Comparative innervation of cephalic photophores of …ckenaley/Kenaley_2010_FarRedHomology.pdfComparative Innervation of Cephalic Photophores ... (Teleostei: Stomiiformes: Stomiidae):

Comparative Innervation of Cephalic Photophoresof the Loosejaw Dragonfishes (Teleostei: Stomiiformes:Stomiidae): Evidence for Parallel Evolution ofLong-Wave Bioluminescence

Christopher P. Kenaley*

School of Aquatic and Fishery Sciences, College of Ocean and Fishery Sciences, University of Washington,Seattle, Washington 98195

ABSTRACT Four genera of the teleost family Stomii-dae, the loosejaw dragonfishes, possess accessorycephalic photophores (AOs). Species of three genera,Aristostomias, Malacosteus, and Pachystomias, arecapable of producing far-red, long-wave emissions(>650 nm) from their AOs, a character unique amongvertebrates. Aristostomias and Malacosteus posses asingle far-red AO, while Pachystomias possesses anteriorand posterior far-red AOs, each with smaller separatephotophores positioned in their ventral margins. Thepurpose of this study was to establish the primaryhomology of the loosejaw AOs based on topological simi-larity of cranial nerve innervation, and subject thesehomology conjectures to tests of congruence under aphylogenetic hypothesis for the loosejaw dragonfishes.On the basis of whole-mount, triple-stained specimens,innervation of the loosejaw AOs is described. The AO ofAristostomias and the anterior AO of Pachystomias areinnervated by the profundal ramus of the trigeminal(Tpr), while the far-red AO of Malacosteus and a smallventral AO of Pachystomias are innervated by the max-illary ramus of the trigeminal (Tmx). The largest far-redAO of Pachystomias, positioned directly below the orbit,and the short-wave AO of Photostomias are innervatedby a branch of the mandibular ramus of the trigeminalnerve. Conjectures of primary homology drawn fromthese neuroanatomical similarities were subjected totests of congruence on a phylogeny of the loosejawsinferred from a reanalysis of a previously publishedmorphological dataset. Optimized for accelerated trans-formation, the AO innervated by the Tpr appears as asingle transformation on the new topology, therebyestablishing secondary homology. The AOs innervatedby the Tmd found in Pachystomias and Photostomiasappear as two transformations in a reconstruction onthe new topology, a result that rejects secondary homol-ogy of this structure. The secondary homology of AOsinnervated by the Tmx found in Malacosteus andPachystomias is rejected on the same grounds. Twoshort-wave cephalic photophores present in all four gen-era, the suborbital (SO) and the postorbital (PO), posi-tioned in the posteroventral margin of the orbit anddirectly posterior to the orbit, respectively, are inner-vated by separate divisions of the Tmd. The primaryhomologies of the loosejaw PO and SO across loosejawtaxa are proposed on the basis of similar innervationpatterns. Because of dissimilar innervation of the loose-jaw SO and SO of basal stomiiforms, primary homology

of these photophores cannot be established. Because ofsimilar function and position, the PO of all other stomiidtaxa is likely homologous with the loosejaw PO. Nonho-mology of loosejaw long-wave photophores is corrobo-rated by previously published histological evidence. Thetotality of evidence suggests that the only known far-redbioluminescent system in vertebrates has evolved asmany as three times in a closely related group ofdeep-sea fishes. J. Morphol. 271:418–437, 2010. � 2009

Wiley-Liss, Inc.

KEY WORDS: homology; phylogeny; trigeminal nerve;facial nerve; lateral-line nerves; osteology

INTRODUCTION

Fishes of the teleost order Stomiiformes arecharacterized by the presence of luciferin-mediatedlight organs on the head and body. The Stomiidae,or dragonfishes, is comprised of over 280 pelagic,deep-sea species in 27 genera (Fink, 1985; Nelson,2006). Stomiid morphology is characterized byspectacular adaptations to life in dark, barren oce-anic waters including huge jaws bearing massive

Additional Supporting Information may be found in the onlineversion of this article.

Contract grant sponsors: Dorothy T. Gilbert MemorialEndowment, University of Washington, Ernst Mayr Grant forAnimal Systematics, Harvard University, Lerner-Gray Grant forMarine Research, American Museum of Natural History, Contractgrant sponsor: National Science Foundation; Contract grantnumber: DEB-0314637.

*Correspondence to: Christopher P. Kenaley, School of Aquaticand Fishery Sciences, College of Ocean and Fishery Sciences, Uni-versity of Washington, Box 355020, Seattle, WA 98195.E-mail: [email protected]

Received 11 March 2009; Revised 8 August 2009;Accepted 11 August 2009

Published online 18 November 2009 inWiley InterScience (www.interscience.wiley.com)DOI: 10.1002/jmor.10807

JOURNAL OF MORPHOLOGY 271:418–437 (2010)

� 2009 WILEY-LISS, INC.

teeth, enormous gapes, and large photophoresassociated with the eye. All stomiid species possessa small suborbital (SO) photophore positionedalong the ventral margin of the fleshy orbit that isused to accommodate the eye during luminescentevents. In addition, stomiids possess a relativelylarge photophore posterior to the orbit that emitsa short-wave (<500 nm) blue or green light (Nicol,1960; Herring, 1983; Mensinger and Case, 1988,1997). In many species, this postorbital photophore(PO) is sexually dimorphic and most likely used insexual communication (Herring, 2000, 2007). Theloosejaw dragonfishes, a monophyletic group of sto-miid fishes, is composed of 16 species in four gen-era: Aristostomias, Malacosteus, Pachystomias,and Photostomias (Fink, 1985). Loosejaw speciesdiffer from other stomiids in having at least oneaccessory orbital (AO) photophore positioned eitherventral or anteroventral to the eye (see Fig. 1).The AO of Photostomias is exposed in males as a

small, blue-glowing organ and covered by a thinlayer of skin in females (Fig. 1A; Kenaley andHartel, 2005; Kenaley, 2009). Species of Aristosto-mias and Malacosteus and Pachystomias microdonpossess larger, far-red glowing AOs (Fig. 1B–D)with emission maxima >650 nm (Widder et al.,1984; Herring and Cope, 2005). Species of Aristos-tomias and Malacosteus possess a single far-redphotophore, while P. microdon, the only knownspecies of the genus, possesses two far-red photo-phores with smaller, separate organs embedded inthe ventral margin of each that glow lower fre-quencies of red light (Herring and Cope, 2005).These fishes are unique among vertebrates in theirability to produce longwave bioluminescence. It ishypothesized that these photophores cast a shortfield of long-wavelength light on their prey(Denton et al., 1970; O’Day and Fernandez, 1974;Fernandez, 1979; Somiya, 1982), typically smallmesopelagic fishes and crustaceans (Sutton and

Fig. 1. Lateral views of heads of the four loosejaw genera. A: Photostomias sp. B: Malacosteus niger. C: Aristostomias grimaldii.D: Pachystomias microdon. Red arrows indicate positions of accessory orbital photophores; blue arrows indicate positions of postor-bital photophores. Arrows are labeled by conventional terms used by most authors and followed in parentheses by terms devisedby Fink (1985). AOP, accessory orbital photophore; PO, postorbital photophore; PRO, pre-orbital photophore; SO, suborbitalphotophore.

FAR-RED PHOTOPHORE INNERVATION 419

Journal of Morphology

Hopkins, 1996; Sutton, 2005) that are unable toperceive long-wave spectra due to their visualadaptations to the blue photoenvironment of thedeep sea (Warrant and Locket, 2004). With retinalpigments tuned to their own red luminescence(O’Day and Fernandez, 1974; Partridge andDouglas, 1995), species of Aristostomias, Malacos-teus, and Pachystomias are afforded a stealthy‘‘night-vision.’’

Although the spectral properties and biochemi-cal and molecular basis of long-wave perceptionhave been the subject of many recent studies, themorphology of loosejaw AOs has received littlescrutiny. Herring and Cope’s (2005) histologicaldescription of the far-red photophores of Aristosto-mias, Malacosteus, and Pachystomias stands asthe only recent morphological investigation sincethe histological descriptions of the orbital photo-phores of A. xenostoma by von Lendenfeld (1887)and P. microdon by Brauer (1908). Herring andCope (2005) reviewed and reported several histo-logical differences in the red-glowing photophore ofM. niger relative to those of Aristostomias andP. microdon. The most striking difference is thepresence of a unique superficial filter in M. nigerthat alters shorter wave-length emissions fromwithin the core to produce long-shifted light at theexterior of the photophore (Denton et al., 1985). Incontrast, Aristostomias and Pachystomias producefar-red light by direct fluorescence from the core ofthe photophore. The basic structure of the far-redAO of M. niger also differs considerably from thoseof Aristostomias and Pachystomias, the formerresembling the structure of plesiomorphic stomiidphotophores, including the blue-emitting PO, andthe latter having a more complicated structure inwhich the bulk of glandular tissue lies outside thereflector of the core, communicating with the cellu-lar material of the core by numerous strands of tis-sue that run through the reflector wall.

Histological and anatomical evidence suggeststhat the far-red photophores of loosejaw generamay not be homologous. Despite presenting a com-pelling body of evidence of nonhomology, Herringand Cope (2005) did not address the terminologyof these organs, maintaining the positional names‘‘suborbital’’ photophore for the large red AOs ofred-emitting species and ‘‘preorbital’’ photophorefor the smaller anterior far-red photophore ofPachystomias. The use of these terms in referenceto the loosejaw AOs by Regan and Trewavas(1930), the earliest in-depth systematic account ofthese taxa, likely led to the proliferation in subse-quent taxonomic literature (e.g., Morrow, 1964),and eventually gained wide use in works focusingon the biochemical and structural aspects or eco-logical importance of these organs (e.g., O’Day andFernandez, 1974; Widder et al., 1984; Campbelland Herring, 1987; Douglas et al., 1998; Douglaset al., 2000; Herring, 2000; Sutton, 2005). In his

phylogenetic analysis of the Stomiidae, Fink(1985) recognized that the term ‘‘suborbital’’ or‘‘preorbital’’ in reference to the loosejaw AOsimplied an a priori assumption of homology withthe photophores of basal stomiiform taxa givensimilar names (see Weitzman, 1974: Table 1). Asa remedy, Fink (1985) designated the AOs as‘‘accessory orbital’’ photophores (AOPs), applyingthe name ‘‘AOPI’’ to the AOs of Aristostomias,Malacosteus, and the largest of Pachystomias;‘‘AOPII’’ and ‘‘AOPIII’’ was applied to the smallanterior AO and organ positioned along theventral margin of AOPI in Pachystomias, respec-tively. The small, circular or oblong blue-emittingorgan anteroventral to the orbit in Photostomiaswas given the name ‘‘AOP IV.’’ A variant of termi-nology used here, AO, has been adopted in recentmorphology-based systematic revisions of loosejawgenera (Kenaley and Hartel, 2005; Kenaley, 2007,2009).

Although Fink (1985) recognized spuriousassumptions of homology and proposed a new ter-minology as a remedy, he coded the far-red AOs ofAristostomias, Malacosteus, and the largest ofPachystomias (‘‘AOPI’’) as apomorphic (his charac-ter 309). Because characters are conceptualizationsof homologies, Fink’s (1985) coding must be inter-preted as a de facto conjecture of homology for thelargest AOs of these taxa. The coding of Photosto-mias as plesiomorphic for this character was possi-bly a mistake (W.L. Fink, personal communication,7 May 2009); however, in a discussion of stomiidphotophores, Fink (1985:117) assigned a uniquename to this structure (‘‘AOP IV’’) and described itquite differently than the AOs of other taxa as anunorganized luminescent mass. Thus, it appearsthat Fink (1985) viewed the AO of Photostomias asautapomorphic.

Studies of teleost morphological homology havefocused almost exclusively on osteological andmyological characters. To assess homology of loose-jaw accessory photophores, however, some otherline of evidence is required. Due in large part tomethodological difficulties, comparative neuroanat-omy remains a relatively untapped source of datain vertebrate morphology, despite being a phyloge-netically informative character complex (Song andBoord, 1993; Song and Parenti, 1995). Cranialnerve innervation of cephalic photophores hasbeen described for the myctophid Stenobrachiusleucopsarus (Ray, 1950) and suggested for thebatrachoidids Porichthys myriaster (Nicol, 1957)and P. notatus (Strum, 1969). Much less is knownabout the innervation of stomiid cephalic photo-phores. Brauer (1904, 1908) and Nicol (1960)described a direct innervation of the postorbitalphotophores of several stomiid species but couldnot identify a particular nerve. Herring and Cope(2005) alluded to trigeminal innervation of thepostorbital and accessory orbital photophores of

420 C.P. KENALEY


Malacosteus niger, but they presented no anatomi-cal evidence.

The aim of this study is to assess the homologyof the loosejaw accessory photophores. Homologyassessment is at least a two-part process (Pat-terson, 1982; de Pinna, 1991; Brower and Scha-waroch, 1996), the most important proceduresbeing conjecture of homology based on similarity(establishing primary homology) and the test ofthat conjecture as congruent with the best phylo-genetic hypothesis at hand (establishing secondaryhomology). Aside from what may be subjectivedecisions by the researcher in the choice of charac-ter optimizations when ambiguous characterreconstructions are inferred (see Agnarsson andMiller, 2008), the establishment of secondaryhomology is based on the decisive and objectivecriterion of congruence. However, the basis of pri-mary homology, similarity, is often subjectivelyevaluated and what is considered similar mayvary by researcher (Rieppel and Kearney, 2002).To reduce subjectivity in the evaluation of similar-ity, it is necessary to use intrinsic discrete featuresthat are part of the structure of the system(Rieppel and Kearney, 2002). Similarity is assessedin this study on the basis of two implicit criteria:(1) topological similarity in the form of shared cra-nial nerve innervation, and (2) Patterson’s (1982)conjunction criterion. To establish secondaryhomology, primary conjectures of homology areevaluated through tests of congruence on a phylo-genetic topology inferred from a reanalysis ofFink’s (1985) morphological dataset. Secondaryaims are to establish primary homology of theloosejaw suborbital and postorbital photophoresand, according to hypothesized homology, addressthe terminology of stomiid cephalic photophores.Terminology that more accurately reflects homol-ogy and removes confusion about evolutionary ori-gin may serve as a framework for further studiesof these unique bioluminescent systems.

MATERIALS AND METHODSMorphological Analysis

Descriptions and illustrations are based on whole specimensprepared with a modified clearing and triple-staining tech-nique described by Song and Parenti (1995) whereby boneand cartilage were stained with alizarin red and alcian blue,respectively (following Dingerkus and Uhler, 1977), andnerves were stained with Sudan Black B (following Filipskiand Wilson, 1984). The following triple-stained specimenswere examined: Aristostomias xenostoma, MCZ 131897, 34mm standard length (SL); UW 118768, 90 mm SL; Malacos-teus niger, MCZ 131819, 138 mm SL; Pachystomias microdon,ZMUC P20591, 91 mm SL; ZMUC P208474, 63 mm SL;ZMUC P208475, 91 mm SL; Photostomias sp., UW 118766, 87mm SL; Ph. guernei, MCZ 131716, 95 mm SL, MCZ 131477,97 mm SL; Ph. liemi, UW 118767, 90 mm SL. Innervation ofAO, SO, and PO photophores was confirmed in the followingby gross microscopical dissections: Aristostomias sp., UW151296, 112 mm SL, A. scintillans, UW 117429, 183 mm SL;116939, 151 mm SL; M. australis, SIO 69-27, 1:152 mm SL;

SIO 61-36-10, 155 mm SL; M. niger, LACM 36046-39, 110mm SL; LACM 11512-11, 172 mm SL; Pa. microdon, UW47216, 195 mm SL; UW 116936, 175 mm SL; Photostomiassp., LACM 36130-8, 111 mm SL; Ph. lucigens, UW 117086, 95mm SL. Descriptions and drawings were made with a ZeissStemi SV 11 dissecting microscope equipped with a cameralucida. Terminology follows Fink (1985) for osteology, Winter-bottom (1974) for myology, and Northcutt et al. (2000) forneuroanatomy. Only those cranial nerves involved in photo-phore innervation or those in close proximity to cephalicphotophores are described. Several lateral-line sensory ramiassociated with the peripheral distribution of the facial nerveare described as components of that nerve. These sensoryrami likely represent anterior extensions of distinct lateral-line nerves such as those described by several descriptive andexperimental studies (e.g., Puzdrowski, 1989; Song and North-cutt, 1991; Piotrowski and Northcutt, 1996; Northcutt et al.,2000); however, descriptions of the central projections of thesesensory nerves were beyond the scope of the study. Althoughno attempt was made to experimentally label sensory ramiand confirm their origin, rami innervating only superficial tis-sues were identified according to nerves with similar periph-eral tracks and target tissues described by Northcutt et al.(2000). Abbreviations used in the text and figure captions arepresented in Table 1. Anterior lateral-line nerves are identi-fied by abbreviations in italics.

Criteria for Establishing Primary Homology

Primary conjectures of homology of the loosejaw cephalic pho-tophores are based on two implicit, discrete criteria:

1. Structural sameness: First order observations of the neuro-anatomy associated with cephalic photophores are used to es-tablish similarity. It is most parsimonious to assume thattwo photophores innervated by the same cranial nerve pro-jections share a common ontogenetic origin.

2. Conjunction: This criterion, as outlined by Patterson (1982),asserts that two structures occurring simultaneously in anorganism cannot be homologous, a corollary of Riedl’s (1978)concept of homologs as anatomical singulars.

TABLE 1. Abbreviations of components of the trigeminal, facial,and anterior lateral-line nerves as discussed in the text and

presented in Figures 2–6

AO Accessory orbital photophoreFhd Hyoid ramus of the facial nerveFhy Hyomandibular trunk of the facial nerveFmd Mandibular ramus of the facial nerveFop Opercular ramus of the facial nerveFpl Palatine ramus of the facial nerveib Inner buccal ramus of the anterodorsal

lateral-line nerveob Outer buccal ramus of the anterodorsal

lateral-line nerveot Otic lateral-line nerveso Superficial ophthalmic ramus of the anterodorsal

lateral-line nerveSO Suborbital photphorePO Postorbital photophoreTmd Mandibular ramus of the trigeminal nerveTmdp Photophore branch of the TmdTmx Maxillary ramus of the trigeminal nerveTmxi Inferior branch of the TmxTmxs Superior branch of the TmxTop Opercular ramus of the trigeminal nerveTpr Profundal ramus of the trigeminal nerve



Establishing Secondary Homology andPhylogenetic Analysis

To establish secondary homology, primary homology conjec-tures were subjected to tests of congruence under a phylogenetichypothesis for the loosejaw dragonfishes. To date, no fullyresolved phylogenetic hypothesis for the Stomiidae exists. Fink(1985) recovered a clade comprised of the loosejaw genera (hisclade ‘‘W’’); but, this clade could not be resolved beyond fourequally parsimonious trees. Fink (1985) analyzed his large data-set, composed of 323 morphological characters of the then known26 stomiid genera, in the early days of computational phyloge-netics, before the development of methodologies to address thelarge number of inapplicable characters, polymorphic taxa, andthe high level of homoplasy in his dataset. To reduce characterlinkage during phylogenetic analysis, Fink’s (1985) matrix wasrecoded from additive binary characters to unordered multistatecharacters where possible (Supporting Information AppendixS1). In Fink’s (1985) analysis, inapplicable and polymorphic char-acters were both coded as ‘‘?’’. In the present study, two measureswere taken to reduce the artifactual influence of ambiguouslycoded character states on tree searching: (1) polymorphic charac-ters were recoded as such (i.e., ‘‘0/1, 0/2, 1/2,’’ etc.); (2) for charac-ters associated with relative development of a structure (e.g., sizeor ossification), taxa with characters states coded as missing (‘‘?’’)in Fink’s (1985) analysis were recoded as absent or lost in a multi-state character. All applicable characters of the four loosejaw gen-era were reevaluated in triple-stained specimens listed above.The recoded matrix, presented in Supporting Information Appen-dix S1, excludes the original AO characters (Fink, 1985:93; chars.309 and 311) and autapomorphic characters. Character descrip-tions are presented in Supporting Information Appendix S2.Eupogonesthes Parin and Borodulina (1993), a rare monotypic ge-nus purportedly sister to Astronesthes (Parin and Borodulina,2004), was described after Fink’s (1985) analysis and was notincluded in the present study.

An unweighted parsimony analysis of 26 stomiid genera andFink’s (1985) composite basal stomiiform outgroup was performedwith TNT version 1.1 (Goloboff et al., 2008b) using ‘‘implicit enu-meration,’’ an exact search strategy. Clade support was evaluatedwith a parametric bootstrap (Felsenstein, 1985) and Bremer sup-port values (Bremer, 1994). To evaluate the overall influence ofhomoplasy on tree searching, Goloboff’s (1993) implied weightingmethod was employed in an additional exact parsimony analysis(using the default concavity constant, k 5 3). Weighting againsthomoplasy has been shown to improve morphology-based phyloge-netic analyses (Goloboff et al., 2008a). Bootstrap values were calcu-lated from 500 pseudoreplicates employing a ‘‘new technologysearch’’ that included five random addition sequences, followed bytree bisection and reconnection (TBR), random and exclusive sec-torial searches, and 10 cycles of tree-drifting. Bremer support val-ues were calculated from suboptimal trees inferred from sixrounds of tree searching applying TBR from the four trees savedin the original analysis. During each round of tree searching, up to10,000 trees were saved with fits of 1, 3, 5, 7, 9, and 11 steps worsethan the optimal trees. To preserve primary homology hypotheses,ambiguous reconstructions of photophore character states on theinferred trees were optimized using accelerated transformation(ACCTRAN) whereby transformations occur as close to the root ofthe tree as possible.

The homology statements proposed in this study are not final.Homology hypotheses may be revised at either the primary orsecondary levels in light of more or better evidence. With addi-tional morphological evidence or a new phylogenetic hypothesis,homology of the cephalic photophores of the study taxa may bereinterpreted in the future.

RESULTSMorphological Analysis

Foraminal projections. In Malacosteus andPhotostomias, four nerve bundles emanate from a

single, large, ventrolateral trigemino-facial fora-men in the prootic: (1) the superficial ophthalmicramus (so) of the anterodorsal lateral-line nerve;(2) a mixed bundle of the palatine ramus and hyo-mandibular trunk of the facial nerve; (3) a mixedbundle of the maxillary and mandibular rami ofthe trigeminal nerve; and (4) the profundal ramusof the trigeminal nerve (Figs. 2A and 3A). In Aris-tostomias and Pachystomias, the trigemino-facialcomplex originates from three separate foramina:(1) the trigeminal components and the so arisingfrom the trigeminal foramen and penetrating amore anterodorsal position in the prootic; (2) thehyomandibular trunk of the facial nerve arisingfrom the facial foramen at a position posteroven-tral to the trigeminal foramen; and (3) the palatineramus of the facial nerve arising from a separateforamen on the anterior ventrolateral surface ofthe prootic (Figs. 4A and 5A).

Facial components. In Malacosteus and Photo-stomias, at the approximate midpoint between thetrigeminal foramen and anterior margin of thehyomandibula, the hyomandibular trunk (Fhy)gives rise to the opercular ramus of the facialnerve (Fop; Figs. 2A and 3A). In Malacosteus, theFop bifurcates immediately to form a more proxi-mal ramulus that projects ventrally to innervatethe adductor arcus palatini, and a more distalramulus coursing to the posteromedial face of thehyomandibula and along the posterior margin ofthe hyomandibula to innervate the levator operculi(Fig. 3A). In Photostomias, the Fop passes alongthe dorsomedial surface of the hyomandibula, dor-sal to the opercular process, bifurcating at a pointjust posterior to the opercular process into dorsaland ventral branches that innervate the adductorarcus palatini and levator operculi, respectively. InAristostomias and Pachystomias, the pattern ismuch the same; however, the Fop arises from theFhy at a point much more proximal to the facialforamen (Figs. 4A and 5A). In addition, the ramuliof the Fop of Aristostomias are joined in anastomo-sis; no such anastomosis was observed in anyother loosejaw taxon.

In all taxa examined, the Fhy enters the medialopening of the hyomandibular foramen andcourses through the thickness of the hyomandib-ula, bifurcating within the hyomandibula, themore dorsal branch projecting through the dorso-lateral opening of the foramen. This dorsal branchbifurcates immediately to produce the otic lateral-line nerve (ot) and mixed trunk of the inner andouter buccal rami of the anterodorsal lateral-linenerve (ib and ob, respectively; Figs. 2A, 3A, 4A,and 5A). The ot courses dorsally along the lateralface of the hyomandibula, bifurcating at its dorsalhead, producing a branch anteriorly, ventral to thesphenotic process of the frontal. In Malacosteus,Photostomias, and Aristostomias, the anteriorbranch of the ot ends in innervation of the anterior

422 C.P. KENALEY


Fig. 2. Innervation of the cephalic photophores of Photostomias. A: Anterodorsal part of neurocranium and suspensorium show-ing the peripheral course of the trigeminal (green) and facial/anterior lateral-line (yellow) nerves; bones are stippled, cartilage isshaded dark gray. B: Triple-stained preparation of P. liemi, UW 118767, 90 mm SL, lateral view of anterodorsal neurocranium anddorsal suspensorium. C: Innervation of accessory photophore. For abbreviations, see Table 1.



Fig. 3. Innervation of the cephalic photophores of Malacosteus. A: Anterodorsal part of neurocranium and suspensorium show-ing the peripheral course of the trigeminal (green) and facial/anterior lateral-line (yellow) nerves; bones are stippled, cartilage isshaded dark gray. B: Medial view of accessory photophore (AO) and associated cranial nerves. C: Triple-stained preparation ofMCZ 131819, 138 mm SL, medial view of AO. For abbreviations, see Table 1.

424 C.P. KENALEY


supraorbital lateral-line canal (Figs. 2A, 3A, and4A). In Pachystomias, after supplying innervationto the posterior portion supraorbital lateral-line

canal, the ot bifurcates, coursing within superficialtissue along the lateral edge of the frontal beforeending in innervation of the anterior portion of the

Fig. 4. Innervation of the cephalic photophores of Aristostomias. A: Anterodorsal part of neurocranium and suspensorium show-ing the peripheral course of the trigeminal (green) and facial/anterior lateral-line (yellow) nerves; bones are stippled, cartilage isshaded dark gray. B: Medial view of accessory photophore (AO) and associated cranial nerves. C: Triple-stained preparation ofA. xenostoma, MCZ 131897, 34 mm SL, medial view of AO. For abbreviations, see Table 1.



supraorbital lateral-line canal (Fig. 5A). In Pachys-tomias, the ot also produces a ventral branch justbeyond the anterior margin of the fleshy orbit thatinnervates superficial tissue surrounding the

anterior fleshy orbit (Fig. 5A). These anteroventralprojections of the ot may indeed exist in otherloosejaws, however, they were not observed in thespecimens examined. In all taxa examined, the

Fig. 5. Innervation of the cephalic photophores of Pachystomias. A: Anterodorsal part of neurocranium and suspensorium show-ing the peripheral course of the trigeminal (green) and facial/anterior lateral-line (yellow) nerves; bones are stippled, cartilage isshaded dark gray. B: Triple-stained preparation of ZMUC P20591, 91 mm SL, lateral view of accessory photophore, ventral antero-dorsal neurocranium, and dorsal suspensorium ventral to orbit. For abbreviations, see Table 1.

426 C.P. KENALEY


posterior branch of the ot extends posterodorsallyending in innervation of the otic canal at a pointmidway between the margin of the elements of theposterior neurocranium and the first ossified verte-bra (Figs. 2A, 3A, 4A, and 5A).

At its origin from the dorsolateral hyomandibu-lar foramen, a mixed bundle consisting of the iband ob courses ventrally and slightly anteriorlyacross the lateral face of the hyomandibula andlateral to the posteroventrally directed branches ofthe mandibular ramus of the trigeminal nerve(Tmd). At a point anterior to the midpoint of theanterior margin of the hyomandibula, the mixedib-ob bundle bifurcates into its constituent rami.In Malacosteus, Aristostomias, and Photostomiasthe ib follows an arching course from its diver-gence with the ob, projecting anteroventrally and,eventually anterodorsally into the snout (Figs. 2A,3A, and 4A). In Malacosteus, the ib arches sharplyover the posterolateral face of the AO (Fig. 3A). Atthe peak of the arch, fibers of the ib anastomosewith and bisect anterodorsal projections of theTmd fibers that innervate the SO. Distal to thisanastomosis, the ib arches ventroposteriorly, thendorsoanteriorly, following the ventrolateral marginof the AO, sending several branches towards themaxilla and premaxilla, eventually ending ininnervation of superficial tissue anterior to theAO. A similar configuration of the proximal sectionof the ib and anterior projections of the Tmd isseen in Photostomias (Fig. 2A). Although Aristosto-mias and Pachystomias lack a bisection of the ibby branches of the Tmd, the ib and Tmd are inclose association and no doubt exchange fibers.Aside from this contrast, the ib maintains muchthe same course in all taxa examined as the mostlateral ramus in the cheek, bifurcating severaltimes on its anterodorsal path to the snout (Figs.2A,B, 4A, and 5A). Several ventral projections ofthe ib supply superficial innervation of the bonesof the upper jaw, snout, and cheek. The first, mostproximal ventral projection is a thin, ventroposter-iorly directed nerve, innervating the skin betweenthe PO and maxilla in Aristostomias, Pachysto-mias, and Malacosteus (Figs. 3A, 4A, and 5A), andthe skin posterodorsal to the PO in Photostomias(Fig. 1A). More distally, the ib of all taxa examinedproduces a nerve that bifurcates over the antero-dorsal margin of the maxilla to form anterodor-sally and posteroventrally directed branches. Moredistally, proximal to a series of complex bifurca-tions in the snout, the ib produces a small nervethat first projects anteroventrally and then archesabruptly to course posteroventrally dorsal to theposterodorsal margin of the premaxilla (Figs. 2A,3A, 4A, and 5A). In Aristostomias, Pachystomias,and Photostomias, the termination of the ib ismore complex than in Malacosteus, bifurcatingseveral times on its course to superficial innerva-tion of tissues of the snout anterior to the AO

(Figs. 2A,B, 3A, and 4A). These projections coursemedial to the AO in Photostomias (see Fig. 2) andlateral to the AO in Malacosteus, Aristostomias,and Pachystomias (Figs. 3A–C, 4A–C, and 5). InMalacosteus, the superficial tissues of the compactspace between the anterior margin of the fleshyorbit and premaxilla is innervated by much lessdendritic terminations of the ib relative to theother taxa examined.

From its bifurcation with the ib, the ob coursesventroposteriorly, over the metapterygoid andquadrate, lateral to posteroventrally directedbranches of the Tmd, to the articulation of poste-rior margin of the anguloarticular, medial to theventral expansion of the supramaxilla (Figs. 2A,B,3A, 4A, and 5). From here, the ob bends sharplyanterodorsally to innervate lateral superficial tis-sues of the lower jaw (not shown). The ob gives offa set of very thin posterolaterally directed nervesthat terminate in innervation of superficial tissuesof the cheek over the hyomandibula and metapter-ygoid.

After giving rise to the ot and the mixed ib-obbundle, the hyomandibular trunk continues withinthe hyomandibula as a mixed Fhd-Fmd bundlethat emerges through the ventrolateral opening ofthe hyomandibular foramen. In Aristostomias,Pachystomias, and Photostomias, this foramen ispositioned in the notch created by the opercularprocess of the hyomandibula (Figs. 2A,B, 3A, and4A), offset slightly to a more ventrolateralposition relative to the dorsolateral opening. InMalacosteus, the hyomandibula does not producean elongate opercular process; thus, the notch isabsent and the ventrolateral opening of the hyo-mandibular foramen is exposed along the face ofthe hyomandibula, nearly directly ventral to thedorsolateral opening (Fig. 3A). At its emergencefrom the ventrolateral foramen, the Fhd-Fmd bun-dle projects posteroventrally, reaching the poste-rior margin of the hyomandibula, continuing ven-trolateral to the cartilage of the opercular processin Malacosteus (Fig. 3A) and ventral to the processin Photostomias (Fig. 2A). Emerging ventroposte-rior to the cartilage of the opercular process of thehyomandibula, above the head of the preopercle,the Fhd-Fmd bundle bifurcates to form its constit-uent rami, the hyoid (Fhd) and mandibular (Fmd)rami of the facial nerve. The Fmd continues ven-troposteriorly between the posterior margins of thehyomandibula and sympletic and the anterior mar-gin of the preopercle. The Fmd bifurcates at vari-ous positions among loosejaw taxa: over the ven-tral third of the hyomandibula, adjacent to theapproximate midpoint of the posterior margin ofthe metapterygoid in Malacosteus and Photosto-mias; over the dorsal third of the symplectic inAristostomias, and over the ventral margin ofthe hyomandibula in Pachystomias (not shown).The posterior branch, no doubt containing fibers of



the deep subdivision of the anteroventral lateral-line nerve, follows a ventroposterior course, pass-ing between the medial face of the preopercle andlateral face of the symplectic, continuing along theposteromedial margin of the preopercle, across themedial face of the quadrate, into the anguloarticu-lar. The anterior branch, no doubt containingfibers of the shallow subdivision of the anteroven-tral lateral-line nerve, courses ventroposteriorlyalong the lateral face of the preopercle to theventromedial margin of the supramaxilla andmedial to the quadrate. Over the posterior notch ofthe anguloarticular, the anterior branch of theFmd joins the ob in anastomosis. The adjoinedFmd-ob bundle courses anteroventrally over thelateral face of the quadrate and passes into thedentary at the articulation of these two bones.

From its bifurcation with the Fmd, the Fhd con-tinues posteroventrally along the posterior marginof the preopercle and anterior margin of theopercle (Figs. 2A, 3A, 4A, and 5A). At a positionventral to the dorsal head of the preopercle, theFhd produces a sensory nerve that innervates theskin of the cheek over the preopercle and opercle.From here, the Fhd continues on its path alongthe posterior margin of the preopercle, breakingfrom this course to pass within the skin betweenthe medial face of the metapterygoid and the firstbranchial arch. The Fhd then wraps around thelateral margin of the junction of the interhyal andposterior ceratohyal, over the medial side of thetwo (in Malacosteus) or three (in Aristostomias,Pachystomias, Photostomias) posteriormost bran-chiostegal rays, between the second and third orthird and fourth posteriormost rays, innervatingthe first branchiostegal photophore, then along thelateral side of the rays, following the ventral mar-gin of the anterior ceratohyal to innervate the pro-tractor hyoideus (not shown). In Aristostomias,Pachystomias, and Photostomias (taxa with well-developed branchiostegal photophores), the Fhdsupplies innervation to the branchiostegal photo-phores. No such innervation of the vestigial bran-chiostegal photophores was observed in Malacos-teus. In Aristostomias and Pachystomias, the Fhdalso supplies innervation to the hyoid barbel.

In Aristostomias and Pachystomias, the palatineramus of the facial nerve (Fpl) arises from a dedi-cated foramen, ventral to the facial foramen (Figs.4A and 5A). In Malacosteus and Photostomias, theFpl is produced outside the neurocranium from thehyomandibular trunk of the facial nerve (Figs. 2Aand 3A). In all taxa examined, the Fpl first coursesanteroventrally and arches anteriorly, lateral tothe parasphenoid. Over the parasphenoid, the Fplbifurcates to form dorsal and ventral branches.The dorsal branch of the Fpl courses into thesnout, innervating unidentified tissues ventral tothe anterior base of the parasphenoid, and endingin close association with projections of the ib. The

ventral branch courses ventrally to innervate thesuperficial tissues of the dorsal buccal cavity, even-tually ending in innervation of tissues of the snoutanterodorsal to the head of the premaxilla.

In Malacosteus and Photostomias, the so arisesfrom the trigemino-facial foramen of the prootic,dorsal to the profundal ramus of the trigeminalnerve (Figs. 2A and 3A). In Aristostomias andPachystomias, the so originates from the trigemi-nal foramen in the same close association with theTpr (Figs. 4A and 5A). From its origin, the so runsanteriorly along the ventrolateral surfaces of theprootic and pterosphenoid, parallel and dorsal tothe Tpr. The so bifurcates over the pterosphenoidin Malacosteus and Pachystomias (Figs. 3A and5A), the supraethmoid in Photostomias (Fig. 2A,B),and the prootic in Aristostomias (Fig. 4A), sendinga dorsally directed branch that enters the ventralsurface of the frontal bone to innervate the supra-orbital sensory canal.

Trigeminal components. In Aristostomias andPachystomias, the trigeminal components emanatefrom the trigeminal foramen as a single trunk con-sisting of the Tmx, Tmd, and Tpr (Figs. 4A and5A). In Malacosteus and Photostomias, the Tprand the remainder of the trigeminal trunk ema-nate from the trigeminal foramen as separate com-ponents (Figs. 2A and 3A). In Aristostomias andPachystomias, the Tpr projects anteriorly from thetrigeminal trunk as it exits the trigeminal fora-men. The opercular ramus (Top) of the trigeminaltrunk is produced lateral or just ventral to theposteroventral elements of the neurocranium(Figs. 4A and 5A). In Malacosteus and Photosto-mias, the Top emanates at a point lateral to thesuture of the prootic and parasphenoid (Figs. 2Aand 3A). In Aristostomias, the Top is produced im-mediately dorsal to the bifurcation of the trigemi-nal trunk, between the anteroventral margin ofthe prootic and dorsal margin of the parasphenoid(Fig. 4A). In Pachystomias, the Top is produced ata point midway between the trigeminal foramenand the posteroventral margin of the prootic(Fig. 5A). In all taxa examined, the Top bifurcatesto form inferior and superior branches. The supe-rior branch bifurcates before its projection aroundthe dorsolateral opening of the dorsal hyomandibu-lar foramen, medial to rami of the hyomandibulartrunk of the facial nerve, terminating posterior tothe hyomandibula in innervation of the dilatoroperculi. The inferior branch of the Top coursesventrally along the anterolateral margin of thehyomandibula ending in innervation of the levatorarcus palatini. In Aristostomias, just anterior tothe anterior margin of the hyomandibula, thesuperior branch produces a ventrally projectingnerve that anostomoses with posterior projectionsof the inferior branch of the Top (Fig. 4A). No suchanastomosis was observed in any of the other taxaexamined.

428 C.P. KENALEY


After producing the Top, the trigeminal trunkbifurcates to form the Tmx and Tmd. In Malacos-teus and Photostomias, the Tmx-Tmd bifurcationoccurs considerably ventral to the parasphenoid(Figs. 2A,B and 3A). In Aristostomias and Pachys-tomias, the trigeminal trunk bifurcates dorsal tothe parasphenoid, over the ventrolateral surface ofthe prootic (Figs. 4A and 5A). In Photostomias andMalacosteus, the Tmd passes medial to the ib at aright angle, arching posteroventrally to courseanteromedially and nearly parallel to the ob(Figs. 2A and 3A).

In all taxa examined, the first major projectionof the Tmd distal to its bifurcation with Tmx is thephotophore branch of Tmd (Tmdp), a ramulus thatends in dendritic innervation of the postorbitalphotophore, supplying the encapsulating reflectivetissue and the inner core. In Photostomias, Tmdpprojects posterioventrally from the Tmd, bifurcat-ing immediately to form posterior and anteriorbranches (Fig. 2A). The anterior branch passesover the lateral surface of the ib, anterior to theposterior branch that passes medial to the ib distalto its split from the ob. Before passing lateral tothe ib, the anterior branch produces a fine nervethat bifurcates immediately, the more dorsal pro-jection arching anteriorly, lateral to the Tmxs, toinnervate the SO. The ventral projection follows acourse along the ib into the snout, breaking fromthe main trunk of the ib as the ib multifurcatesventral to the SO, arching dorsoanteriorly, lateralto fine divisions of the ib, eventually ending ininnervation of the small AO (Fig. 2C). The poste-rior branch of the Tmdp, bifurcates into a smallnerve that courses posteroventrally and a nervethat courses anteroventrally to rejoin the anteriorbranch of the Tmdp. The resulting nerve entersthe dorsal margin of the PO, coursing through thelong axis of the photophore and providing innerva-tion.

The Tmdp of Malacosteus projects from the Tmdjust dorsal to the ib, bifurcating a short distancefrom its origin to form a ventrally directed ramu-lus that innervates the PO and an anterodorsallydirected ramulus that is immediately bisected bythe ib (Fig. 3A). The bisected fibers of this ramu-lus, no doubt now containing fibers from the ib,rejoin to course anterodorsally over the posterolat-eral margin of the core of the AO, bifurcating atseveral points to innervate unidentified tissuesmedial and lateral to the AO, eventually ending ininnervation of the SO.

The Tmdp of Aristostomias is shortest relativeto those of other loosejaw taxa. After originatingfrom the Tmd, the Tmdp sends two successiveramuli anterodorsally, the first ending in anasto-mosis with the ib, the second coursing passinglateral to the ib (Fig. 4A). The latter ramulusbifurcates just dorsal to the ib to form a shortanteroventrally directed fiber that ends in

anastomosis of with the ib and an anterodorsallydirected fiber that ends in innervation of the SO.In Pachystomias, the Tmdp is produced ventrallyand then arches anteroventrally, producing sev-eral fine, laterally coursing fibers that innervatesuperficial tissues posterior to the AO (Fig. 5).Distal to these fibers, the Tmdp produces a largenerve that arches anteroventrally, then postero-ventrally through the cheek to innervate the post-orbital photophore. The main trunk of the Tmdpcontinues anteroventrally, emitting several ven-trolaterally directed fibers that enter the AO,penetrating a reflective layer before innervatingthe core. Proximal to entering the core of the AO,Tmdp produces an anteriorly directed branch thatanastomoses with a branch of the Tmx thatinnervates the glandular mass surrounding thecore of the AO photophore. A fine ramulus origi-nates between the bifurcation of the Tmdp fromthe trunk of the Tmd and the origin of the nerveinnervating the PO. This ramulus courses antero-ventrally in close association with the ib, inner-vating unidentified soft tissue before innervatingthe SO photophore at the bundle’s most distalpoint.

In all taxa examined, the Tmd maintains a post-eroventral course distal to the origin of the Tmdp,producing several posteriorly directed branchesthat end in innervation of the adductor mandibu-lae. The Tmd continues to the articulation of thelower jaw, arching anterodorsally, medial to theventral expansion of the supramaxilla and postero-dorsal expansion of the anguloarticular (notshown). From here, the Tmd passes along themedial margin of the dentary, and, before passingmedial to the anguloarticular, produces a ramulusthat innervates the intermandibularis section ofthe adductor mandibulae.

After breaking from the trigeminal trunk,Tmx arches anteroventrally, bifurcating to formsuperior and inferior branches (the Tmxs andTmxi, respectively). In all taxa examined, theTmxi takes an anteroventral course, bifurcatingover the maxilla to form anterodorsal and ven-troposterior nerves that innervate the teeth andsuperficial tissue surrounding the maxilla. TheTmxs arches anterodorsally from its bifurcationwith the Tmxi, coursing medial to the core ofthe AO and lateral to the parasphenoid whereit produces a dorsally directed ramulus thatends in bifurcation and innervation of tissueanterior to the lateral ethmoid. The main trunkof the Tmxs continues anterodorsally within thesnout, lateral to the supraethmoid, and bifur-cates several times before ending in innervationof tissue anterior to the supraethmoid andlateral and anterior to the dorsal process of thepremaxilla.

Distal to its bifurcation with the Tmxi, theTmxs of Malacosteus courses along the dorsal



margin of the AO (Fig. 3A,B). Along its trackover the dorsal margin of the AO, the Tmxs bearsa series of nerves that innervate the core andouter reflective surface. The main trunk of theTmxs continues along the anterodorsal margin ofthe AO, past the pointed anterior tip of the AO,terminating in innervation of tissue anterior andmedial to the anterodorsal head of the premaxilla.With the exception of lacking innervation to alarge accessory photophore, the Tmx of Photosto-mias takes much the same course as thatobserved in Malacosteus; however, the Tmximaintains a course more proximal to the Tmxsthan in other loosejaw taxa. The Tmxs of Photo-stomias passes along the anteroventral margin ofthe orbit, over the parasphenoid, bifurcating im-mediately posterior to the small accessory photo-phore (Fig. 2A). Medial to the AO, fibers of theTmxs pass over the lateral margin of the supra-ethmoid, medial to fine fibers of the ib, ending ininnervation of general tissues of the snout. At itsbifurcation posterior to the accessory photophore,the Tmxs produces a series of small nerves thatprojects laterally and then dorsally. One of thesesmall nerves, passing posterior to the posteriormargin of the AO, supplies a small fiber thatinnervates the luminescent tissues of the organ(Fig. 2A).

In Pachystomias, the Tmx bifurcates over thedorsoposterior margin of the large accessory photo-phore ventral to the orbit (Fig. 5). The resultingventral projections immediately penetrate the pho-tophore, sending projections posteriorly and anteri-orly into the glandular mass surrounding the coreand reflective layer. Dorsal to the core, more proxi-mal sections of these nerves anastomose with ante-rior fibers of the Tmdp. The most ventral fibers ofthe Tmx supply innervation to glandular tissue ofthe small photophore embedded in the ventrolat-eral margin or the larger, more dorsal photophore(Fig. 5A). Innervation of a core of this small photo-phore was not observed but rather only innerva-tion of outer tissues.

The profundal ramus of the trigeminal nerve(Tpr) emanates from the trigeminal foramen an-terior to the trigeminal trunk in Malacosteusand Photostomias (Figs. 2A and 3A) or directlyfrom the trigeminal trunk in Aristostomias andPachystomias (Figs. 4A and 5A), ventral to andin contact with the so. In Malacosteus, the Tprcourses anterodorsally, parallel to the so, alongthe ventrolateral face of the pterosphenoid, arch-ing anteroventrally over the frontal and supra-ethmoid, and eventually arching posteroventrallyover the ventral supraethmoid (Fig. 3A). The Tprcontinues on a posteroventral path medial to theposterior margin of the premaxilla, eventuallyterminating in innervation of tissues anterior tothe reflective material of the accessory photo-phore. The Tpr of Photostomias takes a more

direct, linear route over the prootic and supra-ethmoid, bifurcating over the supraethmoid andterminating in innervation of tissues of thesnout anterior to the supraethmoid and frontal(Fig. 2A).

The Tpr of Aristostomias and Pachystomiastakes a radically different course after departingfrom the so. In Aristostomias, the Tpr coursesanteriorly along the ventrolateral face of the pter-osphenoid to pass medial to the posterior processof the lateral ethmoid, making a sharp turn ven-trally, lateral to the medial base and anterior pro-cess of the lateral ethmoid (Fig. 4A). After passingmedial to the dorsal margin of the AO, the Tprproduces a small branch anteriorly that innervatestissues lateral to the anterior margin of supraetho-moid. The main trunk continues into the glandularmass of the AO, innervating the organ in a fern-like pattern that extends posteroventrally alongthe anterodorsal-posteroventral axis of the AO (seeFig. 4).

The Tpr of Pachystomias follows much the samecourse as that observed in Aristostomias, followingthe ventrolateral face of the pterosphenoid, ventro-lateral to the ventrolateral margin of the frontal(Fig. 5A). Immediately dorsal to the lateral eth-moid, the Tpr bifurcates, producing an anteriorlydirected branch that innervates tissues ventral tothe lateral ethmoid and dorsal to the parasphe-noid. The more posterior branch courses ventro-posteriorly, along the anteroventral margin of theorbit, to innervate the round accessory photophoredirectly anterior to the larger accessory photo-phore positioned ventral to the orbit. Fibers of theTpr innervate both the core of the AO and the sur-rounding minute glandular mass. No innervationseparate from this was observed for the minuteorgan embedded in the ventrolateral margin of thephotophore.

Phylogenetic Analysis

An unweighted exact parsimony analysis of 202parsimony informative characters in TNT resultedin four most parsimonious trees (length 5 503,CI 5 0.48, RI 5 0.76). Under implied weights, asingle tree of the same unweighted length wasrecovered (score 5 51.7). A strict consensus of thefour unweighted trees is presented and contrastedwith the weighted tree in Fig. 6. The unweightedanalysis recovered 19 of the 22 stomiid cladesrecovered by Fink (1985), including a loosejawclade. Under implied weighting, the loosejaw cladeis identically resolved. The most striking outcomeof these reanalyses is the resolution of the rela-tionships within the loosejaw clade. All analysesrecovered Pachystomias 1 (Aristostomias 1 (Mala-costeus 1 Photostomias)) albeit with low cladesupport and conflicting character evidence for theinternal nodes.

430 C.P. KENALEY


DISCUSSIONPhotophore Innervation

Innervation of the loosejaw cephalic photophoreshas been uncovered. The stomiid AO, PO, and SOare innervated solely by rami of the trigeminalnerve. The SO and PO of all taxa examined areinnervated by branches of the Tmd. The largeanterior far-red photophore of Pachystomias(Fig. 1D; ‘‘AOP III’’ of Fink, 1985; ‘‘preorbital’’ ofother authors) and adjacent minute photophoreare innervated by the Tpr (Figs. 4 and 5) and arehereafter treated as a single photophore. The far-red AO of Aristostomias (Fig. 1C; ‘‘AOP I’’ of Fink,1985; ‘‘suborbital’’ of most other authors) is alsoinnervated by the Tpr. The shortwave AO ofPhotostomias (Fig. 1A; ‘‘AOP IV’’ of Fink, 1985;‘‘preorbital’’ of other authors) and large posteriorfar-red AO of Pachystomias (Fig. 1D; ‘‘AOP I’’ ofFink, 1985; ‘‘suborbital’’ of most other authors) are

innervated by a branch of the Tmd (Figs. 2 and 5).The far-red AO of Malacosteus (Fig. 1B; ‘‘AOP I’’ ofFink, 1985; ‘‘suborbital’’ of other authors) andluminescent mass ventral to the posterior far-redphotophore in Pachystomias (Fig. 1D; ‘‘AOP II’’ ofFink, 1985) are innervated by branches of the Tmx(Figs. 3 and 5). A summary of loosejaw AO inner-vation patterns is presented in Figure 7.

Homology of Cephalic Photophores

On the basis of cranial nerve innervation uncov-ered in this study, the AOs of the stomiidae arenot homologous. In particular, the far-red AO ofMalacosteus is not homologous with that ofAristostomias nor to any of Pachystomias. The far-red AO of Malacosteus is innervated by the Tmx;however, the large single AO of Aristostomias andthe circular AO anteroventral to the orbit inPachystomias are innervated by the Tpr. The coreof the large sausage-shaped AO of Pachystomias,directly ventral to the orbit, (‘‘AOP I’’ of Fink,1985), is the only far-red AO innervated by theTmd. These dissimilar innervation patterns do not

Fig. 6. Strict consensus of four most parsimonious treesfrom an analysis of 202 unweighted morphological characterstaken from Fink’s (1985) analysis of 26 stomiid genera. Bremersupport values and parametric bootstrap values are indicatedabove and below branches, respectively. Bootstrap values below50 are omitted. Nodes labeled with a closed circle indicateclades recovered by Fink (1985). Gray, dashed branches repre-sent resolved clades recovered from an additional analysisunder implied weighting of the same dataset. The loosejawclade is indicated by red branches.

Fig. 7. Schematic drawing of position and innervation ofaccessory orbital photophores (AOs) of the loosejaw dragon-fishes. Far-red AOs are shown in red, short-wave AOs by pur-ple. The shortwave postorbital photophores, innervated by abranch of the mandibular ramus of the trigeminal nerve, areshown in blue.



permit conjecture of primary homology based ontopological similarity. In addition, Patterson’s(1982) conjunction criterion is invoked here as atool to detect non-homology. The occurrence of aan AO innervated by the profundal ramus andanother innervated by the mandibular ramus inPachystomias, photophores also present in Aristos-tomias and Photostomias, respectively, indicatesnon-homology of these structures.

An assessment of topological similarity and con-junction renders the loosejaw AO divisible intothree characters subject to further scrutiny, here-after referred to as the maxillary, profundal, andmandibular AOs (Table 2). Reconstructions of themandibular and profundal AOs on the topologyrecovered through phylogenetic analysis (see Fig. 6)reveals a single evolutionary origins or the profun-dal AOs under ACCTRAN optimization (Fig. 8A). Aprofundal AO is apomorphic for the loosejaw cladeand distally lost in a clade of Photostomias 1 Mala-costeus. The primary homology of the profundal AOof Pachystomias and Aristostomias survives thecongruence test, and thus, the homology of this pho-tophore cannot be rejected. Most parsimoniousreconstructions of mandibular and maxillary AOsdo not confirm homology at the secondary level. Themandibular AO of Pachystomias and Photostomiasand the maxillary AO of Pachystomias and Mala-costeus each appear as parallel transformations inreconstructions on a phylogeny of the loosejaws(Fig. 8B,C). Under the phylogenetic hypothesis pro-posed in this study, alternative reconstructions ofthe mandibular and maxillary AOs as apomorphicfor the loosejaw clade requires three transforma-tions, rather than two: a gain and two losses inAristostomias and Malacosteus for the mandibularAO and a gain and two losses in Aristostomias andPhotostomias for the maxillary AO.

Primary homology based on innervation patternsand subsequent tests of secondary homologyreduce the stomiid far-red AOs to three uniquestructures: (1) the profundal AO of Pachystomiasand Aristostomias; (2) maxillary AO of Malacos-teus, and (3) mandibular AO of Pachystomias.Based on this, it is hypothesized that photophorescapable of producing far-red, longwave lightevolved independently three times in theStomiidae.

The homology of the profundal AO hypothesizedhere would not be so under a different optimiza-tion criterion (i.e., delayed transformation orDELTRAN) for character reconstruction. In thereconstruction of the profundal AO, the applicationof ACCTRAN maintains the primary homologyhypothesis by accounting for homoplasy as a loss.Under DELTRAN, the reconstruction of thesecharacters accounts for homoplasy as parallelism,thereby introducing an ad hoc hypothesis of inde-pendent gains despite the contrary evidence ofshared innervation. On these grounds, ACCTRAN

is the preferable optimization and the homologyof the profundal AO the more conservativehypotheses.

Primary homology of the loosejaw SO ishypothesized on basis of shared innervation by themandibular ramus of the trigeminal nerve. Fink(1985) proposed that the SO was likely homologousin all stomiiform taxa, but the results of this studydo not fully corroborate this hypothesis. Handrick(1901) reported that the preorbital photophorein Argyropelecus hemigymnus (the purportedhomolog of the SO in all other stomiiform genera;Weitzman, 1974), is innervated by a projection ofthe Tmx. Gierse (1904) reported an identicalinnervation pattern for the SO of Cyclothoneacclinidens. Thus, based on these basal stomiiformtaxa, innervation of the SO by the maxillaryramus would appear to be the pleisiomorphic con-dition of stomiiforms. Innervation of the loosejawSO by the mandibular ramus of the trigeminalwould suggest that either (1) trigeminal neuronsof once maxillary peripheral distribution are pres-ent in the mandibular ramus of loosejaw species,rendering the stomiiform SO homologous, butinnervated by neurons of different peripheral posi-tions; or (2) that the SO of the loosejaws is nothomologous with the SO of basal stomiiforms, theformer being innervated by neurons of differentcentral origins. Without experimental labeling ofthe neurons in question, hypotheses of homology ofthe loosejaw SO with the SO of basal stomiiformscannot be completely rejected. Nevertheless, ifcomparative neuroanatomy is to be interpreted asprima fascia evidence of primary homology (as ithas in other phylogenetic studies, e.g., Thewissenand Babcock, 1991; Song and Boord, 1993; Songand Parenti, 1996; Nakae and Sasaki, 2004;Anderson, 2008), innervation of the SO of all loose-jaws by fibers of the Tmd implies nonhomology ofthe SO of basal stomiiform taxa.

The primary homology of the loosejaw PO ishypothesized on the basis of shared innervation bythe Tmdp. All stomiid taxa possess an organ ofsimilar position and bioluminescent capability(Fink, 1985; Herring, 2007). Indeed, the presenceof a postorbital photophore was described byHarold and Weitzman (1996) and Harold (1998) asa synapomorphy of the Photichthya (‘‘Phosich-tyidae’’ 1 Stomiidae) sensu Weitzman (1974).Because these authors reported no intrinsic evi-dence, such as innervation or histology, that estab-lished the primary homology of the photichthyanPO and the current study focused on only four sto-miid genera, the homology of this organ acrossphotichthyan taxa cannot be confirmed. However,as far as species of the Stomiidae are concerned,given innervation patterns uncovered in this study,and its similar size, position, and function, thehypothesis that the stomiid PO is a homologousorgan cannot be rejected.

432 C.P. KENALEY


TA

BL

E2.

Pro

pos

edte

rms

an

dsy

non

ymy

oflo

osej

aw

acc

esso

ryp

hot

oph

ores

Pro

pos

edte

rmIn

ner

vati

onT

axon

Sh

ap

ean

dP

osit

ion

Col

or/k

max

offl

oure

scen

cea

Syn

onym

Au

thor

s

Pro

fun

dal

Acc

esso

ryO

rbit

al

Ph

otop

hor

e(A

Op

r)

Pro

fun

dal

ram

us

oftr

igem

inal

Ari

stos

tom

ias

Tea

r-d

rop

,an

tero

ven

tral

marg

inof

orbit

(Fig

s.1C

,4,

8A

)

Red –

500,�

585,

605,

640–650

nm

‘‘Su

bor

bit

al

Org

an

’’(S

O)

Her

rin

g,

2007;

Her

rin

gan

dC

ope,

2005;

Men

sin

ger

an

dC

ase

,1997;

Mor

row

,1964;

O’D

ay

an

dF

ern

an

dez

,1974;

Wid

der

etal.

,1984

‘‘Acc

esso

ryO

rbit

al

Ph

otop

hor

eI’

’(A

OP

I)

Fin

k,

1985

Pa

chys

tom

ias

Cir

cula

r,an

tero

ven

tral

toor

bit

(Fig

s.1D

,5,

8A

)R

ed –615,

640,

680

nm

‘‘Pre

orbit

al

organ

’’‘‘A

cces

sory

Orb

ital

Ph

otop

hor

eII

I’’

(AO

PII

I)

Her

rin

g,

2007;

Her

rin

gan

dC

ope,

2005

Fin

k,

1985

Maxil

lary

Acc

esso

ryO

rbit

al

Ph

otop

hor

eI

(AO

mx

I)

Maxil

lary

ram

us

oftr

igem

inal

Ma

laco

steu

sT

ear-

dro

p,

an

tero

ven

tral

marg

inof

orbit

(Fig

s.1B

,3,

8C

)

Red –

710

nm

‘‘Su

bor

bit

al

Org

an

’’(S

O)

Her

rin

g.

2007;

Her

rin

gan

dC

ope,

2005;

Den

ton

etal.

,1985;

Dou

gla

set

al.

,1998,

2000;

Mor

row

,1964;

Su

tton

,2005;

Wid

der

etal.

,1984

‘‘Acc

esso

ryO

rbit

al

Ph

otop

hor

eI’

’(A

OP

I)

Fin

k,

1985

‘‘Acc

esso

ryO

rbit

al

Ph

otop

hor

e’’(A

O)

Ken

ale

y,2007

Maxil

lary

Acc

esso

ryO

rbit

al

Ph

otop

hor

eII

(AO

mx

II)

Maxil

lary

ram

us

oftr

igem

inal

Pa

chys

tom

ias

Oblo

ng,

dir

ectl

yven

tral

toA

Om

d(F

ig.

1D

,5,

8C

)

Red –

607,

663

nm

‘‘Acc

esso

ryO

rbit

al

Ph

otop

hor

eII

’’(A

OP

II)

Fin

k,

1985

Man

dib

ula

rA

cces

sory

Orb

ital

Ph

otop

hor

eI

(AO

md

I)

Man

dib

ula

rra

mu

sof

trig

emin

al

Pa

chys

tom

ias

Oblo

ng,

sau

sage-

shap

ed,

dir

ectl

yven

tral

toor

bit

(Fig

s.1D

,5,

8B

)

Red –

590,

650

nm

‘‘Su

bor

bit

al’’

(SO

)H

erri

ng,

2007;

Her

rin

gan

dC

ope,

2005;

Mor

row

,1964;

Wid

der

etal.

,1984

‘‘Acc

esso

ryO

rbit

al

Ph

otop

hor

eI’

’(A

OP

I)

Fin

k,

1985

Man

dib

ula

rA

cces

sory

Orb

ital

Ph

otop

hor

eII

(AO

md

II)

Man

dib

ula

rra

mu

sof

trig

emin

al

Ph

otos

tom

ias

Oblo

ng

orci

rcu

lar,

an

tero

ven

tral

toor

bit

(Fig

s.1A

,2,

8B

)

Vio

let

–N

od

ata

‘‘Pre

orbit

al

ph

otop

hor

e’’

‘‘Acc

esso

ryO

rbit

al

Ph

otop

hor

eIV

’’(A

OP

IV)

Her

rin

g,

2007

Fin

k,

1985

‘‘Acc

esso

ryO

rbit

al

Ph

otop

hor

e’’(A

O)

Ken

ale

y,2009;

Ken

ale

yan

dH

art

el,

2005

aD

ata

from

Her

rin

gan

dC

ope

(2005).



The homology statements proposed in this studyare not final. As hypotheses of homology, thesestatements may be falsified by new datasets. Forexample, several neuroanatomical charactersdescribed in this study may, with outgroup com-parison, be useful in generating additional phylo-genetic hypotheses for the loosejaw genera.

Although fully resolved by a reanalysis of Fink’s(1985) dataset, the relationships of among loosejawgenera are weakly supported (see Fig. 6). A phylo-genetic analysis of these neuroanatomical charac-ters, or some other new dataset (e.g., molecularsequence data), may result in a different phyloge-netic hypotheses, and thus permit a reinterpreta-

Fig. 8. Reconstructions of accessory orbital (AO)characters on the strict consensus tree from Fig. 6. A:Profundal AO under accelerated transformation. B:Mandibular AO. C: Maxillary AO. Mapped AOs of theloosejaw taxa are indicated by color above the tree:red in Aristostomias, Malacosteus, Pachystomias, vio-let in Photostomias.

434 C.P. KENALEY


tion of secondary homology of the loosejaw AOs.Because the primary homology of the loosejaw AOhas been rejected, a topology different than thatpresented here will have no bearing on the hypoth-esis of parallel evolution of far-red bioluminescencein the group.

Additional Evidence for Parallel Evolutionof Far-Red Bioluminescence

The patterns of innervation described and phylo-genetic relationships inferred in this study suggestmultiple evolutionary origins of long-wave biolumi-nescence in the Stomiidae, specifically that theAOs of Aristostomias and Pachystomias are not ho-mologous with the AO of Malacosteus. The neuroa-natomical evidence for multiple origins of long-wave bioluminescence in the group is corroboratedby Herring and Cope’s (2005) histological data.The AOs of Aristostomias and Pachystomias areconsiderably different in their structure than theAO of Malacosteus. The former are composed of atranslucent aperture that leads internally to a coresurrounded by a crystalline reflective layer that is,in turn, surrounded by a large glandular massthat dominates the volume of the photophore;strands of the gland perforate the reflective layer,communicating with the core (Herring and Cope,2005). The maxillary AO of Malacosteus, however,is much simpler and more similar to the PO inthat it is comprised of a glandular core surroundedby a pigmented sack that is in turn lined by areflective layer. In addition, unlike the AOs ofPachystomias and Aristostomias, cells of the aper-ture of the maxillary AO of Malacosteus are heav-ily pigmented (Herring and Cope, 2005). Fluores-cent emissions of the exposed inner core of the AOof Malacosteus have a much shorter maximum(approximately 650 nm) than the surface of anintact photophore (approximately 710 nm; Herringand Cope, 2005). Thus, as proposed by Dentonet al. (1970, 1985), it appears that these heavilypigmented cells in the aperture act as a filter thatconverts shorter-wave emissions from the core tolong-wave emissions. In contrast, the cores of theAOs of Pachystomias and single AO of Aristosto-mias have fluorescence emission maxima thatdiffer little from the emission maxima at their sur-face. These histological, structural, and biochemi-cal properties shared by the photophores of Aris-tostomias and Pachystomias and those unique tothe far-red photophore of Malacosteus can be seenas special qualities (sensu Remane, 1952; Rieppeland Kearney, 2002) that distinguish far-red photo-phores further and corroborate the rejection of pri-mary homology based on innervation patterns.

Herring and Cope (2005) reported that both theAOs and POs of Malacosteus, Aristostomias, andPachystomias were innervated by branches of thetrigeminal nerve; however, no histological evidence

was provided or sources cited. Apart from thisstatement, innervation of the stomiid PO or AOhas not been described in the literature. Despitethe unclear provenance of their statement, theresults of this study, specifically that all cephalicphotophores are innervated by rami of the trigemi-nal, do not conflict with the innervation describedby Herring and Cope (2005).

Although the focus of their work was not toassess the homology of the loosejaw AOs, Herringand Cope (2005) concluded that a common originof far-red AOs in Aristostomias and Pachystomiasis almost certain. The evidence presented hereboth corroborates and conflicts with their state-ment. Aristostomias and Pachystomias share a ho-mologous profundal AO, organs of similar positionsbut dissimilar shape and size (Fig. 1C,D; Table 2).However, the mandibular AO of Pachystomias, anorgan with a shape similar to that of the profundalAO in Aristostomias, has no counterpart in Aris-tostomias, nor in any other stomiid taxon. In sum-mary, the AOs of similar size, but of dissimilarposition (homologous organs according to Herringand Cope, 2005), are nonhomologous; however, theorgans of similar position but of dissimilar size arelikely homologous (Fig. 8A).

Implications for Terminology

Fink’s (1985) photophore terminology has, inlarge part, gone unadopted. Although mostsystematists avoid the terms, ‘‘suborbital’’ and‘‘preorbital’’ are used universally by authors inves-tigating the optical, biochemical, and structuralproperties of these organs. Although the use ofthese terms has remained largely consistent, theiruse is apt to cause confusion in the best case andimply homology with dissimilar structures in theworst. For instance, if a reader encounters theterm ‘‘suborbital’’ in reference to the far-red AO ofAristostomias, and later, in the same or some otherwork, encounters the term in reference to thesmall short-wave organ present along the ventralmargin of the orbit in a more primitive stomiiformspecies, the reader may logically assume theseorgans share structural and functional similarity,as well as phylogenetic history. Thus, if anatomicalterminologies exist to facilitate unambiguous com-munication about similar structures, the use of‘‘suborbital’’ in reference to any stomiid AO isinappropriate. Fink’s (1985) terminology should beapplied to the stomiid AOs with slight modifica-tions based on the results of this study. Reflectingtheir shared innervation and test of congruenceunder the phylogenetic hypothesis proposed in thisstudy, the following terms are proposed (Table 2):The AO of Aristostomias and the anterior AO ofPachystomias should be referred to as the profun-dal AO (AOpr); the AO of Photostomias as mandib-ular AO I (AOmd I); the largest AO of Pachysto-



mias as mandibular AO II (AOmd II); the AO ofMalacosteus as maxillary AO I (AOmx I); and theAO positioned directly under mandibular AO II ofPachystomias as maxillary AO II (AOmx II).

ACKNOWLEDGMENTS

Courtney Kenaley, K. Liem, and T. Pietsch weresources of helpful discussion and consultation con-cerning comparative neuroanatomy. K. Hartel, P.Møller, and K. Maslenikov provided specimens andcuratorial support. T. Pietsch, J. Orr, and D.Stevenson read early versions of the manuscriptand provided insightful comments. I thank W.Fink and E. Hilton for criticial reviews thatimproved the manuscript. P. Herring provided theimage in Fig. 1A; E. Widder provided images inFig. 1B–D.

LITERATURE CITED

Agnarsson I, Miller JA. 2008. Is ACCTRAN better thanDELTRAN? Cladistics 24:1032–1038.

Anderson PSL. 2008. Cranial muscle homology across moderngnathostomes. Biol J Linn Soc 94:195–216.

Brauer A. 1904. Uber die leuchtorgane der knochenfische. VerhDtsch Zoologischen Ges 14:16–35.

Brauer A. 1908. Die Tiefsee-Fische II. Anatomischer Teil. WissErgeb Dtsch Tiefsee-Exped Dampfer ‘‘Valdivia’’ (1898–1899)15:1–266.

Bremer K. 1994. Branch support and tree stability. Cladistics10:295–304.

Brower AVZ, Schawaroch V. 1996. Three steps of homologyassessment. Cladistics 12:265–272.

Campbell AK, Herring PJ. 1987. A novel red fluorescent proteinfrom the deep-sea luminous fish Malacosteus niger. Comp Bio-chem Physiol B Biochem Mol Biol 86:411–417.

de Pinna MCC. 1991. Concepts and tests of homology in thecladistic paradigm. Cladistics 7:367–394.

Denton EJ, Gilpin-Brown JB, Wright PG. 1970. On the ‘filters’in the photophores of mesopelagic fish and on a fish emittingred light and especially sensitive to red light. J Physiol 208:72P–73P.

Denton EJ, Herring PJ, Widder EA, Latz MF, Case JF. 1985.The roles of filters in the photophores of oceanic animals andtheir relation to vision in the oceanic environment. Proc RSoc Lond B Biol Sci 225:63–97.

Dingerkus G, Uhler LD. 1977. Enzyme clearing of alcian bluestained whole small vertebrates for demonstration of carti-lage. Stain Technol 52:229–232.

Douglas RH, Mullineaux CW, Partridge JC. 2000. Long-wavesensitivity in deep-sea stomiid dragonfish with far-red biolu-minescence: Evidence for a dietary origin of the chlorophyll-derived retinal photosensitizer of Malacosteus niger. PhilosTrans R Soc Lond B Biol Sci 355:1269–1272.

Douglas RH, Partridge JC, Dulai K, Hunt D, Mullineaux CW,Tauber AY, Hynninen PH. 1998. Dragon fish see using chloro-phyll. Nature 393:423–424.

Felsenstein J. 1985. Confidence limits on phylogenies: Anapproach using the bootstrap. Evolution 39:783–791.

Fernandez HRC. 1979. Visual pigments of bioluminescent andnonbioluminescent deep-sea fishes. Vis Res 19:589–592.

Filipski GT, Wilson MVH. 1984. Sudan black-B as a nerve stainfor whole cleared fishes. Copeia 1984:204–208.

Fink WL. 1985. Phylogenetic interrelationships of the stomiidfishes (Teleostei: Stomiiformes). Misc Publ Mus Zool UnivMich 171 (i–vii):1–127.

Gierse A. 1904. Untersuchungen uber das Gehirn und die Kopf-nerven von Cyclothone acclinidens. Tafeln MorphologischeJahrbucher 32:602–608.

Goloboff PA. 1993. Estimating character weights during treesearch. Cladistics 9:83–91.

Goloboff PA, Carpenter JM, Arias JS, Esquivel DRM. 2008a.Weighting against homoplasy improves phylogenetic analysisof morphological data sets. Cladistics 24:758–773.

Goloboff PA, Farris JS, Nixon KC. 2008b. TNT, a free programfor phylogenetic analysis. Cladistics 24:774–786.

Handrick K. 1901. Zur Kenntnis des Nervensystems und derLeuchtorgane des Argyropelecus hemigymnus. Zool H 32:602–688.

Harold AS. 1998. Phylogenetic relationships of the Gonostoma-tidae (Teleostei: Stomiiformes). Bull Mar Sci 62:715–741.

Harold AS, Weitzman SH. 1996. Interrelationships of stomii-form fishes. In: Stiassny MLJ, Parenti LR, Johnson GD,editors. Interrelationships of Fishes. San Diego: AcademicPress. pp 333–353.

Herring PJ. 1983. The spectral characteristics of luminous ma-rine organisms. Proc R Soc Lond B Biol Sci 220:183–217.

Herring PJ. 2000. Species abundance, sexual encounter and bio-luminescent signaling in the deep sea. Philos Trans R SocLond B Biol Sci 355:1273–1276.

Herring PJ. 2007. Sex with the lights on? A review of biolumi-nescent sexual dimorphism in the sea. J Mar Biol Assoc UK87:829–842.

Herring PJ, Cope C. 2005. Red bioluminescence in fishes: Onthe suborbital photophores of Malacosteus, Pachystomias andAristostomias. Mar Biol 148:383–394.

Kenaley CP. 2007. Revision of the stoplight loosejaw genusMalacosteus (Teleostei: Stomiidae: Malacosteinae), withdescription of a new species from the temperate southernhemisphere and Indian Ocean. Copeia 2007:886–900.

Kenaley CP. 2009. Revision of Indo-Pacific species of the loose-jaw dragonfish genus Photostomias (Teleostei: Stomiidae:Malacosteinae). Copeia 2009:175–189.

Kenaley CP, Hartel KE. 2005. A revision of Atlantic species ofPhotostomias (Teleostei: Stomiidae: Malacosteinae), with adescription of a new species. Ichthyol Res 52:251–263.

Mensinger AF, Case JF. 1988. Dual role of the orbital photo-phores in the stomiatoid fish, Malacosteus niger. Am Zool 28:90A.

Mensinger AF, Case JF. 1997. Luminescent properties of fishesfrom nearshore California basins. J Exp Mar Biol Ecol 210:75–90.

Morrow JE Jr. 1964. Family Malacosteidae. In: Bigelow HB,Breder CM, Olsen YH, Cohen DM, Schroeder WC, Mead GW,Schultz LP, Merriman D, Tee-Van J, editors. Fishes of theWestern North Atlantic 1, pt. 4. New Haven: Sears Founda-tion for Marine Research. pp. 523–549.

Nakae M, Sasaki K. 2004. Homologies of the adductor mandibu-lae muscles in Tetraodontiformes as indicated by nervebranching patterns. Ichthyol Res 51:327–336.

Nelson JS. 2006. Fishes of the World. New York: Wiley. p 624.Nicol JAC. 1957. Observations on photophores and lumines-

cence in the teleost Porichthys. Q J Microsc Sci 98:179–188.Nicol JAC. 1960. Studies on luminescence on the subocular

light-organs of stomiatoid fishes. J Mar Biol Assoc UK 39:529–548.

Northcutt RG, Holmes PH, Albert JS. 2000. Distribution andinnervation of lateral line organs in the channel catfish.J Comp Neurol 421:570–592.

O’Day WT, Fernandez HR. 1974. Aristostomias scintillans(Malacosteidae): A deep-sea fish with visual pigments appa-rently adapted to its own bioluminescence. Vis Res 14:545–550.

Parin NV, Borodulina OD. 1993. A new mesobenthic fish. Eupo-gonesthes xenicus (Astronesthidae), from the Eastern IndianOcean. Voprosy Ikhtiologii 33:442–445.

Parin NV, Borodulina OD. 2004. Phylogeny, systematics, andzoogeography of the mesopelagic genus Astronesthes (Astro-nesthidae. Stomiiformes). J Ichthyol 43:557–576.

436 C.P. KENALEY


Partridge JC, Douglas RH. 1995. Far-red sensitivity of dragonfish. Nature 375:21–22.

Patterson C. 1982. Morphological characters and homology. In:Joysey KA, Friday AE, editors. Problems of PhylogeneticReconstruction. London: Academic Press. pp 21–74.

Piotrowski T, Northcutt RG. 1996. The cranial nerves of theSenegal bichir. Polypterus senegalus [Osteichthyes: Actino-pterygii: Cladistia]. Brain Behav Evol 47:55–78.

Puzdrowski RL. 1989. Peripheral distribution and central pro-jections of the lateral-line nerves in goldfish. Carasius aura-tus. Brain Behav Evol 34:121–131.

Ray DL. 1950. The peripheral nervous system of Lampanyctusleucopsarus. J Morphol 87:61–178.

Regan CT, Trewavas E. 1930. The fishes of the families Stomia-tidae and Malacosteidae. Danish Dana Expedition 1920(22 6):1–143.

Remane A. 1952. Die Grundlagen des Naturlichen Systems derVergleichenden Anatomie und der Phylogenetik. Leipzig:Geest und Portig K.G.

Riedl R. 1978. Order in living organisms: a system analysis ofevolution. New York: Wiley.

Rieppel O, Kearney M. 2002. Similarity. Biol J Linn Soc 75:59–82.

Somiya H. 1982. Yellow lens eyes of a stomiatoid deep-dea fish,Malacosteus niger. Proc R Soc Lond B Biol Sci 215:481–489.

Song J, Boord RL. 1993. Motor components of the trigeminalnerve and organization of the mandibular arch muscles invertebrates—phylogenetically conservative patterns and theirontogenic basis. Acta Anat 148:139–149.

Song J, Northcutt RG. 1991. Morphology, distribution andinnervation of the lateral-line receptors of the Florida gar,Lepisosteus platyrhincus. Brain Behav Evol 37:10–23.

Song J, Parenti LR. 1996. Phylogenetic significance of the pec-toral-pelvic fin association in Acanthomorph fishes: A reas-

sessment using comparative neuroanatomy. In: StiassnyMLJ, Parenti LR, Johnson GD, editors. Interrelationships ofFishes. San Diego: Academic Press. pp 427–444.

Song JK, Parenti LR. 1995. Clearing and staining whole fishspecimens for simultaneous demonstration of bone, cartilage,and nerves. Copeia 1995:114–118.

Strum JM. 1969. Photophores of Porichthys notatus: Ultrastruc-ture of innervation. Anat Rec 164:463–477.

Sutton TT. 2005. Trophic ecology of the deep-sea fish Malacos-teus niger (Pisces: Stomiidae): An enigmatic feeding ecologyto facilitate a unique visual system? Deep Sea Res Part I:Oceanogr Res Pap 52:2065–2076.

Sutton TT, Hopkins TL. 1996. Trophic ecology of the stomiid(Pisces: Stomiidae) fish assemblage of the eastern Gulf ofMexico: Strategies, selectivity and impact of a top mesopela-gic predator group. Mar Biol 127:179–192.

Thewissen JG, Babcock SK. 1991. Distinctive cranial and cervi-cal innervation of wing muscles: New evidence for bat mono-phyly. Science 251:934–936.

von Lendenfeld R. 1887. Report on the structure of the phos-phorescent organs of fishes. Challenger reports. Zoology 22(Appendix B):277–329.

Warrant EJ, Locket NA. 2004. Vision in the deep sea. Biol Rev79:671–712.

Weitzman SH. 1974. Osteology and evolutionary relationshipsof the Sternoptychidae, with a new classification of stomiatoidfamilies. Bull Am Mus Nat Hist 153:327–478.

Widder EA, Latz MI, Herring PJ, Case JF. 1984. Far red biolu-minescence from two deep-sea fishes. Science 225:512–514.

Winterbottom R. 1974. A descriptive synonymy of the striatedmuscles of the Teleostei. Proc Acad Nat Sci Philadelphia 125:225–317.



Documents

Comparative innervation of cephalic photophores of …ckenaley/Kenaley_2010_FarRedHomology.pdfComparative Innervation of Cephalic Photophores ... (Teleostei: Stomiiformes: Stomiidae):