TARUMANIIDAE, Pinna & Zuanon & Py-Daniel & Petry, 2018

Pinna, Mário De, Zuanon, Jansen, Py-Daniel, Lucia Rapp & Petry, Paulo, 2018, A new family of neotropical freshwater fishes from deep fossorial Amazonian habitat, with a reappraisal of morphological characiform phylogeny (Teleostei: Ostariophysi), Zoological Journal of the Linnean Society 182, pp. 76-106 : 80-82

TARUMANIIDAE FAM. NOV.

Type genus: Tarumania gen. nov. described below.

Species included: Tarumania walkerae sp. nov. ( Figs 1–3) described below.

Diagnosis: Distinguished from all other families of Osteichthyes by the presence of a swimbladder composed of 11 longitudinally arranged interconnected compartments extending along most of the body, immediately ventral to the vertebral column ( Fig. 4; vs. swimbladder with one or two compartments). Also unique among teleosts (except Platytroctidae ) by the presence of reverse-imbricated scales (i.e. with free margins directed anteriorly) covering most of the head ( Fig. 5; vs. scales with normal imbrication throughout body and head). Further distinguished from all other Ostariophysi by having two rows of teeth on the maxilla ( Fig. 6; vs. single row). Uniquely diagnosed from all other Characiformes by each of the following characters: the numerous vertebrae (69–70; vs. 68 or fewer), pleural ribs (41–44; vs. 32 or fewer) and scales (244–267 along midlateral row and 25 rows on caudal peduncle; vs. 162 or fewer); the spatulate caudal peduncle ( Fig. 1; vs. oblong or round in cross-section); the lanceolate caudal fin ( Fig. 1; vs. bifurcated. emarginate or round); and the platybasic skull, with the parasphenoid expanded and conjoined with the remainder of the neurocranium, forming the floor of the braincase ( Fig. 7; vs. skull tropibasic). Other characters not necessarily unique to the taxon but still rare or unusual across a wide taxonomic array include: pelvic fins long and in the living fish deflectable 180 degrees anteriorly ( Fig. 3; arrow); a flexible ‘neck’ that can bend the head at a right angle relative to the trunk; the infraorbital bone series reduced to single plate-like element lacking a sensory canal; the interopercle with a large semicircular notch along the dorsal margin ( Fig. 8); the presence of a single upper pharyngeal toothplate (corresponding to the posterior element in other characiforms) ( Fig. 9); a large hypural-like bone located between haemal spines of second and third ural centra ( Fig. 10). Other characters variably shared with a number of other clades but still useful to diagnose the taxon include the latero-sensory canal system mostly absent on the neurocranium and body but present in the nasal, dentary and preopercle; the teeth all unicuspidate, with two on each jaw hypertrophied and caniniform ( Figs 5, 8); the eyes very small and located on the anterior portion of the head ( Figs 1–3); the external notochord and larval pectoral fin persistent, the former in specimens up to 50-mm SL and the latter up to 30-mm SL; the cranial fontanels closed ( Fig. 7); the central part of the frontals and parietals elevated and exposed on the surface of the head in large specimens, forming an ornamented platform ( Fig. 7); the post-temporal fossae absent ( Fig. 7); the intercalar absent ( Fig. 7); the frontals expanded laterally, forming shelves along the anterior half of the skull ( Fig. 7); the mesethmoid anteriorly expanded ( Fig. 7); the quadrate-metapterygoid foramen absent or not differentiated ( Fig. 8); the postcleithra and suprapreopercle absent; the supraoccipital spine deeply sunk in the anterior trunk musculature.

Remarks: As shown below, our investigation into the phylogenetic position of the new taxon concludes that it is the sister group to the family Erythrinidae . In purely nomenclatural terms, such positioning is compatible either with an expansion of the latter family or with the erection of a new family. Both choices result in exclusively monophyletic named taxa and therefore conform to the principles of phylogenetic classification. Given such flexibility, our choice for a new family rests on considerations additional to criteria of monophyly. First, the inclusion of Tarumania in Erythrinidae would result in profound modifications of the composition and range of morphological and biological variation seen in that family, which has been stable for more than a century. This would create information-retrieval problems with a vast amount of data in previous literature. Second, most of the traditionally recognized diagnostic characters in Erythrinidae are not present in Tarumania , rendering such familial allocation difficult to integrate with established taxonomic practice. Finally, the large phenotypic distance between Tarumania and species of Erythrinidae meets or exceeds that seen among characiform families in general, making a separate family for the genus a solution fitting the established classification of the order. Inclusion of Tarumania into an expanded Erythrinidae is clearly a less satisfactory solution and one that would incur more disturbance than necessary in the classification of characiforms. Recognition of a separate family more closely reflects the biological reality of the entities involved and better promotes nomenclatural stability.