The Tonkin weakfish, Cynoscion similis (Sciaenidae, Perciformes), an endemic species of the Amazonas-

Despite its importance in biogeographical, ecological, and commercial terms, the fish fauna of the northern Brazilian coast is still poorly known, representing the least sampled portion of the Brazilian Exclusive Economic Zone. We collected Tonkin weakfish, Cynoscion similis specimens during extensive surveys of the northern Brazilian coast and concluded that C. similis is common in this region. While the species had not previously been reported for the northern Brazilian state of Pará, it may have been recorded in studies of industrial fisheries, being identified only as Cynoscion sp. or by the common name pescada negra. This reinforces the need for the reliable taxonomical identification of species, to guarantee the collection of accurate data on ecology and fisheries, and ultimately, support the development of effective conservation strategies. Here we provide additional morphological and molecular data to distinguish Cynoscion similis from the closely related Cynoscion jamaicensis, and other congeners.


INTRODUCTION
The Tonkin weakfish or pescada negra, Cynoscion similis Randall & Cervigón, 1968, was described from specimens collected on the coast of Venezuela, Suriname, and French Guiana. Since its description, however, the species has only rarely been mentioned in the literature (Chao 1978(Chao , 2002(Chao , 2003Uyeno et al. 1983;Cervigón 1992Cervigón , 1993Aguillera 1998;Casatti and Menezes 2003). Cynoscion similis is very similar to the Jamaican weakfish, Cynoscion jamaicensis (Vaillant and Bocourt 1883), being distinguished from the latter species by the presence of a sheath of 1-2 scales on the base of the 2nd dorsal fin and the anal fin, which, in C. jamaicensis, is covered with small scales over at least half of its total extension (Chao 2003).
In a review of the classification of the sciaenids of the western North Atlantic, Chao (1978) noted that Cynoscion similis is found from "Venezuela to Brazil", although its occurrence on the Brazilian coast is known only from records from the state of Amapá (http://www.fishnet2.net). This position was perpetuated in subsequent publications, as in the paper of Uyeno et al. (1983) on the fish species trawled off Suriname and French Guyana, and in the list of Brazilian marine fishes (Casatti and Menezes 2003). In the IUCN Red List account for C. similis, however, it is stated that "There (are) no known records of this species from Brazil (F. Lucena-Frédou, pers. comm. 2009)" (Chao and Aguillera 2015) We conducted extensive surveys on the northern Brazilian coast, and, during these surveys, we collected a number of Cynoscion similis specimens. In this context, the present study reports the occurrence of Cynoscion similis on the northern coast of Brazil, and provides a new morphological diagnosis of the species, together with an analysis of its DNA barcode.

Material examined
Material examined in the present study is held by the following Brazilian institutions: the zoological collection (Acervo Zoológico) of Universidade Santa Cecília (AZUSC) in Santos, São Paulo State, Brazil; Museu Paraense Emílio Goeldi (MPEG)

Morphological analysis
Measurements were taken either with a ruler and recorded to the nearest millimeter, or with dial calipers and recorded to the nearest 0.1 mm. The following measurements were obtained through orthogonal projections using a ruler: Caudal -peduncle length (CPl), from the posterior basal margin of the last anal-fin ray to the end of the hypural plate; Distance from snout to anal fin (SAd), from the tip of the snout to the anterior basal portion of the first anal-fin spine; Distance from snout to dorsal fin (SDd), from the tip of the snout to the anterior basal portion of the first dorsal-fin spine; Distance from the tip of the snout to the pectoral fin (SPd), from the tip of the snout to the anterior basal portion of the first pectoralfin ray; Head length (Hl), from the tip of the snout to the end of the operculum, including its membranous portion; Snout length (Sl), from the tip of the snout to the anterior margin of the orbit; Standard length (SL), from the tip of the snout to the end of the hypural plate. The remaining measurements were taken point-to-point: Anal fin height (Ah), from the insertion to the end of the longest spine; Anal-fin base length (Abl), from the anterior basal margin of the first spine to the posterior basal margin of the last ray; Body depth (Bd), the greatest vertical distance between the anterior basal margin of the first dorsal-fin ray to the anterior basal margin of the first anal-fin ray; Body width (Bw), greatest body width at the pectoral girdle; Caudal peduncle height (CPh), least depth of the caudal peduncle; Dorsal fin height (Dh), from the basal insertion to the tip of the longest spine; Dorsal-fin base length (Dbl), from the anterior basal margin of the first dorsal-fin spine to the posterior basal margin of the last dorsal-fin ray; VOL. 49(3) 2019: 197 -207 ACTA AMAZONICA Head depth (Hd), the greatest vertical height of the head at the posterior margin of the operculum; Maxillary height (Mh), greatest vertical height of the maxilla at the mouth angle; Maxillary length (Ml), greatest distance between the anterior and posterior margins of the maxilla; Mouth width (Mw), external distance between mouth angles; Orbital diameter (Od), greatest orbital width; Interorbital distance (IOd), the greatest distance between the upper margins of the orbit; Postorbital length (POl), the distance from the posterior margin of the orbit to the end of the membranous posterior portion of the operculum; Pectoral fin height (Ph), from the basal insertion to the tip of the longest ray. Counts were taken of the elements of the dorsal, anal and pectoral fins, as well as the number of gill rakers on the first gill arches, the number of scales with pores on the lateral line, and the number of scales above and below the lateral line, at the level of the dorsal fin. Whenever possible, the sex of the specimens was determined through the examination of the morphology of the gonads under magnification.
Principal Component Analysis (PCA) was used to analyze the log-transformed measurements in a matrix of co-variance. The values that were constant were excluded from the PCA, so the measurements included were Abl, Bd, CPh, CPl, Dh, Hl, Od, and Ph. Cynoscion similis was differentiated from the closely related C. jamaicensis based on the analysis of the factor loadings for the first and second principal components. Meristic ranges are presented for Cynoscion similis and C. jamaicensis. In the text, measurements are given as percentages of the standard length (SL), unless otherwise stated. The diagnosis of the western Atlantic species was based on direct examination of specimens. The diagnostic characters of the western Atlantic species in relation to species from the eastern Pacific was based on the literature (Chao 1978(Chao , 1995(Chao , 2003Castro-Aguirre 1999;McEachran and Fechhelm 2005) and the Guide to the shore fishes of the Caribbean and adjacent areas of the Smithsonian Tropical Research Institute (www. stri.org/ sfgc and www.stri.org/sftep).

DNA extraction and sequencing
Genomic DNA was isolated from fins or muscle tissues of each specimen with a DNeasy Blood and Tissue Kit (Qiagen, cat. 69506), according to the manufacturer instructions. Amplifications were performed in a total volume of 12.5 μl with 1.25 μl of 10X buffer ( All PCR products were first visually identified on a 1% agarose gel and then purified using ExoSap -IT® (USB Corporation) following manufacturer instructions. The purified PCR products were sequenced using the "Big DyeTM Terminator v 3.1 Cycle Sequencing Ready Reaction Kit" (Applied Biosystems), purified again by ethanol precipitation and loaded on an automatic sequencer 3130 -Genetic Analyzer (Applied Biosystems).

Sequencing analysis
Consensus sequences from forward and reverse strands were obtained using Geneious Pro 8.1.8 (Kearse et al. 2012). Alignments were generated using MUSCLE algorithm (Edgar 2004) under default parameters. After alignment, the matrix was checked by eye for any obvious misalignments, and to detect potential cases of sequencing errors. A quality control step was included in our workflow to detect contamination, paralogous copies or pseudogenes. After that, the presence of stop codons was checked using Geneious. Nucleotide variation, substitution patterns and genetic distances were examined using Mega 6.06 (Tamura et al. 2013). Maximum likelihood (ML) analyses were generated in a partitioned RAxML analysis using the CIPRES web server (Miller et al. 2010). Random starting trees were used for ML tree search and all other parameters were set to default values. All ML analyses were performed under GTR+G since RAxML only applies to this model (Stamatakis et al. 2008). The robustness of the topology was investigated using 1000 bootstrap pseudoreplicates (Felsenstein 1985).
Cynoscion similis is most similar to C. jamaicensis, but both species were clearly distinguished in the Principal Component Analysis (PCA) in relation to the most prominent character traits (Figure 3). The two principal axes of the PCA together explained 85.6% of the variation in the morphometric characters. The first axis was related positively to Od, SDd, CPh, IOd, SPd, SL, Mh, and negatively to Mw and POl ( Figure 3). The second axis was related positively to SPd, Pol, Mh, IOd, Sl and negatively to Mw, CPh, SDh, Od (Figure 3). Otherwise Cynoscion similis can be differentiated by having black pectoral and anal fins (vs. yellowish); soft dorsal fin barely covered with 1 -2 series of scales (vs. dorsal fin covered with scales along more than half its length); pectoral fin usually with 18 rays (vs. usually 16 -17 rays in pectoral fin); rear maxillary border slender, rounded, reaching vertical that pass over posterior eye border (vs. truncate, not reaching vertical through rear eye border); mouth width 9.2 -12.3% of SL (vs. 7.0 -9.2% of SL); and caudal peduncle depth 7.8 -8.9% of SL (vs. 8.9 -10.5% of SL).
Snout pointed, nearly equal to eye length; dorsal profile convex from upper jaw to dorsal fin origin; slightly concave on dorsal margin, immediately anterior to eye. A pair of small, slender pores on the lower jaw separated by symphysis, pores otherwise not visible even under stereomicroscope. Mouth large, terminal, cleft angle when closed 45 o ; lower jaw prognathous; maxillary reaching vertical at posterior margin of eye. Lower jaw heavy; mental process prominent. Lips slender. Anal-fin length 12.5 11.2 -12.0 21 11.0 9.5 -12.8 28 11.2 8.9 -13.8

ACTA AMAZONICA
Upper jaw with two series of caniniform teeth (on inner row largest); 2 stout fangs nearby tip of upper jaw about half size of pupil; lower jaw with two rows of caniniform teeth, one series posteriorly with 5-8 larger teeth about size of posterior nostril. Eye large, diameter slightly greater than interorbital width. Anterior and posterior nostrils separated, anterior one small, oval, second about 2 times larger, elongate or teardrop-shaped, almost half size of pupil. Distance between nares slightly smaller than first nostril. Gill rakers stout, without spines, longest one about half eye size. Scales on head and body. Head with cycloid scales on interorbital region (posterior to that in smaller specimens to 173 mm SL), nape, chin (about 10 vertical rows), opercle (about 15 vertical rows), posterior part of infraorbital region (naked in smaller specimens), ctenoid on belly, abdomen and flanks. Scales absent on dorsal and anal fins except for basal 2 or 3 rows. Pectoral fins much longer than pelvic fins, reaching pelvic fin tips, but not vent. First anal fin spine, second one shorter than longer anal fin rays. Caudal fin tip concave. Gas bladder long, carrot-shaped, similar to that of C. jamaicensis (Figure 2a), but less slender, with a pair of lateral hornlike stout appendages on the anterior border of bladder. Sagitta more slender than in C. jamaicensis (Figure 2b), profile of the upper margin slightly convex, with a small fingerlike projection; lower margin convex with a tiny medial bump, anterior border convex, blunt, or with evident anterodorsal concavity.

Color in life
Fresh specimens are silvery with slender dark brown lines along scale rows on the flanks and upper half (Figure 1a). Snout and lower jaw tip dark, mouth aperture pale with dusky margin near lower jaw. Upper half of pectoral fin with dense black, black spot over pectoral fin, caudal fin rays black, fins otherwise dusky, more so in the first spine of the pelvic fin. Inner operculum margin black, with pale rear border.

Biology
There is a wealth of studies on the biology of the other Cynoscion species of the Atlantic Ocean (Santos and Vianna 2017), but no information is available specifically for C. similis. Randall and Cervigón (1968) described C. similis based on specimens collected at depths of 10 to 30 fathoms (18-54 m), but the specimens we examined were collected in deeper waters (40-70 m), over sandy or muddy bottoms. Randall and Cervigón (1968) described 2-3+7-8 rakers (9-11) on the first gill arch of Cynoscion similis. Our specimens had a larger number of gill rakers (11-14), but there is compelling evidence from the other morphological parameters, as well as the DNA barcode, that this difference is derived from individual variation, or different ways to count gill rakers.

DISCUSSION
While C. similis had not previously been recorded from the northern Brazilian state of Pará, our examination of material obtained from CEPNOR indicated that the fish identified by Maia et al. (2016) as pescada-negra is, in fact, Cynoscion similis, given that this species is the only Atlantic sciaenid that presents dark pigmentation. In the same way, quite possibly the species went undetected in other inventory studies of local industrial fisheries, recorded as Cynoscion sp. (Pinheiro and Frédou 2004;Paiva et al. 2009).
Whereas Randall and Cervigón (1968) reported that C. similis and C. jamaicensis typically occurred in the same area, this was not so common on the northern Brazilian coast. Cynoscion similis showed to be common on the Brazilian continental shelf (at depths of 40-80 m), targeted by industrial trawling fisheries in Pará, while C. jamaicensis is more common in the coastal zone, at depths of less than 40 m, where it is typically exploited by artisanal gillnet fisheries. Quite possibly Cynoscion similis is harvested in Brazil and marketed as goete or pescada gó, the local name for C. jamaicensis. This reinforces the need for reliable taxonomical identification of species, in order to guarantee the collection of accurate data on their ecology and fisheries, and ultimately, support the development of effective conservation strategies.

CONCLUSIONS
Despite its importance in biogeographical, ecological, and commercial terms, the fish fauna of the northern Brazilian coast is still poorly known, and this region is the least well sampled of the Brazilian Exclusive Economic Zone (Marceniuk et al. 2013). Our results indicate that C. similis is common throughout the northern coast of Brazil. The study also provides new morphological and molecular data to distinguish C. similis more reliably from the closely allied C. jamaicensis.  9  10  12  15  18  24  30  33  36  45  45  48  54  57  60  63  66  69  70 72 231 234 237 240 243 246 249 252 255 258 261 262 264 267 270 273 276 279 280 282   G  A  T  T  T  T  C  G  T  G  T  C  T  A  A  A  C  T  C This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.