Elops smithi, Mcbride, Richard S., Rocha, Claudia R., Ruiz-Carus, Ramon & Bowen, Brian W., 2010

Mcbride, Richard S., Rocha, Claudia R., Ruiz-Carus, Ramon & Bowen, Brian W., 2010, A new species of ladyfish, of the genus Elops (Elopiformes: Elopidae), from the western Atlantic Ocean, Zootaxa 2346, pp. 29-41 : 31-38

Elops smithi , new species

Figure 1 ; Table 1

Elops saurus non Linnaeus 1766: Regan, 1909 (in part), Hildebrand, 1943 (in part); Bertin, 1944 (in part); Fowler, 1931; Gehringer, 1959 (in part); Whitehead, 1962 (in part); Hildebrand, 1963 (in part); Eldred and Lyons, 1966 (in part); Carles, 1967; Miller and Jorgenson, 1973 (in part), Santos-Martίnez and Arboleda (1993). These studies examined what we recognize here as E. smithi , either in whole or in part, based on their reports of meristic values, sample locality, or both.

Elops sp.: Smith, 1989 ; Smith-Vaniz et al., 1999; Obermiller and Pfeiler, 2003

Holotype. UF 45683, 303 mm SL, marine waters of Guyana, by trawl, 16 July 1967.

Paratypes. UF 45682, marine waters of Trinidad, 13 December 1967; GCRL 11950, 4, Commewijne River, Suriname, 14 October 1973; GCRL 12733, Gatun Locks, Canal Zone, Panama, 4 March 1974; UF 11269, 2, Tortuguero Lagoon, Costa Rica, August 1963.

Other museum material examined. Elops saurus : FSBC 0 0 258, Florida, Pinellas County, 13 December 1957; FSBC 10270, Florida, Pinellas County, 15 December 1977; FSBC 12589, 3, Florida, Hillsborough County, 3 June 1983; GCRL 0 345, Louisiana, 8 September 1959; GCRL 1938, Mississippi, 8 August 1966; UF 47036, Honduras, 26 April 1967; UF 105715, 2, Florida, Volusia County, 17 November 1972.

Diagnosis. Elops smithi is distinguished from E. saurus in the number of vertebrae (73–80, usually 75–78 versus 79–87, usually 81–85, respectively). According to McBride and Horodysky (2004), vertebrae counts in the overlap range (79–80) occurred in only 94 of 3,255 (2.9%) specimens examined from the coasts of the Americas, the Bahamas, and the Caribbean islands.

As reported by Whitehead (1962) counts of gill rakers and total vertebrae can be used to separate all species of Elops . Western Atlantic Elops , now recognized as E. smithi and E. saurus , have lower gill raker counts (10–15 on the lower part of the first arch) than E. affinis (16–20) and E. lacerta (17–19). Elops smithi and E. saurus have higher vertebrae counts (> 72; see above for ranges) than E. senegalensis (67), E. machnata (63–64), and E. hawaiensis (68–70).

Description. Body elongate (head length 25–29% of SL) and slender (width 7.5–9.4% of SL). Mouth large (maxilla 56–60% of head length) and nearly terminal. Caudal fin deeply forked, with lobes equal. Principal dorsal-fin rays 19–20 (24–27 total); anal rays 12–13 (16–19 total); pectoral rays 17–18; pelvic rays 13–16; branchiostegal rays 30–34; gill rakers on lower arch 13–15 (total 21–23 excluding rudiments); lateralline scales (102–118, but may be higher if specimen is found north of the Caribbean Sea [see below]); and 73– 80 vertebrae (usually 75–78). Data for above description are from Table 1 , Figure 2 , and McBride and Horodysky (2004).

Meristic characters

continued.

The larvae are of leptocephalus form, and the total myomere number of the leptocephalus equals that of total vertebrae ( McBride & Horodysky 2004). Predorsal (61–66) and preanal (68–72) myomeres are also reliable characters for identifying premetamorphic leptocephali ( Smith 1989 ). Others examining Elops from the Caribbean Sea have noted minor variations in these meristic characters ( Carles 1967; Santos-Martίnez and Arboleda 1993), indicating differences in counting methods or geographic variation.

Coloration. Living adults bright silver, particularly on sides; may be bluish-gray on back with yellowish hue on fins. Not as silvery in preservative.

Comparisons. We found no character other than counts of vertebrae that separates adult E. smithi from E. saurus ( Table 1 ). Although lateral-line scale counts appeared to be a diagnostic character, the scales are formed late in larval development so this character is indicative of latitude where larval transformation occurs instead of where spawning occurred. To demonstrate this, it can be shown that lateral-line scale counts are distinct among E. smith and E. saurus that had not been dispersed outside their typical range ( Table 1 , Fig. 2 A), but the lateral-line scale counts are not distinct among a test collection from the southern Indian River Lagoon, Florida, which included E. smithi that had presumably been dispersed from their spawning grounds ( Fig. 2 B). This disconnect can be explained because vertebral number is set during embryogenesis but scales are not developed until about 50 mm in the late metamorphic period ( Gehringer 1959). Thus, lateral-line scale counts will be a misleading diagnostic character when measured from specimens collected within areas of sympatry. The leptocephalus larva, common to this genus, is associated with long-distance dispersal ( McBride & Horodysky 2004), so scale counts may be problematic as a taxonomic character in other species of Elops as well. In addition, the scales of these species are small and easily displaced, further confounding their use as a taxonomic character.

Mitochondrial DNA sequence data provide an independent character for recognizing two species of Elops in the western North Atlantic. We observed 14 haplotypes in two primary lineages corresponding to E. saurus and E. smithi (Table 2). These sequences have been deposited in GenBank under accession numbers GQ183881 – GQ183882 ( E. saurus , haplotypes A, B) and GQ183883 – GQ183894 ( E. smithi , haplotypes C – N). The genetic difference (d) between the E. saurus haplotypes and the E. smithi haplotypes ranged from 0.023 to 0.029 ( Fig. 3 ). Obermiller and Pfeiler (2003) reported a similar level of divergence between what we recognized as E. smithi and E. saurus (d = 0.021 with 12S and 16S rRNA mtDNA sequences). In addition, Obermiller and Pfeiler (2003) observed a similar level of divergence between E. saurus and E. hawaiiensis (d = 0.024), two recognized species that occupy different ocean basins.

TABLE 2. Data associated with fish used in the genetic analysis. Collection date and locale are listed, along with standard length (SL), and total number of vertebrae (Vert.). Fish are sorted as vertebrae morph (high vs. low-count) and haplotype. *Mismatches between phenotype and genotype are indicated with an asterisk. Locality codes are: IR = Indian River Lagoon (northern or southern regions of the lagoon), southeast Florida, CK = Cedar Key, west coast Florida, and FW = Fort Walton, panhandle Florida.

continued next page Although the mtDNA data presented here and data presented previously by Obermiller and Pfeiler (2003) indicate two evolutionary lineages, the mtDNA character was not absolutely diagnostic. In seven of 56 specimens (12.5%), there was a mismatch between classification based on genetics and that on morphology (Table 2). Five of these individuals had morphology similar to E. smithi but haplotypes in the E. saurus lineage, and two had the reverse. These mismatches may indicate ecophenotype plasticity within species, hybridization, or retention of ancestral polymorphisms. All three phenomena have been documented in fishes (see Rocha et al. 2007). The sympatric sturgeons Scaphirhynchus albus and S. platorynchus share mtDNA haplotypes because of their recent speciation and hybridization ( Campton et al. 2000). The marine angelfishes Centropyge argi and C. aurantonotus are sister species that share haplotypes despite diagnostic differences in coloration ( Bowen et al. 2006). And the Atlantic bluefin tuna ( Thunnus thynnus ) has haplotypes derived from both Pacific bluefin ( Thunnus orientalis ) and albacore ( Thunnus alalunga ; Alvarado Bremer et al. 2005). Surveys of multiple nuclear loci have resolved these phenomena in other species and could profitably be applied to Elops . Nonetheless, given these genetic results, it is unlikely that we are observing an intraspecific, or population-level, genetic phenomenon, so we reject the hypothesis that E. smithi and E. saurus are ecophenotypes.

Distribution. Elops smithi occurs along the northern coast of South America, in the Caribbean Sea, and throughout the Bahamas; it also occurs sympatrically with E. saurus in the Gulf of Mexico and along the eastern seaboard of North America ( Smith 1989 ; see McBride & Horodysky [2004] for distribution maps of larvae, juveniles and adults). In addition, there are two records of Elops from Bermuda ( Linton 1907) although no resident population has been found there ( Smith-Vaniz et al. 1999). Only one specimen is available, a 181 mm SL fish (BAMZ 1990-083-037) with 73 vertebrae; this apparently represents a waif from the population inhabiting the Caribbean ( Smith-Vaniz et al. 1999).

Elops smithi is found in a wide range of salinities. Mature adults and early-life-history stages are found in offshore, marine habitats, where spawning presumably occurs ( Gehringer 1959; Santos-Martínez & Arboleda 1993; McBride & Horodysky 2004). Transforming larvae and subadults are found throughout estuaries, as far up as the oligohaline zone, as well as in hypersaline lagoons ( Carles 1967; McBride et al. 2001; McBride & Horodysky 2004).

Etymology. The specific epithet honors David G. Smith , of the Smithsonian Institution, for his thoroughness in examining leptocephali of Elops to reveal that two morphs were present. We recommend the vernacular name Malacho, which is already used for Elops in several countries bordering the Caribbean basin.

Figures

FIGURE 1. Elops smithi, holotype, UF 45683, 303 mm SL, left lateral view.

FIGURE 2. High (or low) counts of vertebrae do not always match to high (or low) counts of lateral line scales with Elops smithi (open symbols) and E. saurus (filled symbols). Two datasets are used here: (A) individuals from Table 1 (various collections), and (B) a single collection of 25 Elops collected March 30, 1998, in the southern Indian River Lagoon, Florida. See text for full explanation as to why we consider lateral-line scale counts to be a misleading diagnostic character.

FIGURE 3. There were 14 haplotypes of Elops in two primary lineages in the western North Atlantic (see Table 2 for data source). Separation between haplotypes occurs by a single base pair substitution, or, if more than one, the number of dots (plus one) placed between each haplotype. Haplotypes A and B are E. saurus; haplotypes C-N are E. smithi (Table 2).

Tables

TABLE 1. Comparison of selected phenotypic characters for Elops smithi (top) and E. saurus (bottom). Abbreviations of meristic characters are: Vert. (total number of vertebrae), D (Dorsal fin elements [total (t) or principal (p)]), A (Anal fin elements [total (t) or principal (p)]), P 1 (pectoral fin elements), P 2 (pelvic fin elements), GR (gillrakers [lower (l) or upper (u)]), LL (lateral line scales), B (branchiostegal rays). Morphometric measurements are in mm; abbreviations used are: SL (standard length), FL (fork length), TL (total length), HL (head length), OD (orbit diameter, LJL (lower jaw length), HWG (head width at gills). * Indicates H for holotype, P for paratype, L for lectotype (data for E. saurus lectotype are from Linnaeus [1766]). No data = nd.

Institution	Lot# (type*)	Vert.	D (t) D (p) A (t) A (p) P1	P2	GR(l)	GR(u)	LL	B
Elops smithi
UF	45683 (H)	76	27 20 nd 13 17	14	14	9	108	32
UF	45682 (P)	76	27 20 17 12 17	14	13	9	116	33
UF	11269 (P)	75	26 20 nd 12 18	14	14	9	112	32
UF	11269 (P)	77	nd 20 nd 12 18	14	14	8	118	34
GCRL	12733 (P)	77	27 20 19 12 18	13	14	8	117	32
GCRL	11950 (P)	78	25 19 16 12 18	14	14	8	102	30
GCRL	11950 (P)	78	24 19 16 12 18	14	13	8	108	31
GCRL	11950 (P)	78	24 20 16 12 18	15	15	8	105	33
GCRL	11950 (P)	76	24 20 16 12 17	16	15	8	109	31
Elops saurus
Linn. Soc.	90 (L)	nd	24 20 16 13 17	14	nd	nd	nd	30
UF	47036	84	nd 20 nd 13 17	14	14	9	119	30
FSBC	10270	84	24 20 18 13 18	14	13	8	120	32
GCRL	1938	84	26 21 18 12 18	15	13	8	123	30
GCRL	345	84	28 20 18 13 17	14	14	9	128	27
UF	105715	83	27 21 18 13 17	13	14	9	124	34
UF	105715	85	25 19 18 14 17	15	14	7	119	30
FSBC	258	85	24 21 17 12 18	14	14	8	123	32
FSBC	12589	84	27 21 17 12 19	14	14	8	121	29
FSBC	12589	85	27 20 18 13 18	14	14	8	122	33
FSBC	12589	84	28 20 17 13 17	13	14	8	119	29