to seven dark transverse bars on dorsum between supraoccipital and dorsal-fin origin in specimens of approximately 60 mm SL), and the absence of a dark spot on the second, third and fourth scales posterior to the gill opening and immediately ventral to the lateral line scale series (vs.present).In addition, L. jamesi is most likely a small-sized species (largest known specimen slightly less than 160 mm SL), whereas L. amazonicus reaches up to 250 mm SL (Santos & Zuanon, 2008).
Nine morphometric characters of Leporinus jamesi and L. amazonicus were compared, in a total of 25 specimens from the central Amazonian region.Differences in morphology between L. jamesi and L. amazonicus were also revealed through a principal component analysis of morphometric

Fig. 2 .
Fig. 2. SEM image of jaws of Leporinus jamesi, lateral view, MZUSP 75639, 89.0 mm SL.The dotted line indicates a fourth tooth that is normally present, but that was lost during preparation of the specimen.Scale bar = 1 mm.

Fig. 3 .
Fig. 3. Map of northern South American showing the distribution of Leporinus jamesi.White symbol represents the type locality.

Fig. 6 .
Fig. 6.Scatter plots of individual scores on first two main axes from the principal component analysis of combined samples of Leporinus jamesi (white squares) and L. amazonicus (black circles) on left; linear regression of the interorbital width (mm) vs. standard length (mm) on center; and linear regression of the orbital diameter (mm) vs. standard length (mm) on right.
Borodin, 1929)se cases, errors appear to be the artist's misrepresentations of the reality, whereas in others, they seem to be Borodin's mistakes.Nevertheless, some lateral-line scale counts are accurate and consistent among the description, figure and the actual specimens, especially with respect to L. garmani and L. crassilabris.The arrangement and shape of teeth in the holotype and Roetter's drawing (plate 9, fig.4ofBorodin,1929)areextremely similar, and conform to the observed in other examined specimens of L. jamesi.A diagnostic feature of L. jamesi, accurately illustrated in Roetter's drawings, is the bicuspid symphyseal premaxillary tooth in adult specimens, which is very uncommon among species of Leporinus.
(MCZ 20372); L. fasciatus, 41 to 44 lateral line scales in the description, 45 in the figure; L. altipinnis, 41 or 42 lateral line scales in the description, 43 in the figure and 42 according to our observations (MCZ 20487); L. cylindriformis, 44 lateral line scales in the description, 46 in the figure and 42 according to our observations (MCZ 20430); L. mulleri, 35 lateral line scales in the description and 38 in the figure; L. pachyurus, 40 lateral line scales in the description and 34 in the figure; L. bahiensis, 32 to 36 lateral line scales in the description and 38 in the figure.The number of series of scales around the caudal peduncle and the transverse series in the drawings is also inaccurate in several cases, including at least L. affinis,L.conirostris, L. crassilabris, and L. garmani.