Rivulus staecki, a new killifish (Teleostei: Cyprinodontiformes: Rivulidae) from the upper Rio Negro drainage in southern Venezuela

Rivulus staecki, una nueva especie del subgénero Owiyeye, se describe desde la parte superior de la cuenca del Río Negro, sur de Venezuela (Estado Amazonas). Se diferencia de las restantes especies del subgénero por una combinación única de los siguientes caracteres: una aleta caudal truncada, aletas pélvicas alargadas en los machos, la presencia de la ‘mancha de rivulus’ en ambos sexos y de 35 a 38 escamas en la serie longitudinal.


Introduction
The neotropical killifi sh genus Rivulus Poey, 1860 is the most species rich within the Rivulidae.Currently it contains more than 130 species which are distributed from Middle America in the north to northern Argentina in the south.The centre of the species diversity and phenetic manifoldness of the genus is situated in the Amazon and Orinoco drainages.
The genus is divided into several species groups and formally recognized subgenera (Costa, 2006(Costa, , 2008a;;Huber, 1992).The new species described here is a member of the recently established subgenus Owiyeye Costa, 2006.This subgenus currently contains ten species (R. altivelis, R. amanapira, R. kirovskyi, R. mahdiaensis, R. nicoi, R. rectocaudatus, R. romeri, R. tecminae, R. uakti and R. uatuman), which are distributed in the Orinoco and Amazon drainages (Costa, 2006(Costa, , 2008b;;Costa & de Souza, 2009).All these taxa were described within the last three decades, with Rivulus rectocaudatus Fels & de Rham, 1981 as the oldest.They are small fi sh usu ally living in forest creeks, shallow ponds and swamps.According to Costa (2006) the subgenus Owiyeye is a monophyletic unit.Its major diagnostic characters are the transversally arranged frontal scales (S-patterned) and a transverse stripe through the chin (Costa, 2006).
During a fi eld trip to the state Amazonas in Venezuela Wolfgang Staeck (Berlin, Germany) collected a new colourful Rivulus possessing the diagnostic character states of the subgenus Owiyeye.The species was introduced in the aquaristic literature as Rivulus spec."San Carlos de Rio Negro" by Staeck (2006).The purpose of this paper is to provide a formal description of this Killifi sh.

Material and Methods
Type specimens were fi xed in formalin and later transferred into 75 % ethanol.The holotype and some paratypes are deposited in Museo de Zoología, Guanare, Venezuela (MCNG-UNELLEZ), the remaining paratypes are stored in MTD (Senckenberg Naturhisto-rische Sammlungen Dresden, Museum für Tierkunde, Germany).
Examined material of Rivulus staecki is listed under the species account.Morphometric data are obtained as described by Huber (1992).According to this procedure we and other authors (e. g.Thomerson & Taphorn 1992;Thomerson et al. 1992) use the snout tip as the most anterior point of reference for the standard length (SL).Costa (e. g. 2004Costa (e. g. , 2006)), however, used the middle of the posterior limit of the depression between the upper jaw and the neurocranium (see Costa, 1988Costa, , 1995)).Thus, the proportional data provided in his papers (e. g.Costa, 2004Costa, , 2006) ) are comparable only to a certain extent with those published by others.Measurements were made with an electronic digital calliper reading to the nearest tenth of a millimetre.All ratios are expressed as percentages of standard length.Counts were made under a dissecting microscope; fi n-ray counts were made with light transmitted through the fi ns.As counting the number of fi n rays is not without diffi culties, it may happen that the very minute fi rst rays of anal, dorsal and caudal fi n are overlooked.Values for a potassium hydroxide cleared and alizarin stained (c&s) specimen are indicated by an asterisk after counts.Terminology  Vertebrate Zoology ■ 61 (1) 2011 nal series (35 -38 versus 41 -44 in R. altivelis, 48 -49 in R. amanapira and 38 -43 in R. rectocaudatus and R. tecminae) and from R. nicoi and R. uakti by having more scales (35 -38 versus 26 -29 in R. nicoi and 32 -33 in R. uakti).

Description.
See Table 1 for morphometric data.Males larger than females, largest male 28.9 mm SL (largest preseved female 19.6 mm SL).Dorsal profi le slightly convex from snout to end of dorsal-fi n base, approximately straight on caudal peduncle.Ventral profi le convex on head, almost straight from anterior portion of venter to end of anal-fi n base, nearly straight on caudal peduncle.Body slender, subcylindrical anteriorly, slightly deeper than wide, to compressed posteriorly.Highest body depth at level of pelvic-fi n base.Posterior tip of dorsal and anal fi ns rounded in both sexes.Dorsal and anal fi n round or slightly pointes in males.Caudal fi n truncate in males, rounded in females.Pectoral fi n rounded.Tip of pelvic fi n reaching between base of 5th and 7th anal-fi n ray in males, and between base of 1st and 3rd anal-fi n ray in females.Dorsal-fi n origin on vertical between base of 8th and 10th anal-fi n ray.Dorsal-fi n rays 8* -9; anal-fi n rays 13* -14; caudal fi n rays 26*; pectoral-fi n rays 13*; pelvic fi n rays 6*.
Scales small, cycloid.Body and head entirely scaled.Frontal squamation S-patterned.Few scales on caudal-fi n base; no scales on dorsal and anal fi n.Frontal squamation scales arranged in a transverse pattern, each scale with posterior margin exposed.Longitudinal series of scales 35 -38.Contact organs on male fl ank.No contact organs on fi n rays.Cephalic neuromasts: supraorbital 3 + 3. Lateral line interrupted.Six branchiostegal rays.Gill-rakers on fi rst branchial arch 1 + 7 -8.Minute dermosphenotic present.Caudal skeleton symmetric; hypural forms two plates, separated by a median gap; single epural narrow, loosely connected with last neural spine at its base; parhypural autogenous, thin, stick like.Total vertebrae 31*.
Colouration in alcohol.Male (Fig. 1): Body side light brown with dark brown stripes, ventrally irregular and interrupted by light dashes.Dorsum uniform dark brown, venter light, whitish.Side of head dark brown, dorsal part darker, branchiostegal area light grey.Dorsal fi n light grey with dark grey irregular dots.Caudal fi n grey with dark brownish or grey elongated interradial dots or short strokes; ventral margin with a broad light band; dark blotch on dorsal portion of caudal fi n base.Anal fi n light grey at base, distally darker with blackish marginal band.Pelvic fi n dark grey with dark blackish border.Pelvic fi n light grey.
Female (Fig. 2): Body and head sides brownish with dark brown stripes, ventrally irregular and interrupted by light strokes.Dorsum darker, venter light-for the arrangement of frontal scales follows Huber (1992), terminology for the cephalic neuromast series Costa (2001).Nomenclature of colour patterns and shapes is according to Huber (1992).
Diagnosis.Rivulus staecki is a member of the subgenus Owiyeye.Similar to R. altivelis, R. nicoi, R. tecminae, R. uakti and distinguished from the remaining congeners by having long pelvic fi ns (pelvic-fi n length up to about 25 % of SL, reaching the middle of anal-fi n base in males versus pelvic fi n short, reaching anterior portion of anal-fi n base).It is distinguished from the other species of the subgenus (except for R. amanapira, R. rectocaudatus and R. tecminae) by a truncate caudal fi n in males (versus round, subtruncate, spatula-or lyre-shaped caudal fi n).It differs from R. altivelis, R. amanapira, R. rectocaudatus and R. tecminae by having fewer scales in the longitudi- Colouration in life.Male (Fig. 3): Dorsum dark brown ish, venter whitish to yellowish.Body sides metallic green, with irregular red stripes, stripes dorsally more regular, ventrally irregular and interrupted.Opercular region metallic greenish with brown or reddish reticulated dots.Ventral part of head silvery whitish.Upper jaw brown, lower jaw dark brown or dark grey.Iris light brown to yellow with golden or greenish iridescence.Dorsal fi n reddish brown with metallic greenish dots.Anal fi n reddish brown with metallic greenish irregular dots or strokes and dark marginal band.Caudal fi n dark grey with iridescent greenish interradial strokes and broad bright yellowish or orange ventral stripe.Black blotch on dorsal portion of caudal fi n base ('rivulus-spot').Pelvic fi ns dark grey with metallic greenish border.Pectoral fi ns pale grey.Female (Fig. 4): Dorsum brown, with dark brown dots, venter whitish.Body sides light brownish with pale olive irregular stripes, which are caudally and ventrally interrupted.Opercular region pale golden with brown spots.Ventral surface of head white.Lower jaw dark brown or brownish grey.Iris light brown to yellow.Dorsal and anal fi n light brown or grey with brown dots.Caudal fi n light brown with brown dots.Black vertically elongated oval spot on dorsal portion of caudal-fi n base ('rivulus-spot').Pelvic and pectoral fi ns yellowish.
Distribution and habitat.Known only from the type locality, a pond in Rio the Negro basin, about 20 km south of San Carlos de Rio Negro at the village Darigua (Amazonas, Venezuela).The water at the collecting site was clear, of light brownish colour and very acid (pH 4.4).Its temperature was 25.4 °C and its electric conductivity < 10 μS/cm.
Etymology.The species is named in honour of WOLF-GANG STAECK (Berlin, Germany), former president of the Deutsche Cichliden-Gesellschaft and collector of the type specimens, to acknowledge his commitment to South American freshwater fi shes and his contributions to the taxonomy of the fi sh families Rivulidae and Cichlidae.

Discussion
The subgenus Owiyeye is defi ned by transversally arranged frontal scales appearing as an S-patterned squamation (Fig. 5) and a transverse stripe through chin (Costa, 2006).The new species R. staecki possesses er.Dorsal-and caudal fi n light grey with dark grey irregularly arranged interradial dots.Dark blotch on dorsal portion of caudal fi n base.Anal fi n grey, at base   & Etzel, 2008) and variability caused by mood or environmental infl uences (Huber, 1992), however, prompt questions about the reliability of differentiation and species limits based only on differences in colour patterns.Rivulus staecki can be delimitated from its relatives both by a combination of meristic character states and differences in the colour patterns.This makes it more likely that it represents a genetically different unit and an independent lineage in the sense of the evolutionary species concept as re-defi ned by Wiley (1978).
Based on fi eld observations on the ecology of Ri vulus species from the Amazon drainage Costa (2006) introduced the hypothesis that species of different clades occupy distinct habitats or distinguishable microhabitats.The preferred habitats of the species of the subgenus Owiyeye are shallow temporary or permanent ponds in the tropical forest area.The type locality of R. staecki is described by Staeck (2006, personal communication) as a riparian biotope and hence similar to those habitats known for the other species of Owiyeye.This supports the hypothesis of divergent ecological specializations during the evolution of this subgenus (Costa, 2006).
In its current concept the genus taxon Rivulus is phy logenetically only weakly supported (Costa, 2006).Based on molecular data it is found to be para phyletic (Hrbek & Larson, 1999) or even polyphyletic (Vermeulen & Hrbek, 2005).Within Rivulus, however, there are several species assemblages which are recognized as well supported monophyletic subgenera (Huber, 1999;Costa, 2006Costa, , 2008a)).The ambivalence between several monophyletic subgenera within a genus Rivulus, which is in a phylogenetical sense probably not a natural unit, raises the question about lifting up these subgenera (or some of them) to genus level.But at present no robust comprehensive phylogenetic analysis based on different kinds of character sets is available.Furthermore, subgenera has the advantage that they are optional parts of the scientifi c names and not in need to be cited.Therefore it seems a good compromise to use subgenera for the classifi cation.This is why we include the new species in the established genus Rivulus instead of using one of the subgenera as a genus.Further phylogenetic studies have to verify whether Rivulus is a natural monophyletic assemblage or whether it has to be split into separate genera.these character states and hence fi ts the characteristics of this subgenus.Costa (2006) mentioned the lack of the dermosphenotic (sixth infraorbital bone) as a possible apomorphic character of Owiyeye.However, the osteology has not been studied in all species of the subgenus yet.Since a minute dermosphenotic is present in R. staecki, we consider this character as polymorphic in Owiyeye and not as an apomorphy of this subgenus.The presence of a minute dermosphenotic is used as a diagnostic character state for the Rivulidae by Costa (1998).Hence, the presence of it is a plesiomorphic character state for Rivulus.
An S-patterned frontal squamation and a transverse stripe through chin also occur in the two species of the subgenus Anablepsoides (see Costa, 2006).Rivulus staecki, however, is readily distinguishable from both species of this subgenus (R. atratus and R. ornatus; for R. obscrus is treated as a synonym of R. ornatus by Costa, 2006) by its truncate caudal fi n (round in Anablesoides), higher number of dorsal-(> 7 in R. staecki versus < 8 in R. atratus and R. ornatus) and anal-fi n rays (> 12 in R. staecki versus < 12 in R. atratus and R. ornatus) and more scales in the longitudinal series (> 34 in R. staecki versus < 33 in species of Anablepsoides).
Rivulus staecki resembles R. tecminae (probably the most closely related species), distributed both in the upper Rio Orinoco and in the upper Rio Negro basin (Thomerson et al., 1992;Costa, 2006).Rivulus tecminae, however, differs from R. staecki by a higher number of scales in the longitudinal series (35 -38, mode 36 in R. staecki versus 38 -43, mode 40 in R. tecminae) and by complete stripes on body fl anks (versus interrupted stripes on fl anks in R. staecki).Rivulus amanapira Costa (2004) is another species of Owiyeye also described from the upper Rio Negro drainage (type locality São Gabriel da Cachoeira, Brazil).It shares with R. staecki the possession of a 'rivulus spot' in the dorsal part of the caudal fi n base in both sexes and a truncate caudal fi n in adult males.Rivulus staecki differs from R. amanapira by a longer pelvic fi n length (> 20 % of SL in males and >10 % in females of R. staecki versus < 10 % in R. amanapira), a lower number of scales in the longitudinal series (35 -38 in R. staecki versus 48 -49 in R. amanapira), a lower number of caudal fi n rays (25 -26 in R. staecki versus 28 -30 in R. amanapira) and the broad yellow stripe on the ventral margin of the caudal fi n (versus entire marginal region of caudal fi n yellowish with narrow reddish line both on dorsal and ventral margins; see Costa, 2004Costa, , 2006)).
Differences in the colour pattern of males of rivulins often are the main or even the sole criterion for species recognition and discrimination (Huber, 1992;Costa, 2006).Observable variation in colour pattern within Rivulus species (discussion in Schindler
lighter, distally darker.Pelvic fi ns grey with dark grey border.Pectoral fi ns light grey.