Teleostei: Leiognathidae: Equulites), with Comments on the Taxonomic Status of Equula Berbis Valenciennes
Total Page:16
File Type:pdf, Size:1020Kb
Zootaxa 2427: 15–24 (2010) ISSN 1175-5326 (print edition) www.mapress.com/zootaxa/ Article ZOOTAXA Copyright © 2010 · Magnolia Press ISSN 1175-5334 (online edition) Geometric morphometrics uncovers a new species of ponyfish (Teleostei: Leiognathidae: Equulites), with comments on the taxonomic status of Equula berbis Valenciennes PROSANTA CHAKRABARTY1,5, JEANETTE CHU2,3, LUTHFUN NAHAR2,4 & JOHN S. SPARKS2 1Museum of Natural Science, Louisiana State University, Baton Rouge; [email protected] 2American Museum of Natural History, New York; [email protected] 3Hunter College High School, New York 4Health Professions & Human Services High School, New York 5Corresponding author Abstract A new species of Equulites is revealed using geometric morphometric techniques and is herein described. Based on features recovered in recent comparative analyses, members of Equulites have been diagnosed on the basis of internal and external male-specific traits related to their light-organ system (LOS; Sparks et al., 2005; Sparks, 2006; Sparks and Chakrabarty, 2007; Chakrabarty and Sparks, 2008). These sexually-dimorphic traits are hypothesized to allow males to signal to conspecific females in photic sexual displays using bacterially-generated luminescence (Woodland et al., 2002; Sasaki et al., 2003; Wada et al., 2005). The holotype and sole name-bearing type of Equulites leuciscus (BMNH 1858.4.21.243, 104.9 mm SL) is an adult female, and therefore lacks the diagnostic external feature of the LOS, a large, translucent flank patch, used to identify species in this genus. Geometric morphometric shape analysis of individuals ascribed to Equulites leuciscus, a traditionally widespread, “catch-all” taxon, reveals two discrete shape groups. Based on the results presented below, members of one of these groups correspond to a morphological variant that represents the new species (Equulites absconditus Chakrabarty & Sparks) described herein, whereas the other group corresponds to traditional E. leuciscus. In addition, the taxonomic status of Equula berbis Valenciennes, to which many female and poorly preserved specimens of the new species have erroneously been attributed, is reviewed and E. berbis is concluded to be a nomen dubium of uncertain placement beyond the family level. Key words: Indo-West Pacific, leiognathids, Photoplagios, taxonomy Introduction Equulites leuciscus Günther, 1860 was originally described from Ambon Island, Indonesia, and has historically been considered to be a geographically widespread species occurring throughout much of the Indo-West Pacific (from Madagascar and the Seychelles, to India, Sri Lanka, Indonesia and Australia; Woodland et al. 2001). Although this taxon has been considered to be one of the most widespread species of ponyfishes, there have been no comparative studies that have examined populations throughout its putative range. This is the first such study and it reveals the existence of geographically localized morphological variants. The female holotype of Equulites leuciscus (Fig. 1) has confounded taxonomic progress because it lacks the diagnostic LOS features typically associated with members of this genus (Sparks et al., 2005; Sparks, 2006; Sparks and Chakrabarty, 2007; Chakrabarty and Sparks, 2008). In addition, much of its original pigmentation pattern has been lost in preservation. To remedy this taxonomic problem and begin to understand geographic variation within this taxon, we compare geographic populations spanning the range of putative E. leuciscus. We use geometric morphometric techniques to determine the existence of different morphotypes and possibly distinct species. Accepted by M.R. de Carvalho: 11 Mar. 2010; published: 15 Apr. 2010 15 Two geographic populations of putative Equulites leuciscus are compared, and are referred to here as groups A (Fig. 2a) and B (Fig. 2b). The range of Group A spans mainly the Western Indian Ocean and South China Sea, from Madagascar to Sri Lanka and extending eastward to Taiwan. Individuals of Group A have relatively small heads, a concavity dorsal to the orbit, and a dorsal flank pigmentation pattern comprising dashed lines. These individuals are often catalogued in collections as E. leuciscus (or under the commonly used older synonyms, Leiognathus leuciscus or Photoplagios leuciscus) if males with a conspicuous translucent flank patch are present. Individuals of Group B occur mainly within the southeastern Indian Ocean, in Australian and Indonesian waters, and have larger heads, lack a concavity dorsal to the orbit, and possess a dorsal flank pigmentation pattern comprising wide vermiculate, circular patterns. Unfortunately, the holotype of E. leuciscus has lost its pigmentation pattern in preservation and is a uniform silver. This lack of pigmentation, and the absence of the male-specific diagnostic external features of the LOS, leaves only external shape as a source of evidence for testing whether Group A or Group B individuals represent Equulites leuciscus sensu strictu. Further confounding the taxonomy of this assemblage is the fact that female specimens (particularly of Group A) are often considered to represent another species, Equula berbis Valenciennes, in Cuvier and Valenciennes, 1835. Equula berbis, for which there is no type material and only a rudimentary original description, is an supposedly widespread taxon to which we hypothesize several distinct species have historically been attributed, including individuals of the new species described herein. In this study, we also review the taxonomic status of Equula berbis. Methods Specimens included in comparative analyses are listed in MATERIAL EXAMINED. Morphometric measurements, following Hubbs and Lagler (2004) were recorded to the nearest 0.1 mm using dial calipers and include: standard length (SL), head length (HL), body depth (BD), predorsal length (PDL), preanal length (PAL), prepelvic length (PPL), head width (HW), caudal peduncle length (CPL), caudal peduncle width (CPW), caudal peduncle depth (CPD), pectoral fin length (PTFL), pelvic fin length (PVFL), snout length (SNL), orbit diameter (OD), upper jaw length (UJL), lower jaw length (LJL), and interorbital width (IOW). Traditional morphometric measurements were taken for 17 different distances measured across the body of 131 individuals. For the geometric morphometric analysis, digital images were taken from the left side of 66 adult specimens. Landmarks (putatively homologous points on anatomical features) were chosen to best represent the external shape of the body (Fig. 3). Information unrelated to shape, including size, orientation and position, was removed by Generalized Least Squared (GLS) Procrustes superimposition. A Canonical Variates Analysis (CVA) was performed to show the axes on which groups are best discriminated by shape. The superimposition and CVA were completed in CVAgen (Sheets, 2001). Institutional abbreviations follow Leviton et al. (1985) except NTUM (National Taiwan University Museum, Taipei). Results There is overlap in the range of each of traditional morphometric measurements taken (Table 1). Notably, Group B individuals had a longer maximum SL (112.7 mm vs. 104.9 mm) and were generally more robust than individuals from Group A. Using traditional measurements, however, we were unable to distinguish between the two geographical groups examined. Canonical variates analysis recovered two distinct shape groups along CV 1 (Fig. 4). Male and female specimens were evenly distributed among the two groups (i.e., there are no gender specific shape trends). Canonical variate axis 1 is significant at the p <.001 level based on the Wilk’s lamda value (the sum of squares within groups divided by the total sum of squares within and between groups). An assignment test performed 16 · Zootaxa 2427 © 2010 Magnolia Press CHAKRABARTY ET AL. in CVAgen (Sheets, 2001) based on CV1 determined that all specimens had been correctly assigned to either Group A or B. The holotype of Equulites leuciscus was assigned to Group B in this test. Based on these data, we conclude that individuals of Group A constitute a new species in the genus Equulites, whereas members of Group B represent Equulites leuciscus sensu strictu. TABLE 1. Comparison of morphometrics for Group A and Group B, means with range in parentheses. A (n=105) B (n=26) Standard Length 82.3 ( 52.8–104.9) 95.3 (73.4–112.7) Head Length (% SL) 26.9 (24.4–30.7) 27.4 (24.0–29.3) Body Depth (% SL) 41.6 (34.0–49.0) 43.6 (38.0–46.6) Predorsal Length (% SL) 44.1 (40.3–48.8) 45.6 (41.9–47.8) Preanal Length (% SL) 55.6 (51.9–58.6) 57.5 (53.7–59.9) Prepelvic Length (% SL) 37.9 (34.6–41.0) 40.1 (34.8–42.5) Head Width (% SL) 12.4 (10.1–14.0) 13.5 (9.0–14.6) CP Length (% SL) 6.9 (3.3–10.1) 6.4 (3.9–10.3) CP Width (% SL) 2.9 (1.7–4.5) 2.9 (2.4–3.9) CP Depth (% SL) 5.5 (3.1–6.6) 5.8 (4.8–6.5) Pectoral Fin Length (% SL) 17.3 (12.0–21.6) 17.1 (12.3–20.0) Pelvic Fin Length (% SL) 11.6 (8.2–16.2) 11.5 (9.4–12.7) Snout Length (% HL) 34.0 (26.0–44.7) 36.2 (32.5–41.7) Orbit Diameter (% HL) 34.0 (27.3–42.7) 33.5 (28.8–35.8) Upper Jaw Length (% HL) 38.1 (30.4–48.4) 42.6 (38.0–50.7) Lower Jaw Length (% HL) 52.0 (42.2–63.4) 52.7 (45.5–58.6) Interorbital Width (% HL) 33.3 (27.6–40.2) 32.9 (28.0–38.7) Equulites Fowler 1904 Equulites is distinguished from all genera of Leiognathidae by the combination of an expansive translucent lateral stripe, triangular, cornucopia-shaped or trapezoidal patch on the flank in males, and a pigmentation pattern on the dorsal flank comprising speckles and vermiculate markings or broad oblong markings that occasionally form open circular patterns. This genus is also diagnosed by the presence of dorsolateral lobes of the light organ that are hypertrophied and extend posteriorly into the gas bladder and lateral clearing of the silvery lining of the gas bladder.