ROLE OF LARVAL STAGES IN SYSTEMATIC INVESTIGATIONS OF MARINE TELEOSTS: THE MYCTOPHIDAE, A CASE STUDY

H. GEOFFREYMOSER AND ELBERTH. AHUXTROM~

ABSTRACT

The lanternfish family Myctophidae is the most speciose and widespread family of mid-water fishes in the world man. As presently recognized it contains about 30 genera and 300 nominal species. Their larvae are highly prominent in the plankton and make up about 509b of all larvae taken in open-oeean plankton tows. Our studies of myctophid larvae, on a worldwide basis, have demonstrated that characters of the larval stages of lanternfishes are of great utility in systematic analysis. The genera and species can be recognized on the basis of eye and body shape, the shape and length of the gut, and pigment pattern and by the sequence of photophore development. In this paper the larvae of 55 species representing 24 genera are illustrated and used to demonstrate the usefulness of larvae in understanding the relation- ships of species within genera. Characters of the larvae provide insight into generic anities of lanternfish, allowing us to construct an evolutionary scheme of tribes and subfamilies that differs in some aspects from those proposed on the basis of adult characters. The concept of using larval characters in combination with adult characters to delineate phylogenetic lines in myctophids is discussed, as is our view of evolutionary strategy in the family.

A major facet of comprehensive systematic inves- played a large role in the taxonomy of anguil- tigations is the search for functionally unrelated liform fishes (Castle, 1969) partly because of the characters. Whether the independence of these conspicuousnessof eel leptocephali and partly be- characters is actual or merely apparent, they con- cause of the unavailability of adults of many of the stitute useful elements in the analysis of systema- families. Bertelsen's (1951) treatment of the tic relationships. Ample evidence of this is the ceratioid fishes is a superb example of the value of higher classification of teleosts (Greenwood et al., utilizing larval stages in a systematic revision of a 1966) generated by the synthesis of a diverse large group of teleosts. Apart from these two array of classical taxonomic characters. The re- groups, it is the larvae of myctophiform fishes cent surge of serological and biochemical studies which have received the most attention from tax- on fish has placed a fresh group of characters in the onomists. Ege (1953,1957) relied heavily on lar- hands of systematic ichthyologists (De Ligny, val stages in his extensive works on the 1969). Likewise, recent advances in fish cytogene- Paralepididae. Johnson (1971) employed larval tics (e.g., Ohno, 1970;Benirschke and Hsu, 1971; characters in defining species and genera of Ebeling, Atkin, and Setzer, 1971) are providing Scopelarchidae. Bertelsen, Marshall, and Krem another group of taxonomic characters. It is likely (pers. commun.) have used larval stages exten- that behavioral science will be still another source sively in their revision of the Scopelosauridae. of taxonomic characters, as exemplified by the Our studies on the family Myctophidae itself growing body of information on the acoustic be- (Moser and Ahlstrom, 1970, 1972) indicated that havior of fishes (Fish and Mowbray, 1970). larval characters can aid significantly in differen- One group of well known taxonomic characters, tiating taxa and defining evolutionary lineages those of the embryonic and larval stages, has re- within this family. ceived scant attention from all but a few systema- The lanternfish family Myctophidae is the most tic ichthyologists. Characters of the larvae have speciose and widespread family of mid-water fishes in the world ocean. As presently recognized it contains about 30 genera and 300 nominal 'This a r was presented at the International Sym ium on the Earg 6fe History of Fish (sponsored by IABO, FE,ICES, species. Their larvae are highly prominent in the -ICNAF. -.. .- , and SCOR) held at Oban. Scotland. 17-23 May 1973. plankton and make up about 50% of all larvae %outhwestFieheries Center, National Marine Fisheries Ser- vice, NOAA, La Jolla, CA 92037. taken in open-ocean plankton tows.

Manuscript accepted Agust 1973. FISHERY BULLETIN: VOL. 72, NO. 2, 1974. 391 FISHERY BULLETIN: VOL. 72, NO. 2 Our studies of the larvae of this family have longations of choroid tissue and some have the included material from all oceans. We have been eyes on stalks. Paxton recognized 11 genera in able to identify larvae from all recognized genera the subfamily Myctophinae and distributed them except Hintonia and Dorsadenu. Larval evidence into two tribes, the Myctophini and the Gonich- supports giving generic status to Metelectronu and thyini. Larvae of the species in each of these Parvilux. Including these, we have developmental genera generally conform to a particular morph series for 29 myctophid genera and for many gen- based on form, pigment, and developmental era we have series for all known species. This has pattern and, although these morphs are remark- afforded a more comprehensive view of the range ably diverse, we can find no character or set of and variability of larval characters, and we are characters that would divide the genera into increasingly impressed with the functional inde- tribes. Within each genus of the subfamily, how- pendence of the larval and adult characters. It is ever, the larval characters are indispensible in apparent that the world of the larvae and the delineating groups of related species or subgenera. world of the adults are two quite separate This is best illustrated by examining the impor- evolutionary theaters. Our studies of larval tant genera of the Myctophinae. lanternfishes have disclosed a full range of charac- Protomyctophum larvae have a slender shape ters, from generalized to specialized and from con- (Figure 1).For all species exceptP. anderssoni, the servative to labile, equal in scope to those of the gut is short during most of the larval period and adults. These characters fall into several characteristically there is a marked interspace categories. An important group is the shape of between the anus and the origin of the anal fin various structures such as the eye, head, trunk, (Figure 1A-D). The gut elongates dramatically in guts, and fins, especially the pectoral fins. Another late larvae, to fill the interspace. Gut development group is the sequence of appearance and the posi- is completely dissimilar in P. anderssoni, where tion of fins, photophores, and bony elements. the gut is long at all larval sizes, in fact longer Another is the size of the larvae when fins and than in most other lanternfish larvae (Figure 1E). other features appear and the size of the larvae Series of ventral tail melanophores are formed in when they transform into juveniles. Pigmentation some species of both recognized subgenera provides an important group of characters based (Heirops and Protomyctophum sensu stricto), for on the position, number, and shape of melano- example in P. Protomyctophum normani (Figure phores. Finally, there are the highly special- 1A) and P. Heirops thompsoni (Moser and Ahl- ized larval characters such as voluminous fin strom, 1970). Larvae of the subgenera can be sepa- folds, elongated and modified fin rays, chin bar- rated, however, on the basis of eye shape, the eyes bels, preopercular spines, etc. It is our purpose ofHeirops (Figure lC, D) being characteristically here to point out some of these characters and narrower than those of Protomyctophum sensu demonstrate how they can be of advantage in stricto (Figure lA, B). Choroid tissue is absent defining taxa and establishing phylogenetic from the ventral surface of the eye in all species of lineages. the genus except P. anderssoni, which has a well- developed “teardrop” (Figure 1E). Larvae of P. THE SUBFAMILY MYCTOPHINAE anderssoni are so markedly different from those of all other species of Protomyctophum, which The most trenchant character of larval myc- otherwise form a rather cohesive group, that this tophids is eye shape. Our studies show that species should be placed in a separate subgenus or perhaps even in a distinct genus. This suggestion lanternfish larvae fall naturally into two groups on the basis of eye shape-those with narrow ellip- is supported by the unique placement of certain photophores and by the structure of the sup- tical eyes and those with round or nearly round racaudal luminous tissue in adults of this species. eyes (Moser and Ahlstrom, 1970). The species composition of these two groups agrees closely Larvae of the genus Electrona are a less with that of the two subfamilies, Myctophinae and homogeneous group but are united by a common- Lampanyctinae, established by Paxton (1972) on ality of body shape and especially gut shape (Fig- the basis of osteological and photophore charac- ure 2). A marked interspace is present between ters of adults. Larvae of the Myctophinae have the end of the gut and the origin of the anal fin. elliptical eyes; some species have ventral pro- This space is closed only at the termination of the

392 MOSER and AHLSTROM: ROLE OF LARVAE IN SYSTEMATICS

FIGURE1.-Larvae of Protomyctophum. A. P. Protomyctophum normani, 15.2 mm; B. P. Protomyctophum teni- soni, 14.5 mm; C.P.Hieropssubparallelum, 15.2 mm; D.P. Hieropschile.lsis. 11.0mm;E.P. anderssoni, 15.7 mm. FISHERY BULLETIN: VOL. 72. NO. 2

--- -

FIGURE2.-La~ae of Electrona and Metelectrona. A. E. antarctica, 12.7 mm; B. E. carlsbergi, 11.1 mm; C. E. subaspem, 10.5 mm; D. M. ahlstrorni, 10.3 mm.

394 MOSER and AHLSTROM: ROLE OF LARVAE IN SYSTEMATICS larval period. None of the species forms photo- are synonyms, however, the uniqueness of the phores during the larval period other than the Br2 larva strongly suggests the resurrection of pair. Metelectrona as a valid genus. The characters that most clearly separate the The genus Benthosemu is the least cohesive of three developmental lines in Electrona are eye any genus in the subfamily Myctophinae, from the shape and the amount of choroid tissue developed viewpoint of larval structure (Figure 3). We can under the eye. Electrona antarctica has an elon- find only four types of larvae in the world ocean, gate choroid mass uniquely divided into two nar- although Nafpaktitis (1973) recognizes five row eyes (Figure 2A). Also, E. antarctica larvae species on adult characters. We cannot distin- attain a large size (20 mm), are the deepest-bodied guish larvae of B. pterota and B. panamense of all Electrona larvae, and have the heaviest pig- although Nafpaktitis has listed a number of con- mentation. The two species in the second de- vincing characters that distinguish the adults of velopmental line transform at a small size (ca. 10 the two species. We find two highly divergent mm in E. rissoi and 12-13 mm in E. carlsbergi), species pairs. One is composed of B. glaciale and have a small choroid mass under a moderately B. suborbitale with a narrow eye subtended by a narrow eye, and develop scant pigment (Figure lunate choroid mass and with a pronounced inter- 2B). In the third line, consisting of E. subaspera space between the anus and the anal fin origin, and E. paucirastra, the eye is the least narrow, reminiscent of Protomyctophum and Electrona has no choroid tissue, and the larvae attain a large (Figure 3A-C). In the other pair, consisting of B. size (20 mm) (Figure 20. panamense-pterota and B. fibulatum, the eye is The larva of the species described as Metelec- wider, is subtended by a mere sliver of choroid trona ahlstromi (Wisner, 1963) is illustrated in tissue and the gut, of moderate length, lacks a Figure 2D. It is more laterally compressed than postanal interspace (Figure 3D, E). any species of Electrona and has no interspace The one feature held in common by the four between the anus and origin of the anal fin. In species is the development of some photophores in some features it resembles the larvae of addition to the Br2 during the larval period. The Hygophum; it has a late-forming dorsal fin and the only other myctophine genera that develop photo- gut is shaped very similarly to that in H. tauningi phores in addition to the ubiquitous Brz during the and H. macrochir. Its pigment is unique and the larval period are Diogenichthys, Myctophum, eye is distinct, with the ventral edge of the scleral and Metelectrona. This feature is much more pre- envelope characteristically squared off. Also, in valent among genera of the Lampanyctinae and is late-stage larvae, in addition to the Br,, a second helpful in defining groups of related genera there pair of photophores (PO,) develops, a feature (Moser and Ahlstrom, 1972). found in neitherHygophum norElectrona. Paxton In B. panamense-pterota and B. fibulatum the (1972) synonymized Metelectrona with Electrona Dn pair is formed soon after the Br2 at about and suggested that M. ahlstromi and E. ventralis 5.0-6.0 mm. The PO5 pair is the third to appear in TABLE1.-Sequence of photophore formation in larvae of three species of Benthosema.

No. of Smallest photophore juvenile Species Photophores pairs (mm)

E. fibularum ca. 4.0 Br2 13.2 5.4 Br2 Dn 6.0 Br2 Dn PO5 6.4 Brz Dn PO5 POI 7.3 Brz Dn Po5 POI AOat 7.7-8.7 Br2 Dn Po5 POI AOar POz ca. 10.0 Brz Dn PO5 PO, AOa, PO2 Opl VLO E. pferofa (panameme) 4.0 Br2 1 11.8 5.0 Br2 Dn 2 6.0 Brz Dn Po5 3 ca. 7.0 Br2 Dn Po5 PVOI 4 7.1 Brz Dn POs PVO, Opt 5 7.5 Brz Dn Po5 PVOt Opz VOI PVOz 7 8.0 Brz Dn PO5 PVOI Opz VO? PVOz POI AOaI 9 E. suborbitale 4.1 Br2 1 10.7 8.59.2 Br2 POI POz 3 9.4 Brz PO, PO2 Brl Br3 Opz 6 11.5 Brz PO, POz Brl Bra Opz POI Po4 Pos AOar AOaz 11

396 FISHERY BULLETIN: VOL. 72, NO. 2

FIGUFCE3.-Larvae of Benthosema. A. B. glociale, 1.2 mm; B. B. glaciale, 10.5 mm; C.B. suborbitale, 9.2 mm; D. B. pterota, 8.5 mm; E. B. fibulaturn, 8.7 mm.

396 MOSER and AHLSTROM: ROLE OF LARVAE IN SYSTEMATICS larvae about 6.0 mm long. Thereafter the pattern dorsally into a prominent enlarged posterior sec- diverges as shown in Table 1, but both species tion. In H. macrochir this enlarged section is gradually add about a dozen pairs during the lar- covered with large melanophores. Larvae of this val period. Specimens of B. pterotu from the Per- group occur only in the Atlantic. sian Gulf off India, formed photophores at some- The genus Hygophum affords an excellent ex- what larger sizes than larvae of B. panamense, ample of the taxonomic utility of larval stages. but in the same sequence. Transformation occurs The juveniles and adults of some species are at a small size, 10-12 mm in B.panamense-pterota notoriously difficult to identify. In contrast, the and 11-13 mm in B. fibulatum. larvae of these species are highly distinct and can Photophores appear relatively late in larvae of be readily identified. We have 11 such distinct B. suborbitale and B. glaciale, however, the Brl, larval types, whereas only 9 species are currently Bm, Ob, and PO series appear in late larvae of known for the adults. Search for adults of the two both species (Table 1). Transformation occurs at remaining larval types has led to the discovery of about 9-11 mm in both species. The larvae of B. two undescribed species. In addition, characters of panamense-pterota and B. fibulatum are close to the adults of this genus reveal little about the the larvae of Diogenichthys in several characters relationships of the member species (Becker, including body shape, gut shape, and early ap- 1965). A study of the larvae, however, shows that pearance of photophores. there are three highly distinct subgeneric groups, As in Benthosema, the larval characters of each containing from two to six closely related Hygophum suggest some divergence within the species. Such an independent view of the complete genus, although all species have a highly charac- species complement of a genus is an invaluable teristic series of isthmal melanophores, form the tool in the formal revision of that genus. dorsal fin late in the larval period, and develop no Larvae of the species of Symbolophorus are photophores other than the Brz, as larvae (Figure perhaps the most cohesive of all myctophine gen- 4). The genus contains three divergent types of era (Figure 5A). In all species known to us the larvae. The most unusual of these are the ex- pectoral fin is large and is supported by an elon- tremely elongate larvae of H. reinhardti and H. gate aliform base. Also, the pelvic fins are large atratum, which have very narrow eyes that are and develop earlier than in any other lanternfish underlain by prominent choroid tissue and are genus. The narrow eyes have choroid tissue and borne on short stalks (Figure 4A). The amount of are borne on short stalks. The amount of pigmen- pigmentation along the gut and tail and on the tation decreases towards the end of the larval myosepta and fin fold increases throughout the period. Most species attain a large size-up to 24 larval period. mm. The species differ principally in the size at A second larval type is represented by the which various larval structures appear. largest number of species, H. proximum, H. The closely related genus, Myctophum, has a hygomi, and H. brunni, all illustrated (Figure diversity of larval form unmatched in the family 4B-D), as well as H. benoiti, H. hanseni, and an (Figures 5,6,7).Before taking up the bulk of the undescribed form in our collection. These larvae species in this genus we must first examine the are only moderately slender and have unstalked most aberrant of all lanternfish larvae, that of M. eyes of moderate width, subtended by prominent auroluternatum (Figure 5B). In this remarkable choroid tissue. Melanophores are located chiefly larva the eyes are borne on long stalks and the free on the head and gut, however some species have trailing section of the gut is almost as long as the pigment on the myosepta and fin fold. The trend in fish itself. The dorsal fin forms at the margin ofthe this group of species is for the early larval stages fin fold. These characters are so bizarre that it to have the heaviest pigment and for melano- would seem preposterous to identify it as a phores to be lost as development proceeds. lanternfish larva, much less that of M. auroluter- A third type of larva is exhibited by H. mac- natum. Nonetheless, A. Vedel Thing first sug- rochir, H. taaningi, and an undescribed form in gested the true identity ofthis larva (E. Bertelsen, our collection (Figure 4E, F). These are relatively pers. commun.) which we can now confirm since deep-bodied, have large, relatively wide eyes with recently receiving the critical transforming little or no choroid tissue, and lack tail pigment. specimens through the courtesy of Warren Also, the gut has a highly distinctive form; the Freihofer (California Academy of Science). The anterior half has a very small diameter and opens uniqueness of this larva would certainly suggest 397 FISHERY BULLETIN: VOL. 72. NO. 2

A

-I__...- '-' -.,>-\

C ,-----.

FIGURE4.-Larvae ofHygophum. A. H. reinhardti, 12.8 mm;B. H. pmrimum, 8.9 mm;C. H. hygomi, 8.1 mm; D. H. brunni, 9.7 mm;E. H. mrochir, 7.3 mm;F. H. toaningi, 6.8 mm.

398 MOSER and AHLSTROM: ROLE OF LARVAE IN SYSTEMATICS

Fi~uirs&-Larvae of Symbolophorus and Myctophum. A. S. californiense, 9.6 mm; B. M. auroluter- mtum, 26.0 mm; C. M.punctatum, 13.6 mm;D. M. nitidulum, 8.2 mm;E. M. phengodes, 9.8 mm.

399

I FISHERY BULLETIN: VOL. 72, NO. 2 the creation of a distinct genus for M. aurolater- Myctophum sp. (possibly brachygnathum) and is natum and it is highly probable that corroborative developed on the jaws, branchistegal membrane characters will appear after careful reexamina- and lower part of the pectoral fin base in Myc- tion of the adults. tophum sp. (possibly fissunoui) as seen in Figure With the removal of M. aurolaternatum, the re- 6B-D. The latter three species form the PLO maining larvae of Myctophum form a diverse, yet photophores on the pectoral fin base soon after the recognizable, group. All have large broad pectoral appearance of the Dn organs (Table 2). fins supported on a highly characteristic fan- Nafpaktitis (1973) has listed a number of shaped base. The species may be divided into two characters for distinguishing adult M. ob- groups, those which form only the Brz photophores \ tusirostre from M. brachygnathum. He showed and those which develop additional photophores that M. pristilepis is a synonym of M. brachyg- during the larval period. In the first group the nathum. The status of M. imperceptum Bekker elongate larva of M. punctatum (Figure 5C) has and Borodulina has yet to be determined. stalked eyes and a slightly aliform pectoral fin A second larval type is represented by M. base, reminiscent of Symbolophorus larvae, and selenops (Figure 7A) and a closely related species may be the closest relative of that genus among restricted to the Indian Ocean and Persian Gulf the species of Myctophum. A closely related for which we can find no adult (Figure 7B). In species, M. nitidulum, is also stalk-eyed, but is these rotund species, the head is relatively longer deeper-bodied,more heavily pigmented, and has a and narrower than in the previous group and the more fan-shaped pectoral fin base (Figure 5D). It is slightly stalked eyes are narrower and bear more obvious from our studies that M. nitidulum is one elongate choroid tissue. The two species differ in member of a complex, that includes M. affine (not that the eyes of the unnamed larva are more illustrated) and several other species. The lightly definitely stalked tha-n in M. selenops. Also the pigmented larva of M. phengodes has only a sug- pigment pattern is markedly different, as is the gestion of stalked eyes but is similar in body shape size at which photophores appear. We have care- to M. nitidulum (Figure 5E). The larval characters fully examined larvae of M. selenops from the At- substantiate Paxton’s (1972) decision to abolish lantic, Indian, and Pacific Oceans, find them to be the genus Ctenoscopelus, established for this identical in all three oceans, and seriously ques- species by Fraser-Brunner (1949). tion Wisner’s (1971) allocation of the Hawaiian The other major group of Myctophum is charac- population as a distinct species, based an slight terized by the appearance of the Dn photophore differences in relative eye size and SA0 photo- anterior to the eye, usually early in the larval phore orientation. period. These species fall into three rather distinct The third type of larvae that develop the Dn species groups on the basis of body and eye shape. photophores is represented by M. spinosum The first is a group of four rotund broad-headed (Figure 7C) and M.lychnobium (Figure 7D). These species, which have large unstalked eyes sub- are elongate fusiform larvae with moderately nar- tended by a short mass of choroid tissue. Ofthese, row unstalked eyes, underlain by a pronounced the larvae of M. asperum are the most heavily choroid mass. M. spinosum is the more slender of pigmented, particularly on the body (Figure 6A). the two and is extremely heavily pigmented, espe- Pigment is confined to the head in M. obtusirostre, cially in older larvae. Pigmentation in M. lych- is heavy under the posterior part of the gut in nobium is confined to that in the illustrated

TABLE2.--Sequence of photophore formation in species ofMyctophum that form two or more pairs during the larval stage.

Size range Size at first formation Size at Species (mm) (mm) transformation Br2 Dn PLO PO? (mm) M. asperum ca. 3.09.8 4.2 4.6 9.8 - Early transf. 11.4 M. obtusirostre ca. 3.0-8.9 3.8 4.0 ca. 7.1 8.9 Late transf. 12.5 M. sp. (possibly fissunovi) ca. 3.07.1 4.1 4.1 7.1 - - M. sp. (possibly brachygnathum) 6.0-11.4 6.0 6.0 ca. 9.0 ca. 9.0 Late transf. 13.8 M. lychnobium 3.5-12.1 ca. 6.0 6.3 12.1 - Late transf. 14.2 M. spinosum 3.5-13.7 ca. 5.5 7.2 13.7 - Late transf. 14.5 M. selenops 3.5-7.5 5.1 5.1 6.2 7.5 Late transf. 11.4 M. sp. 4.0-9.1 ca. 7.0 9 1 - - -

400 MOSER and AHLSTROM: ROLE OF LARVAE IN SYSTEMATICS

FIGURE6.-Larvae of Myctophum. A. M. asperum, 6.8 mm; B. M. obtusirostre, 7.6 mm; C.M. sp. (possibly brachygnathum), 7.5 mm; D. M. sp. (possibly fissunoul), 7.4 mm.

401

_- FISHERY BULLETIN: VOL. 72. NO. 2

FIGURE7.-La~ae of Myctophum. A. M. selenops, 7.8 mm; B. M. sp., 9.1 mm; C.M. spinosum, 9.0 mm; D. M. lychnobium, 9.5 mm.

402 MOSER and AHLSTROM: ROLE OF LARVAE IN SYSTEMATICS

specimen. Only larvae of M. lychnobium have genera and about 75 species in the Myctophinae. been taken in the eastern Pacific, whereas both Paxton (1972) divided the genera into four tribes species have been taken in the central and west- on the basis of a combination of osteological fea- ern Pacific and in the Indian Ocean. Taxonomists tures and characters of the photophores. In a pre- dealing with adult characters only, have placed M. vious paper (Moser and Ahlstrom, 1972) we dis- lychnobium in synonymy with M. spinosum but cussed Paxton’s placement of genera in these the distinctiveness of the larvae suggests that the tribes and indicated that the larval characters adult characters should be reexamined. suggested a somewhat different distribution of The larvae ofM. spinosum and M. lychnobium, genera among the four tribes. For the purposes of although clearly developing the Dn pair of photo- this discussion the tribes referred to here are those phores, resemble the larvae of M. punctatum in suggested by the larval characters. body shape and pigmentation, a species which In general, the larvae ofthe Lampanyctinae are does not develop the Dn as larvae. Actually, there much less diverse in larval characters and are some common characters of pigmentation and specializations than are the larvae of the Myc- eye structure which appear in all of the groups of tophinae, although exceptions to this may be Myctophum species described above. What we ap- found in two of the lampanyctine tribes, the pear to be dealing with is a mosaic of larval Diaphini and the Lampanyctini. characters in a highly complex genus. The tax- The tribe Diaphini is made up of two genera onomy of Myctophum presently is confused; our -Diaphus contains about 50 species and Lobian- work on the larvae should help define the number chia has 3 species. Both genera develop photo- of species in the genus and, perhaps, adult charac- phores, in addition to the Br2, during the larval ters will emerge that can be combined with larval period; in fact more are developed in Diaphus characters to define the phyletic lines within the than in any other lanternfish genus. genus. There are two basic larval types in Diaphus Larvae of the four genera known collectively as (Figure 9A, B). One has a slender body, small the slendertailed myctophids are shown in Figure head, and a series of persistent melanophores on 8. Quite obviously there are two highly divergent the ventral midline of the tail. The other type has generic pairs. Loweina and Tarletonbeania are a deeper body, bulbous head, and a single persis- characterized by large oval eyes, posterior place- tent ventral tail melanophore, or none. It is excep- ment of median fins to accommodate the immense tional for larvae of either type to develop pigment fin fold, and elongated lower pectoral rays bearing on the head and it never occurs between the eyes, spatulate processes. Gonichthys and Centrobran- as is common in Lampanyctus. Both types do form chus are characterized by a deep but markedly embedded melanophores at the base of the caudal compressed head and body and small narrow eyes fin rays. with extremely elongate choroid tissue. As stated The slender type is restricted to the species that earlier, the larval characters suggest strongly develop a suborbital photophore as adults that the two generic pairs are not closely related (Diaphus sensu stricto of Fraser-Brunner, 1949) and should not be grouped into a tribe. The and is represented in Figure 9A by D. theta. The Gonichthys-Centrobranchus pair is similar in eye stubby type is represented by D. pacificus (Figure shape and gut shape to some species of Myc- 9B). The specimens illustrated for the two species tophum, however no species of Myctophum even are rather early larval stages which have not yet approaches this pair in body shape. The characters formed their larval photophores, other than the of the other pair are so divergent as to give no Brz. The first additional pair to form in both types clue of their affinities within the subfamily is the POs and then the PO, (Table 3). The large Myctophinae. genus Diaphus, except for the Atlantic species ably reviewed by Nafpaktitis (1968),is in a state of THE SUBFAMILY taxonomic confusion. Various workers (Fraser- LAMPANYCTINAE Brunner, 1949; Bolin, 1959) have attempted to split the genus into smaller, more cohesive groups; The subfamily Lampanyctinae is considerably the larval evidence would suggest that at least two larger than the Myctophinae; it contains about 19 divergent groups are present. genera and 200-250 species compared with 12 The larvae of the three species ofLobianchia are

403 FISHERY BULLETIN: VOL. 72. NO. 2

FIGURE8.-La~ae of Gonichthyini. A. Loweim rara, 11.6 mm; B. Tarletonbeania crenularis, 18.9 mm; C. Gonichthys tenuiculus, 1.7 mm; D. Centrobranchus chwrocephalus. 1.3 mm.

404 MOSER and AHLSTROM: ROLE OF LARVAE IN SYSTEMATICS

FIGURE9.-Larvae of Diaphus and Lobianchia. A. D. theta, 6.9 mm;B. D. pacificus, 5.2 mm; C. L. urolampus, 7.2 mm;D. L. gemllari, 6.7 mm;E. L. dofleini, 8.2 mm.

405 FISHERY BULLETIN: VOL. 72. NO. 2

TABLE3.-Sequence of photophore formation in larvae of two species of Diaphus.

Size N~.of Smallest larva photophore juvenile Species (mm) Photophores pairs (mm) D. theta 6.2 Br2 PO5 2 ca. 12.0 7.6 Br2 PO5 POI 3 8.2 Brz Po5 PO? VOI 4 8.6 Br2 PO5 POI VOI PO2 Opz 6 9.0 Brz PO5 POI VO, VOt PO2 Opz Po3 PO, VOs 9 9.2 Br2 POSPOI VO, PO2 Op2 POI PO4 VO5 VLO 10

D. oaciiicus~- 57 Er7 PO- PO? 3 9.8 6 2 Eri PO; POI PO2 PVO, 5 6 5 Br2 POSPO? PO2 PVOt PO, 6 7 5 Br2 PO5 PO, PO2 PVOI PO, VO, 7

deep-bodied,have large broad heads, and are eas- pectoral fin, internally in the region of the ily identified by their unique wing-shaped pec- cleithra, and along the myosepta. The pigment toral fins (Figure 9C-E). The larvae of all three patterns are of prime importance in identifying species are heavily pigmented and develop the the larvae to species. Brz,POI, and PO5 photophores sequentially. In L. There are several rather distinct larval types in urolampus (Figure 9C) and L. gemellari (Figure Lampanyctus. One of these consists of a group of 9D) the eyes are large and nearly round and the species whose adults are characterized by having lower pectoral rays are delayed in developing. In the pectoral fins much reduced or even absent, and L. dofleini the lower pectoral rays develop along has been separated recently as a distinct genus with the produced upper rays and the eye is com- Paralampanyctus by Kotthaus (1972) with P. pletely different (Figure 9E). With its narrow el- niger as type. Previously, Giinther (1887) had liptical shape and unique squarish mass of choroid proposed the generic name Nannobrachium for tissue, it is the single obvious exception to the rule this species and this has priority over Paralam- of narrow eyes in the subfamily Myctophinae and panyctus (John Paxton, pers. commun.). There is a rounded eyes in the Lampanyctinae. All other lar- remarkable trend of jaw specialization in the lar- val characters identify this species as a Lobian- vae of this group (Figure 10). The larva of L. ritteri chia, and we conclude that the narrowing of the has jaws of moderate length and the other species eye in this species occurred independently as a shown have progressively longer jaws with more secondary adaptation. prominent teeth, particularly anteriorly. This In our view the tribe Lampanyctini contains the trend culminates in the larva of Lampanyctus sp. genera Lampanyctus, Triphoturus, Steno- (possibly achirus) which somewhat resembles a brachius, and Paruilux. As recently as Fraser- larval billfish. This species will lack the pectoral Brunner’s (1949) review of the family Myc- fin in juveniles and adults, even though it is well tophidae, Lampanyctus was still a catchall genus developed in the larvae. The pectoral fins are also with a number of disparate subgenera. Since then large in L. regalis and L. niger larvae, but will be the subgenera Stenobrachius, Triphoturus, and small and weakly developed in adults. This dis- Lepidophanes have been removed from Lam- parity is even more apparent in another eastern panyctus and afforded generic status. Lepido- Pacific species, which lacks pectoral fins as an phanes has been further split into the genera adult, but whose larvae have the largest pectoral Lepidophanes and Bolinichthys. All of the fins with the highest number of rays that we have separated genera have distinctive larval morphs. encountered among Lampanyctus larvae. Other With their removal, the species of Lampanyctus less spectacular specializations appear in the form a more coherent assemblage of 40-50 species, other subgroups of Lampanyctus, but it appears and despite the diversity of larval specializations that the larval characters will be helpful in encountered in the genus, there is a central morph defining the species composition of the several and pattern of larval development. subgenera. Lampanyctus larvae are deep-bodied and Representatives of other genera of Lampanyc- bigheaded. In older larvae characteristic pigment tini, Triphoturus, Stenobrachius, andParvillwc are can develop at a variety of locations such as the tip illustrated in Figure 11A-C. Small larvae of of the lower jaw, between the eyes, the back of the Triphoturus and Stenobrachius have a row of head, the side of the head, the adipose fin, the melanophores along the ventral margin of the tail

406 MOSER and AHLSTROM: ROLE OF LARVAE IN SYSTEMATICS

FIGURElO.-L.arvae of Lampanyctcrs. A. L. ritteri, 10.1 mm;B. L. regalis, 9.1 mm;C. L. niger, 8.7 mm; D. L. Sp. (possibly aehircrs), 13.4 mm.

407 FISHERY BULLETIN: VOL. 72, NO. 2

FIGURE11.-Larvae of Lampanyctini and Gymnoscopelini. A. Triphoturus mexicanus, 10.5 mm; B. Stenobrachius leucopsarus, 10.4 mm; C. Paruilw ingens, 14.4 mm; D.Bolinichthys supralateralis, 9.4 mm;E. Ceratoscopelus townsendi, 16.6 mm.

408 MOSER and AHLSTROM: ROLE OF LARVAE IN SYSTEMATICS but these coalesce into one or two spots in mid- variable from specimen to specimen; it can be nar- stage larvae. Triphoturus larvae are character- row and elliptical or nearly round, but most typi- ized further by their distinctive head shape and by cally would be classified as irregular in shape. The the series of melanophores along the ventral mid- shape of the head, body, and gut is unusual and line below the gut. Stenobrachius larvae add con- distinctive. The larval characters are of little help siderable pigment late in the larval period, par- in elucidating the affinities of this species within ticularly along the dorsum and on the myosepta the Myctbphidae and, when added to the list of of the trunk. The larvae of Parvilux are distinct unique adult Characters, only magnify the prob- in shape and pigmentation. Paxton (1972) placed lem. It would seem to make just as much sense to this genus in Lampanyctus based on osteological establish a separate subfamily for Notolychnus as characters. In certain photophore arrangements, to place it in a monotypic tribe in the subfamily however, particularly in the posterior placement Lampanyctinae. of the VLO and the nonangulate arrangement of The larvae illustrated in this paper comprise 55 the SAO, the genus appears to us to be more species representing 24 genera. Illustrations are closely related to Stenobrachius than to Lam- included for larvae of 11 of the 12 genera in the panyctus. These characters in addition to the dis- subfamily Myctophinae; not included are illustra- tinctness of the larvae would suggest that the va- tions of Diogenichthys (see Moser and Ahlstrom, lidity of Parvilux should be reconsidered. 1970 for D. laternatus and D. atlanticus). In the The tribe Gymnoscopelini judged from larval subfamily Lampanyctinae larvae are illustrated and/or adult characters contains a dozen genera for 13 of the 19 genera. The omitted genera (Lam- (Notoscopelus,Lampichthys, Scopelopsis, Cerato- pichthys, Lampanyctodes, Gymnoscopelus, and scopelus, Lepidophanes, Bolinichthys, Lampadena, Tanningichthys),all from the tribe Gymnoscope- Taaningichthys, Dorsadena, Lampanyctodes, lini, are illustrated in Moser and Ahlstrom (1972). Gymnoscopelus, and Hintonia). The larvae Larvae of Hintonia and Dorsadena have not yet of these genera are united by a group of common been identified. characters, including a distinctive, usually slender, body outline, a series of melanophores SOME EVOLUTIONARY on the dorsal and ventral midlines of the tail CONSIDERATIONS (in most genera), and the development of a group of photophores during the larval period, With this brief review of lanternfish larvae most notably the POS, PLO, and Vn. The larvae of completed, let us now turn to an interesting prob- this tribe were treated extensively in a previous lem of myctophid evolution to which study of the paper with representative larvae illustrated for 10 larvae may contribute importantly-the evolu- of the 12 genera (Moser and Ahlstrom, 1972). Ad- tion of photophore pattern. With a single excep- ditional species ofBolinichthys (B. supralateralis, tion, all adult myctophids have two conspicuous Figure 11D),Ceratoscopelus (C.townsendi, Figure rows of photophores that extend the length of the llE), Lampadena (L. luminosa, Figure 12B), body on each side of the ventral midline. The Lepidophanes (L. gaussi, Figure 12C) are illus- photophores are grouped and positioned in a trated herein. Illustrations of Notoscopelus re- definite and often diagnostic pattern. Also, splendens (Figure 12A) and Scopelopsis mul- lanternfishes have a specific pattern of photo- tipunctatus (Figure 12D) are included for com- phores on the sides of the body, below the lateral parative purposes. It need only be mentioned here line, and on the ventral aspect of the head. The that the clusters of closely related genera within exception is Tauningichthys puurolychnus, which this tribe are readily apparent from examining appears to lack body photophores entirely. In ad- the larval characters, especially the sequence of dition to these photophores, some lanternfish gen- photophore development, and these groupings era have photophores positioned in, a pattern agree closely with those established on the basis of above the lateral line and some have small “sec- adult characters. ondary” photophores distributed more generally The species Notolychnus valdiviae has so many over regions of the body and head. Another type of unique adult characters that Paxton (1972) as- luminous structure present on most myctophids signed it to the monotypic tribe Notolychnini. are discrete glands located at the caudal peduncle. Likewise the larva has a number of unusual Typically, these are sexually dimorphic in charac- characters (Figure 12E). The shape of the eye is ter and, doubtless, play some part in courtship 409 FISHERY BULLETIN: VOL. 72, NO. 2

.. . -......

E

FIGURE lZ.-Larvae of Gymnoscopelini and Notolychnini. A. Notoscopelus resplendens, 11.2 mm; B. Lampaaha hrninosa, 12.8 mm; C.Lepidophues gaussi, 13.5 mm; D.Scopelopsis rnultipunctutus, 17.5 mm; E. Notolychnus valdiviae, 9.2 mm.

410 MOSER and AHLSTROM: ROLE OF LARVAE IN SYSTEMATICS behavior. Finally, some myctophids have small The individual “primary” photophores are typi- patches of soft whitish, apparently luminous, tis- cally highly developed and concentrated ventrally sue located at various regions of the body. on the body. The ventral location of photophores in The most popular speculation as to the possible Myctophinae is probably related to their habitat. function of the patterns of photophores and lumin- That is, they are generally shallow-living active ous scales is that they function in species recogni- fishes that have well-developed gas bladders and tion (see McAllister, 1967).An explanation for the it is plausible that concentration of photophores universality of the two ventral rows was postu- ventrally on the body evolved as an adaptation for lated by Clarke (1963). His suggestion that these countershading and protection from deeper-living downward directed rows emit a continuous light of predators. This view of the Myctophinae is com- ambient wavelength, which conceals the fish from pletely contrary to those previously held for this deeper-living predators by countershading, has subfamily. On the basis ofthe “upward migration” much appeal. theory of photophore evolution, myctophines were We have long been interested in the mechanism thought to be primitive unspecialized forms. For- by which such patterns of photophores could have merly, we too subscribed to this view, and con- evolved and believe we have gained some insight trasting the then supposed primitive features into this mechanism through our studies of the such as low photophore position and short jaws of larval stages. Our proposal, as expressed in an the adults with the highly specialized and diverse earlier paper (Moser and Ahlstrom, 1972), is that features of the larvae, we proposed that the ancestral myctophids had a generalized arrange- evolutionary pace had differed in the larval and ment of unspecialized photophores, one at the adult stages of the subfamily. Our altered opinion posterior margin of each scale pocket, and a group would view both larvae and adults of the Myc- of similar photophores on the head. We further tophinae as highly advanced and would interpret proposed that the specific photophore patterns of the low photophores, prominent gas bladders, contemporary myctophids evolved through pro- short jaws, and often narrow caudal peduncles as gressive enlargement and specialization of certain specialized adaptations of active surface-dwelling photophores of the generalized pattern and con- fishes. current diminution or loss of the unspecialized Our view of the Lampanyctinae must also be photophores. This idea came to us upon discover- revamped. Formerly we considered the diverse ing the remarkable transforming specimens of and often dorsally oriented pattern of photophores Scopelopsis multipunctatus, the adults of which and accessary luminous tissue to be highly have a small photophore at each scale pocket and a specialized features. Possibly, the luminous group of photophores on the head. In the adults, scalelike patches and luminous glands are the “primary” organs can be distinguished only by specialized adaptations, but we feel that the pres- their modified lens-bearing scales, but in the ence of small secondary photophores and the dor- transforming specimens (Figure 11D) the primary sal positioning of primary photophores in many of photophores stand out clearly as enlarged mem- the genera, indicate a retention of the ancestral bers of the meristic series of light organs. It struck condition. The Lampanyctinae are generally us that a similar arrangement of photophores deeper-living than the Myctophinae and many might have existed in the adults of an ancestral genera are lethargic fishes that rest vertically in species, and led to the development of our ideas on the water column (Barham, 1970). In deeper- the evolution of photophore pattern. Our theory living fishes with such a behavior pattern there was further strengthened by neurological findings would be little evolutionary advantage in having and by what we feel are inherent weaknesses in ventrally concentrated photophores, and the fat- Bolin’s (1939) and Fraser-Brunner’s (1949)theory invested swim bladders and long jaws typical of that photophore patterns evolved by the upward many genera could have evolved in ,relation to migration of organs from ventral rows of photo- habitat and activity pattern. It is interesting that phores. the most obvious exception in the subfamily, the Viewed from the standpoint of our theory the Diaphini, are active, often surface-dwelling fishes subfamily Myctophinae would be considered with relatively short jaws and ventrally concen- highly specialized, since it is here that diminution trated photophores. It is obvious from the present of secondary photophores has reached its highest paper that the larvae of Lampanyctinae exhibit degree; they are totally lacking in the subfamily. much less diversity than the Myctophinae, but we 411 FISHERY BULLETIN: VOL. 72. NO, 2 no longer view the adult myctophines as being G. B. Farquhar (editor), Proceedings of an international more “primitive” than the adult lampanyctines. symposium on biological sound scattering in the ocean, p. We feel that the adults of both subfamilies are 100-118. Maury Center for Ocean Science, Department of equally specialized and that these specializations the Navy, Wash., D.C. MC Rep. 005. BECKEU,V. E. are basically related to their particular habitat. 1965. Lanternfishes of the genus Hygophum (Myctophidae, In summary, thorough study of the larvae of a Pisces). Systematics and distribution. Tr. Inst. Okeanol. teleost family such as the Myctophidae can be Akad. Nauk SSSR8062-103. (In Russ., Engl. Transl. No. most helpful in species validation, in analyzing 45, Natl. Mar. Fish. Serv., Syst. Lab., Wash., D.C.) affinities at all taxon levels, and in assessing BENIR~CHKE,K., AND T. C. Hsu (editors). 1971. An atlas of mammalian chromosomes, Vol. 5 and phylogenetic lineages. Also, the above discussion 6. Springer-Verlag, N.Y., 200 p. would indicate that larval studies can provide in- BERTELBEN,E. teresting insights into the major directions of 1951. The ceratioid fishes. Ontogeny, taxonomy, distribu- evolution within a family of fish. tion and biology. Dana Rep., Carlsberg Found. 39,276 p. BOLIN, R. L. 1939. A review of the myctophid fishes of the Pacific coast of ACKNOWLEDGMENTS the United States and of Lower California. Stanford Ichthyol. Bull. 1:89-156. George Mattson executed 15 of the illustrations 1959. Iniomi Myctophidae. Rep. Sei. Results “Michael (Figures lE, 4A, 5A and D, 8A-D, 10A, B, and D, Sars” North Atl. DeepSea Exped. 1910. 4, 2(7):1-45. 11A and E, and 12A and E) and we thank him for CASTLE,P. H. J. 1969. An index and bibliography of eel larvae. J. L. B. his efforts. The remaining illustrations were made Smith Inst. Ichthyol., Rhodes Univ., S. Afr. Spec. Publ. 7, by one of us (H. G. Moser). Larval specimens came 121 p. from a variety of sources and we are especially CLARKE,W. D. indebted to the following persons for their gener- 1963. Function of bioluminescence in mesopelagic ous provision of material E. Bertelsen and J. organisms. Nature (Land.) 198:1244-1246. DE LIGNY,W. Nielsen, Zoological Museum, Copenhagen; N. B. 1969. Serological and biochemical studies on fish Marshall and A. Wheeler, British Museum; W. populations. Oceanogr. Mar. Biol. Annu. Rev. 7:411-513. Nellen, Institute fur meereskunde, Kiel, Ger- EBELING,A. W., N. B. ATXIN, AND P. Y. SETZER. many; R. McGinnis and B. Nafpaktitis, Univer- 1971. Genome sizes of teleostean fishes: increases in some deep-sea species. Am. Nat. 105549-561. sity of Southern California (USC); R. J. Laven- EGE,V. berg, Los Angeles County Museum; R. Rosenblatt 1953. Paralepididae I. fParalepis and Lestidium). and R. Wisner, Scripps Institution of Oceanog- Taxonomy, ontogeny, phylogency and distribution. Dana raphy (SIO); T. Clarke and J. Miller, University of Rep., Carlsberg Found. 40, 184 p. 1957. Paralepididae 11. (Macroparalepis). Taxonomy, on- Hawaii; W. Freihofer, California Academy of Sci- togeny, phylogeny and distribution. Dana Rep., Carlsberg ences. We would like to thank B. Nafpaktitis and Found. 43, 101 p. R. McGinnis, USC; R. Wisner, SIO; J. Paxton, FISH, M.P., AND W. H. MOWBRAY. Australian Museum, Sydney; and G. Krefft, 1970. Sounds of Western North Atlantic fishes. Johns Fisheries Institute, Hamburg, Germany, for shar- Hopkins Press, Baltimore, 207 p. ing their vast knowledge of lanternfishes with us FRASER-BRUNNER,A. ’ 1949. A classification of the fishes of the family in numerous discussions. Discussions with N. B. Myctophidae. Proc. 2001. Soc. Lond. 118:1019-1106. Marshall, British Museum; B. Robison, Stanford GREENWOOD,P. H., D. E. &SEN, S. H. WEITZMAN,AND G. s. University; and A. Kendall, Middle Atlantic MYERS. Coastal Fisheries Center, National Marine 1966. Phyletic studies of teleostean fishes, with a provi- sional classification of living forms. Bull. Am. Mus. Nat. Fisheries Service (NMFS);Sandy Hook, have been Hist. 131341-455. helpful in stimulating some of the ideas put forth G~~NTEIER,A. herein. We appreciate the able technical assis- 1887. Report on the deep-sea fishes collected by H.M.S. tance of Elaine Sandknop, Elizabeth Stevens, and Challenger during the years 1873-76. Rep. Sci. Res. Voy- Patricia Lowery, Southwest Fisheries Center La age H.M.S. Challenger 22:335 p., 73 plates. JOHNSON,R. K. Jolla Laboratory, NMFS. Kenneth Raymond 1971. A revision of the alepisauroid family Scopelarchidae kindly lettered the illustrations. (Pisces: Myctophiformes). Ph.D. Thesis, Scripps Inst. Oceanogr., La Jolla, 474 p. LITERATURE CITED Karnuvs, A. 1972. Die meso-und bathypelagischen Fische der BARHAM,E. G. Meteor-Rossbreiten-Expedition 1970 (2. und 3. Fahr- 1970. Deep-sea fishes lethargy and vertical orientation. In tabschnitt). “Meteor” Forsch.-Ergeb. Dl1:l-28.

412 MOSER and AHLSTROM: ROLE OF LARVAE IN SYSTEMATICS

McALLlmn, D. E. 1973. A review of the lantedshes (family Myctophidae) 1967. The significance of ventral bioluminescence in described by A. Vedel Thing. Dana Rep., Carlsberg fishes. J. Fish. Res. Board Can. 24537-554. Found. 83,46 p. MO~ER,H. G., AND E. H. AHISTROM. OHNO,S. 1970. Development of lanternfishes (family Myctophidae) 1970. The enormous diversity in genome sizes of fish as a in the California Current. Part I. Species with narrow- reflection of nature’s extensive experiments with gene eyed larvae. Bull. Los Ang. Cty. Mus. Nat. Hist. Sci. 7, duplication. Trans. Am. Fish. Soc. 99:120-130. 145 p. PAXTON,J. R. 1972. Development of the lanternfish, Scopelopsis mul- 1972. Osteology and relationships of the lanternfishes tipunctatus Brauer 1906, with a discussion of its (Family Myctophidae). Bull. Los Ang. Cty. Mus. Nat. phylogenetic position in the family Myetophidae and its Hist. Sci. 13, 81 p. role in a proposed mechanism for the evolution of photo- WIENER,R. L. phore patterns in lanternfishes. Fish. Bull., U.S. 1963. A new genus and species of myetophid fish from the 70641-564. South-Central Pacific Ocean, with notes on related genera NAFPAKTITIS,B. G. and the designation of a new tribe, Electronini. Copeia 1968. Taxonomy and distribution of the lantemfishes, gen- 196394-28. era Lobianchia and Diaphus, in the North 1971. Descriptions of eight new species of myctophid fishes Atlantic. Dana Rep., Carlsberg Found. 73, 131 p. from the eastern Pacific Ocean. Copeia 1971:39-54.

413