Notarius Gill

Ricardo Betancur-R. & Arturo Acero P., 2004, Description of Notarius biffi n. sp. and redescription of N. insculptus (Jordan and Gilbert) (Siluriformes: Ariidae) from the eastern Pacific, with evidence of monophyly and limits of Notarius., Zootaxa 703, pp. 1-20 : 11-18

[[ Genus Notarius Gill ]]

Discussion

The genus Notarius was originally described by T. N. Gill in 1863 to accommodate the WA Arius grandicassis . Marceniuk and Ferraris (2003) resurrected this generic name and also placed in it Arius planiceps Steindachner , Sciades troschelii Gill , and Tachisurus lentiginosus Eigenmann and Eigenmann . Following the well-supported phylogenetic hypothesis presented in Fig. 8, which is based on the combined mitochondrial data set cyt b and ATPase 8/6 (1937 bp), we believe that Notarius comprises at least 11 species (Table 3), most of which have been previously included in Arius or other genera (e.g. Sciadeops Fowler and Aspistor Jordan and Evermann ). Furthermore, other neotropical species not sequenced by us, such as A. phrygiatus (similar to N. rugispinis ), A. luniscutis (similar to N. quadriscutis ), and T. lentiginosus , are likely to be included in Notarius . However, it is noteworthy that Notarius is a complex taxonomic entity and possibly comprises two more undescribed EP species.

The monophyly of Notarius is supported due to the nested position of its type species, N. grandicassis , and the high bootstrap value of the clade (100%). From topology it is also clear that the neotropical sea catfish species treated herein under Notarius are not closely related to A. arius . This fact gives support to Betancur-R.'s (2003) hypothesis, which anticipated on morphological grounds that the genus Arius should not be used in the New World. Betancur-R. (2003) also proposed that the presence of a cranial fontanelle posteriorly limited by the frontals and the supraoccipital constituted an osteological synapomorphy of the Notarius group. However, because in A. grandicassis the supraoccipital does not participate in the cranial fontanele (unpublished data), this species exhibits the putative plesiomorphic state. Therefore, a morphological synapomorphy for Notarius species is still lacking. In any case, although Betancur-R. (2003) did not analyze either the osteology or molecular data of N. grandicassis , the monophyletic status of the branch conformed by several Notarius species, among different ariid lineages, was consistent with both mitochondrial and nuclear markers (»3900 bp).

Under the phylogenetic assumption presented in Fig. 8, the systematic scheme of Notarius sensu Marceniuk and Ferraris (2003) is evidently paraphyletic. Those authors also accepted the genus Aspistor for A. luniscutis and A. quadriscutis , included A. cookei , A. neogranatensis , A. phrygiatus , and A. rugispinis in Arius , and A. kessleri and A. osculus Jordan and Gilbert in the polyphyletic genus Hexanematichthys Bleeker (see a detailed discussion about the nonmonophyly of Hexanematichthys in Betancur-R., 2003). The placement of the mentioned species in Arius and Hexanematichthys is rejected on the basis of molecular evidence. Our results show that the genera Aspistor and Sciadeops should be considered as junior synonyms of Notarius . Alternatively, it would be possible to accept Aspistor as the sister genus of Notarius . However, the low bootstrap value of such scenario (<60%) implies a weakly supported monophyletic Notarius . Moreover, in three of four mitochondrial topologies presented in Betancur-R. (2003), after combining two data sets (cyt b and ATPase 8/6 vs. cyt b, ATPase 8/6, 12S and 16S) and two reconstruction criteria (maximum parsimony vs. Bayesian inference), A. quadriscutis appears in a nested position within a clade of several Notarius species. Therefore, we reject the liberal action of accepting Aspistor as a valid genus, and accepting at least two other genera exclusive of Notarius . We herein opt for an inclusive Notarius and would accept Aspistor and Sciadeops only at a subgeneric level.

There are four EP ariid species listed as inquirendae in recent literature (see Kailola and Bussing, 1995; Marceniuk and Ferraris, 2003). The types of these species were examined to avoid nomenclatural chaos. The holotype of Arius hassleriana Borodin , described from Panamá, displays a large mouth, small eyes, relatively wide and triangular-shaped supraoccipital process, and numerous granulations on the rear surface of the skull. These features are similar to N. kessleri and suggest that this species is a senior synonym of the former. On the other hand, the presence of molariform teeth on the palatal tooth patches in the unique type of A. festae Boulenger , from Naranjal in Ecuador, indicates that this species is a member of the genus Cathorops . In addition, these teeth are large and globular, which suggests that it is a senior synonym of C. tuyra (Meek and Hildebrand) . Finally, the types of A. labiatus Boulenger and Hexanematichthys henni Eigenmann , from Peripa and Daule rivers in Ecuador, lack inner palatine tooth patches and possess only rudimentary lateral palatine patches with villiform teeth, display a narrow and elongated snout, and present numerous gill rakers on rear surfaces of first two gill arches. Therefore, neither A. labiatus or H. henni are species of Notarius ; they seem to be species of the freshwater genus Potamarius Hubbs and Miller , which is so far unknown from the EP. In conclusion, our new species is distinct from any of the above species, poorly diagnosed in the literature.

After reading the original description of N. planiceps by Steindachner (1877) and examining several of the types of this species from Panamá and Altata, we conclude that its correct identity has been misunderstood, at least in recent literature (see Bussing and López, 1994; Kailola and Bussing, 1995; Robertson and Allen, 2002). The studied type specimens have small mouth (33.9-39.4% HL), thick lips (8.6-9.2% HL), and low gill raker counts on first arch (2-3+6-7). These are features that correspond mostly to the species identified by recent authors as A. osculus . However, a direct comparison with A. osculus cannot be accomplished, because Jordan and Gilbert´s (1883) original description is obscure, the type locality is not precise (Pacific Panamá) and the only type specimen (USNM 29476) have been lost for more than two decades. Therefore, due to the lack of reliable evidence, the status of A. osculus should be considered uncertain.

Although N. biffi had not been formally described, it is known to scientists working on the fish fauna of the tropical EP. Bussing and López (1994) presented a sketch of the head of Arius species A and a short description. Kailola and Bussing (1995) also gave a description, showed sketches of head and palatine teeth, and included it in their key to the EP ariids. Finally, Robertson and Allen (2002) presented key features and two pictures of the species. The phylogenetic hypothesis presented herein indicates a close affinity between N. biffi and the transisthmian lineage conformed by N. kessleri and N. cookei from the EP, and N. neogranatensis and N. sp. from the WA. Comparisons of select features distinguishing N. biffi from six other EP species of Notarius are summarized in Table 4.

In their summary of the EP ariids, Jordan and Gilbert (1883), described three new species, two of which were Arius insculptus and A. elatturus . They justified their separation on the basis of the continuity of the palatine teeth patches (fully confluent in A. insculptus vs. separated by a narrow interspace in A. elatturus ) and on the size of the humeral process (more developed in A. insculptus ). However, they did not notice at that time that both features in fact reflect sexual dimorphism, since two of the three types of A. insculptus are females (the smaller specimen remains unsexed) and the existing paratype of A. elatturus is a male. N. insculptus , as probably all sea catfishes, can be easily sexed by the size of the pelvic fins, which are larger in females (18.3-20.9% SL) than in males (13.2-15.4% SL). After examining the type series of A. insculptus / elatturus and additional material (one female and two males), sexual differences in adults associated with the shape of the palatine teeth patches (Fig. 7) and with the relative area of the humeral process (Ihp 1.5- 1.6 in females vs. 1.0-1.1 in males) were consistent. Furthermore, HL seems to be larger in males (29.6-30.3% SL vs. 25.4-26.4% SL in females) and the pelvic fin bases larger (4.7-5.0% SL vs. 3.3% SL in one male) and lips thicker (4.9-7.2% HL vs. 3.4% HL in one male) in females. In their review of the marine fishes of Panamá, Meek and Hildebrand (1923) were apparently the last authors who validated Netuma insculpta and N. elattura . After that, the species remained forgotten to science until Kailola and Bussing (1995) and subsequent authors (see Acero and Betancur-R., 2002; Marceniuk and Ferraris, 2003) treated both names together with Netuma insularum as junior synonyms of Notarius kessleri . Gilbert and Starks (1904) commented that N. insculptus was a rare species. Additionally, we located few specimens deposited in museums. This probably explains its omission in the literature through most of the 20th and early 21st centuries. As Fig. 8 clearly indicates, N. insculptus is sister species of N. planiceps clade. Comparisons of select features distinguishing N. insculptus from six other EP species of Notarius are summarized in Table 5.

Key to described species of the genus Notarius from the eastern Pacific

The species of the genus Notarius are distinguished from other EP ariid taxa by the following combination of features: humeral process pointed, triangular to elongated, but never fan-shaped; three pairs of barbels present; fleshy furrow between posterior nostrils absent; fleshy groove in median depression of head absent; coarse to sharp granules or spinulations on anterior surface of head shield absent; gill rakers on rear surfaces of first two gill arches absent. Some of the data ranges showed below are based on wider ranges proposed by Kailola and Bussing (1995).

1 Predorsal plate large, square or hexagonal and shaped like a forward pointing arrow .. .................................................................................................................... N. troschelii

- Predorsal plate narrow and crescent-shaped.................................................................2

2 Gill rakers on second arch 5-6; anal fin rays 23-28.............................. N. lentiginosus

- Gill rakers on second arch 8 or more; anal fin rays 17-22...........................................3

3 Epioccipital bones extensively invasive over skull surface, and forming with the supraoccipital a basally wide complex process which tapers drastically posteriorly (Fig. 6); supraoccipital process length shorter than base of complex process width; maxillary barbels relatively long, their length in adult specimens 26.7-30.3% SL........ ................................................................................................................... N. insculptus

- Epioccipital bones not invasive or only slightly invasive over skull surface (Fig. 2); supraoccipital process length as long as or longer than its width at base; maxillary barbels relatively short, their length in adult specimens 26.1% SL or less....................... 4

4 Mouth small, its width 33.9-42.5%) HL; anterior internarial distance 17.9-24.0%) HL; eye relatively large, its diameter 3.5-4.5% SL .............................................................5

- Mouth large, its width 44.1-54.2% HL; anterior internarial distance 25.3-32.2% HL; eye relatively small, its diameter 2.5-3.7% SL.............................................................6

5 Gill rakers on first arch 11-13; lips thin; mandibulary barbels comparatively short, their length 10.2-13.1% (mean 11.6%) SL; caudal peduncle relatively slender, its depth 6.1-6.7% (mean 6.4%) SL........................................................................ N. biffi

- Gill rakers on first arch 8-10; lips usually thick; mandibulary barbels comparatively long, their length 13.7-17.7% (mean 16.1%) SL; caudal peduncle relatively deep, its depth 6.8-7.4% (mean 7.1%) SL ........................................ N. planiceps /aff. planiceps

6 Supraoccipital process elongated, its base width 1.6-1.7 in its length............ N. cookei

- Supraoccipital process relatively wide and triangular-shaped, its base width 1.0-1.3 in its length ....................................................................................................... N. kessleri