New Records ofthe Fish Genus Grammatonotus (Teleostei: Percifonnes: Percoidei: Callanthiidae) from the Central Pacific, Including a Spectacular Species in the Northwestern Hawaiian Islands l

Bruce C. Mundy and Frank A. Parrish 2

Abstract: A second species of Grammatonotzts from the Hawaiian Islands, ten­ tatively identified as G. macrophthalmus Katayama, Yamamoto & Yamakawa (Callanthiidae), is recorded from French Frigate Shoals and the Northampton Seamount based on observations from a research submersible. In the absence of collected specimens, identification was made by comparing characters visible in video images with previously published images and descriptions. The fish were observed from 340 to 440 m at or near rocky habitats with crevices. All of the observations were near current-swept areas that supported gold coral (Gerardia sp.) colonies, although the fish were never seen within the colonies. A habitat feature important for both Grammatonotus and Gerardia, such as current or planktonic food supply, may therefore influence distribution of the fish. Exten­ sive fish surveys conducted in comparable depths at other areas of the archipel­ ago have not encountered this species, with one poorly documented exception from trawling surveys. Two other range extensions of Grammatonotzts are in­ cluded herein: Grammatonotzts laysanus Gilbert from the Line Islands with a specimen collected at Christmas Island at 274 m and an unidentified Gramma­ tonotzts juvenile from the Tuamotu Archipelago at 705 m. Our examination of specimens and review of previous records of Grammatonotus indicate that this genus needs taxonomic revision.

THE FISH FAUNA of the Hawaiian Islands is ored examples of one such species (Plate better known than that of any other Pacific I) tentatively identified as Grammatonotus island group. Despite this, species previously macrophthalmus Katayama, Yamamoto & Ya­ unrecorded from these Islands continue to be makawa, 1982. The fish were encountered found. Many of these are small and/or deep­ during submersible dives in the Northwestern water species from poorly sampled habitats. Hawaiian Islands but could not be collected. In this paper, we document spectacularly col- Our identification is therefore based on com­ parisons of video images with published de­

1 Supported in part by an award from the National scriptions, illustrations, and photographs. Undersea Research Program's Hawai'i Undersea Re­ The only previous reference found to G. search Laboratory to the NOAA Fisheries Honolulu macrophthalmus in the Hawaiian Islands is Laboratory's Ecosystems and Environment Investigation. Fujii's (1986) brief review of seamount fish Dives at the Northampton Seamount were funded by an award from the NOAA Ocean Explorations Initiative. biogeography, which includes the species in Participation in this cruise by B. Mundy was part of the a table with the annotation "ESC, HI" (for NOAA Fisheries Honolulu Laboratory Coral Reef Eco­ the Emperor Seamount Chain and Hawaiian system Investigation multivessel 2001-2002 surveys of Islands). No other information was given. reef and reef-associated ecosystems in the Northwestern Hawaiian Islands. Manuscript accepted 22 July 2003. Given the scarcity of information on this 2 National Marine Fisheries Service (NOAA Fish­ species, the publication of our observations is eries), Pacific Islands Fisheries Science Center, Honolulu warranted. Laboratory, 2570 Dole Street, Honolulu, Hawai'i 96822. MATERIALS AND METHODS Pacific Science (2004), vol. 58, no. 3:403-417 Work of the U.S. Government Observations were made during Hawai'i Un­ Not under copyright dersea Research Laboratory (HURL) Pisces

403 404 PACIFIC SCIENCE· July 2004

165' 1!5S·W o Fujii (1986),specimens trawled from an o o unknown NWHI locality o DC 0 -~ RaitaBank ~ 7 dlveS Brooks Bank ~ 118-657 m 5 dives cfiJ ~ 0 50-550 m 0 0 ~~~l 25- ~~~o

NorthamPt~ 0 St Rogatien Bank Seamount 0 5 dives 12 sightings 110-567 m in 4 dives 302-1044 m East French o KahO'olal..w~e---:-:'::":;:~j;~~f;:~ o 0 Frigate Shoals 13 dlves- o 4 sightings ---__152·459 m Hawsi'; Island 20' in 5 dives /-:J 0 ~ross Seamount 2 dives 171·675 m '/ ~6dives ~ 385-593 m _ c...... r-- 343-675m ~ no· 16!)" ---- 160*

FIGURE 1. Locations in the Hawaiian Islands where Grammatonotus macrophthalmus has been observed from HURL Pisces submersibles (asterisks) and locations where other Pisces submersible fish census dives were conducted during 1997-2002. The number of dives and their combined depth ranges (meters) are given for each location.

V submersible dives in the Northwestern with the precious corals and adjacent habitats. Hawaiian Islands (NWHI) (Figure 1) on 8-9 These dives were conducted in conjunction September 2001 at the east extension of the with another set of submersible surveys that French Frigate Shoals (FFS) platform (23 0 looked at fish assemblages at the same depths 55' N, 165 0 23' W) and on 20 September and in similar habitats at several other 2002 at the Northampton Seamount (25 0 Northwestern and main Hawaiian Islands lo­ 16.9' N, 172 0 01.3' W). The dives (P5-455­ cations (Figure 1). 456, P4-055) were conducted to survey Collections of fishes were not possible, so deepwater foraging areas of the endangered all counts and measurements were taken from Hawaiian monk seal, Monachus schauinslandi video images and are therefore approximate. Matschie. The FFS location contains a gold The images had no calibrated scale for mea­ coral (Gerardia species) bed, where seals from surements. Therefore, morphometric data are the FFS colony are known to feed (Parrish expressed as proportions of a relative refer­ et al. 2002), with colonies encrusting pin­ ence measurement (typically standard length nacles spread over a 2-ha area. Submersible [SL], head length [HL], or postorbital length) surveys of this bed were conducted in 1998 taken in the same video frame. Measurements (one dive), 2000 (one dive), and 2001 (two were those given as diagnostic for Gramma­ dives). The Northampton Seamount site is tonotus species by Katayama et al. (1982) and also a foraging area for seals (Brent Stewart, Anderson (1999), using landmarks specified Hubbs Sea World, pers. comm., October therein. Measurements were taken with di­ 2002), but the deepwater habitat was unex­ viders from individual frames on the video plored before the three dives of our 2002 screen and checked to the nearest 0.5 mm survey. All submersible dives lasted 6-7 hr against a ruler. Measurements were taken at the bottom. The dive objectives were to from only those frames that best showed a identify, count, and obtain size estimates of fish viewed either laterally for most measure­ potential monk-seal prey species associated ments or dorsally for head, interorbital, and Central Pacific Grammatonotus . Mundy and Parrish 405

body widths, without foreshortening of the were seen together on 8 September 2001, and fish's image from turning. Images in which three fish were seen individually on 9 Sep­ individual fish were filmed at close range in tember 2001. The first sighting of the second the frame's center were used to avoid image day was at the same location as the initial distortion at frame edges and to reduce mea­ sighting and was possibly one of the in­ surement error due to small image size rela­ dividuals seen on 8 September. We therefore tive to measurement precision. Only two suspect that we observed four individuals at individuals were filmed adequately for mea­ three sites within the same general area of the surements to be taken. Images of one of bank. these allowed us to obtain many body mea­ On 20 September 2002, 12 individuals of surements, but the other allowed only head this species were found at a second location, measurements. Therefore, single values are the Northampton Seamount (Figure 1). The given for most body proportions, but a range first sighting at this location was of a single of values is given for some head proportions. individual at 440 m (9SC) on a flat carbo­ Grammatonotus specimens at the NOAA nate substrate (that at first appeared not to be Fisheries Honolulu Laboratory (NMFS) and near high-relief habitat). The fish took cover the Bernice P. Bishop Museum (BPBM) col­ inside a small crevice as the submersible ap­ lections were examined to obtain compara­ proached. The substrate changed to a sheer tive data, as were photographs of G. crosnieri vertical wall, estimated to be over 100 m high, (Fourmanoir) provided by W. Anderson a few dozen meters away from the site of this (Grice Marine Biological Laboratory). first observation. A second sighting of the species on this dive occurred at a ledge and crevice on the wall face at 420 m (9.8°C), just RESULTS up-current of a ridge with colonies of gold Observations coral. One individual was again seen briefly as it took cover under a small arch at the crevice Our first sighting of G. macrophthalmus was on the cliff face. Ten more individuals were of two individuals at FFS (Figure 1) at 352 m seen after the submersible moved up to 340 m by submersible pilot T. Kerby and FAP. (12.1°C) and traveled downcurrent over flat The fish were obviously unlike any previously carbonate substrate at the top of the cliff face. seen during thousands of hours of submers­ All of the fish were seen close to small crev­ ible observations in the Hawaiian Islands at ices within which they took shelter. As with comparable depths, including previous sur­ the first sighting of the species at this loca­ veys in the same area. The most distinctive tion, no high-relief substrate was visible im­ characteristics of the fish were their striking mediately adjacent to the crevices where the color patterns and elongated tails. The slow fish were seen, but the drop-off to the cliff movement of one of the fish stunned by was just beyond our range of vision. All of the bright light of the submersible enabled these individuals quickly hid in crevices so F.A.P. to record close-up video images with that video images could not be obtained at the submersible's digital video system (Plate the Northampton Seamount. I). Examination of the video by the scientists aboard the R/V Ka'imikai-o-Kanaloa in the Description evening after the dive led to unanimous agreement that the fish was not a species re­ The sizes of the fish were not measured di­ corded from the Hawaiian Islands in taxo­ rectly but were approximately 20 cm total nomic reviews of the fishes available at that length and 12 cm SL. time (e.g., Chave and Mundy 1994, Randall The greatest body depth was 3.6 in SL. 1996, Mundy in press, and references cited The eye diameter (2.0-2.3 in HL) was therein). greater than the postorbital length (2.4-2.9 in Four sightings of the fish at FFS, at 350­ HL). The body was elongate and compressed. 360 m, occurred in 2 days of diving. Two fish The body width at the pectoral-fin bases was 406 PACIFIC SCIENCE· July 2004

6.4 in SL. The caudal peduncle length was middle three areas extended as bars ventrally greater than its depth, 1.7 in HL. The dorsal on the sides, becoming faint on the lower profile of head was gently convex from the sides. There was a stripe of diffuse magenta dorsal origin to the blunt, rounded snout. behind the eye to the posterior opercular The snout was short, ca. 4.3-5.1 in HL. The edge. The iris was silvery except lavender interorbital width was narrower than the eye dorsally. The chin, branchial region, thorax, diameter, 2.6 in HL. The head width was and anterior ventral abdominal region were 4.8 in SL and 1.6 in HL. The mouth was diffuse light blue, appearing gray in shadow moderately large and oblique. The maxilla when viewed from above. The pectoral fin reached below the anterior edge of the pupil, was transparent. The pelvic fin was reflective 2.3 in HL. The nostrils, teeth, opercular bluish white anteroventrally with bright white morphology, scale morphology, and lateral tips, becoming translucent greenish white line pores were not visible. The dorsal fin was posterodorsally. The dorsal fin was golden unnotched, with XI spines and 9 soft rays. ocher proximally, with a light blue stripe on The other fin elements were not countable. its upper half and a purple stripe with pink The insertion positions of the fins with re­ spinous and soft ray tips distally. There was a spect to other landmarks were not accurately second light blue stripe on lower half to one­ determinable. The third anal spine was third of the spinous dorsal-fin rays posteri­ longer than the first dorsal spine but not ac­ orly. The anal fin was light ocher proximally, curately measurable. The shortest dorsal fin with a light blue stripe on the proximal one­ spine was the first, 4.4 in HL. The longest third to half, a distal purple stripe, and laven­ dorsal fin spine was probably the seventh, 2.3 der pink tips on the spinous and soft rays. in HL. The longest dorsal soft ray and The caudal fin base was golden ocher, the longest anal soft ray were each ca. 1.1 in HL. free filaments of the central rays were blue, The pectoral fin length was not measur:lble. the outer rays were purple with magenta The pelvic fin length was 3.8 in SL. There outer edges, and the tips of the filamentous were 10 caudal rays produced into filaments, rays were pale bluish white at some angles. the longest 1.2 in SL. About 27 scale rows were counted from the nape to the caudal fin Habitat and Behavior when viewed dorsally. The fish were observed in a variety of hab­ Coloration itats at 340-440 m, but all of the sites were near high-relief areas free of sediments, in­ Under dim illumination when first seen at dicative of strong currents. Colonies of gold distance, the fish were uniformly pale whitish coral (Gerardia sp.) were found near the areas blue with brighter bluish white pelvic fins where the species was observed, but the fish and distal tips of the filamentous caudal fin were never seen within the immediate area of rays. When illuminated next to the submers­ the colonies. ible (Plate I), the fish showed the following The fish at FFS were observed at mixed colors: The body was golden yellow from the basalt cobble and sand patches at or near posterior head to the abdomen, grading pos­ undercut basalt outcroppings of moderate teriorly to golden ocher on the tail to the (0.5-1 m) relief, adjacent to high-relief basalt base of the caudal fin. The snout and inter­ outcrops that had bushes of gold coral. The orbital area to the insertion of the dorsal fin 2-ha coral bed is the only known habitat of were magenta. The dorsum under the dorsal this composition at FFS. Four of the sight­ fin and on the caudal peduncle was lavender ings were close to outcrops, but one observa­ pink. Five triangular pink areas extended tion was of a fish swimming ca. 1 m above from the dorsal stripe ventrally on the dorsal flat, semiconsolidated substrate many meters one-third of the sides, approximately beneath away from the nearest shelter. The species spinous dorsal-fin rays 2-5, 7-9, 3-5, 8-9 was not observed during extensive surveys (2­ and beneath middle of caudal peduncle. The 4 km in length) radiating N, S, E, and W Central Pacific Grammatonotus . Mundy and Parrish 407

from the coral bed in 1998, 2000, and 2001, DISCUSSION over uniform carbonate and manganese crust Identification bottom that composes the east extension of the FFS platform. The gold coral bed was The morphology of the fish suggested that the shallowest point, at 350-360 m, on the they were a member of the family Call­ platform, which had depths of 370-400 m at anthiidae, although many diagnostic charac­ most of its highest area. The fish were seen ters that distinguish the Callanthiidae from near the coral beds or on the area of bottom superficially similar taxa (Anthiinae, Symphy­ immediately adjacent (ecotonal) to the coral sanodontidae) found in this depth range were pinnacle. Temperatures at this site were 9.1­ not visible in the video images. Those char­ 14.0°C, logged by temperature recorders left acters include the lateral line morphology, the there for a year. number of spines on the opercular bone, The fish at the Northampton Seamount the internal morphology of the nasal organ, were mostly seen on flat carbonate substrate the branchiostegal ray number, and the mi­ covered with a thin layer of sediment and croornamentation of the scales. Key mor­ with small crevices, adjacent to a vertical car­ phological features used to identify the fish as bonate wall. One individual was seen at a callanthiids include the presence of spines on larger unsedimented crevice with an arch the fins, thoracic pelvic fins, an undivided and on this wall. No basalt was present at the unnotched dorsal fin, the relative lengths of Northampton Seamount site. Gold coral col­ the dorsal and anal fins, the number of dorsal onies on this wall were growing on a sharp­ spines as indicated by incised notches along edged ridge, several meters downcurrent of the dorsal edge of that fin's membrane, the the arched crevice where one of the fish was shapes and sizes of the fins, the number of found and several dozens of meters below the scale rows, the short rounded snout, and the top of the wall where most of the fish were eye size (Plate I, Table 1). Comparison of seen. The fish at the Northampton Seamount these characters with Carpenter's (1999) re­ were seen at 340-440 m where temperatures view of characters diagnostic for central Pa­ were 9.5-12.1°C. No long-term temperature cific teleost families eliminates all families records are available from the Northampton except the Callanthiidae for these fish. Seamount. This fish species was not seen on The characters that unequivocally dif­ two other dives at the Northampton Sea­ ferentiate callanthiid genera (Anderson and mount, even though those dives were in the Johnson 1984) are not visible without speci­ same depth range and included habitats with mens in hand, although our tentative dorsal­ gold coral colonies similar to the habitats where fin soft ray count agrees with Grammatonotus the fish were seen on 20 September 2002. (8-10, usually 9 rays) instead of Callanthias The fish either were stunned by the sub­ (10-11 rays, rarely 12). Most Callanthias spe­ mersible's lights or, in most sightings, moved cies are deeper bodied with shorter snouts quickly to shelter within crevices. All indivi­ than Grammatonotus (see Masuda et al. 1984, duals swam slowly with labriform motion, Anderson 1999). We also infer that Pacific sculling with their pectoral fins, keeping their Callanthias species have more scale rows bodies straight and their caudal fins wide­ (>33) than Grammatonotus (25-26 [Table 1]), spread, until alarmed by the submersible's although we have found no direct compari­ proximity. When startled, they darted rapidly son of this in the literature. We reached this with tail flexions. conclusion from the following comparison of The fish were seen singly or in pairs, but the various methods of counting longitudinal not in larger groups. However, we cannot body scales used in various taxonomic treat­ discount the possibility that other Gramma­ ments of callanthiids. The number of tubed tonotus with different color patterns and lateral-line scales is 25-42 in Callanthias (see shorter caudal fins, seen near the long-tailed Anderson and Johnson 1984) versus 14-17 in fish, could have been other life-history stages Grammatonotus (see Katayama et al. 1980, of the same species. Katayama et al. 1982, Anderson 1999). The 408 PACIFIC SCIENCE· July 2004

TABLE 1 Counts and Proportions of the Putative Grammatonotus macrophthalm1lS from French Frigate Shoals and Other Pacific Grmnmatonotlts Species

French Frigate Character Shoals Fish' G. macrophthalmus G. laysanzts G. Slwllgaensis G. crosnieri

Standard length (mm) ca. 120? 95-119 129 63-81 35-119 Dorsal fin elements XI,9? XI,9 XI,9 XI,9 XI,9 Scale rows 27 (25?) 25 26 25-26 25-26 Body depth in SL 3.6 3.00-3.50 3.07-3.31 2.79-2.93 2.4-3.3' Body width in SL 6.4 5.89-7.19 6.59-7.17 5.73-6.41 No data Pelvic fin length in SL 3.8 3.63-4.46 3.63-4.05 2.8-5.4?' Caudal fin length in SL 1.2 1.18-1.85 1.79-1.87 2.25-3.06 1.01-1.03' Snout length in HL 4.3-5.1 4.43-5.70 4.19-4.67 4.80-5.25 5.44-8.00' Orbit diameter in HL 2.0-2.3 2.17-2.43 2.55-2.59 2.63-2.68 2.17-2.82' Upper jaw length in HL 2.3 2.24-2.50 2.20-2.21 2.18-2.92 2.17-2.80' Interorbital width in HL 2.6 3.11-3.57 3.26-3.50 2.74-2.92 No data Postorbital head 2.4-2.9, ;£ orbit 2.30-2.67, > orbit > orbit 1.86-2.63,* length in HL diameter < orbit diameter diameter diameter ;£ orbit diameter Caudal peduncle 1.7 1.53-1.74 1.33-1.45 1.78-1.94 1.16-1.72' length in HL Caudal peduncle 2.2 1.89-2.24 1.83-2.08 ca. 1.78-1.94 1.63-1.98' depth in HL Length of 1st dorsal 4.4 5.42-6.84 3.76-7.00' spine in HL Length of longest dorsal 2.3 2.13-2.53 2.44-2.33 2.24-2.40 1.46-2.18' spine in HL Length of longest dorsal 1.1 1.09-2.23 1.33-1.38 1.63-1.85 0.96-1.41' ray in HL Length of longest anal 1.1 1.09-1.77 1.22-1.40 2.08-2.63 0.81-1.58' ray in HL

Sources: Our data for the French Frigare Shoals fish; Katayama et al. (1980) for G. surogaensis; Katayama et al. (1982) and Matsuura and Tachikawa (1994) for G. 11lacrophthat",us; Katayama et al. (1982) for G. mysa"us; and Fourmanoir (1981), Anderson (1999), and unpublished photographs provided by W. Anderson for G. crosnieri. *, data from photographs. lateral line in callanthiids ends near the from specimens. The fish observed by us thus dorsal-fin insertion or on the caudal peduncle agree with Pacific Grammatonotus species in­ (Anderson 1999), so the number of lateral­ stead of Pacific Callanthias species in body line scales is less than or equal to the number depth, snout length, dorsal-fin spine count, ofmidbody lateral scales or scale rows. Pacific and lateral-scale row count. Callanthias species have >33 lateral-line scales The identification of our fish to species with the exception of C. parini Anderson & proved more problematic. Four species, G. Johnson, which, although it has only 25-30 laysanus Gilbert, 1905; G. surugaensis Kata­ lateral-line scales, has 38-46 midbody lateral yama, Yamakawa & Suzuki, 1980; G. macro­ scales (Anderson and Johnson 1984). We phthalmus Katayama, Yamamoto & Yama­ were unable to count tubed lateral-line scales kawa, 1982; and G. cromieri (Fourmanoir, in our images, but 27 scale rows were ob­ 1981), were included in the most recent served dorsally on the FFS fish. We consider reviews of the genus (Katayama et al. 1982, this to be congruent with the 25-26 scale Kanou et al. 2003). Grammatonotus cromieri rows reported for Grammatonotus species was not included in Katayama et al. (1982), (Katayama et al. 1980, Katayama et al. 1982) likely because it was originally described as because our count included scales at the nape being in a different genus, Callanthias, at ap­ not included in the scale row counts taken proximately the same time as their paper was Central Pacific Grammatonotus . Mundy and Parrish 409 written. Two other species have been omitted those in our images. The color values of the from most publications on callanthiids. Heli­ markings of the photograph in Okamura and astes roseus Giinther, 1880, described from Amaoka (1997) differ from our images in be­ the Ki (Kai) Islands, Arafura Sea, Indonesia, ing more red, but the pattern is almost iden­ is also a Grammatonotus species of uncertain tical. Differences in the head and anterior status (W. Anderson, Grice Marine Biolog­ body shape can be explained as the difference ical Laboratory, pers. comm., November between our specimen photographed alive 2002). (Kanou et al. [2003] stated that C. and Okamura and Amaoka's (1997) specimen roseus is a pomacentrid, but only the genus photographed after death, probably by suf­ Heliastes is a synonym ofa pomacentrid taxon.) focation when brought to the surface. Sim­ Fourmanoir (1976) gave G. laysanus as a ju­ ilarities in the color pattern of this image with nior synonym of G. roseus (as Zabulon roseus), ours include a dorsal stripe with bars extend­ but this hypothesis needs reevaluation. Para­ ing ventrally down the upper one-third of barossia lanceolata Kotthaus, 1976, described body, with four of these bars under the dorsal from Socotra Island, northwestern Indian fin and the fifth on the upper center of caudal Ocean, was listed by Eschmeyer (1998) as a peduncle; a bluish lavender upper edge on the valid callanthiid species in that third genus, iris; dark edges to the tips of the dorsal and but in Eschmeyer's (1998) list of genera, anal fin elements and membrane edges, with Parabarossia was considered a synonym of these fins otherwise yellow or yellow orange; Grammatonotus following Springer (1982) and a yellow orange ground color to most of the Mooi (1993). This also needs more investiga­ lateral body and caudal peduncle below the tion (W. Anderson, Grice Marine Biological dorsal stripe and bars; a yellow orange proxi­ Laboratory, pers. comm., November 2002). mal center of the caudal fin; and purple, blue, The Grammatonotus species most similar and bluish lavender greatly prolonged free to the one in our images is G. maerophthalmus, caudal rays. This unusual shared color pat­ as shown by a photograph of a badly dam­ tern, along with the general appearance and aged trawled specimen in Okamura et al. fin morphology of the fish in our images and (1982 :220), reprinted in Masuda et al. (1992). in Okamura and Amaoka (1997) convinced us This fish has a general appearance that that our fish are best identified as G. maero­ matched that of the fish in our images, in­ phthalmus, pending collection of specimens. cluding greatly produced caudal rays with a The fish observed at FFS shared morpho­ violet or bluish lavender color. But the fish metric characters only with G. maerophthal­ illustrated in Okamura et al. (1982) lacks the mus among the species reviewed by Katayama distinctive color pattern of our fish, described et al. (1982) (Table 1). In addition, the eye on their p. 373 as "Body yellowish pink. diameter described for G. roseus (scarcely two­ Dorsal and anal fins with pink margins. Cau­ fifths of head length [Gunther 1880]) does dal fin pink. Pectoral and pelvic fins pale not match that of either the fish observed at pink." This description does not match FFS or G. macrophthalmus (Table 1). The eye our images. The original description of the diameter of G. lanceolatus is 2.3 in HL but is color of G. maerophthalmus (Katayama et al. less than the postorbital length as calculated 1982:373) also does not match the fish we from other head measurements given in the observed: "Body yellowish pink; iris pink; description of the species (Kotthaus 1976). dorsal and anal fins with pink margins; caudal (Osteological and larval pigment characters fin pink; pectoral and pelvic fins pale pink." used by Kanou et al. [2003] to identify juve­ The identity of our fish is revealed by the nile Grammatonotus were not available from photograph of G. maerophthalmus in Okamura our visual observations.) and Amaoka (1997:280). Their photograph Sufficient descriptive detail was therefore also seems to be of a damaged, trawled speci­ given for all nominal Grammatonotus species men, but unlike the photograph in Okamura to make our identification with confidence, et al. (1982), it shows a dorsal stripe and bars with one exception. The original description on the upper one-third of the body similar to of G. crosnieri gave only body depth as mor- PLATE 2. Top, An unidentified Grammatonotus species with a diamond-shaped caudal fin photographed from the Hawai'i Undersea Research Laboratory Pisces V submersible at 248 m, Penguin Banks, Moloka'i, Hawai'i (HURL dive PS-426; submersible pilot, C. Holloway; camera operator, C. Kelley). Bottom,. A color variety of Grammatonotlls laysanus? photographed from the Hawai'i Undersea Research Laboratory Pisces V submersible at 372 m, Cross Sea­ mount, southwest of Hawai'i Island, Hawai'i (HURL dive PS-SOl; submersible pilot, T. Kerby; camera operator, C. Kelley). Central Pacific Grammatonotus . Mundy and Parrish 411 small juveniles. It is noteworthy, however, but were unsuccessful.) Fujii (1986) cited Ja­ that larval and pelagic juvenile Grammatonotus pan Fisheries Agency (1974) as a source of from O'ahu and the Hancock Seamounts in information for his record. Grammatonotus our collection (NMFS HLIP-l, 18.7 mm SL; macrophthalmus was not mentioned in that HLIP-34, 4.2 mm SL) have orbit diameters report, but unidentified Serranidae were, and shorter than the postorbital head length, as Grammatonotus was placed in the Serranidae has been described for larvae and juveniles at that time. Within the Hawaiian Ridge, un­ from other regions (Fourmanoir 1976, Mis­ identified serranids were collected at an "un­ kiewicz et al. 2000, Kanou et al. 2003). The surveyed bank" (26° 18-19' N, 174° 31-32' use of this character for species identification W = Bank 8), the Ladd Seamount (28° 31­ of juvenile Grammatonotus therefore needs 34' N, 176° 35-43' W), and the Hancock further evaluation. Seamounts (29° 46'-30° 18' N, 179° 02-48' E), all well to the northwest of FFS and Ecology and Biogeography the Northampton Seamount (Japan Fisheries Agency 1974). Grammatonotus macrophthalmus Previous records of G. macrophthalmus are has not been seen in numerous other sub­ from the Kyushu-Palau Ridge at 300-500 m mersible and remotely operated vehicle sur­ (Okamura and Amaoka 1997), from a veys south of FFS in the Northwestern and beached specimen in the Ogasawara Islands main Hawaiian Islands (Chave and Mundy (Matsuura and Tachikawa 1994), and from 1994, Chave and Malahoff 1998) (Figure 1). unspecified depths in the Emperor Sea­ Twenty-six other fish species are recorded mounts and Hawaiian Islands (Fujii 1986). in the archipelago from only the NWHI Our observations from 340 to 440 mare (Mundy in press). therefore within the known depth range of Grammatonotus macrophthalmus undoubt­ the species. edly occurs at locations within the NWHI A habitat attribute shared by the Hawaiian other than the two from which our observa­ Archipelago sites where G. macrophthalmus tions came. Its patchy distribution, limited was observed is proximity to high-relief depth range, and habitat preferences make substrate with gold coral colonies. Gold cor­ observation and sampling of it difficult. As als grow at ca. 300-500 m in areas where suggested before, the proximity of G. macro­ currents sweep planktonic food organisms phthalmus to high-relief substrate would limit through the colonies and sweep sediments its susceptibility to capture by trawling. away from the habitat (Grigg 1974 as Para­ Grammatonotus species are not taken in traps zoanthus sp., 1984, Parrish et al. 2002). Little and are too small to be taken by conventional else is known about the habitat requirements hook and line fishing methods. Numerous of gold corals. The fish were not seen within submersible observations in the Northwest­ the colonies, only near them. We therefore ern Hawaiian Islands have not found G. suggest that there are shared habitat require­ macrophthalmus, but many of the submersible ments such as currents and food supply that surveys in which fishes were the primary tar­ promote the proximity of the fish and the get have focused on depth ranges other than coral. We also note that both species are very those at which the species occurs (Chave patchy and clumped in their known distribu­ and Mundy 1994, Chave and Malahoff 1998) tions within the Hawaiian Archipelago. Gold (Figure 1). corals are known from many more areas than Grammatonotus macrophthalmus is other­ are G. macrophthalmus. wise known only from the Kyushu-Palau Fujii's (1986) records of G. macrophthalmus Ridge at ca. 26° 13.5-47.2' N, 135° 20.7­ from the Hawaiian Islands are likely from 52.1' E (Katayama et al. 1982) and the Oga­ the northwestern part of the ridge, given the sawara Islands (Matsuura and Tachikawa known distributions of the other species 1994, Kanou et al. 2003). It is one of 57 spe­ included in his paper. (We attempted to cies known only from the western North Pa­ contact E. Fujii to obtain more information, cific and the Hawaiian Islands, including 37 412 PACIFIC SCIENCE· July 2004

demersal or benthic species (Mundy in press). (Grigg and Bayer 1976, Grigg 1981, Springer Other examples of western North Pacific 1982, Randall et al. 1985, Randall 1996, and Hawaiian Islands deep-dwelling demer­ Mundy in press). Two pathways proposed sal fishes are Gymnothorax ypsilon Hatooka in the literature seem possible to connect & Randall, Lophioides miacanthus (Gilbert), the Hawaiian Archipelago with the western Hoplostethus erassispinus Kotlyar, Satyrichthys North Pacific: (1) the Subtropical Coun­ engyceros (Giinther), Parapercis roseoviridis tercurrent (STCC) moving over the region (Gilbert), and an unnamed Argyripnus species of the Marcus-Necker Ridge and (2) the that was also found for the first time in the Kuroshio/North Pacific Current moving Hawaiian Islands during our surveys. It is from the Kyushu-Palau Ridge past Japan and possible that the ranges of some deepwater the Ogasawara Islands and then eastward to species with distribution patterns limited to the Emperor Seamounts and northernmost the western North Pacific, the Emperor Sea­ end of the Hawaiian Ridge. mounts, and Hawaiian Islands are artifacts of The Marcus-Necker Ridge extends south­ limited deepwater sampling at Pacific islands. westward of and includes Necker Island, the As an example, G. laysanus was reported as island nearest to FFS (Grigg and Bayer 1976). a Johnston Atoll, Hawaiian Islands, and Em­ The STCC, a seasonably variable current peror Seamount endemic species (e.g., Ka­ of broad latitudinal extent and poorly defined tayama et al. 1982, Randall et al. 1985) until relative to the Kuroshio Extension current it was found at the Sala-y-G6mez Ridge dur­ to the north (Qiu 1999), is assumed to flow ing deepwater trawling (Parin et al. 1997). intermittently over this ridge (Grigg 1981). (Earlier reports of G. laysanus juveniles from The Marcus-Necker Ridge extends westward midwater sampling at New Caledonia [Four­ from the NWHI to the Line Island Ridge manoir 1971, 1976] were overlooked, but north ofJohnston Atoll, giving an easily seen the identifications of the New Caledonian connection at least that far. But westward of specimens need reevaluation considering the Johnston Atoll only isolated groups of deep problematic taxonomy of the genus.) How­ seamounts continue the connection to the ever, some of the 57 western North Pacific Ogasawara Islands or Kyushu-Palau Ridge. In and Hawaiian Islands species are well sampled addition, even the shallowest portion of the and easily identified, indicating that their Marcus-Necker Ridge near the Hawaiian Is­ known distribution is not a sampling artifact. lands is at depths of 1,660-2,693 m, deeper When G. maerophthalmus was found in a than the habitat for G. maerophthalmus. This limited area at FFS in 2001, the question would have been true even during known arose as to whether the fish seen were waifs sea level regressions (Grigg and Bayer 1976). or part of a previously unrecorded resident The distances between stepping-stones along species with a reproducing population. The the Marcus-Necker Ridge seem too great for discovery of Fujii's (1986) records and the the ridge to have been the most likely conduit sightings of more individuals at the North­ for dispersal of G. maerophthalmus to the Ha­ ampton Seamount support our contention waiian Islands on any time scale. that G. maerophthalmus is a resident Hawaiian In contrast, the stronger and more consis­ Archipelago species. tent Kuroshio Extension connects the west­ The discovery of G. maerophthalmus in ern Pacific to the Hawaiian Islands by flowing the Hawaiian Islands prompts inquiry into from southern Japan to the northernmost the origins of its occurrence there. Contem­ Hawaiian Islands. This has been suggested as porary biogeographic theory suggests that a mechanism for colonization of the Hawai­ the distributions of Hawaiian Archipelago ian Islands by western North Pacific fish marine organisms have been formed pri­ species (Randall et al. 1985). This pathway marily from vicariance linked with changing is likely less general for Hawaiian Islands ocean circulation and colonization of the ar­ fishes than has been suggested, but its efficacy chipelago from other areas via larval dispersal is supported by the 4.6% of the fauna (57 and stepping-stones of suitable adult habitat species) with ranges restricted to the west- Central Pacific Grammatonotus . Mundy and Parrish 413 ern North Pacific and the Hawaiian Islands .0rg/Photos/ThumbnailsSummary.cfrn?ID= (Mundy in press). Grammatonotus macro­ 27474) show fish with (1) a damaged caudal phthalmus is one of those species. We are not fin that appears to be double emarginate with implying by this that G. macrophthalmus is a a filamentous upper principal ray and (2) a recent colonist of the Hawaiian Islands, how­ diamond-shaped caudal fin without filaments. ever. The genetic differentiation of the spe­ The image of the fish with a double emargi­ cies throughout its range would have to be nate caudal fin with filaments matches G. studied before the antiquity of its distribution laysanus by having an eye diameter less than pattern could be determined. the postorbital head length. The image of the Little is known about the early life his­ fish with the lanceolate tail has an eye diam­ tories of callanthiids to support or refute eter equal to the postorbital length, casting hypotheses about their dispersal (Trnski and doubt on that specimen's identity. Miskiewicz 1998, Miskiewicz et al. 2000). A It has been suggested that callanthiids prolonged larval duration for the genus might are protogynous hermaphrodites (Anderson be indicated by the relatively large size of 1999). It was also suggested that Callanthias pelagic juveniles (23 mm SL [Miskiewicz et al. parini is sexually dimorphic, with larger 2000)), as well as by the size of newly settled individuals having longer dorsal, anal, and juveniles that retained larval pigment patterns caudal fin rays than smaller individuals (An­ (32 and 39 mm SL: BPBM 37191 and 37135). derson and Johnson 1984). One hypothesis But such inferences from larval and juve­ for the identity of the diamond-tailed pheno­ nile sizes need verification from otolith daily type is that it could be the juvenile and/or growth ring analysis before they can be con­ female morph of one or both of the other sidered valid. And even with such precise in­ species. The diamond-tailed Grammatonotus formation, pelagic larval durations are poorly and G. laysanus are sometimes seen together correlated with dispersal abilities, as measured but are more often seen in groups of only one by geographic ranges (Jones et al. 2002). type (c. Kelley, HURL, pers. comm., No­ vember 2001). The G. macrophthalmus in­ Additional Taxonomic Problems with dividuals that we observed are much larger Grammatonotus Species in the Hawaiian than the diamond-tailed fish, but we have Islands seen G. laysanus that are smaller, the same size, and larger than the diamond-tailed phe­ Three morphological phenotypes of Gram­ notype. matonotus have been observed during Pisces V Seven specimens of Grammatonotus were submersible surveys in the Hawaiian Islands, available to us, but of these two are recently raising the question ofhow many species exist settled juveniles, one is a pelagic juvenile, and in the archipelago. These phenotypes are one is a larva. The three largest specimens distinguished by the shape and extensions of (BPBM 22757, 123 mm SLj NMFsH P-I13, rays of the caudal fin: (1) double-emarginate 145 mm SLj BPBM 39292, 62 mm SL) all with filamentous outer principal rays (these have double emarginate tails with filamentous appear as round tails without filaments when outer principal rays. The smaller specimens viewed at a distance from the submersible), (BPBM 37191, 39 mm SLj BPBM 37135, 32 mm identified as G. laysanus; (2) broomlike, with SLj NMFS HLIP-1, 18.7 mm SLj NMFS HLIP­ extremely long, filamentous caudal fin rays, 34, 4.2 mm SL) have broken caudal fin rays identified as G. macrophthalmus; and (3) dia­ but appear to have had slightly lanceolate to mond-shaped or lanceolate (Plate IT, top). truncate caudal fins. This does not refute the The diamond-tailed phenotype has not been suggestion that caudal fin shape is related to identified to species. reproductive state in this genus, but insuffi­ Two photographs by]. E. Randall identi­ cient specimens were available to investigate fied as G. laysanus on the International Center this further. We cannot reject the hypothesis for Living Aquatic Marine Resources Fish­ that the diamond-tailed Grammatonotus is a Base online database (http://www.fishbase third Hawaiian Islands species. 414 PACIFIC SCIENCE· July 2004

An additional complication in our efforts our knowledge, this is the first record of the to identify Grammatonotus from submersible genus from French Polynesia. As mentioned observations arose from color and fin mor­ before, this specimen has an orbit diameter phology differences in fish that otherwise diagnostic of G. maerophthalmus, but we are appear to be G. laysanus. Randall et al. (1985) reluctant to assign a species identification to described and published a photograph of G. such a small specimen in a group diagnosed laysanus at Johnston Atoll with a white or almost entirely by subtle morphometric dif­ red spot on the back and elongate pelvic, soft ferences and differences in adult color pat­ dorsal, and anal fins. Although those authors tern, fin morphology, tooth morphology, and stated that this morph was also seen in the other characters likely not to be fully de­ main Hawaiian Islands, it was not found in veloped in larvae and juveniles. Osteological numerous main Hawaiian Islands observa­ characters are useful for identification of ju­ tions in recent years (Figure 1). It was ob­ venile Grammatonotus (see Kanou et al. 2003), served again in November 2002 at Cross but we were unable to examine the osteology Seamount (18 0 40' N, 158 0 17' W), south­ of the specimens on loan to us. west of Hawai'i Island but far northeast The NOAA Fisheries Honolulu Labora­ of Johnston Atoll (c. Kelley, HURL, pers. tory had in its collection one specimen (now comm., December 2002) (Plate II, bottom). BPBM 39292, 62 mm SL) from Christmas Is­ We can say no more about this morph at land, Line Islands, taken in a shrimp trawl at present, except to note that it adds yet an­ 274 m. This specimen agrees with G. laysanus other mystery to be solved in Grammatonotus in all key characters: the orbit diameter (2.75 taxonomy and life history. in HL) is less than the postorbital length; the teeth are of moderate size; the caudal fin Other New Records ofGrammatonotus Species is rounded (i.e., double emarginate) with produced outer rays; the caudal peduncle is Two specimens of Grammatonotus that we longer than its depth; and the third anal spine examined were collected from the Line Is­ is longer than the first dorsal spine (Katayama lands and the Tuamotu Archipelago, island et al. 1982). The known range of the species groups for which we found no previous is now the Emperor Seamounts, the Hawai­ records of the genus. A third specimen was ian Islands, Johnston Atoll, the Line Islands, collected in the Hawaiian Islands, but in ex­ New Caledonia, and the Sala-y-G6mez Ridge ceptionally shallow water. These records are (Gilbert 1905, Fourmanoir 1971, Chen 1980, therefore also included in this paper. Katayama et al. 1982, Randall et al. 1985, A juvenile G.laysanus (BPBM 37191, 39 mm Parin et al. 1997). SL) was collected at Kahe Point, O'ahu, by J. During this study we attempted to obtain Culp with a hand net on sand at the remark­ additional literature, specimens, and visual ably shallow depth of 7.6 m. The fish was not sightings to clarify the identification of the remembered as being noticeably unhealthy Grammatonotus that we observed from the despite being so far above its usual depth Pisces submersibles. To our dismay, most of range (J. Culp, University ofHawai'i HURL, these additional sources of information in­ pers. comm., June 2003). It had remnants of voke additional taxonomic questions rather the larval pigment pattern, suggesting that than clarifications. We conclude that the ge­ it was recently settled. We know of no other nus Grammatonotus is greatly in need of taxo­ record of the genus from such shallow water, nomic revision. The potential for allometry except for the G. maerophthalmus that washed in eye size and other characters, for sexual up on a beach in the Ogasawara Islands dimorphism, and for geographic variability in (Matsuura and Tachikawa 1994). color pattern will likely necessitate the exam­ The other newly settled juvenile (BPBM ination of numerous specimens over a large 37135) was collected by J. Poupin with a size range each from a number of localities to dredge at 705 m at Hao Atoll, Tuamotu Ar­ resolve the species composition of the genus. chipelago (18 0 4' 12" S, 141 0 20' 12" W). To Unfortunately, the small size of fishes in the Central Pacific Granzmatonotus . Mundy and Pa1'77sh 415 genus, .the rugged habitats in which they unknown, specimen illustrated in Anderson often live, and the absence of sampling for (1999). deepwater fishes at Pacific island groups Uncatalo~ed?, 90 mm SL, locality un­ other than the Hawaiian Islands will make known, receIved from P. Fourmanoir. collection of the specimens necessary for this .Uncataloged?, 71 mm SL, Philippines, re­ revision very difficult. ceIved from P. Fourmanoir.

Material Examined ACKNOWLEDGMENTS SPECIMENS. BPBM 22757. Grammatonotus We thank the Hawai'i Undersea Research laysanus (1 of 9 specimens, the others on Lab~ratory (HURL) for the opportunity and loan). 123 mm SL. Hancock Seamount phySIcal support that enabled us to do this Hawaiian Archipelago. NOAA ship Townsend research. Terrr Kerby and Charles Holloway Cromwell cruise TC7702, Sta. 52. 26 May (HURL) proVIded expert submersible pilot­ 1977. 144-177 m. Nor'eastern bottom trawl. ing, found the G. macrophthalmus for us, and . BPBM 37191. Grammatonotus laysanus juve­ enabled us to obtain the video images. Chris­ mle. 39 mm SL. Kahe Point O'ahu Hawai­ topher Kelley (HURL), Jane Culp (HURL), ian Islands. 19 May 1989. 25 [7.6 ~]. Hand it Walter Ikehara (State of Hawai'i Department net on sand. Collector, Culp. J. of Land and Natural Resources), Robert BPBM 37135. Grammatonotus sp. juvenile. Moffitt (NOAA Fisheries Honolulu Labora­ 32 mm SL. Hao Atoll, Tuamotu Archipelago. tory), and Sean Corson (NOAA National 18 0 4' 12" S, 141 0 20' 12" W. 2 June 1990. Ocean Service) provided scientific assistance 705 m. Dredge. Collector, Poupin. J. and discussions about the identification of the NMFSH P-113. Grammatonotus laysanus. G. macr~phthalmus. Jane Culp provided copies 145 mm SL. SE Hancock Seamount Ha­ of the VIdeo and other data from these dives. waiian Archipelago. NOAA ship To./vnsend Rac~el Shackelford and Leping Hu (HURL) Cromwell cruise TC8205, Sta. 18. 6 Novem­ p~ovlded data on HURL bottomfish survey ber 1982. 140 fathoms [256 m]. Nor'eastern dIves. The crew and other scientists aboard trawl. the R/V Ko'imikai-o-Konaloa enabled us BPBM 39292. Grammatonotus laysanus. to complete our research. Ani Au NOAA 62 mm SL. Christmas Island, Line Islands. Fi?heries ~onolulu Laboratory libr;rian, ob­ NOAA ship Townsend Cromwell cruise TC60 tamed copIes of references. Keiichi Nemoto Sta. 53. 22 August 1972. 150 fathoms [274 (University of Hawai'i Joint Institute for m]. Shrimp trawl. Marine and Atmospheric Research) translated NMFS HLIP-34. Grammatonotus sp. larva. references from Japanese to English for us 4.2 mm SL. 20 km west of Hancock Sea­ ~nd provided assistance in contacting E. Fujii mount, Hawaiian Archipelago. NOAA ship m Japan. R. Moffitt, Michael Seki (NOAA Townsend Cromwell cruise TC8405, Sta. 130. Fisheries Honolulu Laboratory), and Ar­ 27 July 1984. 0-25 m. Tucker trawl net 3. nold Suzumoto (Bernice P. Bishop Museum, NMFS HLIP-1. Grammatonotus sp. pelagic Honolulu) made specimens available for ex­ juvenile. 18.7 mm SL. South of O'ahu Ha­ amination. William D. Anderson Jr. (Grice waiian Islands. NOAA ship Townsend Crom­ Marine Biological Laboratory) provided un­ well cruise TC32, Sta. 31. 80-121 m. Cobb published information on nominal species rnidwater trawl. of Grammatonotus. W. Anderson, R. Moffitt, Photographs of Grammatonotus crosnieri and C. Kelley reviewed the manuscript. provided by W. Anderson (Grice Marine Biological Laboratory, December 2002): MNHN 1978-79, 117 mm SL, Philippines, Literature Cited holotype of Callanthias crosnieri Fourmanoir (1981). Anderson, W. D., Jr. 1999. Callanthiidae. MUSORSTOM 3, Sta. 12, Philippines, length Groppos (also goldies and splendid 416 PACIFIC SCIENCE· July 2004

perches). Pages 2553-2556 in K E. Car­ Fujii, E. 1986. Zoogeographical features of penter and V. H. Niem, eds. FAO species fishes in the vicinity of seamounts. Pages identification guide for fishery purposes. 67 -69 in R. N. Uchida, S. Hayashi, and The living marine resources of the western G. W. BoeWert, eds. Environments and central Pacific. Vol. 4. Bony fishes part 2 resources of seamounts in the North Pa­ (Mugilidae to Carangidae). FAO, Rome. cific. NOAA Tech. Rep. NMFS 43. Anderson, W. D., Jr., and G. D. Johnson. Gilbert, C. H. 1905. The aquatic resources 1984. A new species of Callanthias (Pisces: of the Hawaiian Islands. Section II. The Perciformes: Percoidei: Callanthiidae) deep-sea fishes. Bull. U. S. Fish Comm. 23 from the southeastern Pacific Ocean. Proc. (2): 577-713. BioI. Soc. Wash. 97 (4): 942-950. Grigg, R. W. 1974. Distribution and abun­ Carpenter, K E. 1999. Guide to orders and dance of precious corals in Hawaii. Pages families. Pages 1548-1617 in K E. Car­ 235-240 in Proc. 2nd Int. Coral Reef penter and V. H. Niem, eds. FAO species Symp. Great Barrier Reef Committee, identification guide for fishery purposes. Brisbane, Australia. The living marine resources of the western ---. 1981. Acropora in Hawaii. Part 2. central Pacific. Vol. 13. Batoid fishes, chi­ Zoogeography. Pac. Sci. 35:15-24. maeras and bony fishes part 1 (Elopidae to ---. 1984. Resource management of pre­ Linophrynidae). FAO, Rome. cious corals: A review and application to Chave, E. H., and A. Malahoff. 1998. In shallow water reef building corals. Pubbl. deeper waters: Photographic studies of Stn. Zool. Napoli I Mar. Ecol. 5 (1): 57­ Hawaiian deep-sea habitats and life-forms. 74. University of Hawai'i Press, Honolulu. Grigg, R. W., and F. M. Bayer. 1976. Present Chave, E. H., and B. C. Mundy. 1994. Deep­ knowledge of the systematics and zooge­ sea benthic fish of the Hawaiian Archipel­ ography of the order Gorgonacea in Ha­ ago, Cross Seamount, and Johnston Atoll. waii. Pac. Sci. 30:167-175. Pac. Sci. 48:367-409. Giinther, A. 1880. Report on the shore fishes Chen, c.-H. 1980. Primary studies of de­ procured during the voyage of the H.M.S. mersal fish resources investigation on trawl Challenger in the years 1873-1876. In Re­ grounds at Kanmu Seamount. Bull. Tai­ port on the scientific results of the voyage wan Fish. Res. Inst. No. 32:317-337. [In of the H.M.S. Challenger during the years Chinese. English translation by W. G. Van 1873-1876. Zoology, 1 (pt. 6): 1-82, pIs. Campen, 1983, NMFS SWFSC Honolulu 1-32. [Reprint by J. Cramer, 1963.J Laboratory Translation No. 79.J Japan Fisheries Agency. 1974. Survey of the Eschmeyer, W. N., ed. 1998. Catalog of North Pacific seamounts. Report of Kaiyo fishes. Calif. Acad. Sci. Cent. Biodiversity Maru survey cruises, 1972 season. [In Res. Inf. Spec. Publ. No. 1. 3 vols. (recent Japanese. English translation by T. Otsu. revisions at http://www.calacademy.org/ NMFS SWFSC Honolulu Laboratory research/ichthyology/catalog/ Translation No. 54.J fishcatsearch.html). Jones, G. P., M. J. Caley, and P. L. Munday. Fourmanoir, P. 1971. Notes Ichtyologiques 2002. Rarity in coral reeffish communities. (IV). Cah. O.R.S.T.O.M., Ser. Oceanogr. Pages 81-101 in P. F. Sale, ed. Coral reef 9 (4): 491-500. fishes: Dynamics and diversity in a com­ ---. 1976. Formes post-Iarvaires et juve­ plex ecosystem. Academic Press, San Di­ niles de poissons cotiers pris au chalut ego. pelagique dans Ie sud-ouest Pacifique. Kanou, K, G. Shinohara, and K Shibukawa. Cah. Pac. 19:47-88. 2003. Juveniles of a rare callanthiid fish ---. 1981. Poissons (premiere liste). genus Grammatonotzts (Pisces: Perci­ Resultats de Campagnes MUSORSTOM, formes) from Japan. Bull. Nat!. Sci. Mus., I-Philippines (18-28 Mars 1976): 85­ Tokyo, Ser. A (Zool.) 29 (1): 39-46. 102. Katayama, M., T. Yamakawa, and K Suzuki. Central Pacific Grammatonotus . Mundy and Parrish 417

1980. Grammatonotus surugaensis, a new Mundy, B. C. In press. A checklist of the serranid fish from Suruga Bay and the fishes of the Hawaiian Archipelago. Ber­ Straits of Osumi, Japan. Biogeogr. Soc. nice P. Bishop Mus. Bull. Zool. Jpn. 35 (4): 45-48. Okamura, 0., and K Amaoka, eds. 1997. Sea Katayama, M., E. Yamamoto, and T. Yama­ fishes ofJapan. Yama-Kei Publishers Co., kawa. 1982. A review of the serranid fish Tokyo. genus Grammatonotus, with description of Okamura, 0., K Amaoka, and F. Mitani, eds. a new species. Jpn. J. Ichthyol. 28 (4): 368­ 1982. Fishes of the Kyushu-Palau Ridge 374. and Tosa Bay, Japan. Fisheries Resource Kotthaus, A 1976. Fische des Indischen Conservation Association, Tokyo. Ozeans. Ergebnisse der ichthyologischen Parin, N. v., A. N. Mironov, and K N. Ne­ Untersuchungen wahrend der Expedi­ sis. 1997. Biology of the Nazca and Sala tion des Forschungsschiffes "Meteor" in y Gomez submarine ridges, an outpost of den Indischen Ozean, Oktober 1964 bis the Indo-West Pacific fauna in the eastern Mai 1963. A Systematischer Teil, XVII. Pacific Ocean: Composition and distribu­ Percomorphi (7). "Meteor" Forsch.­ tion of the fauna, its communities and his­ Ergebnisse, Reihe D, No. 23:45-61. tory. Adv. Mar. BioI. 32:145-242. Masuda, H., K Amaoka, C. Araga, T. Uyeno, Parrish, F. A, K Abernathy, G. J. Marshall, and T Yoshino, eds. 1984. The fishes and B. M. Buhleier. 2002. Hawaiian monk of the Japanese Archipelago. Tokai Uni­ seals (Monachus schauinslandi) foraging in versity Press, Tokyo. Vol. 1 (text), Vol. 2 deepwater coral beds. Mar. Mam. Sci. 18 (plates). (1): 244-258. ---. 1992. The fishes of the Japanese Qui, B. 1999. Seasonal eddy field modulation Archipelago. 3rd ed. Tokai University of the North Pacific Subtropical Counter­ Press, Tokyo. Vol. 1 (text), Vol. 2 current: TOPEX/Poseidon observations (plates). and theory. J. Phys. Oceanogr. 29:2471­ Matsuura, K, and H. Tachikawa. 1994. 2486. Fishes washed up on beaches of Chichi­ Randall, J. E. 1996. Shore fishes of Hawai'i. jima, Ogasawara Islands. Bull. Nat!. Sci. Natural World Press, Vida, Oregon. Mus., Tokyo, Ser. A (Zool.) 20 (3): 131­ Randall, J. E., P. S. Lobel, and E. H. Chave. 147. 1985. Annotated checklist of the fishes of Miskiewicz, A G., C. C. Baldwin, J. M. Leis, Johnston Island. Pac. Sci. 39:24-80. and D. S. Rennis. 2000. Callanthiidae Springer, V. G. 1982. Pacific Plate biogeog­ (yellow-fin basses, splendid perches). Pages raphy, with special reference to shore­ 280-284 in J. M. Leis and B. M. Carson­ fishes. Smithson. Contrib. Zool. 367. Ewert, eds. The larvae of Indo-Pacific Trnski, T, and A G. Miskiewicz. 1998. coastal fishes: An identification guide to Callanthiidae: Yellow-fin basses. Pages marine fish larvae. Fauna Malesiana 189-191 in F. J. Neira, A G. Miskiewicz, Handbook 2. Brill, Leiden. and T Trnski, eds. Larvae of temperate Mooi, R. D. 1993. Phylogeny of the Plesio­ Australian fishes: Laboratory guide for pidae (Pisces: Perciformes) with evidence larval fish identification. University of for the inclusion of the Acanthoclinidae. Western Australia Press, Nedlands, West­ Bull. Mar. Sci. 52 (1): 282-326. ern Australia.