RELATIONSHIPS OF SUNDASA LA NX 



19 



Fig. 10 Schematic illustration of the opercular and branchiostegal series 

 of Sundasalanx malleti (anterior to left); A = opercle, B = subopercle, C 

 = interopercle, D = branchiostegals, E = preopercle, F = infraorbital 

 bone. 



resolution of the staining procedures used, or the materials prepared 

 as whole mount preparations. 



Dentition 



Roberts (1981, 1984) reported teeth on the premaxilla and maxilla, 

 and that teeth are embedded directly in cartilage on the lower jaw 

 (Meckel's cartilage), ceratobranchial 5, and pharyngobranchial 4 of 

 S. praecox and S. microps. Kottelat (1991) reported a pharyngo- 

 branchial 4 tooth plate in the species here named S. platyrhynchus, 

 rather than teeth embedded directly in cartilage. Materials prepared 

 only as alizarin preparations in this study indicate the conclusion 

 that some teeth in Sundasalanx are directly embedded into cartilage 

 is erroneous, and most likely due to examination of specimens 

 prepared by counter-staining for cartilage and bone. 



In addition to those structures listed above, all species of 

 Sundasalanx examined were found to also have a tooth plate and 

 teeth associated with the underside of the posteromedial part of 

 pharyngobranchial 3. The size of the tooth plate appears to vary 

 among species. Sundasalanx praecox was found also to possess 

 teeth on the posterior end of basibranchial 1-2) in the floor of the 

 mouth, and small 'vomerine' tooth patches were found also in the 

 roof of the mouth, one on either side of the anterior tip of the 

 parasphenoid. The largest cleared and stained specimen of S. malleti 

 appears to have a tooth plate (toothless) associated with the posterior 

 end of the basihyal. 



Suspensorium 



The suspensorium is weakly ossified and not well differentiated. 

 Independent quadrate, metapterygoid, symplectic, or hyomandibula 

 ossifications were not detected in any species of Sundasalanx. The 

 posterior articulation of the suspensorium with the skull is broad. An 

 independent cartilaginous palatine was observed in all examined 

 species except S. praecox, the smallest species, in which a thin strut 

 of cartilage communicates between the palatine and the hyomandi- 

 bular complex. An extremely thin, sheet-like ossification under the 

 eye between the palatine and hyomandibular is here identified as the 

 mesopteryogoid. 



Lower hyoid arch 



The basihyal is a large expansive structure that essentially fills the 

 floor of the mouth anterior to the hyoid arch. A cartilaginous 

 hypohyal is present, but apparently without differentiation into 

 upper and lower hypohyoid elements. Ossified anterior and poste- 

 rior ceratohyals are present, as is an unossified interhyal. Two, three, 



or four branchiostegals, subequal in size, were found to be associ- 

 ated with the ventrolateral side of the posterior ceratohyal. The 

 number of branchiostegals present is not constant in any species of 

 Sundasalanx examined. The anteriormost branchiostegal is posi- 

 tioned at the anterior edge of the posterior ceratohyal, partially 

 overlapping the gap between the posterior- and anterior ceratohyal. 

 Succeeding branchiostegals are positioned posteriorly along the 

 ceratohyal and there is a distinct gap between the branchiostegal 

 series and the interopercle, which is associated with the posterola- 

 teral side of the posterior ceratohyal near the articulation of the 

 interhyal. The interopercle is distinctly larger than any member of 

 the branchiostegal series (Fig. 10). 



Opercular series 



An opercle, subopercle, interopercule, and preopercle were found in 

 all species examined (Fig. 10). All elements are very lightly ossified 

 and are easily overlooked, especially the preopercle. The opercle, 

 subopercle and interopercle form an overlapping series, with the 

 ventroanterior corner of the opercle lying external to the posterior 

 edge of the subopercle and the anterior edge of the subopercle lying 

 external to the posterior edge of the interopercle. The interopercle is 

 associated with the posterolateral side of the posterior ceratohyal, 

 near the interhyal . The preopercle is an elongate element. Its anterior 

 end lies lateral to the undifferentiated quadrate. More posteriorly it 

 lies lateral to the ventral portion of the cartilaginous structure that is 

 the undifferinated hyomandibula. Its posterior portion occupies the 

 space between the infraorbital and the opercle. No lateral sensory 

 canal was observed to be associated with the preopercle. 



Infraorbital series 



A single infraorbital element was detected. It is an extremely thin 

 ossification located in the interior of the bend of where the infraorbital 

 laterosensory canal turns in an anterior direction. The infraorbital 

 canal emerges from the recessus lateralis well behind the eye. 

 Consequently the sole infraorbital is positioned well posterior to the 

 eye. 



Skull Roofing Bones 



The only roofing bone identified is an ossification associated with 

 the supraorbital laterosensory canal and recessus lateralis. The bone 

 is very weakly ossified and difficult to detect. The pair, one on each 

 side, are here identified as frontals, because of the association 

 between the supraorbital laterosensory canal, the recessus lateralis 

 and the frontal in clupeiforms. Parietals, and dermal ethmoid ele- 

 ments were not detected. 



RELATIONSHIPS 



In the discussion that follows, the informal term salangid refers to 

 icefishes (=Salangidae of Roberts, 1984). This does not imply 

 criticism of the rank ascribed to them by Johnson and Patterson 

 (1996). Features identified as suggesting a relationship between 

 Sundasalanx and salangids (Roberts 1981, 1984; listed previously 

 [p. 5]) are all paedomorphic, with the possible expection of the 

 orientation of the posterior part of the maxilla. If maxilla orientation 

 in other larval clupeoids is found to be like that Sundasalanx then 

 this feature also is paedomorphic. The features are plesiomorphic 

 too, being features found in larvae of lower teleosts, and in larvae of 

 some euteleosts. Such a list, no matter how long, is nothing more 

 than an appeal to sy mplesiomorphy as evidence for the hypothesis of 

 relationship. However, as Roberts (1984) indicated the question 



