THE OSTEOLOGY AND RELATIONSHIP OF THE PERCOLDEAN FISH, DINOLESTES LEWINI.
By Edwin Chapin Starks,Assisiant Professor in the Department of Biology and Curator of the Museum, Universityof Washington, Seattle, Washington.
As a preliminary to this paper I will quote a paragraph, from a pa])erby Dr. Theodore Gill, " On the identity of Esox lewini with the Dino-lestes millleri of Klnnzinger," published about twenty-five years agoin the Annals and Magazine of Natural History. In it he gives thefollowing concise history of Dinolestes:In the tenth volume ("the class Pisces") of the "Animal Kingdom" of Cuvier,edited by Edward Griffith (1834), are a figure (plate 60) and a brief notice (p. 465)of a fish which has long been a puzzle to me. It is called Esox lewini and only-noticed as follows : " Our figure of Esox letvini is from a drawing by Mr, Lewin, madein New Holland, of a species not hitherto noticed." It was evident that the speciesthus named belonged neither to the genua -Esox nor anywhere near it; and (1) therelations of the fins, (2) the position of the ventrals with a spine and five rays each,(3) the form of the head, and (4) the teeth indicated for it affinity to Chilodipterus andallied forms; but no first dorsal fin was represented. The question then arosewhether that fin had been atrophied (as in Aspidophoroides, GoMopus, etc.) or (as wasmore likely) had been overlooked. After nearly forty years the species has beenrecovered, and singularly enough, after having escaped the observation of the num-erous collectors in the Australian seas for so long a time,' has in the same year beenobtained and described by three difterent naturalists under as many names, viz, Dino-lestes millleri by Klunzinger, Neosphyrcena multiradiata by Castelnau, and Laniopercamordax by Giinther. There can be no question about at least the generic identity ofthe Esox lewini with the fishes described by the three contemporaries ; and it nowappears that the first dorsal fin exists, but is quite small, and sustained by only fouror five spines. Klunzinger and Castelnau refer the type to the family Sphyrsenidse,and Giinther (with more justification I think) to the "Apogonina," i. e., Chilodip-teridse.This work was undertaken at Dr. Gill's suggestion in the hope offinding, in a comparative study of the skeleton of Dinolestes with thoseof the Sphyrsenidse and the Cheilodipteridce, some characters that
' It is said, however, by Castelnau to be common [at Melbourne] in the months ofMay, June, and July ; it attains 2 feet in length. The fishermen call it " Shij) Jack,"but that name is more particularly applied to Temnodon saltator.Proceedings U. S. National Museum, Vol. XXII-No. 1186.Proc. N. M. vol. xxii 8 113
114 PROCEEDINGS OF THE NATIONAL MUSEUM. vol.xxh.would decide the question on which the authorities disagreed, as shownin the last sentence quoted from the above paper.It would appear an easy matter to place this form under one or theother of two families arranged in different suborders, as the Sphyrsen-idse and Oheilodii^teridae usually are: but the fact that two of theauthorities consider it under one family and two under the other, indi-cates how close the resemblance must be to either.Sphyrmna argentea is the form chosen to represent the family Sphy-rsenidse, while the only representative procurable of the Oheilodipter-idse was Apogon maculatiis.The skeleton of Apogon was found to be of but little assistance in thiscomparison. Though being undoubtedly Percoid, it differs as muchfrom Dinolestes as they both differ from the more generalized Percoids,such as the bass or perch. Considering the difference between Dino-lestes and Apogon in external appearance, we have probably little reasonto expect the internal resemblance to be otherwise. Perhaps if one ofthe 8p]iyrwna-]ike Oheilodipteroids could be examined there would bea closer resemblance.To be sure Dinolestes differs from Apogon only in shape and compar-ative size of elements (that is, comparative between correspondingelements of each species), and not in arrangement, or lack or posses-sion of elements; but it is so very different in shape of cranium, formof body, and shape of fins that it would seem better to consider theCheilodipteroid side of the question partly by considering the Percoidfishes more or less as a whole. Though, of course, if it is ijlaced withthe Percoids it is only under the family Oheilodipteridse, as the Per-coids are now arranged, that Dinolestes could be admitted.Were it not that the ventrals oi Dinolestes are apparently thoracic, itmight appear after a superficial external examination to be related toSphyrcena. The long head, projecting lower jaw, fanglike teeth, andelongate preorbital region are very Sphyrcenalike. Q'he shape of thebody and dorsal fins are also suggestive of that relationship.In internal characters we find that the ethmoid is wide and flat, some-what overlying the vomer and prefrontals instead of being interposedbetween them. Tbis is the condition found in iSphyrcena. The nasalsare very much like those of Sphyrcena, being long and channeled andattached by their sides to the ethmoid for nearly their whole length.This, however, is probably caused by the elongate snout, and goes withit as a part in keeping with the surrounding conditions. It disposes ofthe characters by which an alliance with Sphyrcena covtM be proved.Though the shajDC of the body and head, the canine teeth, and dorsalfins exhibit perhaps a closer superficial resemblance to the Sphyraenidsethan to the Cheilodipteridiie, there are forms to be found in the latterfamily which approach this Sphyrcena-like appearance also. Thesecharacters therefore denote nothing in favor of either relationship.The characters of the ethmoid and nasals are the only characters pos-sessed by Sphyrcena in common with Dinolestes that are not also shared
NO. 1186. RELATIONSHIP OF DINOLESTES LEWINI?STAEES. 115in by members of the family Cheilodiptericlne. They are of uo greatimportance.This, as has been said, is without considering the position of theventrals. As the Percoid fishes have thoracic ventrals, and the mem-bers of the suborder Percesoces, under which the Sphyrtenidce isplaced, have abdominal ventrals, it is cUfiQcult to see how Binolestescould have been thought to be related to Sphyrcvna unless the ventralswere interpreted as being abnormally anterior abdominal ventrals.The ventrals, however, prove to be typical thoracic ventrals with theanterior point of the pelvic girdle interposed and attached betweenthe opposing clavicles near their lower end above their symphysis.Anterior abdominal ventrals might have the point of the pelvic g-irdletouching* the clavicles or even extending slightly under them, butnever interposed between them.Another character that refutes the Sphyraenoid relationship of I) ino-lestes is the lack of the long processes developed backward from theepiotics and supraoccipital crest, which are possessed in a greater orless degree by all the Percesoces and reaches its greater developmentin Sphyrcena.A more important difference is the structure of the teeth. Thoughboth Binolestes and Sphyrcena have large backward-directed canines,they are entirely different in the way in which they are attached tothe bone of the jaw. The calcified tooth substance of the teeth ofBinolestes reaches only to the bone, where it is anchylosed or so incor-porated with the bone as to make it difficult to distinguish the line ofjunction. This attachment is effected by what Tomes calls "bone ofattachment." A substance resembling cement, but unlike true cement,is developed from the periosteum rather than from the dental capsule.Such teeth may be developed from sockets, but as the calcified toothsubstance is jjushed out the cavity behind fills with the bone of attach-ment and becomes obliterated. The dentine never extends into acavity of the bone in the mature tooth. This is a Percoid character.The teeth of Sphyrcena, on the other hand, are set in sockets. Theattachment of teeth in alveoli is of such rare occurrence among fishesand must be so deep seated that we can hardly interpret it as less thana family character. Mr. W. G-. Eidewood, in a paper,' has this to sayin regard to this class of teeth
:
The tooth and bone are in organic continuity by means of a periosteal layer com-mon to the tooth and the jaw; and this layer may remain nncalcified so that theteeth can be pulled out of their sockets, as in some Caracinoid fishes; or "bone ofattachment" may, except in young teeth, anchylose the tooth to the wall of thesocket, e. g,, Sphyriena.But whether or not the tooth becomes cemented in, the enamel anddentine extend into a cavity and do not become incorporated with thebone.
'Natural Science, YIII, June, 1896, p. 383.
116 FEOCEEDINGS OF THE NATIONAL MUSEUM.There is also a difference in the shape of the teeth of Sphyrwna andDinolestes. Those of the former are laucelike?that is, compressedlaterally and with cutting edges. Those of the latter are round intransverse section.The vertebrae of Dinolestes are typical, or iu general resembling thePercoids and most bouy fishes iu that they are of moderate length andnot much constricted in the middle. They have parapophyses devel-oped behind the fourth vertebra, two pits on the side of each vertebraseparated by a longitudinal ridge, and the abdominal vertebraj with apit on the ventral side with ridges on each side of it. The vertebrae ofSphyraina are long and smooth with scarcely any pits, much constrictedin the middle, making them hourglass sliaped, and with only one ortwo pairs of parapophyses.The shape of the cranium of Dinolestes is also more typically Percoidin appearance than SphyrfFuoid with the slightly rising supraoccipitalcrest and more wedge-shaped lateral view.A recapitulation with these points condensed will show at once theaflBnity of Dinolestes to the Cheilodipterid?e.1. The Percoid appearing cranium.2. The thoracic ventrals.3. The anchylosed teeth rather than teeth in sockets.4. The character of the vertebrii^, typical; not specialized as inSpliyrcvna.5. The lack of the long posterior i)rocesses from the epiotics.These conclusions are fortified by the following description.DINOLESTES LEWINI.Esox lewini Griffith (?), Cuvier's Animal Kingdom, Griffith eel., X (1834),p. 46,5, pi. 60.Dinolestes miiUeri Klunzinger, Arcliiv fiir Nat., 38. Jahrg. 1 (1872), p. 30; HobsonBay, South Australia.Neosphyrwna muUiradiata Castelnau, Proc. Zool. and Acclim. Soc. Victoria, I(1872), p. 96; Melbourne.LaiHoperca mordax Gunther, Ann. and Mag. Nat. Hist., 4th ser., X, (September,1872), p. 183; Tasmania.DinoJesies lewini Gill, Ann. and Mag. Nat. Hist., 4th ser., XIV (1874), p. 160.DIAGNOSIS.Body rather elongate ; preorbital region produced ; mouth large, thelower jaw projecting; canine teeth on lower jaw posteriorly and onpremaxillaries at their symphysis; sharp, cardiform teeth in a singlerow on jaws, vomer and palatines; an inner row of villiform teeth onpremaxillaries; three toothed superior pharyngeals; lower pharyngealsrather narrow, separate j gill rakers long and slender, about 4 + 135opercles without spines or ridges; branchiostegals 7; maxillary withsupplemental bone; nasals elongate, attached by their sides for nearlytheir whole length; parietals separated; ethmoid somewhat overlyingprefrontals and vomer j posttemporal forked ; postclavicle of two parts;
NO. 1186. BELA TIONSHIP OF DINOLESTES LEWINI?STARKS. 117basispheuoid with, a descendiug process; myodome present with asmall pore to the exterior posteriorly; no suborbital shelf; vertebrae27; parapophyses present on all abdominal vertebnie except first 3;scales cycloid; maxillaries, cheeks, opercles, and lower jaw with scales;lateral line straight, ranning well out on base of caudal, scales alongits course systematically crowded; base of anal, soft dorsal, and caudalwith small scales; anal with 2 spines and about 25 rays; dorsals remote,the first of about 5 slender spines, the second with 1 spine and about18 rays; ventrals with 1 spine aud 5 rays each; inters pinous bones notextending between vertebral spines.OSTEOLOGY.Cranium, as viewed from above, rather elongate and narrow. Inter-orbital region a flattish area with the sides nearly parallel the greaterpart of its length, occupying at least a third of length of cranium.Preorbital region elongate, not tapering till near end, occupyinganother third of length of cranium. Region surrounding foramenmagnum slightly produced. Lateral view of cranium taj)ering ratherregularly to vomer.Supraoccipital interposed between exoccipitals nearly to their poste-rior ends; its crest developed superiorly and posteriorly, scarcelyextending past exoccipitals iDosteriorly.Exoccipitals broadly meeting above basioccii^ital, entirely surround-ing foramen magnum.Parietals widely separated by supraoccipital, not extending overepiotics.Bpiotics with scarcely any process.Prootics, opisthotics, sphenotics, and pterotics typical; that is,Percoid-like.Alispheuoids widely separated. The anterior opening into braincase large.Basispheuoid present; a foramen between it and basis crauii. Aslender process descending from it and attached to parasphenoid,Myodome present; opening to the exterior at its posterior endthrough an extremely small foramen.Parasphenoid spreading out wide ])osteriorly.Vomer bearing sharp short teeth in a single row around its anterioredges. Teeth becoming smaller anteriorly.Prefrontals large and rather elongate, the olfactory foramen scarcelybehind middle.Ethmoid entirely superior to prefrontals aud vomer, widely overly-ing them aud extending to edge of rostrum, A raised area along itsmiddle.Nasals thin elongate rods of bone attached by their sides to ethmoidfor nearly their whole length; their length over a fourth that ofcranium.Preorbital longer than wide; its outline triangular.
118 PROCEEDINGS OF THE NATIONAL MUSEUM. vouxxii.Suborbital ring of the usual number of five bones with a sensorycanal through them. No suborbital shelf.^Opercle without ridges or spines on outer surface. On inner sur-face a sharp horizontal ridge runs posteriorly from its condyle withhyomandibular.Subopercle extending around lower corner of opercle, upward andbackward, forming lower part of posterior opercular angle.Intero])ercle very broadly attached to subopercle at its upper posteriorside.Preopercle with ridge and sensory canal as usual.Lower limb of hyomandibular rather long and rod-like.Palato-pterygoid process very long and stout. A single row of smallteeth along lower edge extending anteriorly upon a process beyondmain part of palatine somewhat similar to the usual process from upjieredge of that bone. Suspensorium otherwise typical.Lower jaw heavy and long. The articular half as wide as long.Teeth in a single row upon dentary, three or four canines presentposteriorly.Angular present, rather small.Maxillary with long supplemental bone along posterior edge.Premaxillary rather slender, much widened at middle into a wideprocess which extends behind maxillary; its lower end very slender.A single row of elongate pointed or small canine teeth along its edge,largest medially, growing gradually smaller toward each end. Insideof this row a villiform band, widest medially. At upper end of eachmaxillary are two large canines anchylosed immovably, the posteriorpair much hooked back.Clavicle and hypercoracoid typical, or as in the Percoids.Hypocoracoid as usual broadly joined at upper end to clavicle andhypercoracoid, thence arching away and touching lower end of clavicleagain with a rather slender i^rocess. Besides this, from its middle, run-ning through the usual interspace between it and clavicle, is anotherprocess flat and thin, but strengthened through its middle, reaching toand joined to clavicle.Actiuosts four, rather short.Pectoral not nearer upjDcr end of clavicle than is usual in the Percoids,its upper ray working directly upon hypercoracoid,Postclavicle in two parts, the inferior very long.Supraclavicle of moderate length.Post-temporal widely forked; its articulation with skull typical.Inferior hypohyal scarcely visible on outer surface of hyoid arch,being covered by superior hypohyal, which forms the greater part offront of arch. Hypohyals of about equal size on inner surface of arch.Ceratohyals, ephyals, and interhyals typical.
' Suborbital shelf: a small sbelf of bone extending inward from the suborbitalring and conforming to the rotundity of the eyeball. Possessed by many of thehigher bony iishes.
NO. 1186. BELATIONSniP OF DINOLESTES LE WINI?STABKS. 119Branchiostegals seven, four being borne by ceratohyal and three byepiliyal; tlie tliree anterior ones attached to inner surface of byoidarch,Glossohyal wide, flat, or slightly concave above.Urohyal elongate, thin, without double heel below, except at extremeanterior end.Basibranchials three?the first not supporting any arch, the secondsupporting the first arch, and to the third the second and third archesare joined. Ko ossified basibranchial to fourth arch.Epibranchials of third arch meet behind last basibranchial; epi-branchials of fourth arch absent.Suspensory pharyngeals present on first arch.Superior pharyngeals three in number, that is, on second, third, andfourth arches ; the second largest, and with the third forming an elon-gate patch.JFirst two interspinal rays of dorsal and of anal not coalesced. Noneof the interspinals interposed between neural or h?emal spines.The interval between first and second dorsals occupied by two freeauxiliary iuterneural spines.First interhaemal spine not enlarged or in any way differentiated fromits fellows. Interhsemals equally graduated from behind forward.Ventral fins truly thoracic. The pelvic girdle long and tapering to aslender point, which is interi^osed between the clavicles above theirunion.The vertebral column composed of 10 abdominal and 16 caudal,which, with the usual hypural, number 27 vertebrne.Superior zygapophyses both posteriorly and anteriorly present, butsmall, as is usual in the Percoid fishes.Inferior zygapophyses well developed posteriorly near middle of col-umn ; anteriorly scarcely discernible.Parapophyses developed on fourth and succeeding abdominal verte-briie, growing larger posteriorly; the last pair connected near theirbases by a bridge of bone.Hypural assisted in supporting caudal fin by spines from, two preced-ing vertebrae.Kibs and ei)ipleurals typical.
120 PROCEEDINGS OF TRE NATIONAL MUSEUM. vol.xxii.EXPLANATION OF PLATES.[Drawn by Chloe Lesley Starks.]Plates VIII, IX, and X, superior, lateral, and posterior views of the cranium ofDinolestes lewini, slseleton No. 47877, U.S.N.M., from Port Jaclsson, Australia.Plate XI, Dinolestes lewini, No. 47929, U.S.N.M., from Tasmania.SIGNIFICANCE OF REFERENCE LETTERS USED ON PLATES.
als. Alispbehoid. jp. Parietalhas. Basisphenoid. pas. Parasphenoid.io. Basioccipital. pf. Prefrontal.e. Ethmoid. pro. Prootic.eo. Exoccipital. pto. Pterotic.epo. Epiotic. so. Supraoccipital./>?-. Frontal. spo. Sphenotic.opo. Opisthotic. V. Vomer.
U. S. NATIONAL MUSEUM PROCEEDINGS, VOL XXII PL. VII!
Superior View of Cranium of Dinolestes lewini.For explanation of plate see page 120.

U. S. NATIONAL MUSEUM PROCEEDINGS, VOL. XXII PL. IX
'fux^
Lateral View of Cranium of Dinolestes lewini.For explanation of plate see page 120.

U. S. NATIONAL MUSEUM PROCEEDINGS, VOL. XXII PL. X
iuQ SMjO
Posterior View of Cranium of Dinolestes lewini.For fxplanation of plate see page 120.

U. S. NATIONAL MUSEUM PROCEEDINGS, VOL. XXII PL. XI
Q i
1