Tropical Transpacific Shore Fishes' D. Ross Robertson,^ Jack S. Grove,^ and John E. McCosker^ Abstract: Tropical transpacific fishes occur on both sides of the world's largest deep-water barrier to the migration of marine shore organisms, the 4,000- to 7,000-km-wide Eastern Pacific Barrier (EPB). They include 64 epipelagic oce- anic species and 126 species of shore fishes known from both the tropical eastern Pacific (TEP) and the central and West Pacific. The broad distributions of 19 of 39 circumglobal transpacific species of shore fishes oifer no clues to the origin of their TEP populations; TEP populations of another 19 with disjunct Pacific distributions may represent isthmian relicts that originated from New World populations separated by the closure of the Central American isthmus. Eighty species of transpacific shore fishes likely migrated eastward to the TEP, and 22 species of shore fishes (12 of them isthmian relicts) and one oceanic species likely migrated westward from the TEP. Transpacific species constitute ~12% of the TEP's tropical shore fishes and 15-20% of shore fishes at islands on the western edge of the EPB. Eastward migrants constitute ~7% of the TEP's shore-fish fauna, and a similar proportion of TEP endemics may be derived from recent eastward immigration. Representation of transpacific species in different elements of the TEP fauna relates strongly to adult pelagic dispersal ability?they constitute almost all the epipelagic oceanic species, ~25% of the inshore pelagic species, but only 10% of the demersal shore fishes. Taxa that have multiple pelagic life-history stages are best represented among the transpacific species. Among demersal teleosts that have pelagic larvae, pelagic spawners are better represented than demersal spawners among transpacific species, perhaps because offshore larval development and longer pelagic larval durations provide the former with greater dispersal capabilities. There are strong phylogenetic effects on representation in the transpacific fauna: (1) elas- mobranchs are proportionally better represented than teleosts, even teleosts with more pelagic life-history stages; (2) a pelagic juvenile stage with great dis- persal potential allows tetraodontiforms that produce demersal or pelagic eggs to be well represented; and (3) various speciose central Pacific families with "adequate" larval dispersal characteristics lack transpacific species. El Ni?os potentially enhance eastward migration by increasing eastward flow and halving transit times across the EPB. However, that effect may be offset by low pro- ductivity and high temperatures in those eastbound flows. There is little clear evidence of strongly increased migration across the EPB during El Ni?os, in- cluding recent extreme events (1982-1983 and 1997-1998). During such events shore fishes in the TEP experience range expansions and become locally abun- ? This research was supported by a National Geo- ^ Smithsonian Tropical Research Institute (Panama), graphic Society Grant (5831-96) to D.R.R. and general Unit 0948, APO AA 34002-0948 (phone: (+507) 212 research funds from STRI. Manuscript accepted 20 Oc- 8704; fax: (+507) 212 8714; e-mail: drr@stti.org). tober2003. 'Section of Ichthyology, Los Angeles County Mu- seum of Natural History, 900 Exposition Boulevard, Los Angeles, California 90007 (e-mail: jsgIMAGES@aol Pacific Science (2004), vol. 58, no. 4:507-565 .com). ? 2004 by University of Hawai'i Press '''California Academy of Sciences, San Francisco, All rights reserved California 94118 (e-mail: jmccosker@calacademy.org). 507 508 PACIFIC SCIENCE ? October 2004 dant at marginal areas such as the Gal?pagos, changes that can be confused with increased migration across the EPB. Although there is a strong bias toward eastward migration among the transpacific shore fishes, there likely is much more westward migration than previously realized: 20-25% of transpacific spe- cies may have migrated in that direction. Stronger eastboimd than westbound currents can accoimt for this bias. Westward migrants have better developed pelagic dispersal characteristics than many eastward migrants, suggesting that westward migration is more difficult. Many westward migrants associate with flotsam and flotsam-mediated migration is more likely to be westward. All westward migrants occur at Hawai'i, but only about one-fifth of them at the Marquesas. This bias may be due to: Hawai'i being a larger target and in the path of most of the flotsam dispersal from the TEP; an eastward current that impinges on the Marquesas, reducing westward arrivals; and most propagules dispersing toward the tropical Marquesas originating in the temperate eastern Pacific. However, the Hawaiian Islands also are much better sampled than the Marquesas. Although the TEP reef-fish fauna may be depauperate relative to that of the Indo-Malayan "center of diversity," it is as rich as the faunas of is- lands on the western side of the EPB. Hence a preponderance of eastward mi- gration does not represent a response to a richness gradient across that barrier. There is little evidence that a paucity of ecological groups in the native TEP fauna is primarily responsible for the structure of the eastward-migrant fauna. Rather, eastward migrants may simply represent a cross section of those in the donor faima, tempered by phylogenetic variation in dispersal ability. Because few central Pacific fishes can live only on uve corals and coral reefs, the rarity of such reefs in the TEP is unlikely to strongly limit eastward migration. Diifer- ences between oceanic and adjacent continental reef-fish faunas in the West Pacific indicate that each is strongly tied to its respective habitat. Hence, the rarity in the TEP of the (overwhelmingly) most abundant habitat present in the central Pacific?tropical oceanic reefs?may strongly limit migration in both directions across the EPB: there is little suitable habitat for eastward migrants in the TEP and few suitable species and tiny source populations for westward mi- grants. The global effects that oceanic/continental habitat differences have on reef-fish biogeography need further assessment. Genetic data on ~18% of the transpacific species indicate: that conspecific populations of oceanic species (especially) and shore fishes are genetically well connected across the EPB; that circumtropical taxa in the TEP include isolated isthmian relicts and recent eastward migrants; that all five TEP species of one circumtropical genus {Tha- lassomd) were derived by several eastward invasions after the closure of the Isthmus of Panama; that some isolated Hawaiian central Pacific populations were established by postisthmian invasion from the TEP; and that Indo-central Pacific species unsuspectedly can co-occur with their endemic sibling sisters in the TEP. Genetic data support distributional data that indicate a strong preponderance of eastward migration across the EPB but also more west- ward migration than previously thought. Euture genetic studies should resolve a question that distributional data cannot: how many widespread presimied eastward-migrant transpacific species actually originated by westward migration from the TEP? THE TROPICAL EASTERN PACIFIC (TEP) cal and subtropical Pacific coast and oflihore biogeographic region encompasses the tropi- islands of central and northern South Amer- Tropical Transpacific Shore Fishes ? Robertson et al. 509 ao'-N t^ ? VVP^V Main Hawaiian . ^ M ? ^^^ Islands '~~~-., '*., ^^ ^^L ^^^ -^^^1^^ 20?N ?': Revillagigedos ? " ^ ^^ ^M" ^B Johnston Island Eastern ^^^fc " 10"N Central Line '?.. islands?*?'?., Pacific Barrier ;?Clipperton TEP ^g-kJ Cocos ^H Malpelo? ^^^ Pacific ??, , Galapagos ^^^^M 1000 Km 10?N 160?W Marquesas * \ iio'-w 85?W ^1 FIGURE 1. The tropical eastern Pacific (TEP) biogeographic region and its isolation fi'om other regions by the Eastern Pacific Barrier and the Central American isthmus. ica. The continental portion of tiiis region extends from the tip of Baja Cahfornia and the lower two-thirds of the Gulf of California (south of ~29? N) to the northern coast of Peru, at about 5? S (Briggs 1974, Hastings 2000, Robertson and Allen 2002). The lat- itudinal limits of the region are set by western extensions of the continental coastiine, where cool currents that flow toward the equator turn oifshore toward the central Pacific. In addition to the continental shore there are five oceanic islands and archipelagos scattered through most of the latitudinal range of this region (see Figure 1): the Revillagigedo Is- lands (at -18? N, -400 km S of Baja Cah- fornia), CHpperton Atoll (at 10? N, ~ 1,100 km SW of central Mexico), Cocos Island (at 5.5? N, 450 km W of Costa Rica), Malpelo Island (at 3.9? N, 350 km W of Colombia), and the Gal?pagos Archipelago (between 1.7? N and 1.5? S, -1,000 km W of Ecuador). The TEP is the most isolated tropical marine biogeographic region. It has been separate from the tropical Atiantic since the closure of the isthmus of Panama, about 3.1 million yr ago (Ma) (Coates and Obando 1996). It also is isolated from the central and western Pacific by the world's widest deep- water marine barrier to the dispersal of ma- rine shore organisms, the 4,000- to 7,000-km Eastern Pacific Barrier (EPB) (Ekman 1953; see Figure 1), which may have existed for the past 65 myr (Grigg and Hey 1992). Although initially the EPB was thought to be com- pletely impenetrable to dispersal by con- temporary marine taxa (Darwin 1872, Mayr 1954), many members of various major taxa now are known to occur on both sides of this barrier. These include corals (Glynn and Ault 2000), sea urchins (Ekman 1953, Lessios et al. 1998, 1999, 2003), moUusks (Vermeij 1987, Emerson 1994), polychaetes (Fauchald 1977), and fishes (Briggs 1961, 1964, Rosenblatt et al. 1972, Leis 1984). There is direct evidence of the transport of invertebrate larvae across the EPB (Scheltema 1988), and accumulating genetic data demonstrate ongoing connec- tions between conspecific populations sepa- rated by the EPB in some invertebrates and fishes (Rosenblatt and Waples 1986, Lessios et al. 1998). Briggs (1961) compiled the first list of transpacific fishes, species that occur on both sides of the EPB. That list has expanded considerably (Briggs 1964, Rosenblatt 1967, Rosenblatt et al. 1972, Springer 1982), and in the most recent review Leis (1984) listed 54 transpacific shore fishes (five of them errone- ous). Most attention to movements across the EPB has been given to eastward migration, due largely to the assumption that species 510 PACIFIC SCIENCE ? October 2004 tend to migrate from areas of higher diversity (the Indo-Malayan area in the western Pa- cific) to areas of lower diversity in the cen- tral and eastern Pacific (e.g., Briggs 1995, I999a,b, Mora et al. 2003). This paper has several purposes: (1) to summarize current information on the oc- currence, distributions, and population status of tropical transpacific shore fishes on each side of the EPB, and the contribution that they make to the faunas on each side of that barrier; (2) to review genetic studies dealing with patterns of connections of populations across the EPB and the role of transpacific migration in the origin of TEP populations; and (3) to examine factors potentially aifect- ing the current level, direction, and routes of migration across the EPB and the establish- ment of resident populations of immigrants in the TEP. MATERIALS AND METHODS Tropical Transpacific Shore Fishes Transpacific shore fishes occur in the central and/or West Pacific as well as the TEP. The great majority of the species discussed here have all or most of their ranges within the Tropics; however, a few mainly warm tem- perate species with ranges extending into at least one tropical area are included. Trans- pacific warm temperate species that appar- ently enter the Tropics only as vagrants (e.g., the sharks Hexanchus griseus, Carcharodon car- char?as, and Galeorhinus galeus and the tuna Thunnus orientalis) are excluded. Here, shore fishes include demersal and coastal pelagic species living at <100 m depth. Oceanic epi- pelagic species also are included mainly to assess the effectiveness of the EPB as a barrier to shore fishes and how dispersal potential affects representation of different types of fishes in the transpacific fauna. RESIDENTS AND VAGRANTS. A vagrant on one side of the EPB likely does not have a self-replenishing population there because only a few isolated individuals have been seen at scattered locations and times or a few same-sized individuals once at a single site. A regional resident likely has a self-sustaining population because it is relatively common at at least one location, has been recorded repeatedly at the same site(s) over intervals of decades, and at least one local popula- tion includes juveniles and adults of various sizes. EASTWARD AND WESTWARD MIGRANTS, AND ISTHMIAN RELICTS. Eastward migrants to the TEP are widely distributed in the cen- tral Pacific, including the Line Islands at the upcurrent end of eastbound currents on the western edge of the EPB, and are either residents or vagrants in the TEP. Residents restricted to areas at the termini of the east- bound currents or to the offshore islands likely are such immigrants. Except for Hetero- priacanthus cruentatus (see section on Genetic Studies in Discussion), eastward migrants are exclusively Indo-Pacific. Most also lack a "look-alike" likely sister species in the West Atlantic because such would indicate that the TEP populations might be isthmian relicts (cf Vermeij and Rosenberg 1993). Westward migrants to the central Pacific are residents in the TEP, where they occur near the upcur- rent ends of westbound currents and at the offshore islands. Elsewhere in the Pacific they are restricted to islands at or near the down- current ends of those currents (e.g., Hawai'i, Johnston Island, and the Marquesas). Isth- mian relicts: Before the rise of the Isthmus of Panama (~3.1 Ma) (Coates and Obando 1996), the biotas of the TEP and the Greater Caribbean are presumed to have had exten- sive connections. This common history is re- flected in the similarities of their fish faunas (Rosenblatt 1967, Thresher 1991, Robertson 1998). In many marine organisms, popula- tions on the Atlantic and Pacific sides of the isthmus represent geminate species pairs (Jordan 1908) (i.e., sister species formed subsequent to the separation of TEP and Caribbean populations of a single species). Isthmian relicts in the TEP arose through the division of a neotropical population by the rise of the isthmus. Sources of Data As well as older material on transpacific fishes (Briggs 1960, 1961, 1964, Rosenblatt et al. Tropical Transpacific Shore Fishes ? Robertson et al. 511 1972, Leis 1984), we draw on information summarized in Robertson and Allen (2002) from botii primary sources and regional works on the shore-fish fauna of the TEP (Thomp- son et al. 1979, Rubio et al. 1992, Allen and Robertson 1994, Fischer et al. 1995, Bearez 1996, Grove and Lavenberg 1997, Garrison 2000); a checklist of fishes from the Hawaiian Islands (Mundy in press); unpublished data on Line Islands fishes (R. Pyle, pers. comm., 1996; B. C. Mundy, pers. comm., 2003); a checklist of Marquesan fishes (Randall and Earle 2000, and see Randall et al. 2003); and more general references for sharks (Com- pagno \9Ma,b), clupeids (Whitehead 1985), lutjanids (Allen 1985), scombrids (Collette and Nauen 1983), and istiophorids (Naka- mura 1985). In addition, D.R.R. has recorded shore fishes at the Hawaiian Islands (at vari- ous times between 1987 and 2003), Kiritimati (Christmas Island) in the Line Islands (1996), die Marquesas (1996, 1999, and 2003), and in die TEP: Chpperton (1994, 1998), Cocos (1992 and 1997), die Gal?pagos (1990, 1992), Malpelo (1992), tiie ReviUagigedos (1991, 1994), Mexico (1990, 1998, 2002), most of the coast of Panama (many times between 1977 and 2003), and El Salvador (2000). J.S.G. has extensive diving experience in the Gal?pagos since 1975. J.E.M. has been in- volved in numerous ichthyological surveys (using scuba and submersibles to 1,000 m depth) since 1967 oif Baja California, Costa Rica, Panama, Cocos Island, the Gal?pagos, and various island groups spread throughout Oceania and the West Pacific. A few records of transpacific fishes already known to occur in the TEP but not previously recorded in the Gal?pagos are included from work by J.E.M. (unpubl. data). Sampling effort has been unevenly spread in areas of interest on both sides of the EPB. There now exists reasonably comprehensive information on the fishes of the continental shore of all countries and offshore islands in the TEP except for the ReviUagigedos Islands and the continental shoreline between Co- lombia and northern Peru. In the central Pacific the fauna of the Hawaiian Islands is well characterized thanks to a long history of intensive research (Mundy in press). Less information is available on the fishes of the Line Islands, and what is known about the Marquesan fauna (Randall and Earle 2000) derives from a few short visits by ichthyolo- gists to a few of those islands. Hence the lack of records at those two island groups of sev- eral widespread species of transpacific shore fishes found in adjacent parts of the central Pacific could well reflect inadequate sam- pling. G tests of independence were used to compare the relative abundances of different types of species in the faunas of the TEP and the central Pacific. RESULTS Distributions of Transpacific Shore Fishes in the TEP and Elsewhere Currentiy 126 species of transpacific shore fishes are known from the TEP, including 114 tropical species, 10 species with tropical to warm temperate distributions, and two species with largely warm temperate distri- butions. In addition there are 64 species of transpacific epipelagic oceanic fishes that enter nearshore waters. The distributions of transpacific species on both sides of the EPB and their general ecology are summarized in Table 1, with supplementary information on some species in Appendix 1. The 126 shore-fish species include 22 elasmobranchs and 104 bony fishes. Three of the elasmobranchs and 28 of the bony fishes have resident populations on only one side of the EPB: three elasmobranchs and five bony fishes occur as vagrants on the western side of that barrier and 23 teleosts as vagrants on the eastern side (Table 1). Ten species of trans- pacific shore fishes are found in the West Pacific but not the central Pacific, and the remainder occur in both areas (Table 1). The 116 species found in both the central and eastern Pacific belong to 55.8% of 77 families of shore fishes present in both the TEP and central Pacific, 44.3% of 97 shore- fish families found in the central Pacific, and 38.7% of 111 famihes present in the TEP (Table 2). U Q ?- H bo -c &? u +-S ?: o "" " .? 2 bo ?- C V ?< U ;? .y ^ S > 5 .y s ^ ^ .y * s .&?? ^ ^ g -g y .fr S ?? S^SS ?^Q S ?y O c>5 ?iq w O fiq :o Q u u "B Q Q u^i i> .y + o Q s S s S u u u 0 M o ^ c ,M lic t, se m i J J? c b? ?m io ce an i th m ia n re m ig ra nt , M C nia n re gr an t, C g 1 ?^ 2 0O? ?i? w u u u o o o pH w &? t? o? t? t? t? t? c c c c C c c c c c c OJ PH > W ^ O? O? ^ I l" + PH "? i, ? "-a I I I" r + r c g + .S,.2. I ^ ^ jh o o -o u s 'S OH W I .i S2 0 1 T3 rt y 3 u o w ? 3O ^ 5 r- ?H-H rt OJ ^ ^I^M a -N^ 3 1 4^ "rt S = ??^f^ 13 ^S^ ^ t? ^^ ^ ^o + S + o?" e4S + .a cS ?I I C o? 0 _o p 3 3 5 3 3 PH w w y i 1 I 0-3 uy 9 e ? sil O O > h o? .0 ? 'S- 'u ? PH e u s 3 o -O o "bn S -O o "bb e 3 OH &, DH b b b ses 3 3 3 .y M c^ -9 8 -2 b To B B u uuu u u .a 13 ? 0^-9 a o o ? b'Sn^ S E U 3 3 ? uu OH 0 'T^ -D ^-^ ^ g 1. 3 1 ?50 ? B r g 3> 3 + 3 u .y M c^ -9 o _o bo B U B ? 3 G ?S u u ^'i?? g CK iJ ? iJ -y ^11 o iJ fi ? ? -at?)-? ?cC ? ? Sa ^ +3 ?? <^ p. ^ ^ -o ?a c U -o I o. o -o o OH O J3 + O U Q o O w W O u Q J J OH OH o o" OJ "bb 'S) s a! M C 11 1 g B ? OJ u ^ q? "^ 5 4-' ?M rt _o 'S-S 'S OH a! u" ?* ?g 1 ?g^ ^ ? -C _S ^ TU -a ??? o ^<; u Q U ? uuu uuuuuuu QQ?QQQQ :SHJ OH O ? "o" P^ W 'bb o tT il'? 2 ?3 ~"B bp O eq w OH UUU QQQ t? t? C ? c c c c c ? C f? OJ OJ OJ -T3 Td 13 C?? ct? ct? ci t? c (?1 i-H c G c c OJ b b > C?? t? Ct? C?? ? ? c c G OJ OJ QJ Df? t? OH ? ? t? o s S^ u ^ o. ou o Soi 3 , I rt o ?5^ , I, bo is? y g, wu ? o-? >u >? 13 O: E _o ; "o ii ?eu t? ^ wO 0 i o OH ?^U ,0,^ .0, U^U o U I ^ I ? s t? ? 2 -I- SS o o I ^ I I ^UU^t?^^ I u y- w c5 s ,1 i" O? J, 6 o ? o I, M T3 yo g. >u 1^ OJ 3 < ?^ o UUt? " + r r 1 a ?0 1" + + 0 sju nc t (? ict ? + + + t-^ + +1 + ^r++ + + + + ? + ^ O? &? .H. 1 St? + ^ ^ j^.aE?, t? >t? t?t?t? + + :s^ 8?T" t? J^t? S;? ^^ ^-^13 "rt T3 T"? ?^-g "rt 'rt "rt l?^i g 1 ^g <^ ? 0 CJ u CJ u 0 CJ u ^ ? O? U 0 CJ CJ K - S u 0 u "a- 1? s. 'S, 'DH s. OH U ^ CJ > ^^'G 'o 'u S 'S 'S 'S 'S 'S 'S 'S -9 ??> u 'u 'o 'cj 0 ^-^ 0 t? 0 0 0 0 0 rt OJ ?^ .C ^ rt C3 rt c^ w w w cd C^ w cfl 0 C P a :if ic ( W P -V C3 C3 rt B ,y fi c b b g c go..>OH Eu Su ? 0 OH OH OH c: ?uuu OH OH OJ OH OH OH OH e^-a 1 OH OH OH e OH s e S E E E ? uuuuuuu U E fuuu 3 CJ u 1 3 0 g g 3 g 3 ?i? ? ? -2 ? ? ? ? 0 ? 0 0 ? 0 ? g ? ? ? 1' 0 i 0 ? ^ OH ? 13 .1- u^ .!H T3 U-T3 CL,-T3 T3 T3 T3 T3 13 T3 13 T3 T3 T3 "O .b 13 1^13- -^13 13 13 J uu au ? uu u^s a 3. ?a^^s ^?^?^5^^ au ^U^ aaa .S~?3 e^' ?S 6'g t^. I ti j ^?^ ?3 C?? Ct? t? > t? Ct? ?3 1 3 1 1 1 c 1 \ u 3 o SS^ ? t? 0 -a ;-i rt ?d O ;-i "O OJ ^ P? 1 1 rt rt 'rt 'rt 1 M I m M > ^^ ^ ^ ^ (2 ^^^ ^ -S i- ? 'S u S w 5 I ^ y 4 I w I C PH PH L f^ < ri ??? m ? I q M O ^ i > _L W i i -3 m o P? ^ ? S.B O o '^, ?? o o fS fq ?? S. ~?S ..Si i = o u 1-1 PH fe W O? M O? ? PL, pi + + R ,R ) R, R) S~p? + o? + + n-. + ? + t?t? t? t? t?t? t? t? 'P OJ 3 o u u 3 ^33 u u O u u o CJ o u u "rt 3 i" 'en U M= "^ ^C .? o 3 ?? ?? "^ "^ 3 3 ? ?^ ? ?^ ? <.g u o L S ?S-'o 'G 'G DH un O OH DH 'G 'G o 'u 'G 'G 'G 'G 'G 'G 'S, 'DH o? ?3 o "^ rt 0 -T; "^ O 0 rt o! rt rt C3 rt C3 rt rt 0 o O b^ P? u 0^ e 5^ c^ 6 e P? P? OH OH PH PH PH OH PH PH c s o & eo u '-f? U E c^ o s E u U u u ? ?OU O?U ? ? '^ 1^ B ? ? f^ ? g O ? 3 3 ? ? ? 3 ? 0 ? o ? 0 ? 3 3 r? (2 b '^ "O PH -? u ^ >5u? "O "O 13 'u -a -a 13 -O T3) -a "O OH u^ ^ U^ ? h5 c 1?1 PH ^ ?S^^ a^^ ? ? O I, s ^S OJ -C rt a, T3 E ?g C OJ "o M W ^ rt -o O n OJ is rt rt < (?H C/5 T3 2 u ? o pH 5 Pu uuuuuu CJO UCJCJU CJ-r; S S Eoj 'u ^ ^ CJ rt o CJ rt o ^ rt CJ rt n ^ C3 rt cfl ? ^o ?^ ? ,o ?^ "3 CJ ? CJ o O u o r'H CJ 'u 'S-'o 'cj 'o-'o -O 'SH'U 'SH ^ 'SH ? 'SH 'SH'SH '-o rt o c? i^ o c? o o ? O O r-. '^ o 0 o PH j3 C^ PH ?3 C^ -a faPn b "bn ?3 1 o g b b U EUU Eu E EU g g E r E E E PH 0 CJ o o CJ o u 3 ? 3 3 3 i ?3 3 3 u T3 .?H T3 13 .!H T3 .tH u -a ^ ^Su^^u^u as u u u '~"u uu ^ fis O? C? CK O? Cj s S' ? r^ o ^ ij CJ ^ i^ i^ Q P P II 1 i 1 ^ ? g ?^ ^ 1 "?? ? ?5 -g c w T3 c -a -C bo C C 0 n U u S3 pH W c o '0 Pu &? t? ^ c PH pi t? S ta U Q PH + t?p? p?p? p?p? pH pH o o o. o ?> OOOO?U Sc^ M m M U U Q W PH PH Q p -? -o PH -g bpf^ + Q M CO C/5 u 1-1 PH uu So de m ra n t de m ? ? t; ? ? c c s S S s S u oj 2 S S S S 3 3 c c c c c C c C OJ OJ c C C c OJ (U 3 3 3 3 3 3 3 3 3 3 3 3 ?s bo-S u rt OJ rt rt rt rt rt u OJ P? >>>>>Ofi Pi t?p? t?p? Pi Pi Pi Pi Pi O? Pi O? Ou I I o TO I ^ w o u w -s. I o U ?a "o s U PJ 1^ Pi ? U o ??CJ CO - o o q? -13 1> i= t? C/2 i ?' "^ I > ^ t? ^ ?L- PiU ^^ Pi I o? P? P? o? P? P? o? P? o? ^ P? P?O?PI rp?o? L > I + P? + ? I I e?'.^ r^rp? P?P? P? P? P? P? I P?O? P? I P? P? P? u u o u u o ^_ ig ^g <^ ? "^ <^ 77 'u 'G 'o 'u 'G 'G (TH rt rt C3 rt rt C3 .^ PH PH OH PH PH OH ^ uuu o ? ? -o -o T3) UUCJP^ o ? ? U -o -o T3) . " ? ?^ ? "rt f J u CJ PH PH PH "hr uu U a A <^ A H -n -n -O c 1?1 ^ c 1?1 ? o ? T3 -o P? ? S ? PH hS ? ? UU G -O rt O U S ? 3 n3 _!-j S? ^ ^ M i ?^ 1 "s ??ffe fe y s .+2 -?H K .'S 5 oooood; uuu uuu cj" u" o" u" o" u" CJUOUOU CJU oooooo oo g c c c c a OJ OJ t? t? t? t? C G G C G G ^ ;_! ?H i-l i-H ?H i-i ^^ fao bo b? b? bo b?'cfi ,rt .TO ,TO ,rt .rt ,TO t? O? t?o? O? > C f? c c c c c c m M 1 u DH\ \ 1 \ G^^^^ o U?UU??U ^ ? i? U 3 OH I ^ O! ^ S u I " _L OH S S PH OH ii B OJ OH I _rt O! U ^ ^ u ? ^ -^ Z::::^ ^ JH ^ ^t? ^ I U ? I o lo C C ai ?' ^ OH OH OH ? ii S S S ?u OJ OJ pH CLH pH ,1,1,1 J -5 rt UCJU eQ ^ ^ ;?- 2 ;?- : ii S" 1 ?? + + + s~ + + + ? 1 -a o? o? + o?" + 0?" + 0?'~+ + + ? r &? t?t?t? t? J^g,> vT ^ ? > "O CJ CJ O CJ CJ 'rt '^ "rt "TO "rt 'TO u 1" 12 u u ?il ? cj "rt 0 u 0 u u -a LC L? (^ LC L? _0 ^C o u u "rt 0 ? > s ? ^g ? ,u ? ?3 ? OJ e e o U u PH OH u u .y" .y" d o! u o o o OH y J Q OH ?^? C O , M ( ar d, m ig an ic t ?H S +^.2 O O ^ O + o u >! o " ? J j J O O OR O ?a o ?O OO O Q^a?i OH OH OH W M M U m OH c)^ m M M M OH W t? t? C c C C c c c OJ (U (U OJ bo bit OH -a 'S ? ?S I o? o o? ^ ^ I I o? O 3 !^ ,3 s??S ?" o ^ ?OH gW-a^^si ^ O ?^^^ O? o y o r?o? ;? MOI,. ^ ^ u ? M -3 u ??-? U g o O? MO I Sou i U_L OH^? o o U m W ^ ^ o, o r 1 a ;3 T3 y S Vi?l ?-^ ->-,. [TJ ??-? ^ '?1 >-^ '?1 ri "rt 'rt ^ ^? ?? w M h OH O? -O o ?50 a, ? o o b b E S g 3 u u o -O o "bn S 3 o? o?" o? o? o? + + o? + o? o?" ? SH 'S + '3^ OH "0 0 rt ^-^ 0 0 -"? ? OH 0 b C OH e uS g S U 3^1.3 3 ? 3 OH U I O? .a + '^ "rt "rt OH O u ?uo ^ w ^- ^ a ^ o? (l? &0? OH U O ." faOH SU 3 ? u -a ?a O? o? OH u o ?~ ?~o? t ? o? o? o? + o?o? t-?- OH OH u u OH OH uo 00 00 "O "O "O Td OH OH uo ? ? -o -o a g ^f ?;S CQ E^ P .1=0 & ?g IH ?-* s 1 ?S .|o s ?:? s ?5 J ?I 1 i ?S^ S p i^P o J3 en o ^ M M -O ?a 0 o -o o t? o U ,1 u s w S)^ + PH PH QUO OO ? ? c C C V QJ QJ QJ t? t? o? o? t? ^ i w 3 O ^ I OH PH 13 ;::a ^ ^ 111 .3 U^ > -3 o? > t- + + t? OJ t?t? U o! '-O rt rt C3 u O u o o "3 "3 a. 'OH TO 'SH'SH'S^ -9 -9 o b g OH Tgsg^^ y 6 1 a e s s e g 3 i ?3 3 3 3 3 ^ "O ij ? ??u ij? i?j, J -S ?- S |5. S ^1 ? ?>^ e QQ ?B g -o o 'n, c cu V 1) -TJi .s * irs ? D-CJ I s ^ o S E CO ^ O O. U .3 US Is gy 5pn;5 n u s s ? ?5 O?; w tr ? g 1-^ j' 3 .S ^. "^ D-CJ ? ? - ~ " c ? ?a ?? -I " ?a s ^ o h I s s ^ C C ,^ g ? J3 OJ O y i/; SH & ?g s;-^ i? ? bu SrS.S q- ?j ?O?O-I u Cl- .2 o u = Sl?g 3 5 S -^ E ? o ? ? 1^ 3 o -o a Si o Td s ; 1^ O- oj ^ ; g-a.a ? ' s ? o CQ -^ bu P .s i S" c 2 % 1^5 b ; t? S ^ .U IJ^oi ?i |l 1 -o -C bu ^ gj gj TABLE 2 Numbers of Indigenous, Circumtropical and Transpacific Shallow-Water Shore Fishes per Family in die TEP, die Greater Caribbean*, and the Central Pacific Family TEP: Indigenous Caribbean (Indo-Pacific/ Native circumtropical) (circumtropical) Central Pacific: Total (Hawaiian, Line, Marquesas) Characteristics Affecting Tropical Dispersal Potential" Hexanchidae - (-/I) ? 1(H1) Live birth, no pelagic phase, antitropical Heterodontidae 3 ? ? Demersal egg, no pelagic phase Ginglymostomidae 1 1 1 Live birth, no pelagic phase Odontaspidae - (-/I) 1 1(H1) Live birth, no pelagic phase, warm-temperate Scyliorhinidae 1 ? ? Live birth, no pelagic phase Triakidae 10 ? ? Live birth, semipelagic, mainly antitropical Carcharhinidae 4 (2/7) 7(8) 12 (H8, L5, M6) Live birth, demersal to semipelagic Sphyrnidae 3 (-/3) 3(3) 3 (H3, L1,M1) Live birth, semipelagic to pelagic Echinorhinidae - (1/-) 1(H1) Live birth, no pelagic phase, anti tropical Pristidae 1-2 2 ? Live birth, no pelagic phase, esmarine Rhinobatidae 7 2 ? Live birth, no pelagic phase Torpedinidae 1 1 ? Live birth, some semipelagic Narcinidae 5 1 ? Live birth, no pelagic phase Rajidae 5 8 ? Live birth, no pelagic phase Urolophidae 10 2 ? Live birth, no pelagic phase Dasyatidae 3 (1/-) 7 5 (H2, LI, M2) Live birth, no pelagic phase, soft bottom; most demersal Gymnuridae 2 2 ? Live birth, no pelagic phase, soft bottom Myliobatidae 2 (-/4) 4(2) 4 (H3, L2, M2) Live birth, benthopelagic or pelagic Elopidae 1 2 3 (H2, LI) Benthopelagic Megalopidae ? 1 1(M1) Benthopelagic Albulidae 3 3 2 (H2, LI, M2) Soft bottom AnguiUidae -(1) 1 2 (LI, Ml) Estuarine, freshwater, marine larvae Heterenchelyidae 1 1 ? Soft bottom Moringuidae 1 2 3 (HI, L1,M2) Soft bottom Chlopsidae 3 5 5(H3,L1,M2) Myrocongridae 1 ? ? Muraenidae 15 (10/-) 28 >50 (H49, L26, M34) Most reef, some soft bottom Ophichthidae 37 52 40(H21, L7,M14) Most soft bottom Congridae 12 18 8 (H6, L2, M2) Soft bottom Muraenesocidae 1 1 ? Soft bottom Engraulidae 20 21 4(H2) Pelagic Clupeidae 18 (-/I) 18(1) 4(H2,L1,M1) Pelagic; TEP: demersal eggs; CP: pelagic eggs Chanidae - (1/-) ? 1 (HI, LI, Ml) Semipelagic Ariidae 20 15 Brooded egg, no pelagic phase, soft bottom, many esmarine Argentinidae 2 3 ? Benthopelagic Synodontidae 5 8(1) 17**(H15, L2,M4) Soft bottom TABLE 2 (continued) TEP: Indigenous Caribbean Central Pacific: Characteristics (Indo-Pacific/ Native Total (Hawaiian, Aflfecting Tropical Family circumtropical) (circumtropical) Line, Marquesas) Dispersal Potential" Carapidae 3 (2/-) 2 8** (H8, L4, Ml) Transpacifics demersal Ophidiidae 18 (1/-) 27 2(H2, L1,M1) Bythitidae 14 28 10 (H2, LI) Live birth, pelagic larvae Moridae 2 1 3(H3) Soft bottom Batrachoididae 14 12 ? Demersal egg, no pelagic phase, soft bottom Lophiidae 2 7 ? Soft bottom Antennariidae 3 (2/-) 7 10 (HID, LI, Ml) Ogcocephalidae 2 9 2 (HI, L2) Reef and soft bottom Gobiesocidae 31 20 1(L1) Demersal egg Adierinidae 13 23 6 (HI) Pelagic, demersal egg Belonidae 6 (1/1) 5(1) 5 (H4, L3, M3) Pelagic, demersal egg Hemirhamphidae 5 5 9 (H2, L4, M2) Pelagic Anomalopidae 1 1 2 Holocentridae 4 (2/-) 11 29 (H19, L20, M15) Pegasidae ? ? 2** (HI, Ml) Soft bottom Syngnadiidae 5 (1/-) 24 25 (HID, L4, M4) Live birth, some with pelagic young Aulostomidae - (1/-) 1 1 (HI, LI, Ml) Fistulariidae 1 (1/-) 2 2 (H2, L2, Ml) Reef and soft bottom Dactylopteridae ? 1 1 (HI, Ml) Soft bottom Scorpaenidae 11(1/-) 20 35 (H22, L12, M9) Reef and soft bottom Triglidae 10 18 1(L1) Soft bottom Platycephalidae ? ? 5 (L2, M2) Soft bottom Centropomidae 6 5 ? Soft bottom, many estuarine Serranidae 51 84+ 83** (H17, L40, M16) Some genera soft bottom Pseudochromidae ? ? 5 (LI) Demersal egg Opistognadiidae 12 17 ? Demersal egg, soft bottom Priacanthidae 2 (1/1) 3(1) 5 (H4, L2, M2) Apogonidae 6 21(1) 52** (HIO, L18, M14) Brooded egg Epigonidae 1 ? 2(H2) Brooded egg? Malacandiidae 3 (1/-) 10 5(H1, L2,M1) Soft bottom Nematistiidae 1 ? ? Semipelagic Carangidae 19 (5/9) 25(6) 29 (H24, L12, M14) Semipelagic to pelagic Lutjanidae 10 (2/-) 14 15(H11, L7,M9) Transpacifics benthopelagic Caesionidae ? ? 10 (L4, M2) Benthopelagic Ledirinidae ? ? 7 (HI, L7, M5) Soft bottom Lobotidae 1 -(1) 1(H1) Semipelagic Gerreidae 11 12 3 Soft bottom, many estuarine Haemulidae 35 23 5 Many soft bottom Sparidae 3 19 ? Soft bottom Polynemidae 2 3 3 (H1,M2) Soft bottom Sciaenidae 76 58 ? Many soft bottom, many estuarine MuUidae 2 (1/-) 4 20 (HID, LIO, Mil) Soft bottom but reef associated Pempheridae ? 2 3 (LI, Ml) Benthopelagic Chaetodontidae 3 (6/-) 7 41**(H22, L23,M14) Pomacandiidae 4 7 25 (H9, L9, M3) Kyphosidae 7 (1/-) 2 4 (H4, L2, M4) Benthopelagic to pelagic Scorpididae ? ? 1(H1) Antiequatorial Kuhliidae - (1/-) ? 5(H2,L1,M1) Pelagic Cirrhitidae 1 (2/-) 1 17(H6, L8,M6) Pomacentridae 19 17 72 (H17, L29, M22) Demersal egg TABLE 2 (continued) TEP: Indigenous Caribbean Central Pacific: Characteristics (Indo-Pacific/ Native Total (Hawaiian, Affecting Tropical Family circumtropical) (circumtropical) Line, Marquesas) Dispersal Potential" Mugilidae 5 (-/I) 7(1) 10 (H2, L5, MS) Benthopelagic Labridae 26 (4/-) 31 103" (H44, LSD, M30) Reef and soft bottom Scaridae 3 (3/-) 14 31 (H7, L19,M6) Pinguipedidae ? ? 7 (HI, L4, Ml) Soft bottom Creediidae ? ? 4(H2) Soft bottom Ammodytidae 1 ? 3** (HI) Soft bottom Uranoscopidae 2 1 ? Soft bottom Trypterygiidae IS 9 10(H1, L3,M2) Demersal egg Dactyloscopidae 24 15 ? Soft bottom Labrisomidae 36 45 ? Demersal egg or live birth + pelagic young Chaenopsidae 34 52 ? Demersal egg Blenniidae 12 15 62(H13, L21,M16) Demersal egg Callionymidae 1 4 13** (H8, M3) Soft bottom Gobiidae 89 108 156**(H26, L27,M23) Demersal egg, many soft bottom Microdesmidae 14 8 5** (H3, L4, MS) Demersal egg Schindleridae ? ? 2(H2) Spawning mode unknown Ephippididae 2 1 2 Reef and soft bottom Siganidae ? ? 5 (LI, Ml) Demersal egg Zanclidae - (1/-) ? 1 (HI, LI, Ml) Acandiuridae 2 (10/-) 3 41**(H2S, L30,M21) Mainly reef, some soft bottom Sphyraenidae 4 (1/1) 3(1) 5 (H3, L3, M4) Pelagic Trichuridae - (-/I) -(1) ? Benthopelagic Scombridae 4 (2/-) 6 1(H1) Pelagic Stromateidae 3 3 ? Pelagic Bothidae 4 (1/-) 14 8** (H8, L4, M3) Soft bottom Sammaridae ? ? 3 (HI, Ml) Soft bottom Paralichdiyidae 22 31 ? Soft bottom Pleuronectidae ? ? 2(H1,L2,M1) Soft bottom Achiridae 7 9 ? Soft bottom Soleidae 1 ? 7**(H3, L1,M1) Soft bottom Cynoglossidae IS 12 ? Soft bottom, deep-living in central Pacific Balisddae 2 (4/2) 4(2) 19(H11, L13,M11) Demersal egg Monacandiidae - (3/-) 8(2) 17(H8, L6,M4) Demersal egg Ostraciidae - (1/-) 5 6 (H4, L2, M4) Tetraodontidae 9 (5/-) 10 21**(H13,L8,M6) Demersal egg, reef and soft bottom Diodontidae - (1/3) 4(2) S(H3, L1,M1) * Caribbean plus Gulf of Mexico, Bahamas, Florida, and Bermuda. Data on numbers of central Pacific species are from Springer (1982), except for families indicated by **. ** Numbers are increased due to new information from the Hawaiian, Line, or Marquesas Islands. " Unless otherwise indicated members of the family are tropical, demersal, shallow-water reef fishes that produce pelagic eggs and pelagic larvae. Exceptions to this pattern: (1) no pelagic phase: live-bearing demersal fishes; (2) adults nondemersal (benthopelagic to pelagic); (3) adults use habitats other than shallow reefs: deep-living, demersal on soft bottoms, in estuaries and freshwater; (4) demersal/brooded eggs: species with benthic or parentally brooded eggs and pelagic larvae; (5) families predominantly contain non- tropical species: temperate, antitropical. Tropical Transpacific Shore Fishes ? Robertson et al. 523 Transpacific Species with Endemic Subspecies in the TEP Sardinops sagax: Genetic analyses (Grant and Bowen 1998) indicate that two antitropical populations in the eastern Pacific represent a single endemic subspecies (S. sagax sagax) of a single widespread Indo-Pacific species. Platybelone argalus and Tylosurus crocodilus: P. argalus pterura and T. crocodilus fodiator are TEP endemic subspecies of circumtropical inshore-pelagic species (Collette 1995). Auxis rochei and A. thazard: A. rochei eudorax and A. thazard brachydorax are TEP endemic sub- species of two circumtropical oceanic species (Collette and Aadland 1996). Doryrhamphus excisus: D. e. paultis is an endemic subspecies of an Indo-Pacific species that is restricted to the Revillagigedo Islands. The widespread Indo-Pacific subspecies, D. e. excisus, occurs throughout the remainder of the TEP (Daw- son 1985). Recent Erroneous and Questionable Records of Transpacific Species in the TEP Carcharhinus amblyrhynchos: Photographs of sharks in the Gal?pagos that were thought to be this species are of melanistic individuals of C. galapagensis (see Grove and Lavenberg 1997; J. E. Randall, pers. comm., 1999). No specimens have been collected anywhere in the region. Carcharhinus plumbeus: Records of the possible occurrence of this species in the TEP, at the Revillagigedos and the Gal?pagos (Compagno 1984^"), are not based on col- lected specimens. Pristis spp.: The systematics of Pristis are confused. Compagno and Cook (1995) briefly discussed the status of the TEP population of large-tooth sawfish, P. zephy- reus, and suggested that it is synonymous with either the Atlantic species P. perotetti or P. microdon from the Indo-West Pacific. Be- cause Pristis spp. are continental fishes found in estuaries and freshwater, and do not occur between the western and eastern continental margins of the Pacific, the TEP population most likely is an isthmian relict. Compagno and Cook (1995) also listed the circum- tropical species P. pectinata from the TEP. However, records of that seem to be based entirely on saws of P. pectinata purchased at sites in the TEP (e.g., Acapulco), to which they may have been imported for the tourist trade. Albula "vulpes": Erom genetic evidence, Colborn et al. (2001) concluded that the TEP contains two endemic members of the A. vulpes species complex, whereas A. neoguinaica, the name previously used for the TEP popu- lation, refers to a West and central Pacific species, and A. vulpes is restricted to the Ca- ribbean. Gymnothorax eurostus: J. E. Randall (pers. comm., 1998) identified as G. buroensis the specimen from Cocos Island on which the report of G. eurostus in the TEP by Ran- dall and McCosker (1975) is based. Gymno- thorax panamensis: Lavenberg (1992) described specimens from Easter Island identified by Randall and McCosker (1975) as G. pana- mensis as a new species, G. australicola, which does not occur in the TEP. Gymnothorax rueppellii: Grove and Lavenberg's (1997) re- port of this species from the Gal?pagos is an error that was resolved by McCosker and Rosenblatt (1975?). Myrichthys maculosus: The report of this Indo-West Pacific species in the TEP by Allen and Robertson (1994) refers to M. tigrinus, a TEP endemic (McCosker and Rosenblatt 1993, Eischer et al. 1995). Scorpaenopsis diabolus: Grove and Lavenberg's (1997) report of this Indo-central Pacific species from the Gal?pagos is based on the misidentification of a juvenile of the TEP endemic Scorpaenodes xyris (McCosker 1998). Priacanthus meeki: The record of this Hawai- ian species from the Gal?pagos (Grove and Lavenberg 1997) is based on a single juvenile specimen of uncertain identity that Starnes (1988) said resembles P. meeki. Lobotes suri- namensis: Allen and Robertson (1994) used this name for the TEP population, which also has been described as an endemic, L. pacificus. According to Springer (1982), L. surinamensis is an Indo-West Pacific and Adantic species that has populations as far east as Eiji and is thought to occur as vagrants at Hawai'i and Tahiti. Because the taxonomy of this ge- nus is uncertain (P. Heemstra, pers. comm., 1997), the identity of not only the TEP pop- ulation but also the vagrants at Hawai'i and Tahiti needs clarification. We follow Heem- stra (1995) and treat the TEP population as 524 PACIFIC SCIENCE ? October 2004 TABLE 3 Amounts of Shallow Habitat" at the Oceanic Islands and the Continental Shore of the Tropical Eastern Pacific* Substrate <100m deep Substrate <200 m deep % of Island % of TEP % of Island % of TEP Location km^ Total Total km^ Total Total Revillagigedos 158 3.0 0.08 222 2.1 0.08 Clipperton 7 0.1 <0.01 18 0.2 <0.01 Cocos 111 2.1 0.05 235 2.2 0.08 Malpelo 6 0.1 <0.01 12 0.1 <0.01 G?al?pagos 4,900 94.7 2.3 10,150 95.4 3.6 Continental shore 205,300 NA 97.5 270,700 NA 96.2 Total -210,500 -281,500 '^ Calculated from nine digitized medium-scale (1:^600,000) nautical charts spanning the continental portion of the region and the Galapagos, plus a smaller-scale chart (1:50,000) covering each of the remaining oceanic islands. ^ Lower Baja California and the lower two-thirds of the Gulf of California to Cabo Blanco, northern Peru. L. pacificus. Cirrhitkhthys serratus: Specimens from the TEP reported as this species by Randall (1963) are C. oxycephalus, and C. ser- ratus is a synonym of C. falco from the West Pacific (Randall 1997). Pseudojuloides cerasinus: Allen and Robertson's (1994) report of this species from Cocos Island in the TEP was based on the incorrect identification of an tmidentified labrid that is not in that genus (Allen and Robertson 2002). Pseudojuloides ce- rasinus is not known from Cocos (W. Bus- sing, pers. comm., 1999) or elsewhere in the TEP. DISCUSSION Transpacific Fishes in the Faunas of Each Side of the EPB Currently known transpacific species consti- tute ~12.5% of the 1,088 tropical shore fishes known from the TEP (Robertson and Allen 2002) and higher percentages of the shore- fish faunas of islands on the western side of the EPB: 19.7% of the Hawaiian fauna, 14.7% of that of the Line Islands, and 20% of that of the Marquesas Islands (Table 2). According to Leis (1984), -7% of the TEP shore-fish fauna consists of endemics that are derived from Indo-central Pacific species that migrated eastward after the rise of the isth- mus of Panama; if correct, that figure in- dicates that ~20% of the TEP shore fishes have recent Indo-central Pacific origins. All shore-fish habitat in the central Pacific consists of oceanic islands, which are spread over an enormous area. In contrast there are few such islands in the TEP. The over- whelming majority of shore-fish habitat in the TEP consists of continental shoreline and a few clusters of continental islands, mainly at Panama and the Gulf of California (Table 3). It was first thought that eastward-migrant transpacific fishes are largely restricted to the oceanic islands in the TEP (Briggs 1961, 1964). However, Rosenblatt et al. (1972) showed that the inshore islands of western Panama support many such species and that nearly half of them are widely distributed in the TEP. The distributions of the 80 east- ward migrants discussed here follow the pat- tern described by Rosenblatt et al. (1972): 30% are widely distributed on the mainland and the oceanic islands. However, almost half the species are limited exclusively (32.5%) or largely (16.3%) to those islands and small parts of the mainland (Table 4). Further, only a fifth of those species are primarily conti- nental, and only one (Decapterus macrosoma) is known solely from the continental shore. Because oceanic islands provide such a tiny proportion (~2.5%, Table 3) of the shore- fish habitat in the TEP, they are dispropor- tionately much more important, in terms of the abtmdance of transpacific species per tmit area, as habitat for transpacific immigrants than is the mainland. All five oceanic islands Tropical Transpacific Shore Fishes ? Robertson et al. 525 TABLE 4 Occurrence of Putative Eastward-Migrant Tropical Transpacific Shore Eishes" at the Offshore Oceanic Islands* and Mainland of the TEP % of Species Found at Transpacifics Islands Only Mainly at Islands'" Islands + Mainland'' Mostly Mainland' Mainland Only Residents in TEP (57) Vagrants in TEP (23) Total (80) 10.5 87.0 32.5 22.8 0 16.3 36.0 28.1 13.0 0 30.0 20.0 1.8 0 1.3 '* Tropical species listed in Table 1. ^ The Revillagigedos, Clipperton, Cocos, Malpelo, and the Gal?pagos. ' Present at most of the islands and only a limited part of the mainland; with notably dense populations at one or more islands. '' Present at most of the islands and much of the mainland or a few islands and a limited part of the mainland; not notably more abundant at islands. ^ Present at most of the mainland and few islands. in the TEP host many eastward-migrant transpacific shore fishes: 60 (including 12 vagrants) at the Gal?pagos, 51 (5 vagrants) at Cocos, 46 (5 vagrants) at Clipperton, 39 (1 vagrant) at the Revillagigedos, and 35 at Malpelo. Those transpacific species are much more important components of the oceanic- island fauna than they are in the continental fatma (Table 5). Although the Gal?pagos contain 95% of the oceanic-island habitat for shore fishes in the TEP (Table 3) and may be in the best position to receive migrants on the main eastbound currents, they support fewer eastward migrants than the other islands to- gether (60 versus 78). Eurther, the Gal?pagos host a resident population of only one trans- pacific species {Enchelychore Ikhenosa) found TABLE 5 Relative Importance of Transpacific Species in the Shore-Fish Faunas of Oceanic Islands and the Continental Shore in the TEP No. % % Eastward Island species Transpacifics Migrants" Revillagigedos Clipperton Cocos 176 102 235 38.6 60.8 32.3 22.7 45.1 21.7 Malpelo Gal?pagos Combined islands 213 349 510 29.1 26.6 23.9 16.4 16.9 14.9 Mainland 971 10.3 5.4 * From Table 6. nowhere else in the TEP, but then so do Clipperton {Uropterygius supraforatus), Cocos (Aphareus fund), and the Revillagigedos {Scomber australaskus). Thus tiny oceanic is- lands that provide <0.1% of the shore-fish habitat in the TEP (Table 3) are of major regional importance for eastward-migrant shore fishes. Transpacifics in the TEP Shore Biota: Fishes versus Other Groups Eor shore fishes, eastward migrants constitute ~7% of the TEP fauna. Similarly low levels occur in most groups of invertebrates: the TEP moUuscan and crustacean faunas each contain <5% transpacific species (Garth 1974, Shasky 1985, Paulay 1997). Ekman (1953) noted the paucity of transpacific echi- noderms (~2% of 240 Indo-central Pacific species [see also Lessios et al. 1996]). In strong contrast, 95% of the TEP coral fauna is the result of recent eastward migration: ~75% of 41 zooxanthellate corals in the TEP likely are such transpacific species (Glynn and Ault 2000) and <5% of the TEP coral fauna appears to be derived from isthmian relicts. The rarity of coral reefs in the TEP (only ~25 km^), the small regional coral fauna, and the importance of transpacific species in that fauna reflect postisthmian extinction of the native coral fauna and destruction of reefs by a variety of environmental stresses (Glynn and Ault 2000). Such stresses, to which reef 526 PACIFIC SCIENCE ? October 2004 fishes may be relatively resistant (see Mora and Ospina 2001, Nilsson and Ostiund- Nilsson 2004), evidendy had much less e?Fect on the TEP's reef fishes: the latter include over 300 regional endemics, and eastward migrants represent only ~15% of the region's resident reef fishes. If Leis' (1984) estimate that ~7% of the endemic TEP shore fishes are derived from past eastward transpacific migration is correct, then ~85% of the shore- fish fauna of the region (tropical and temper- ate species combined) is derived from taxa present before the rise of the isthmus. Among the 31 transpacific corals, 38.7% are known only or mainly from the mainland, 38.7% are well represented on both the oce- anic islands and the mainland, and 22.6% are entirely or largely restricted to the oceanic islands, which provide ~4% of the region's habitat suitable for coral reef development (Glynn and Ault 2000). Most of the trans- pacific moUusks also are most common at the oceanic islands of the TEP (Paulay 1997). Thus, the few small oceanic islands in the TEP are disproportionately important as sup- porters of transpacific species of a range of major taxa. Eastward versus Westward Migration by Shore Fishes across the EPB Early studies of transpacific shore fishes em- phasized eastward migration (Briggs 1961, Hubbs and Rosenblatt 1961, Rosenblatt and Walker 1963, Rosenblatt et al. 1972), and Briggs (1961, 1974, 2003) asserted that there was no evidence of any westward dispersal. Rosenblatt et al. (1972) concurred in gen- eral with this conclusion, but they noted one probable westward migration, by Sectator ocyurus to the Hawaiian and Society Islands. Later, Jokiel (1984) found evidence of west- ward dispersal of TEP corals by rafting. Rosenblatt and Waples (1986) maintained that connections between the TEP and the Hawaiian populations of 10 transpacific fish species indicated by their genetic data are eastward and indirect, with another central Pacific site acting as a stepping-stone. Briggs (1961, \999a,b, and see Mora et al. 2003) pro- posed that eastward dispersal is more likely because it reflects dispersal from the Indo- Malayan center of diversity and speciation to more depauperate areas such as the TEP. Rosenblatt et al. (1972) thought that dis- persal across the EPB predominandy should be eastward because of adverse effects of low temperatures in the higher-latitude west- bound currents on tropical organisms. How- ever, Scheltema (1988) found larvae of tropical shore invertebrates in westbound currents, which he noted have surface tem- peratures experienced by organisms in the TEP. Vermeij (1991) stated that all trans- EPB invasions are eastward, even though westward currents can support immigration. Here we address the issue of directions of migration across the EPB by shore fishes. EASTWARD MIGRANTS. There are 80 shore-fish species whose current distributions and, in several cases, genetic evidence (see section on Genetic Studies later in Discus- sion) indicate that they likely migrated to the TEP from the central or western Pacific (Table 6). Those include 73 widespread Indo- Pacific species and one widespread circum- tropical species (Heteropriacanthus cruentatus) found at the western edge of the EPB. Three of the remainder also occur at the western edge of the EPB: two from the central and western Pacific {Stethojulis bandanensis and Acanthurus achilks) and one {Ctenochaetus mar- ginatus) from the central Pacific. Three other eastward migrants {Encheliophis vermicular is, Cyclichthys spilostylus, and Scomber japonicus) are entirely absent from the central Pacific, and another (Canthigaster valentini) occurs in the central Pacific but ~2,000 km west of the western edge of the EPB. Among the 57 eastward migrants that have resident populations in the TEP, 54 are present at one or more of the Hawaiian, Line, and Marquesas Islands, 42 of them at all three sites (Table 6). Ten are absent from the Ha- waiian fauna (three others occur there only as vagrants), 10 are absent from the Marquesas, and only one {Carcharhinus albimarginatus, a large, readily recognizable shallow-water reef shark) almost certainly does not occur at the Line Islands. The apparent absences of four others from the Line Islands may reflect in- adequate sampling. Twenty-three species of TABLE 6 Occurrence of Supposed Eastward-Migrant" Tropical Transpacific Shore Fishes at Hawai'i, Johnston Island, the Line Islands, and the Marquesas Islands Eamily Transpacific Species Hawaiian Is. Johnston I. Line Is. Marquesas Is. Carcharhinidae Carcharhinus alhimarginatus Vagrant? No No + Triamodon obesus + + + + Dasyatidae Taeniura meyeni No No + + Anguillidae Anguilla marmorata No No + + Muraenidae Echidna nebulosa + + + + Enchelynassa canina + No + + Gymnomuraena zebra + + + + Gymnothorax buroensis Vagrant + + + G. flavimarginatm + + + + G. javanictis + + + + G. mekagris + + + + G.pictus + + + + G. undulatm + + + No Scuticaria tigrina + + + + Uropterygius macrocephalm + + + + U. supraforatus + + + No Chanidae Chanos chunos + No + + Carapidae Encheliophis vermicularis No No No No Ophidiidae Brotula multibarbata + + + + Antennariidae Antennarius coccineus No No + + A. commerson + No 5 No Belonidae Tylosurus imperialis melanotus + No + + Holocentridae Myripristis bemdti + + + + Plectrypops lima + + 5 + Syngnathidae Horyrhamphus excisas excisus + + + + Eistulariidae Eistularia commersonii + + + + Aulostomidae Aulostomus chinensis + + + + Scorpaenidae Taenianotus triacanthus + No + + Priacanthidae Heteropriacanthus cruentatus'' + + + + Malacanthidae Malacanthus brevirostris + + + + Carangidae C. orthogrammiis + + + + Caranx melampygus + + + + C. sexfasciatus + + + + Decapterus macrosoma + No + No Gnathanodon speciosus + No + + Lutjanidae Aphareus furca + + + + Pristipomoides zonatus + + ? + MuUidae Mulloidichthys vanicolensis + + + + Chaetodontidae Chaetodon auriga + + + + C. kleinii + No + No C. l?nula + + + + C. meyeri No No + No C. unimaculatus + + + + Forcipiger flavissimus + + + + Kuhliidae Kuhlia mugil No No + No Cirrhitidae Cirrhitichthys oxycephalus No No + + Oxycirrhites typus + No + + Labridae Iniistius pavo + + ? + NovacuUchthys taeniourus + + + + Stethojulis bandanensis No No + No Thalassoma purpureum + + + + Scaridae Calotomus carolinus + + + + Scarus ghobban No No + No S. rubroviolaceus + + + + Zanclidae Zanclus comutus + + + + 528 PACIFIC SCIENCE ? October 2004 TABLE 6 (continued) Family Transpacific Species Hawaiian Is. Johnston I. Line Is. Marquesas Is. Acanthuridae Acanthurus achules + + + + A. nifficans + + + + A. triostegus triostegus No (difFerent No (difFerent + No (difFerent subspecies) subspecies) subspecies) A. xanthopterm + No + + Ctenochaetus marginatus No + + + Naso annulatus + + + + N. brevirostris + + + + N. hexacanthm + + + + N. lituratm + + + + N. vlamingii No No + No Sphyraenidae Sphyraena barracuda + + + + S. qenie Vagrant No + + Scombridae Scomber australasicus + No No No Scomber japonicus No No No No Bothidae Bothus man?is + + + + BaKstidae M. vidua + + + + Xanthichthys caeruleolineatus + No + + Monacanthidae Cantherhines dumerilii + + + + Ostraciidae Ostracion mekagris mekag-is No (difFerent subspecies) No (difFerent subspecies) + + Tetraodontidae Arothron hispidus + + + + A. mekagris + + + + Canthigaster amboinensis + No + + C. janthinoptera No No + + C. valentini No No No No Diodontidae C. spilostylus No No No No Total 80 59 (+3 vagrants) 49 70 (+4?) 64 '^ Species other than those labeled as circuintropical or westward migrants in Table 1. ^ Circumtropical, but see text on this species. transpacific species (Table 6) that are va- grants in the TEP must be very recent east- ward migrants. Twelve of them occur at the Hawaiian, Line, and Marquesan Islands and two at none of those islands. One (Chaetodon meyeri) occurs only at the Line Islands, and the apparent absence of one of those species from the Lines may be a sampling artifact. Four do not occur in the Hawaiian Islands and three are absent from the Marquesas. Thus there are more occurrences at the Line Islands than at either the Hawaiian or Mar- quesas Islands, and there is only one major absence from the Lines (C. albimarginatus, which is also absent from Hawai'i). This pattern is generally consistent with eastward migrants being transported from the Line Islands on the only major eastbound warm surface current that flows across the EPB, the North Equatorial Counter Current (NECC). The few eastward migrants not fotmd at any of the islands at the western edge of the EPB occur farther west in the path of the NECC. The only notable inconsistency with this pattern that is not readily attributable to a sampling artifact is the distribution of C albimarginatus. The two Scomber species are basically temperate species, only one of which {S. australasicus) occurs at Hawai'i. Both could have taken the same route eastward from Japan, on extensions of the cool Kuroshio Current that reach to Baja California. Four of the 80 species (Heteropriacanthus cruentatus, Fistularia commersonii, Plectrypops lima, and Bothus mancus) that we include as eastward immigrants are widely distributed in the Indo-central Pacific and have Atlantic sister species with similar morphology and Tropical Transpacific Shore Fishes ? Robertson et al. 529 ecology. Thus their TEP populations could represent either isthmian relicts that have remained connected to Indo-central Pacific populations or recent (postisthmus) eastward immigrants. Heteropriacanthus cruentatus cur- rendy is recognized as a circumtropical species in a monotypic genus. The genus Pkctrypops includes only P. lima and P. retro- spinnis (tropical Northwest Adantic). The fact that P. lima has only been recorded three times in the TEP suggests that it is a vagrant there and hence an eastward migrant. Fistu- laria commersonii has an Adantic sister, F. ta- bacaria. If Fritsche's (1976) scenario is correct, the TEP population of F. commersonii is an isthmian relict. Bothus mancus has a West Adantic look-alike congener, B. lunula- tus, that, like B. mancus, lives in reef environ- ments. There is no information bearing on the origin of the TEP population of B. man- cus. Vermeij and Rosenberg (1993) noted that tropical marine invaders often have limited distributions in areas they invade and occur at offshore islands or small sections of the avail- able mainland. The distribution of B. mancus in the TEP fits this scenario. Although gene flow evidendy occurs between TEP and cen- tral Pacific populations of H. cruentatus (see section on Genetic Studies later in Discus- sion), the origin of its TEP population re- mains unclear. No genetic studies have been done on the other three genera that would clarify the origins of their TEP populations. WESTWARD MIGRANTS. The distributions of 22 shore-fish species and one oceanic species, supplemented in one case by genetic data, indicate that they are or may well be westward migrants (Table 1). The only central Pacific area at which 16 of them are known are the Hawaiian Islands: 10 shore-fish species as likely Hawaiian residents {Carcharhinus altimus, Sphyrna zygaena, Echi- norhinus cookei, Odontaspis ferox, Etrumeus teres, Mugil cephalus, Priacanthus alalaua, Ca- rangoides caballus, Decapterus muroadsi, Uraspis helvola) and five shore-fish species {Sphyrna mokarran, Dasyatis dipterura, Encheliophis du- bius. Seri?la lalandi, and Seri?la rivoliana) and one oceanic species {Euthynnus lineatus) as vagrants. Six species are present at Hawai'i and the Marquesas (or nearby islands) but not in the remainder of the central Pacific: Carcharhinus limbatus and Sphyrna lewini (res- idents at both sites), Sectator ocyurus (resident at the Marquesas and a vagrant on the equa- tor west of the Line Islands), Balistes polylepis (vagrant at both sites), Xanthichthys mento as a resident of Hawai'i and islands south of the Marquesas, and Chilomycterus reticulatus as a resident in Hawai'i and a vagrant at the Mar- quesas, Line, and other islands. The case for westward migration from the TEP is strong- est for 11 species that are residents of the TEP and occur only at central Pacific sites at the western side of the EPB, particularly as vagrants. However, the other 12 species have populations in the West Pacific that may have provided central Pacific migrants or those central Pacific populations may be remnants of a once more-continuous popula- tion. Hawai'i, for example, has connections with the Japanese fauna, mediated by dispersal on the eastward-flowing Kuroshio Current to the NW Hawaiian Islands (e.g., Fritsche 1976, Mundy in press). If the TEP form of Ostracion mekagris is shown to be the sub- species currendy thought to be endemic to Hawai'i (see Appendix 1), the number of westward-migrant shore fishes would increase to 23 and decrease the number of eastward migrants to 79. If vagrants of Lobotes from Hawai'i and French Polynesia prove to be L. pacificas (see section on Recent Erroneous and Questionable Records in Results), the num- ber of westward migrants would increase to 24. Transpacific Species with Highly Disjunct Pacific Distributions Twelve species of transpacific shore fishes occur in the TEP and the West Pacific but nowhere in between (Table 1). TEP pop- ulations of seven of those, all circumtropical species, likely represent isthmian relicts (see next section) that lack ongoing contact across the Pacific. The remaining five most probably are not isthmian relicts and we attribute their occurrence in the TEP to ongoing or recent eastward migration: Two of those five {Cy- clichthys spilostylus and Canthigaster valentini) occur in the TEP as vagrants and two others 530 PACIFIC SCIENCE ? October 2004 that either have {Encheliophis vennicularis) or may have {Enchelychore lichenosa) resident populations in the TEP lack Atlantic look- alike sister species. Genetic data on the fifth, S. japonicus (see under Shore Fishes in Ge- netic Studies later in Discussion), indicate that it arrived in the eastern Pacific well after the rise of the isthmus. Canthigaster valentini and C. spilostylus evidently have long-lived pelagic larvae or juveniles able to migrate 9,000-11,000 km to the TEP. The range of S. japonicus is primarily temperate (Grant and Bowen 1998), and its pelagic adults likely made the migration between Japan and the eastern Pacific on extensions of the Kuroshio Current that span the northern Pacific and extend to off Baja California. There is no in- formation on the pelagic larval durations of carapids, although they do have large larvae and a pelagic juvenile stage that has been collected far beyond adult ranges (Markle and Olney 1990). Hence, whether ongoing con- nections occur between western and eastern populations of E. vermkularis is an open question; they may represent currently dis- junct populations of a previously more wide- spread species. The situation is unclear for E. lichenosa. ft is a resident in the NW Pa- cific, around Taiwan and Japan, but has an uncertain population status in the Gal?pagos, where specimens have been collected over 28 yr (see Appendix 1). If it is resident in both areas those populations could repre- sent disjunct survivors of a previously more widespread species. However, moray eels are thought to generally have pelagic larval dura- tions on the order of many months (see later in Discussion under What Types of Eishes Are Transpacifics?), and it is conceivable that ongoing pelagic dispersal across the entire Pacific could lead to the repeated arrival of vagrants of ?. lichenosa in the Gal?pagos. This could be accomplished by two main surface routes: first, extensions of the cool Kuroshio Current could carry larvae eastward to Baja California (cf. the situation with Scomber), which then make their way south to the Gal?pagos. Alternatively, propagules could make their way south from Taiwan to the eas- tern Philippines, the origin of the main east- ward equatorial currents flowing to the TEP, the Equatorial Undercurrent (EUC) and the North Equatorial Counter Current (NECC) (cf Anguilla, see later in this section). There are difficulties with both those routes, which are roughly similar in distance (~ 16,000 km). The Kuroshio Current is a cool one, and larvae carried by it to Baja would have to move south 30? of latitude through a region dominated by westward-flowing currents and lacking major southerly currents. On the sec- ond route the main surface oceanic flow across the 12? of latitude separating the Philippines and Taiwan is counter to the direction larvae would need to move to get to the eastbound equatorial flow. Another substantial problem for the ongoing-dispersal scenario is why such activity has not resulted in records of this species at any of the many sites scattered along either route between the NW Pacific and the Gal?pagos. Hence, currently it seems more likely that the Gal?pagos harbor a relict population of E. lichenosa. Although Anguilla mannoratus occurs in the central Pacific, in- cluding islands on the western edge of the EPB, it probably eifectively has a disjunct distribution, because its only known Pacific spawning grounds lie immediately east of the Philippines and of Papua (Aral et al. 2002). The NECC and EUC both originate near the northern spawning groimd and to reach the TEP from there larvae of this species, which have very long pelagic larval durations (Table 9), would have to travel ~15,000 km. The eastboimd current originating around the southern spawning area (the South Equa- torial Counter Current [SECC]) likely only delivers larvae to the central Pacific (see sec- tion on Transport Potential later in Discus- sion). Transpacific Fishes as Isthmian Relicts in the TEP The best candidates for transpacific species whose TEP populations represent isthmian relicts are circumtropical species that have disjimct Pacific distributions, Atlantic gemi- nates, and are likely westward migrants. Twenty of the 39 circumglobal shore fishes have disjunct Pacific distributions and either are absent from the central Pacific (Carcharhi- Tropical Transpacific Shore Fishes ? Robertson et al. 531 nus brachyurus, C. lencas, C. obscurus, Sphryna mokkaran, Notorynchus cepedianus, Mobula ja- panica, Mobula thurstoni, and Trichiurus lep- turus) or may be westward migrants to there {Carcharinus altimus, C. limbatus, Sphyma zygaena, Odontaspis ferox, Echinorhinus cookei, Etrumeus teres, Mugil cephalus, Decapterus mur- oadsi. Seri?la lalandi, S. rivoliana, and Uraspis helvola) (Table 1). Chilomycterus reticulatus also falls into the latter group as a resident of Ha- wai'i and a vagrant in the rest of the central Pacific. However, its large pelagic juvenile may also have allowed long-distance eastward migration from the West Pacific similar to that by its confamilial Cyclichthys spilostylus. One "circumtropical" species (Sphyraena bar- racuda) is a vagrant in the TEP and hence not an isthmian relict. The broad Pacific dis- tributions of another 18 circumglobal species provide no clues to the status of their TEP populations, which may be clarified by genetic data (e.g., Heteropriacanthus cruentatus, see under Genetic Studies later in Discussion). Ocean Current Systems and Migration aa'oss the EPB TRANSPORT POTENTIAL OF EASTWARD AND WESTWARD CURRENTS. BeCaUSC ttOpi- cal shore-fish larvae seem to be restricted to the upper 100 m, and probably much shal- lower levels (Leis and McCormick 2002), surface and shallow subsurface currents are those most relevant to trans-EPB migration. Information on such currents in the EPB is derived from published analyses plus post- 1993 satellite data (see http://www.oscar. noaa.gov). There are two eastbound currents in the central part of the EPB: the NECC, a warm, surface current that flows between ~5? and 10? N and carries about 30% of the eastward flow, and the EUC, a narrow (~2? of latitude wide), cool, subsurface current flowing along the equator that gradually shoals as it progresses eastward. The EUC carries 70% of the eastward flow (Wyrtki 1967) and is about twice as fast as the NECC (Wyrtki 1965, 1967, Firing and Lukas 1983, Chavez et al. 1999). Surface eastward flow through the central Pacific also occurs on the SECC. In the central Pacific this current is narrow (~1? wide) and slow, with flows wan- dering between ~8 and 17? S to the Mar- quesas (7-10? S). The SECC seems to have little influence in the EPB, although it may extend east to the longitude of the Gal?pagos but well south of those islands (Eldin 1983, and see http://www.oscar.noaa.gov). West- bound currents include the North Equatorial Current (NEC) and South Equatorial Cur- rent (SEC). Divergence associated with a thin band of upwelling from the EUC along the equator cuts the flow of the SEC into two sections, to the north and south of that band, with stronger flow on the northern arm. Al- though the SEC extends down to at least 17? S, its strongest flow is above 8? S. The NEC is notably slower than the SEC, and the speed of the northern arm of the SEC is similar to that of the NECC. Flows of the NEC and SEC are weaker at higher latitudes and stronger on the western side of the EPB than on the eastern side. These trans-EPB current systems are dy- namic, changing seasonally and in relation to the El Ni?o/Southern OsciUation (ENSO) cycle, as well as over periods of decades (Chavez et al. 2003). Seasonal changes in- clude more vigorous surface flow in both di- rections in the second half of the year, when the NECC extends eastward into the TEP (early in the year it terminates well to the west). Current flows during the two extremes of the ENSO cycle (El Ni?o and La Ni?a) differ as follows: During El Ni?os eastward flow is greatly enhanced and westward flow reduced, particularly during the latter three- fourths of the year. The speed of the NECC can double, and during a strong El Ni?o, the transit time for a water mass across the EPB can be cut from a normal ~100-155 days to ~50-80 days (Wyrtki et al. 1981, Firing et al. 1983, Wyrtki 1985, Richmond 1990). During a strong El Ni?o the latitudinal spread of the NECC also increases southward, to ~5? S, and eastward flow replaces the usual west- ward flow by the northern arm of the SEC. In addition, during a strong El Ni?o the EUC (and its associated upwelling) can disappear (Firing et al. 1983, Chavez et al. 1999). Dur- ing a La Ni?a, westward flow of the SEC strengthens, as does flow of the EUC and 532 PACIFIC SCIENCE ? October 2004 the upwelling above it (Chavez et al. 1999). The increase in eastward flow was greater during the intense 1997-1998 El Ni?o than during the weaker 2002-2003 event, the in- crease in flow of the SEC during the 1999 La Ni?a was less than that of the NECC during the 1997-1998 El Ni?o, and there was litde change in the NEC during either type of event (http://www.oscar.noaa.gov). The 1999 La Ni?a was much weaker than six other La Ni?as that occurred between 1950 and 1990 (http://www.cdc.noaa.gov/ENSO/enso. mei_index.htnil). Satellite data on currents are available only since 1993 (http://www. oscar.noaa.gov), so it is unclear what elfect strong La Ni?as have on westward flow. Oc?anographie and biologic conditions vary gready in different parts of the EPB that contain the major currents and the islands on each side of that barrier, to the extent that Longhurst (1998) included diff?rent lat- itudinal zones of the barrier in four separate pelagic biogeographic provinces. A thin band (<5? of latitude wide) of cool upwelling from the EUC that extends west from the Gal?pagos along the equator across most of the EPB produces an overlying strip of increased productivity (http://seawifs.gsfc. nasa.gov/ SEAWIES / IMAGES / SEAWIES_ GALLERY.html, Chavez and Smidi 1995, Chavez et al. 1999). The level of productivity in that upwelling band, although higher than in areas to the north and south, is much lower than that of productivity resulting from coastal upwelling, due to iron limitation (Chavez and Smidi 1995, Longhurst 1998). Productivity in the EPB declines with in- creasing latitude away from the equatorial upwelling band. In addition there are two zones of notably low productivity in the northern part of the EPB, one along the center of the NECC and the other, which is very low, spanning half or more (depending on the season) of the area between the Ha- waiian Islands and the TEP. Thus productiv- ity is higher in the SEC than in most of the NEC and very low in the entire zone occu- pied by the greatly expanded NECC during an intense El Ni?o. The NECC is a warm, surface current, but the EUC is cool for tropical organisms (15-20?C at its center, at ~100 m depth), although warmer near the surface. Temperatures are cooler in both the NEC and die SEC dian in die NECC, al- though (see Scheltema 1988) they are suitable for tropical organisms. However, because the northern arm of the SEC is fed pardy by warm water from the equatorial TEP and the southern arm by flow from the cold Peru current, the southern arm is distincdy cooler (at ~22?C near the southern Gal?pagos) than the northern arm when it leaves the TEP during the season of peak flow. Whether currents carry pelagic propagules of shore organisms across the EPB depends not only on transit times and the duration of the larval life (and the ability of larvae to ex- tend their pelagic lives) but also on conditions within the currents for larval sustainment. How might variation in conditions in east- and westbound currents affect all these as- pects of transit potential for fish propagules? In most accounts eastward migration is as- sumed to occur on the NECC (e.g., Rosen- blatt et al. 1972, Briggs 1974, Scheltema 1988, Vermeij 1991), altiiough die EUC route has also been considered (e.g., Grigg and Hey 1992, Clarke 1995). Under non-El Ni?o conditions the EUC could deliver a propagule to the TEP in about half the time it would take via the NECC. The limited information available on depth distributions of tropical shore-fish larvae (Leis and Mc- Cormick 2002) suggests that they remain above 100 m depth and often much shal- lower. Although shallow larvae could take advantage of warmer, higher productivity conditions above the core of the EUC, they would do so away from the high speed of the cold core in a zone where they could be advected into the westbound SEC, to the north and south of the upwelling strip. Thus eastward transport on the EUC seems prob- lematic. More information on the depth dis- tributions of shore-fish propagules should indicate whether it is a realistic possibility. The SECC likely is irrelevant for eastward migration to the TEP; it is very slow and whatever propagules of tropical species it picked up from the Marquesas would be carried to an unsuitable, temperate area of the eastern Pacific. Rather, it may act more Tropical Transpacific Shore Fishes ? Robertson et al. 533 as a hindrance to the Marquesas receiving propagules from the TEP on the SEC. For westbound transits the combination of higher speeds and productivity in the SEC than in the NEC suggests that potential for transits should be higher in the SEC. During a La Ni?a, a transit from the Gal?pagos to the southern Line Islands on the faster northern arm of the SEC likely would be shorter than one to the Marquesas on the slower southern arm of that current, despite the substantially greater distance of the former route (see Fig- ure 1). Westward migration from the equa- torial section of the TEP to the Line Islands, which has not been considered previously, is a realistic possibility given that westward mi- gration does occur to both Hawai'i and the Marquesas in current systems that seem less conducive to such events. POTENTIAL EFFECTS OF EL NI?OS ON TRANs-EPB MIGRATION. El Ni?os are thought to enhance migration of tropical shore fishes and other organisms across the EPB to the TEP, largely due to the doubling in the velocity of the NECC during such events (e.g., Richmond 1990, Leis 1991, Grigg and Hey 1992). This increased flow rate should increase the number of potential migrants to include species with moderate pelagic larval durations, and the tripling of the latitudinal spread of that current should mean that more propagules can be drawn from a larger area of reefs. However, con- ditions in the center of the EPB change dur- ing strong El Ni?os in ways that may limit eastward dispersal. First, during an El Ni?o the equatorial productivity produced at other times by upwelling from the EUC (Chavez et al. 1999) is nearly eliminated. Second, during El Ni?os the surge of water moving east- ward across the EPB has temperatures 2-3?C above normal. The development and growth of pelagic larvae of tropical reef fishes is strongly temperature dependent (McCormick and Molony 1995, Leis and McCormick 2002), with faster-growing individuals having shorter pelagic larval durations (Wellington and Victor 1992). The age at metamorphosis of fishes decreases with increasing tempera- ture (Benoit et al. 2000) and tropical reef-fish larvae have increased growth rates and de- creased pelagic larval durations at higher temperatures (McCormick and Molony 1995, Wilson and Meekan 2002, Meekan et al. 2003). Further, marine fishes tend to produce smaller eggs at higher temperatures (e.g., Chambers 1997, Kokita 2003). It is unclear how long the larvae of tropical fishes can delay settlement after achieving competency. Although some widely distributed taxa (in- cluding acanthurids, labrids, and flatfishes) can do so, extensive delays may not be the rule (McCormick 1994, Leis and McCormick 2002). Further, <1% of individuals surviving to settlement may have extended larval dura- tions (Leis and McCormick 2002), and such extension may reduce subsequent survival (McCormick 1999). Thus, although eastward transit times may be halved during strong El Ni?os, that enhanced transit potential may be oifset by reduced larval provisioning in the egg, low food availability in the eastbound flow, and "premature" achievement of settle- ment potential. A direct test of this scenario may not be feasible, but comparison of varia- tion in egg content and pelagic larval dura- tions of reef fishes under normal and El Ni?o conditions at the likely origin of eastbound propagules (the Line Islands) could act as a partial proxy; partial because conditions for larval development around an island diifer from those in the middle of the EPB. Based on observations of vagrant trans- pacific fishes in the TEP around the time of an El Ni?o event, it has been proposed that transpacific recruitment to the TEP is en- hanced by such events (Grove 1984, 1986, Robertson and Allen 1996, Glynn and Ault 2000). What evidence is there of El Nino- enhanced migration across the EPB? Wide- spread observations on reef fishes in the TEP only began ~25 yr ago. Since then there have been two intense El Ni?os (1982-1983 and 1997-1998), plus several weaker events (1987, 1991-1995, 2003). There are few records of the precise timing of appearances of trans- pacific shore fishes on either side of the EPB relative to those events. Arrivals of transpacific species in the TEP: El Ni?o related? The dates of the first record of each of 10 species in the TEP are relevant to this question. Two arrivals are associated 534 PACIFIC SCIENCE ? October 2004 with El Ni?o events: Anguilla marmorata: McCosker et al. (2003) estimated that the in- dividual of this vagrant freshwater eel they examined arrived in the Gal?pagos in 1993 (i.e., in the center of a 5-yr period of sustained El Ni?o conditions [see http://www.cdc.noaa. gov/ENSO/enso.niei_index.html]). Brotula multibarbata: The first record for this species in the TEP consists of two small juveniles collected by D.R.R. at Clipperton in March 1998, at the end of the 1997-1998 El Ni?o. This is the best record of an El Nino- associated arrival in the TEP. However, other records do not clearly demonstrate such as- sociations: Uropterygius supraforatus: The first specimens of this species were taken at Clip- perton by D.R.R. in March 1998; however, because they included both adults and juve- niles it is unlikely that this species first arrived during the 1997-1998 El Ni?o. Chaetodon auriga and C. l?nula: Individuals of these spe- cies were first observed in the TEP in the northern Gal?pagos in mid-1987 (Merlen 1988); however lack of size data makes it un- clear whether they arrived during the 1987 El Ni?o. Chaetodon kleinii: The first record of this species in the TEP was by J.S.G. in the Gal?pagos in mid-1984 (Grove 1986) (i.e., more than a year after the 1982-1983 El Ni?o had finished). Three other individuals were observed at other sites in the Gal?pagos in 1988. Canthigaster amboinensis: A single medium-sized adult was collected in the northern Gal?pagos in Eebruary 1967, a year after the 1965-1966 El Ni?o. Canthigaster janthinoptera: A single individual of this spe- cies was found in the Gal?pagos in mid-1984, over a year after the 1982-1983 El Ni?o. Canthigaster valentini: Two individuals of this species were found in dilferent parts of the Gal?pagos, in late 1983 and late 1984. The earlier record could represent an El Nino- mediated arrival. Naso vlamingii: A small adult of this species was photographed in the Gal?pagos in 1990 (Robertson and Allen 2002), several years after an El Ni?o. Arrivals at HawaiH: El Ni?o related? Mundy (in press) suggested that changes in the abundance of two carangids in Hawai'i have resulted from El Nino-enhanced westward migration. First, Carangoides cahallus, which was first collected in Hawai'i early in the twentieth century (Fowler 1928), became very abundant there at the end of the 1997- 1998 ENSO (Randall and Carlson 1999), but within several years was no longer being caught by fishers (B. C. Mundy, pers. comm., 2003). Second, records of vagrant adult Se- ri?la lalandi at the southern Hawaiian islands are during El Ni?o years (Mundy in press). Effects of El Ni?os on intraregional distri- butions. Rather than being due to El Nino- enhanced migration across the EPB, local increases in abundances of transpacific species on either side of the EPB could arise through other mechanisms. El Nino-related environ- mental changes strongly aifect the fish faunas of peripheral parts of the TEP, with the Gal?pagos providing the best documented example. These islands are at the junction of the temperate Peruvian province and the TEP, and have both a strong north/south gradient and strong seasonal variation in environmen- tal conditions and local faunal composition (Rosenblatt and Walker 1963, McCosker et al. 1978). In the Gal?pagos during El Ni?os there are both declines of local endemics and increases in the abundance of species widely distributed in the TEP, including both re- gional endemics and transpacific species (Grove 1984, 1986, Ruttenberg 2000, Victor et al. 2001). Similar changes in resident transpacific populations have occurred at the northern edge of the TEP (Victor et al. 2001). Dramatic environmental changes that occur in the TEP during El Ni?os evidently aifect both larval and adult survivorship and pro- duce local changes in abundances and shifts in distributions by established residents within a region. Thus, rather than being due to en- hanced trans-EPB migration, local increases in transpacific species during El Ni?os could be due to either (a) environmental changes that favor short-term population increases of usually tmcommon resident transpacific spe- cies, or (b) species whose propagules transit across the EPB at other stages of the ENSO cycle as well as El Ni?os experiencing condi- tions favorable for their development to adulthood at the arrival point only during an El Ni?o. Interpretation of new records of vagrant Tropical Transpacific Shore Fishes ? Robertson et al. 535 transpacific species in a region immediately after El Ni?os as evidence of enhanced trans- EPB migration needs to be treated with cau- tion for another reason. The intense 1982- 1983 El Ni?o produced large changes in many TEP reef environments, particularly at the Gal?pagos, which have large areas of temperate habitat (Glynn and Ault 2000). That was the first El Ni?o during which large-scale changes in oc?anographie condi- tions and current flows in the central and eastern Pacific were revealed by satellite data. This event stimulated a heightened aware- ness of the impact of El Ni?os that has undoubtedly affected the eifort expended in looking for changes in fish populations (e.g., Victor et al. 2001) and for "new" arrivals. The intensity of scuba diving in the Gal?- pagos and elsewhere in the TEP also gready increased beginning shortly before the 1982- 1983 El Ni?o and continues to do so. As a result divers are more likely to expect immi- grants during El Ni?os and interpret increases in the abundance of transpacific species dur- ing them to enhanced trans-EPB migration. In conclusion, contrary to previous im- pressions (e.g., Grigg and Hey 1992, Glynn and Ault 2000) equivocal evidence suggests at most a limited enhancement of migration by shore fishes in both directions across the EPB during seven El Ni?os since 1982-1983, a period when the frequency and intensity of those events has been increasing. Given that El Ni?os typically recur at 2- to 7-yr intervals and last for 1-5 yr, some sightings of vagrant immigrants are bound to be associated with such events by chance alone. Even though some vagrants may well have arrived during El Ni?os, there is no evidence of a large influx of immigrants of any species leading to the establishment of a "new" transpacific population in the TEP during a period that included the two most intense El Ni?os (1982-1983 and 1997-1998) of the past cen- tury. The 1997-1998 event is particularly informative because ample prior warning of its occurrence undoubtedly led to increased surveillance of its eifects in the TEP. Thus the long-term impact of such events on the composition of the TEP shore-fish fauna re- mains unclear. Until more is known about the "background" level of immigration it will be hard to determine the effect El Ni?os may have on trans-EPB immigration of new fau- nal elements and thus on long-term changes in regional faunas. If only about 50-60 east- ward migrants have established themselves in the TEP so recently that they retain specific continuity with central Pacific populations, then influxes of central Pacific species in suf- ficient numbers to establish a new resident population in the TEP must occur very in- frequendy. What Types of Fishes Are Transpacifics? TAxoNOMic BIASES. Prcvious authors noted strong taxonomic biases in the repre- sentation of different families of Indo-central Pacific fishes among transpacifics (e.g., Rosen- blatt et al. 1972, Leis 1984). However, when those papers were written comprehensive lists of the shore fishes of the central Pacific and the three island groups at the western edge of the EPB were not available. New information makes such a compilation possible (see Table 2). Three families (Carcharhinidae, Murae- nidae, and Carangidae) that are well repre- sented in the TEP and central Pacific faunas, including the faunas of the Hawaiian, Line, and Marquesas Islands, contribute many transpacific species, and transpacifics repre- sent major proportions of the members of those families in the faunas on both sides of the EPB. In contrast, other families that have many species in the central Pacific and Hawaiian, Line, and Marquesan faunas lack transpacific representatives, most nota- bly Ophichthidae, Scorpaenidae, Serranidae, Apogonidae, Pomacanthidae, Pomacentridae, Blenniidae, and Gobiidae. Most of those families also have substantial numbers of spe- cies in the TEP (Table 2). In further contrast, all five families in one order (Tetrao- dontiformes: Balistidae, Monacanthidae, Os- traciidae, Tetraodontidae, and Diodontidae) that have relatively few species on each side of the EPB also are well represented among the transpacifics. Thus these additional data strengthen previous conclusions of strong taxonomic biases in the representation of different potential "donor" families in the 536 PACIFIC SCIENCE ? October 2004 Representation TABLE 7 of DiiFerent Fishes in the Transpacific and Indigenous Faunas of the TEP Fauna and of Transpacifics in Each Part of the TEP Fauna Fish Type % of Transpacifics % of Indigenous Transpacifics as % of Each Represented by Fauna Represented by Group in Total TEP Fauna Oceanic epipelagic Inshore pelagic Inshore demersal n 33.7 0.1 98.4 19.5 9.5 26.1 47.9 90.4 9.7 190 949 1,139 transpacific fatina: the occurrence of trans- pacific species in di?Ferent families is largely tinrelated to family richness on either side of the EPB. Although much of this pattern of biases is real, some part may represent sampling artifacts. Members of two genera {Chaetodon and Naso) include a substantial number of the transpacifics recorded as vagrants in the TEP. These fishes are large, diurnally active, conspicuous, and readily recognizable as dif- ferent from indigenous TEP species. Vagrant individuals of those and various other taxa are likely to be readily noticed on eastern Pacific reefs and oifshore islands frequented by scuba divers. In contrast, other families for which no transpacifics are known have character- istics that greatly reduce the likelihood that any vagrants present in the TEP or the cen- tral Pacific would be recorded. Eor example synodontids are cryptic soft-bottom fishes that look very similar to one another. Ophichthids and scorpaenids are cryptic, sedentary fishes and most of the former bur- row within soft bottoms. In other cases diifer- ent species on each side of the EPB resemble one another closely enough that migrants from one side of the EPB might well pass unrecognized on the other side (e.g., spe- cies of Kyphosus, which often associate with flotsam [e.g.. Hunter and Mitchell 1966] and are good candidates to be such migrants, look much the same worldwide). Mulhidich- thys vankoknsis and M. dentatus are so simi- lar, morphologically, that the unsuspected penetration of the former into the TEP has only been revealed by genetic analyses (see section on Genetic Studies later in the Dis- cussion). PELAGIC DISPERSAL POTENTIAL: TRANS- PACIFIC SPECIES VERSUS NONTRANS- PAciFics. Distances across the EPB are vast (4,500-7,000 km) and minimal transit times under the best of conditions are on the order of a couple of months. Rosenblatt et al. (1972) pointed out that transpacific fishes generally belong to taxa that have either pe- lagic adults or pelagic larvae that likely are relatively long-lived. Leis (1984) extended this analysis and found that, although trans- pacific species do generally have long-lived larvae many families that have similar larvae lack transpacifics. To assess how variation in dispersal characteristics may have shaped the structure of the transpacific fauna we com- pared such characteristics of the transpacific fauna and the nontranspacific faunas on both sides of the EPB. Effects of different pelagic life-history stages. The representation of difl^erent fishes in the transpacific fauna and diiferent segments of the TEP fauna relates strongly to the degree to which their adults are pelagic (Table 7). At one extreme, oceanic epipelagic species represent < 1 % of the indigenous TEP fauna but almost one-third of the transpacific fauna and almost all the oceanic species in the TEP fauna. In contrast, demersal shore fishes rep- resent 90% of the indigenous TEP fatma and about half the transpacific fauna, but only ~10% of demersal species in the TEP are transpacifics. Inshore pelagic species are in- termediate between these two extremes. Hence, the pelagic habits of adults enhance Tropical Transpacific Shore Fishes ? Robertson et al. 537 TABLE 8 Occurrence of DiiFerent Pelagic Life-History Phases among Transpacific and Nontranspacific Shore Fishes % of Fauna" Represented by Each Group Indigenous hidigenous TEP + Central Dispersal Characteristics Transpacifics* TEP Central Pacific Pacific No pelagic phase 3.3 8.4 1.3 4.5 Demersal adult and egg"", pelagic larva 1L6 33.9 37.1 35.7 Demersal adult, pelagic egg and larva 52.1 44.9 58.7 52.6 Pelagic adult, no pelagic larva 10.7 2.4 0.5 1.4 Pelagic adult, demersal egg, pelagic larva 1.7 4.0 0.9 2.2 Pelagic adult, egg, and larva 20.7 6.4 2.0 3.6 n 121 949 1,194 2,143 ^ Faunal data from Table 2. ^ Excludes transisdimian relicts not occurring in the central Pacific. ' Includes species with brooded eggs and live bearers. their long-distance dispersal capabilities and increase their representation in the trans- pacific fauna and their section of the TEP fauna. The vast majority of tropical reef teleosts have pelagic larvae, which are released from either pelagic eggs (most species) or demersal (including brooded) eggs (Leis and Carson- Ewart 2000). In the TEP, -94% of the shore fishes have pelagic larvae and 61% of such species produce them from pelagic eggs (Robertson and Allen 2002). Indo-Pacific reef fishes that produce pelagic larvae from de- mersal eggs tend to have smaller geographic ranges than do pelagic spawners (Thresher and Brothers 1985, Thresher 1991, Pyle 1999, Hughes et al. 2002, Jones et al. 2002). Demersal spawners may have a reduced dis- persal ability due to their larvae (1) tending to remain inshore (while those of pelagic spawners often develop oifshore) (Leis 1986, Brogan 1994), and (2) having short pelagic larval durations (Thresher and Brothers 1985, 1989, Thresher 1991). Although most analy- ses of Zoogeographie patterns of tropical fishes in relation to dispersal modes have focused on contrasts between demersal- and pelagic-spawning teleosts that produce pe- lagic larvae, Pyle (1999) included species that lack any pelagic life-history phase in his analysis of the distributions of shore fishes in Oceania. He found the largest ranges to be among pelagic-spawning teleosts but no difference between the range sizes of demer- sal spawners and species that lack a pelagic phase. Because all of the latter are elasmo- branchs he suggested that the more ancient origins of elasmobranchs had effectively "compensated" for the lack of a pelagic dis- persal phase by giving them more time to enlarge their ranges. Regardless of the causes of this pattern, his study demonstrates the need to include entire faunas in such analyses. We made an entire-fauna comparison of the reproductive and dispersal modes of the transpacific fauna and nontranspacific fauna of the fringes of the EPB (i.e., the three central Pacific island groups plus the TEP). There are strong differences between the structure of those two faunas (Table 8: G test, G = 14.3, P < 0.001): relative to their abun- dance in the nontranspacific fauna (1) de- mersal species with demersal eggs and pelagic larvae are strongly underrepresented; (2) de- mersal species with pelagic eggs and larvae are neutrally weighted (in terms of their pro- portional abundance); (3) species with pelagic adults (including types with and without pelagic eggs and larvae) are strongly over- represented; (4) species that lack a pelagic stage and pelagic species that have demer- sal eggs and pelagic larvae are about neu- trally weighted (sample sizes for these two groups are very small). As with Pyle's (1999) comparison, there is a strong phylogenetic component to these patterns?all demersal 538 PACIFIC SCIENCE ? October 2004 transpacific species that lack a pelagic stage or that have pelagic adults but not pelagic larvae are elasmobranchs, and all species that have pelagic larvae are teleosts. Among teleosts, having pelagic adults has a strong additive ef- fect over having only pelagic eggs and larvae: the ratio of the abundances of the former in the transpacific versus the nontranspacific fauna is much higher than the ratio for de- mersal species with pelagic eggs and larvae (Table 8: 5.8:1 versus <1:1). The ratio of the relative abundance of transpacifics to nontranspacifics among species in which all three stages are pelagic (Table 8: teleosts, 5.8:1) is less than the ratio among species with pelagic adults but not larvae (elasmo- branchs, 7.8:1). Thus having a pelagic adult provides a dispersal advantage for both tele- osts and elasmobranches, but teleosts that have pelagic larvae and adults do not have a greater achieved dispersal than elasmo- branchs with pelagic adults. These patterns extend Pyle's (1999) finding by indicating that elasmobranchs have achieved greater dispersal than teleosts despite the latter hav- ing additional dispersive life-history stages. That diiference could be due to the time available for dispersal (as mediated by the age of a species), as Pyle suggested, and/or to biological attributes of elasmobranchs (e.g., large size and mobility of adults and greater ability for adults to survive in oifshore waters). Diiferences in reproductive mode gener- ally are family specific, and families of teleosts vary greatly in their contribution to the fau- nas on both sides of the EPB and the trans- pacific fauna. Thus assessment of the effect of egg type on dispersal using total numbers of species with each spawning mode is con- founded by a phylogenetic effect. To control for that effect on dispersal potential among demersal shore fishes that produce pelagic larvae from the two types of eggs, we com- pared the proportions of demersal-spawning and pelagic-spawning families found on both sides of the EPB that have transpacific repre- sentatives (see Table 2). The difference, 0.21 of 17 versus 0.34 of 62, is in the same direc- tion as that involving total numbers of species in Table 8 but is not significant (G test of in- dependence: G =1.8, ns). Tetraodontiform fishes, which produce demersal eggs (balis- tids, monacanthids, and tetraodontids) or pe- lagic eggs (ostraciids and diodontids), have a pelagic juvenile stage capable of an extended pelagic existence and long-distance dispersal. All five tetraodontiform families are well rep- resented among the transpacifics (Table 1). Among nontetraodontiform teleosts there is a clear difference between pelagic- and demersal-spawning families in terms of rep- resentation in the transpacific fauna: 0.0 of 14 demersal-spawning families versus 0.32 of 60 pelagic-spawning families, G = 4.72, P < 0.01). Thus dispersal-related larval at- tributes associated with egg type do affect which families are represented in the trans- pacific fauna. Larval dispersal characteristics of transpacific species versus other teleosts. Leis (1984) pointed out that certain speciose families of pelagic- spawning Indo-central Pacific fishes (chaeto- dontids, pomacanthids, muUids, synodontids, and scorpaenids) that were not then known to have transpacific representatives probably have pelagic larval durations and larval dis- persal capabilities broadly similar to those of families that do. Newer analyses are consis- tent with that finding: Hughes et al.'s (2002) analysis of distribution patterns of Indo- Pacific reef fishes showed that five families of pelagic spawners (acanthurids, chaetodontids, holocentrids, lutjanids, and serranids) have proportionately more widespread species than expected, and that two others (pomacanthids and lethrinids) are neutrally weighted in this characteristic. Such patterns indicate that members of those 10 families have good powers of larval dispersal relative to other reef-fish families. However, although all 10 families have many species in the Line Is- land fauna (Table 2), only five of them have eastward-migrant transpacifics, and those mostly occur in the TEP as vagrants. Leis (1984) was obliged to rely on in- formed guesstimates of the pelagic larval du- rations of transpacific fishes; figures he gave ranged from <1 to >6 months (for moray eels), with most ~2 months. The greatly in- creased availability of otolith-based data on pelagic larval durations (Table 9) allows us to TABLE 9 Pelagic Larval Durations of Transpacific Fishes, Members of Genera of Nontranspacifics Present at the Line Islands, and Species of Shore Fishes Endemic to the TEP Group Family Species Range in Pelagic Larval Duration (days) Transpacifics at AnguilHdae Anguilla marrmrata 145-160 Line Islands Chanidae Chanos chatios 14-29 MugiKdae Mugil cephalus 29-67 Aulostomidae Aulostomus chinensis 71-106 Chaetodontidae Chaetodon auriga 40-53 Chaetodon kleinii 56 Forcipiger flavissimus 41-57 Cirrhitidae Cirrhitichthys oxycephalus 36-45 (mear i39) Labridae Novaculichthys taeniourus 44-74 Iniistius pavo 42-77 Stethojulis bandanensis 22-44 (mear i29) Thalassoma purpureum 50-62 Scombridae Scomber australasicus and S. japonicus 21-28 Acanthuridae Acanthurus triostegus triostegus 44-60 Naso brevirostris 80-120 Naso hexacanthus 82-111 Tetraodontidae Canthigaster valmtini 64-113 [found west of the Line Is.] Range in maximum PLD Nontranspacifics Genus Any species [n] Species at Line Islands (Indo-centtal Pacific) [with data/total present] Serranidae Epinephelus 39-50 [4] 39-41 [2/15] Apogonidae Apogon <30 [11] - [-/13] Lutjanidae Lutjanus 25-62 [5] 62 [1/7] Caesionidae Caesio 41 [1] - [-/2] Pterocaesio 32-47 [2] - [-/2] Lethrinidae Lethrinus 30-37 [6] 30 [1/4] MuUidae Eupenem 37 [1] - [-/I] Mulloidichthys 65 [1] 65 [1/3] Parupeneus 44-50 [3] 44-50 [3/5] Chaetodontidae Chaetodon 20-90 [17] 30-54 [5/22] Henioclms 44 [1] 44 [1/1] Pomacanthidae Centropyge 27-39 [11] 28-34 [3/5] Pomacanthus <30 [6] <30 [1/1] Pomacentridae Ahudefduf, Chrysiptera, Dascyllus, Lepidozygous, Pomacentnts <30 [4,7,4,1,17] <30 [1/4,2/2,1/3,1/1,2/3] Chromis 24-46 [17] 24-46 [5/9] Plectroglyphidodon 28-33 [4] 28-33 [2/5] Stegastes 23-35 [4] 24-32 [3/5] Labridae Anampses 28-52 [2] 38 [1/4] Bodianus 23-78 [4] <30 [1/5] Cheilinus 26-60 [8] 26-42 [3/4] Cirrhilabrus <30 [1] - [-/I] Coris 25-54 [5] 52-53 [2/3] Cymolutes 91 [1] 91 [1/1] Epibulus 38 [1] 38 [1/1] Gomphosus 72 [1] 72 [1/1] Halichoeres 22-56 [14] 22-56 [3/8] Hemigymnm 29-34 [2] 34 [1/1] Labroides 24-38 [4] 24-26 [2/3] Macropharyngodon 30-43 [3] 36 [1/1] 540 PACIFIC SCIENCE ? October 2004 TABLE 9 (continued) Range ir 1 Maximum PLD Species at Line Islands Group Family Genus Any species [B] [with data/total present] Pseudocheilinus 46- -72 [4] 46-55 [2/2] Pseudofuloides 52 [1] 52 [1/1] Thalassoma 55- -121 [8] 55-99 [5/7] Scaridae Chlorurus 40 [1] 40 [1/3] Scams 45 [1] - [-/12] Blenniidae Petroscirtes <30 [1] - [-/I] Gobiidae Amblygobius 41 [1] - [-/3] Gohiodon 22- -40 [3] - [-/I] Paragobiodon 36- -47 [3] - [-/2] Acandiuridae Acanthurus 34- -64 [4] 34-64 [2/17] Ctenochaetm 60 [2] 60 [2/5] Naso 94 [1] 94 [1/4] Zehrasoma 62 [1? 62 [1/3] TEP endemics Mugilidae Mugil 31 I [data for 1 of/5 species present in TEP] Serranidae Paralahrax <30 [1/8] Apogonidae Apogon 21-49 [2/6] Lutjanidae Hoplopagrus Lutjan-m <30 [1/1] 24-44 [4/9] Gerreidae Eucinostomtis 33 [1/4] Haemulidae Haemulon <30 [1/5] Pomacentridae Abadefdaf Chromis Nexilosiis Microspathodon Stegastes <30 [3/3] 20-43 [3/3] 33 [1/1] 30-34 [2/2] 25-48 [8/8] Labridae Bodianus Halichoeres Thalassoma Xyrichtys 32-48 [2/2] 29-80 [9/11] 59-104 [4/4] 59-131 [5/6] Tripterygiidae Axoclinus <30 [1/6] Labrisomidae Malacoctenus <30 [2/9] Blenniidae Ophioblennim 58 [1/1] Gobiidae Coryphopterus Bathygobiiis 51 [1/1] 39 [1/3] Sources: Arai et al. (2002); Bagarinao (1994); Bowen et al. (2001); Chang et al. (2000); Chen (1999); Doherty et al. (1995); Fowler (1989); Graves (1998); Leis (1989); McCormick (1994, 1999; unpubl. data on Mulhidichthys flavolineatus); C. Mora, F. Zapata, and A. Ospina. (unpubl. data on Cirrhitichthys oxycephalus and eastern Pacific species o? Apogon, Malacoctenus, Haemulon, Eucinostomus, Mugil, ChroTnis, Coryphopterus, and Bathygobius); Planes (1993, 1998 [sources of pelagic larval duration values given in these papers are not indicated in many cases], 2002); Planes et al. (1998, 2002); Riginos and Victor (2001); Robinet et al. (2003); Stroud et al. (1989); S. Swearer and D.R.R. (unpubl. data on Ophioblennius steindachner); Thresher and Brothers (1985); Thresher et al. (1989; Victor 1986, 1987); Victor and Wellington (2000); Victor et al. (2001); Wellington and Robertson (2001); Wellington and Victor (1989, 1992); G. Wilson (2001); Wilson and McCormick (1999); Zapata and Herron (2002). examine the relationship between pelagic lar- val duration variation and representation in the transpacific fatina. In 13 of 18 transpacific representatives of 10 families maximum pe- lagic larval durations exceed 50 days, and, in all but five cases, minima vary upward from 40 days. The exceptions are the millcfish Chanos chanos, the hawkfish Cirrhitichthys oxy- cephalus, the mackerels Scomber australasicus and S. japonicus, and the labrid Stethojulis ban- danensis. The millcfish (pelagic larval duration 2-4 weeks) is widely distributed throughout the TEP and the rest of the tropical and subtropical Indo-Pacific. Although juvenile settlers develop in estuarine and other shel- tered inshore habitats (Bagarinao 1994), adults occur in extremely isolated situations where conditions likely are lacking for setders (e.g., oceanic islands such as Easter and Clip- perton). Adult milkfish are large (to 1.8 m), Tropical Transpacific Shore Fishes ? Robertson et al. 541 fast swimming, pelagic fish, and adult dis- persal could compensate for limited larval dispersal capabilities. However, larger juve- niles of this species also associate with flotsam (J.E.M., pers. obs.), which could also facili- tate long-distance dispersal. Adults of the two Scomber species (pelagic larval durations of 3- 4 weeks) are nearshore pelagics that could well have accomplished the transpacific mi- gration. Cirrhitkhthys oxycephalus and S. ban- danensis, in contrast, are small (to 10 cm and 15 cm, respectively), demersal reef fishes. Al- though C. oxycephalus, and other members of its genus, occasionally associate with flotsam (Randall 1997; C. Mora, pers. comm., 2003), such behavior is not known for S. bandanensis. Both species are resident in and widely dis- tributed in the TEP. Data on S. bandanensis from three studies (Victor 1986, Victor and Wellington 2000, Victor et al. 2001) that in- cluded individuals from both the western Pa- cific and TEP showed a maximum pelagic larval duration of 44 days and an average of 29 days; the equivalent figures for C. oxy- cephalus are 45 and 39 days, respectively (C. Mora, pers. comm., 2003). The maximum pelagic larval duration of these two species is a little less than the minimum estimated time to transit the EPB on the NECC during an El Ni?o (~50 days). If we use the pelagic larval duration of S. bandanensis as the bench- mark for the minimum pelagic larval duration needed for larvae of a demersal species to transit the EPB and establish a population in the TEP, then data in Table 9 indicate that many species and genera present in the Line Islands have not realized a pelagic larval du- ration capacity to migrate to and establish populations in the TEP. These include lutja- nids, caesionids, mullids, chaetodontids, vari- ous genera of labrids, as well as acanthurids. On the other hand, some pelagic spawners (e.g., lethrinids, pomacanthids) as well as demersal spawners (e.g., apogonids and most pomacentrids) lack pelagic larval durations as long as that benchmark. The exception among the pomacentrids may be Chromis, which can have relatively long pelagic larval durations and which Clarke (1995) collected ~ 1,000 km east of the western side of the EPB. There are insufficient data on the pe- lagic larval durations of other speciose fami- lies in the Line Islands and central Pacific (serranids, scarids [with two species of trans- pacifies], blenniids, and gobiids) although some are close to the benchmark. The large goby fauna of the western side of the EPB includes one genus (Gnatholepis) whose Ca- ribbean member has a maximum pelagic lar- val duration of at least 120 days (Sponaugle and Cowen 1994), so some likely have pelagic larval durations exceeding the benchmark. Thus while short pelagic larval durations can account for the lack of transpacific species in some taxa, particularly among the demersal spawners, many other taxa lacking trans- pacifies have or likely have pelagic larval du- rations that greatly exceed the benchmark minimum. Variation in the representation in the transpacific fauna of five families that have a leptocephalus larva, which typically has a very long pelagic duration (elopids, albulids, muraenids, congrids, and ophichthids), pro- vides the strongest case in point: moray eels have more transpacific species than any other family of demersal fishes (Table 1), but there are no transpacifics in any of the other four families. This is particularly notable for ophichthids, which have substantial numbers of species and genera on both sides of the EPB (Table 2). Large differences in rates of endemism among Hawaiian ophichthids and muraenids (22.7 versus 8.9% [B?hlke and McCosker 2001, McCosker 2002]) suggest that ophichthids in general have more limited dispersal potential than muraenids. Indirect estimates of the larval durations of muraenids (based on seasonal change in abundance of larvae at different stages of development) range from 3 to 10 months (Leis 1984, B?hlke 1989, B?hlke and McCosker 2001). Unfortunately there are no comparable esti- mates for ophichthids. Estimates of pelagic larval durations (using the same methodol- ogy) of West Atlantic congrids and morin- guids from genera that occur in the TEP range from 3 to 12 months (B?hlke 1989), which are congruent with otolith-based de- terminations of a European Conger species (Correial et al. 2003). Estimates of the pelagic larval duration of western Atlantic Elops are 2- 3 months, and those for members of the Al- bula "vulpes" circumtropical species complex range from 2 to 24 months (B?hlke 1989, 542 PACIFIC SCIENCE ? October 2004 Colborn et al. 2001), including 6-7 months for one TEP species (Pfeiler et al. 1988). Yet genetic data indicate that there have been no connections between TEP and central Pacific members of the A. "vulpes" complex for about 10 myr (Colborn et al. 2001). D. G. Smith (pers. comm., 2003) noted that although adult populations of many marine eels have precise distributional boundaries, their larvae range well beyond those boundaries. Eor example, Clarke (1995) collected muraenid and con- grid larvae ~800 km east of the Line Islands on the western side of the EPB. Thus factors other than pelagic larval duration limitation clearly are involved in producing these dif- ferences in abundances of transpacific species among groups that share a pelagic larval form that characteristically has a very long pelagic duration. Among the group of 22 likely westward migrants, 17 are elasmobranchs and/or have pelagic adults. The demersal remainder are members of families (Carapidae, Priacanthi- dae, Balistidae, and Diodontidae) known to have large, specialized pelagic larval stages and, in most cases, pelagic juvenile stages (Maride and Olney 1990, Leis and Carson- Ewart 2000). There are pelagic larval dura- tion data for a limited range of nontrans- pacific TEP shore fishes (Table 9). Pelagic larval durations of TEP species are not shorter than those of congeners from the central Pacific in the two families (Labridae and Pomacentridae) where that has been examined (Mora et al. 2003). In a few cases such pelagic larval durations reach >100 days. However, because westward transits of the EPB may take twice as long as eastward tran- sits, pelagic larval duration limitations proba- bly have much stronger effects on westward migration. There is evidence (see section on Genetic Studies later in Discussion) that the Hawaiian population of Mugil cephalus could have been derived by westward migration from the TEP. Although this species has a pelagic larval duration of 29-67 days (Table 9), its adults reach 1 m in length, are fast swimming and semipelagic, and may have ac- complished a transit to Hawai'i. Since Leis' (1984) report there have been only two studies involving sampling for shore- fish larvae within the EPB but near its edges rather than in its center. Victor (1987) col- lected larvae of three genera of labrids that were >80 days old between the Gal?pagos and mainland Ecuador. Clarke (1995) sam- pled shore-fish larvae on the western side of the EPB and collected larvae of 14 fami- lies 750-1,100 km from the nearest shore- fish habitat: muraenids, congrids, holocen- trids, scorpaenids, acropomatids, carangids, lethrinids, muUids, serranids, labrids, poma- centrids {Chromis), acanthurids, callionymids, and ostraciids. These two studies support the idea that a range of reef fish taxa that lack transpacific representatives have long- distance dispersal capabilities. As Clarke (1995) noted, net avoidance is likely to result in underestimates of the diversity and abun- dance of large fish larvae and pelagic juveniles far from their adult habitat. Flotsam-mediated dispersal of shore fishes. Dispersal of large juveniles and adults of shore fishes across the EPB in association with flotsam likely is involved in some cases. Various known or potential westward mi- grants have been found with flotsam in the TEP: Lobotes pacificus, Carangoides caballus, Se- ri?la rivoliana, Uraspis helvola, Sectator ocyurus, Euthynnus lineatus, and Balistes polylepis (see Himter and Mitchell 1966). Other proba- ble eastward migrants have also been found with durable flotsam, including Chanos cha?as (J.E.M., pers. obs), Fistularia commtersonii and Cirrhitichthys oxycephalus (C. Mora, pers. comm., 2003), Sphyraena barracuda (Gooding and Magnuson 1967), and Caranx sexfasciatus (Hunter and MitcheU 1967). Other trans- pacific species with unclear direction(s) of migration that associate with durable flotsam include Elagatis bipinnulata (Gooding and Magnuson 1967) and Aluterus monocerus and A. scriptus (Hunter and Alitchell 1966). Ages of demersal stages of TEP species associated with flotsam can reach over a year (Hypso- blennius [Mora et al. 2001]), which should make them capable of a westward transit of the EPB. Because long transit times would strongly limit westward larval dispersal, flot- sam dispersal may be proportionately much more important for westward migration than for eastward migration. Tropical Transpacific Shore Fishes ? Robertson et al. 543 POTENTIAL ECOLOGICAL LIMITATIONS ON IMMIGRANTS. Availability of equivalent habitats on both sides of the EPB. Access to equivalent habitats aifects the colonization ability of organisms moving in either direc- tion across the EPB (Rosenblatt et al. 1972, Scheltema 1988). This includes conditions for larvae as well as adults. Leis (1986) noted that taxa whose larvae typically occur in conti- nental habitats on the Australian Great Bar- rier Reef are less well represented on the (central) Pacific Plate than those whose larvae occur in oceanic habitats on the Great Barrier Reef. There are strong diiferences between the fish faunas of continental and outer bar- rier (oceanic) reefs of the Great Barrier Reef in terms of both species composition and the abundances of fishes in different feeding groups (Williams 1991). Because those dif- ferences occur across tens of kilometers, they are much more likely to be related to inter- specific differences in adult and/or larval re- quirements than to variation in pelagic larval durations and larval dispersal capabilities. Leis (1986) recognized the difficulty of distinguishing between eifects of adult versus larval habitat requirements in limiting faunal exchange between continental and oceanic reefs. Adult-habitat limitations relating to the paucity of coral reefs in the TEP (cf. Rosen- blatt et al. 1972) can reasonably be invoked to explain lack of establishment of a few eastward migrants that reach the TEP as va- grants. First, central-Pacific Chaetodon species that feed on live corals regularly disperse to the TEP but lack established populations. Second, if scarids are strongly tied to a food base provided by carbonate substrata on coral reefs (cf. Streelman et al. 2002), this could explain why the TEP and the other tropical Pacific region that has few coral reefs (the Marquesas) have small parrotfish faunas (see Robertson 1998). However, there are many noncorallivorous Chaetodon species in the Line Islands (Table 2), and there is no evi- dence that the vast majority of "coral-reef" fish species are directiy dependent on live corals or coral reefs. Overall similarities in taxonomic structure of the reef-fish faunas of the TEP and the Caribbean (which has twice as many coral species and a ~ 1,000 times greater area of coral reefs as the TEP) indi- cate that a lack of coral reefs in the TEP is not likely to be a major limiting factor on successful eastward migration for most taxa. The occurrence of vagrants of various species of Naso and various noncorallivorous species of Chaetodon at diff?rent locations in the TEP does suggest that their ability to establish is limited more by suitability of conditions for self-sustaining populations in the TEP than by their ability to disperse across the EPB. What those conditions might be is far from obvious. Large fluctuations in the abundance of the westward-migrant carangid Carangoides caballus in Hawai'i, where it increased greatiy following tiie 1997-1998 El Ni?o, dien de- clined sharply, clearly demonstrate environ- mental limitations on successful invasion of Hawai'i in addition to any effects El Ni?os might have on its rate of immigration. If this species experiences a similar increase at Ha- wai'i during a future El Ni?o, then assessing the reproductive state of adults there and comparing their growth rates (using otolith analyses) with those occurring in the TEP could help clarify why it fails to establish in Hawai'i. Another Hawaiian example of possible island-environment limitations on establishment of a continental TEP species derives from the fate of three species of non- native lutjanids introduced in large num- bers in the 1950s. Two Indo-central Pacific reef-dwelling species successfully established substantial resident populations, but a con- tinental TEP species that lives mainly on soft bottoms did not (Oda and Parrish 1981). The biggest difference between the central Pacific and the TEP in terms of habitat availability is in the abundance of equivalent offshore habitats. The pelagic and benthic environments of the Line Islands and the oceanic islands of the TEP share physical and biological characteristics that are very differ- ent from those along the mainland of the TEP (Longhurst 1998). Oceanic reefs pro- vide 100% of habitats for shore fishes in the equatorial part of the central Pacific, and all potential immigrants from there to the TEP are adapted to life in that type of environment. In contrast, in the TEP the 544 PACIFIC SCIENCE ? October 2004 amount of oceanic-island habitat is very small in both absolute and relative terms, where such islands provide only ~2.5% of shallow, inshore habitat (Table 5). Further, ~97.5% of the oceanic-island habitat in the TEP is situated at the cooler latitudinal limits of the region: the Revillagigedos in the north and the Gal?pagos in the south (Table 5). The largest (southern and western) islands in the Gal?pagos have a marine environment that is warm temperate rather than equatorial (Glynn and Wellington 1983, Grove and Lavenberg 1997). Further, the second-largest area of oceanic-reef habitat (the subtropical Revillagigedos) have a reduced ability to re- ceive eastward migrants because they lie well to the north of the flow path of the NFCC, in an area where westbound currents predomi- nate. Consequently the amount of tropical oceanic-reef habitat likely not only to receive immigrants on the NFCC but also to sup- port continuing populations of those immi- grants is miniscule: it consists of the northern Gal?pagos, Cocos, Clipperton, and Malpelo, which together provide ~0.5% of the shallow inshore habitat in the TFP. Some suitable habitat for oceanic-island species occurs at a moderate-sized cluster of coastal islands in western Panama (see Glynn et al. 1972, Rosenblatt et al. 1972) and at few small similar islands on the coast of Costa Rica. However, inclusion of those islands only marginally increases the total amount of reef habitat suitable for resident populations of central Pacific immigrants. That the TFP's oceanic islands are disproportionately impor- tant (relative to their number and size) as supporters of transpacific shore fishes is con- sistent with their requiring equivalent habi- tat to that on the western side of the EPB. However, the offshore islands of the TFP also are substantially closer to the western edge of the FPB (by ~ 1,000 km in the cases of the Gal?pagos and Clipperton) and more directly influenced by eastbound currents than is the continental coast of that region and particularly the continental islands at Panama and Costa Rica. Hence greater ac- cessibility of the oflihore islands may also aSect the extent to which they support trans- pacific fishes. The paucity of tropical oceanic-reef habi- tat in the TFP can also be expected to limit westward emigration in two ways. First, it reduces the pool of TFP species adapted to ocean-island life and thus capable of estab- lishing populations in the central Pacific: only 99 TFP species of shore fishes (11.4% of the regional fauna) are ocean-island endemics and <30% of the remainder of the region's endemic shore fishes probably have self- sustaining insular populations (Robertson and Allen 1996, 2002). Second, the smaU size of the TFP's tropical oceanic islands must be reflected in small adult populations of fishes and relatively small numbers of emi- grant propagules being released. How differences in oceanic-reef and continental-reef environments aifect the faunal composition of reef-fish assemblages throughout the Indo-Pacific and what differ- ent adaptations are required of members of the fish fauna of each environment are ques- tions of major biogeographic significance that need to be addressed. Explanations of lat- itudinal gradients in diversity across the Indo- Pacific have emphasized the importance of dispersal ability and the geologic history of the Indo-Philippine center of diversity (e.g., Hughes et al. 2002, Mora et al. 2003). Yet although the West Pacific contains an abun- dance of both continental and oceanic reefs, the vast majority of the reef habitat in the central Pacific and central Indian Oceans, and particularly the habitat that can act as stepping-stones for dispersal, consists of oce- anic reefs. Interactions between the indigenous fauna and potential immigrants? Fco-evolutionary factors that are thought to affect the level and pattern of invasion of a fauna by immigrants include (1) the relative size of the recipient fauna, (2) the extent to which it has become depau- perated by mass-extinction events, and (3) whether invaders have biological character- istics that provide them with an advantage over members of the native fauna. Vermeij (1978, 1987, 1991) and Vermeij and Rosen- berg (1993) examined the pattern of trans- pacific invasion of the TFP moUuscan fauna from this perspective. They attributed a pre- ponderance of eastward migration of mol- Tropical Transpacific Shore Fishes ? Robertson et al. 545 lusks across the EPB to higher levels of prior extinction in the TEP and the possession by the western invaders of ecological character- istics possessed by extinct members of the TEP but lacking in the contemporary native fauna. Lacking a recent fossil record for trop- ical fishes in the TEP, West Atlantic, and central Pacific, we cannot assess the relative level of any extinction in the TEP versus those other regions or whether the TEP fish fauna has become depauperate. Although the reef-fish fauna of the TEP is much smaller than that of the area used as the global stan- dard (the Indo-Malayan center of diversity [e.g., see Hughes et al. 2002, Mora et al. 2003]), it is similar in size to the fauna of its sister region, the tropical Northwest Atlantic (e.g., Robertson 1998, Robertson and Allen 2002, Smith 1997). Those two regions shared a common history and marine biota until the closure of the isthmus of Panama, and their shore-fish faunas share a substantial number of Neotropical endemic genera and families (Rosenblatt 1967, McCosker and Robertson 2001). To provide an indication of which fami- lies in the indigenous TEP shore-fish fauna might be functionally depauperate we com- pared the relative abundance of each in the faunas of the TEP and the greater Carib- bean, and assumed that families with pro- portionately fewer (or no) species in the TEP are depauperate. To assess whether eastward migration is more common in those depau- perate families we examined the occurrence of successful (i.e., resident) eastward-migrant transpacific species in each family present in the TEP, the Caribbean, and the Line Islands (the most likely source of such migrants). These data (Table 2) show that the propor- tion of "depauperate" TEP families that have resident transpacific species does not differ from the proportion of nondepauperate fam- ilies that do: 0.54 of 50 versus 0.47 of 15, re- spectively (G test: G = 0.33, P > 0.05). It is also possible that immigration occurs more often in families that have relatively fewer species in the TEP than in the central Pa- cific. In that comparison eastward-immigrant transpacific species are present in 0.55 of 49 depauperate families and 0.31 of 16 non- depauperate families, a nonsignificant differ- ence (G test: G = 2.71, P > 0.05). Vermeij and Rosenberg (1993) found that in the Caribbean, where the moUusk fauna was subject to a mass extinction 2-5 Ma, recent immigrants are more common and widespread than in other nearby regions (including the TEP) that experienced fewer extinctions. There are 37 species of eastward- migrant shore fishes that are widely distri- buted residents in the TEP and 20 species of vagrants (Table 1). If interactions with indig- enous species determine whether a trans- pacific species establishes a population in the TEP and how widespread it becomes, then the proportion of widely distributed resident transpacific species that are members of de- pauperate families should be greater than the proportion of vagrants that are members of such families. Because there is virtually no difference in those proportions (0.76 and 0.75, respectively), this is not the case. To assess whether interactions with the indigenous TEP shore-fish fauna may have affected the success of eastward invasions we also compared the ecological composition (habitat usage and feeding groups) of the eastward-migrant transpacific fauna with the composition of the donor (Line Islands) and recipient (native TEP) faunas. As previously noted, oceanic and inshore pelagics are both overrepresented in the transpacific fauna rel- ative to their abundances in the donor and recipient faunas. Among the demersal species, the relative abundances of reef fishes and soft- bottom fishes are almost identical in the transpacific and Line Island faunas. In con- trast, soft-bottom species are much less abun- dant in the transpacific fauna than in the TEP fauna (Table 10: G test: G = 25.5, P < 0.01), in which they predominate over reef spe- cies. The relative abundances of species in eight feeding groups (Table 11) differs in the transpacific fauna from that in both the TEP and Line Islands faunas (G tests: G = 39.1, P < 0.01, and G = 16.8, P < 0.05, respec- tively). However, the overall structure of the transpacific fauna is more similar to that of the Line Islands fauna than the TEP fauna; Czekanowski similarity indices (Bloom 1981) are 0.78 and 0.66, respectively. These pat- 546 PACIFIC SCIENCE ? October 2004 TABLE 10 Percentage of DiiFerent Groups of Demersal Shore Fishes among Eastward-Migrant Transpacific Species and in the Faunas on Both Sides of the EPB TEP Line Indigenous Islands Habitat Type Transpacifics Fauna Fauna Demersal: 10.4 56.7 13.0 soft bottom" Demersal: reef* 89.6 43.3 87.0 No. of species 67 858 487 '^ Mud, sand, and gravel. * Rock, rubble, and coral. terns of dissimilarity between the structures of the transpacific and TEP faunas are what would be expected if transpacific species were being excluded from groups with many native species. However, the similarity of the trans- pacific and Line Islands faunas suggests a simpler explanation: that transpacific species are drawn largely at random, at least in terms of ecological groupings, from the Line Is- lands fauna, with differences between the transpacific faima and the faunas of both the TEP and the Line Islands arising from phy- logenetic diff?rences in dispersal potential. Greater dispersal potential leads to over- TABLE 11 Representation of Different Feeding Groups among Eastward-Migrant Transpacific Shore Fishes and the Shore-Fish Faunas on Each Side of the EPB Feeding TEP Line Group" Transpacifics Indigenous Islands Midwater 12.5 3.3 2.1 carnivores Demersal 44.5 66.8 60.9 carnivores Planktivores 9.4 22.5 13.8 Bentbic 11.9 2.7 7.4 herbivores Bentbic 15.0 4.1 9.3 omnivores Bentbic 2.5 0.8 3.9 detridvores Parasitic 1.3 0.2 1.2 Corallivores 1.9 0 1.4 No. of species 80 949 515 '^ Species in two feeding groups (see Table 1) were assigned balf value for each. weighting of the following groups in the transpacific fauna: pelagic species among the inshore fishes, and among the trophic groups: midwater carnivores (all but one are pelagic or semipelagic species) and most of the ben- tbic herbivores and omnivores (which are members of taxa with high dispersal capa- bilities: acanthurids, chaetodontids, and tetraodontiforms). We conclude that there are no clear in- dications that interactions with the native fauna have played a role in determining what taxa or ecological groups of eastward- immigrant fishes have established resident populations in the TEP. In only a few cases do eastward-migrant shore fishes have nota- ble ecological characteristics lacking in the native TEP fauna: the TEP includes no in- digenous members of the herbivorous sur- geonfish genera Acanthurus or Ctenochaetus (both are abundant in the Indo-central Pa- cific and there are three species o? Acanthu- rus in the Caribbean) and only one native member of the scraping/excavating herbivore/ detritivore genus Scarus, which is abundant in both the Indo-central Pacific and North- west Atlantic. In addition there are no native corallivores in the TEP, but two among the transpacific species. However, together those species that have characteristics lacking in the native TEP faima represent only 15.5% of the resident eastward-migrant transpacific species in the TEP. Genetic Studies and the Status of TEP Populations of Transpacific Taxa The overall similarities of the shore-fish faunas of the Caribbean and the TEP and the occurrence of numerous transisthmian gemi- nate species pairs suggest that TEP popula- tions of many circumtropical species may be isthmian relicts. The alternative to that scenario is an eastward range expansion and immigration to the TEP, perhaps during pe- riods with different climate conditions (cf. Bowen and Grant 1997). Because fossils of reef fishes are rare and not known from the Caribbean and eastern Pacific, a major source of historical information bearing on rela- tionships within the coral and molluscan fau- Tropical Transpacific Shore Fishes ? Robertson et al. 547 ?as of the New World is lacking for shore fishes. Hence data on genetic relationships of Neotropical and central Pacific fish taxa are vital for establishing how the TEP shore- fish fauna was formed. Such data should help answer questions about the origin of TEP populations of transpacific fishes: Were TEP populations separated from those of putative West Atlantic sister taxa before or during the rise of the Isthmus of Panama? Are TEP and central Pacific populations of transpacific species currently in genetic contact? What is the relative importance of eastward versus westward migration across the EPB? To what extent have widely distributed Indo-Pacific species originated in the TEP and spread westward? Are the Line Islands both receiv- ing westward migrants from and providing eastward migrants to the TEP (see discussion in the conclusions on directions of migration across the EPB)? There are data available relevant to these questions for a range of transpacific oceanic and shore fishes. OCEANIC FISHES. IsuTus oxyriiichus: Heist et al. (1996) examined relations between Pa- cific and Atlantic populations (including those in the Northeast Pacific) using mtDNA They found "no evidence of multiple sub- species ... nor of past genetic isolation be- tween ... populations." Coryphaena hippurus: Rosenblatt and Waples (1986) found virtually no difference (in allozymes) between pop- ulations in the TEP, Hawai'i, and the Atlan- tic, indicating gene flow across the EPB, as well as between the Pacific and the Atlantic. Thunnus alalunga: Graves and Dizon (1989: mtDNA) found no differences that could dis- tinguish populations at South Africa and the Northeast Pacific and concluded that they are either connected or only recently isolated. Thunnus albacares: Ward et al. (1997: allo- zymes and mtDNA) examined global popula- tion structure of this species. They found three subdivisions (Atlantic, Indian Ocean, and Pacific) with no evidence of separation between East and West Pacific populations. Xiphias gladius: Chow and Takeyama (2000: mtDNA and nDNA) compared populations from locations scattered around the world. They identified four groupings (Mediterra- nean, Northwest Atlantic, South Atlantic, and Indo-Pacific) but no subdivisions within the Indo-Pacific. Istiophorid billfishes: Graves and McDowell (1995: mtDNA) examined relationships between Atlantic and Pacific sister species of three istiophorids: blue mar- lin (Makaira nigricans and M. mazara), sailfish {Istiophorus albicans and 7. platypterus), and white and striped marlin (Tetrapterus albidus and T. audax). The TEP population was in- cluded among samples of sailfish and striped marlin. They concluded that all three Atlantic and Indo-Pacific pairs probably do not war- rant the status of separate species and that populations in the two oceans are connected by recent gene flow. Although sampling was insufficient to assess the status of the TEP population versus that of the rest of the Pa- cific, connections between the Indo-Pacific and Atlantic indicate that isolation of the TEP population is unlikely. SHORE FISHES. Manta birostris: Limited sampling of central Pacific, TEP, and Carib- bean populations of M. birostris by T. Clark (pers. comm., 2003: mtDNA) indicated that the TEP population is an isthmian relict that has not had contact with central Pacific pop- ulations since before the rise of the Isthmus of Panama. Sphyma lewini: Analysis of pop- ulations in the West Atlantic and the eastern, central, and West Pacific (K. Duncan, pers. comm., 2003: mtDNA) indicated that the TEP population is not an isolated isthmian relict and is genetically connected to cen- tral and western Pacific populations. Albula "vulpes": Colborn et al. (2001: mtDNA) assessed relationships among Indo-Pacific and Atlantic populations in the pantropical "vulpes" species complex. The TEP pre- viously was thought to have one representa- tive of this group, referred to as either A. vulpes (Eschmeyer et al. 1983, Whitehead and Rodriguez-Sanchez 1995, Grove and Laven- berg 1997) or A. neoguinaica (Allen and Rob- ertson 1994). Colborn et al. (2001) concluded that (1) there might be up to eight species in this complex; (2) A. forsteri (a senior synonym of A. neoguinaica according to Randall and Bauchot [1999]) is restricted to the West and central Pacific and A. vulpes to the Caribbean; and (3) the TEP lineage contains a pair of sister species. They also noted (4) that the 548 PACIFIC SCIENCE ? October 2004 TEP lineage separated from a lineage that contains two Atlantic and one Indo-central Pacific species ~10 Ma. This complexity means that it is not possible to determine whether the TEP lineage originated by an invasion from the central Pacific or from an Adanto-East Pacific stock. The TEP lin- eage was formed well before the Isthmus of Panama closed, and the two TEP species separated ~4 Ma (about when that closure occurred) and represent isthmian relicts. Gymnomuraena zebra: Rosenblatt and Waples (1986: allozymes) found virtually no diifer- ence between populations of this species in the TEP and Hawai'i, indicating re- cent gene flow. Sardinops sagax: Bowen and Grant's (1997) and Grant and Bowen's (1998: mtDNA) analysis led them to propose that this genus contains a single widespread Indo- Pacific species that spread throughout its current range only within the last few hun- dred thousand years. They cited paleonto- logical evidence indicating that the eastern Pacific endemic subspecies {S. s. sagax), which has an antitropical distribution that largely excludes the TEP, may have arrived in the northern part of its range less than 100,000 yr ago. Where the eastern Pacific population came from is unclear. Fistularia commersonii: Rosenblatt and Waples (1986:279: allozymes) found evidence of "relatively recent inter- ruption of previously significant levels of geneflow" between Hawai'i and the TEP. Tylosurus imperialis: Collette and Banford (2001: mtDNA) foimd htde differentiation between western and eastern Pacific pop- ulations of the Indo-Pacific subspecies (T. /'. melanotus) of this circumtropical, inshore pe- lagic species. The deepest split in the T. im- perialis lineage is between the Indo-Pacific and Atlantic, which was followed by separa- tion of T. pacificus (a TEP endemic) and T i. melanotus. The simplest explanation for this pattern is that the Indo-Pacific Hneage invaded the TEP twice after the Adantic lineage became isolated, and neither TEP population is an isthmian relict. Tylosurus crocodilus: Currently two subspecies of this circumtropical species are recognized: T. c. fodiator from the TEP and T. c. crocodilus from the Atlantic and Indo-central Pacific (to the Hawaiian, Line, and Marquesas Islands). Collette and Banford (2001: mtDNA) con- cluded that there are three allopatric lineages (Atlantic, TEP, and Indo-central Pacific) of T. crocodilus with an initial separation of line- ages between the New World and the Indo- central Pacific, followed by separation of the TEP and West Adantic lineages. This se- quence of events indicates that the TEP sub- species likely is an isthmian relict. Aulostomus chinensis: Bowen et al. (2001: mtDNA) found weak genetic structuring across the Indo- Pacific (East Africa to the TEP) and no sub- stantial separation across the EPB. These data (and see also Rosenblatt and Waples [1986]) indicate ongoing to recent gene flow across the EPB. Levels of genetic difference be- tween A. chinensis and the Caribbean species {A. maculatus) suggest that they were sepa- rated by the closure of the Isthmus of Pan- ama. The fact that genetic diversity in A. chinensis is low in the TEP and that TEP haplotypes are a subset of those found else- where in the Indo-Pacific indicate that the TEP population originated by recent eastward migration and remains in genetic contact with central Pacific populations. Hippocampus ingens: Teske et al.'s (2003: nu- clear and mtDNA) global phylogeny of the genus indicates that H. ingens, a TEP en- demic (and the only member of its genus in that region), is a member of an Indo-Pacific Hneage that invaded the Atiantic/eastern Pa- cific before the closure of the Central Ameri- can isthmus and that H. ingens is an isthmian reHct with a Caribbean geminate. Their data do not show how that lineage invaded the New World: from the central Pacific or from the western Indian Ocean around southern Africa. Heteropriacanthus cruentatus: Rosen- blatt and Waples (1986: allozymes) found virtually no difference between populations of this species in the TEP and Hawai'i (indicat- ing ongoing gene flow) but a large difference between Pacific and Adantic populations of what is currentiy regarded as a circumtropical species. Bermingham et al. (1997: mtDNA) reported a large divergence between TEP and Adantic populations of this species. These two data sets indicate that the TEP population is not an isolated isthmian relict Tropical Transpacific Shore Fishes ? Robertson et al. 549 and either was derived from an eastward mi- gration after the rise of the Isthmus of Pan- ama or was in the TEP before that event and has remained in genetic contact with the central Pacific population. Gnathanodon spe- ciosus: Rosenblatt and Waples (1986:279: al- lozymes) found evidence of "relatively recent interruption of previously significant levels of geneflow" between Hawai'i and the TEP. Trachurus murphyi: The eastern Pacific pop- ulations of this primarily warm-temperate fish have a disjunct, antitropical distribution and sometimes are referred to as separate species. The southern form is a transpacific that ex- tends westward to New Zealand. Stepien and Rosenblatt (1996: mtDNA) found indications of ongoing gene flow across the TEP and suggested that the two forms are conspecific. Mulloidichthys vankoknsis: The widespread Indo-central Pacific species M. vankoknsis is extremely similar in morphology to its TEP congener, M dentatus. Stepien et al. (1994: allozymes) examined relationships between populations of M. vankoknsis from five Indo- central Pacific sites (including Hawai'i) and a population of M. dentatus from the Gulf of California. They found M. dentatus to be ge- netically distinguishable from M. vanicoknsis. Bermingham et al. (1997) found very low mtDNA sequence divergence between M. dentatus and M. martinicus from the Carib- bean, indicating that they separated very recently. Relationships between the three species and the timing of their separations (before or after the rise of the Isthmus of Panama) remain unclear. Lessios, D.R.R., and Kessing (unpubl. mtDNA sequence data) compared presumed M. dentatus at Baja Cali- fornia, Clipperton, and Panama in the TEP with M. vanicoknsis at sites throughout the Indo-central Pacific, including Hawai'i, the Line Islands, and the Marquesas. They found that all three TEP populations contained se- quences found only in that region but also sequences commonly found in Indo-central Pacific populations of M vanicoknsis. These results indicate that although the TEP popu- lation was previously isolated, M. vanicoknsis recently invaded from the central Pacific. Given that it was found in all three pop- ulations sampled by those workers it likely is widespread in the TEP. Mugil cephalus: This circumtropical species has a disjunct distri- bution in the Pacific; the only central Pa- cific population is at Hawai'i. Crosetti et al. (1994: mtDNA) examined populations cover- ing much of its range. Their data indicate that the TEP population is most closely related to the Hawaiian population and that the sister population of those two is in the Atlan- tic and Mediterranean. This pattern is con- sistent with westward migration to Hawai'i by an isthmian-relict TEP population, al- though further data are needed to determine whether the ancestral population was in the TEP. Thalassoma spp.: Bernardi et al. (2003: mtDNA and nDN?) constructed a phylog- eny of this pantropical genus, which includes four endemics and one resident transpacific species, T. purpureum, in the TEP. The most ancient split occurred between the Atlantic and Indo-Pacific lineages, and the TEP endemics are much more recently derived from Indo-Pacific lineages. There are three lineages in the TEP: one containing the TEP sister endemics T. lucasanum and T. robert- soni plus the Indo-central Pacific T. am- blycephalum, one consisting of T. virens (a TEP endemic) and T. purpureum (perhaps the most widespread Indo-Pacific wrasse), and one that includes the TEP endemic T. grammaticum and the Indo-central Pacific T. lutescens. The entire TEP fauna evidently is derived from four separate invasions from the Indo-central Pacific (the T. purpureum lineage having invaded twice) and lacks any isthmian relicts. Acanthurus triostegus: Planes and Eau- velot (2002: allozymes) examined relationships between Pacific populations of this species. Planes (2002) proposed that the species orig- inated in Erench Polynesia and then spread east and west. The Clipperton population was most closely allied to (although distinct from) that of the Tuamotus among those they sam- pled, and the Marquesas and Hawaiian pop- ulations each were very distinct from all others. This pattern of genetic relationships parallels that of the three recognized subspe- cies of A. triostegus, which occur in Hawai'i, the Marquesas, and the remainder of the Indo-Pacific (Randall 1956). It also is con- sistent with the TEP being populated by 550 PACIFIC SCIENCE ? October 2004 migrants from the Line Islands. However, sampling of the continental TEP, the Line Islands, and islands to the west of the Line Islands is needed to clarify this situation and determine the extent of gene flow across the EPB. Zanclus comutus: Rosenblatt and Waples (1986: allozymes) found low levels of diifer- entiation between Hawaiian and TEP pop- ulations of this species, indicating ongoing or very recent gene flow. Scomber australasicus: This Indo-Pacific species occurs in the West Pacific, Hawai'i, and the Revillagigedo Is- lands in the TEP. Distributions of mtDNA haplotypes (Scoles et al. 1998) show that the Revillagigedos population has low genetic diversity and evidently was established by eastward immigration some time ago (it has unique haplotypes), but has experienced recent/ongoing eastward gene flow. Scomber japonicus: What was once regarded as a cir- cumglobal species is now restricted to the Indo-Pacific with an Atlantic sister species (Collette et al. 2001). Geographically inade- quate sampling precludes a clear assessment of the relationship of the eastern Pacific populations to those elsewhere in its range, although disjunct northern and southern populations in the eastern Pacific have been isolated from each other for several hundred thousand years (Stepien and Rosenblatt 1996: mtDNA). Aluterus scriptus: Rosenblatt and Waples' (1986:279: allozymes) data indicated "relatively recent interruption of previously significant levels of geneflow" between Ha- wai'i and the TEP. Bermingham et al. (1997: mtDNA) found very low divergence between populations in the TEP and the Caribbean. Together these data suggest that this is a cir- cumtropical species with gene flow through- out its range. Melichthys niger: Bermingham et al. (1997: mtDNA) found a very low level of divergence between TEP and Caribbean populations, indicating a circumtropical spe- cies with gene flow throughout its range. Arothron hispidus and Arothron m,eleagris: Rosenblatt and Waples (1986:279: allozymes) found virtually no diiferences between the TEP and Hawai'i populations of the former species, indicating ongoing gene flow, and indications of "relatively recent interrup- tion of previously significant levels of gene- flow" between such populations in the latter. Diodon holacanthus: Rosenblatt and Waples (1986:279: allozymes) found evidence of "relatively recent interruption of previously significant levels of geneflow" between pop- ulations in Hawai'i and the TEP. Diodon hys- trix: Using mtDNA sequences Bermingham et al. (1997) found extremely low divergence between TEP and Atlantic members of this species; this circumtropical species exhibits pantropical gene flow. PATTERNS OF HISTORICAL RELATION- SHIPS. In summary these genetic data have shown the following: (1) TEP populations of eight large, highly mobile oceanic species (in four different families) with circumglobal dis- tributions are connected to those elsewhere in the Pacific, and often the Atlantic, and do not represent isolated isthmian relicts. (2) Among the shore fishes there is good evidence of ongoing to recent genetic connections be- tween the TEP and the central Pacific in 10 Indo-Pacific species from nine families, be- tween the TEP and the western Pacific in three Indo-Pacific species, and between the TEP and the rest of the Indo-Pacific and the Atlantic in two circumtropical species. (3) Six TEP representatives of five circumtropical genera {Manta, Albula, Tylosurus, Mugil, Hip- pocampus) include isthmian relicts, and TEP populations in 10 species in five other such genera (Sardinops, Tylosurus, Aulostomus, Mul- loidichthys, Thalassoma, Scomber) apparently are the result of eastward migration since the closure of the isthmus. The latter include species whose TEP populations range from having been isolated for a substantial amount of time, to having intermittent or ongoing contact with populations on the western side of the EPB. (4) In one genus {Thalassoma), multiple eastward invasions after the rise of the isthmus by several different Indo-Pacific lineages established the entire TEP fauna of four endemic and one transpacific species. (5) An isolated central Pacific population of a circumglobal species {Mugil cephalus) likely was established by westward migration from an isthmian-relict population in the TEP. (6) Indo-central Pacific species can co-occur with endemic siblings in the TEP (e.g., Mullodich- thys). (7) Species with disjunct, antitropical Tropical Transpacific Shore Fishes ? Robertson et al. 551 distributions may have less gene flow between TEP populations and those elsewhere than do species with continuous tropical distribu- tions (cf. Graves 1998). Northern and south- ern populations of antitropical species in the eastern Pacific range from being well con- nected to well isolated. (8) Genetic data are consistent with distributional data indicating a strong preponderance of eastward migra- tion across the EPB but more westward mi- gration than previously thought. However, existing genetic data in many cases are very sketchy, with few samples per site and a lack of samples from substantial parts of the geo- graphic range. Hence there is a largely unre- alized potential for such data to resolve many questions about the extent, timing, and routes of migration in both directions across that barrier and to assess historical patterns in- dicated by current geographic distributions. In particular, such data could determine the extent to which transpacific fishes that are widespread in the Indo-Pacific, including circumtropical species and many of those treated here as eastward migrants, originated by westward migration from the TEP. OTHER TAXA. Bcsidcs shorc fishes, there are genetic data on the status of TEP repre- sentatives of circumtropical taxa in only one other group of marine shore organisms: the echinoids. Data on five pantropical genera of sea urchins, which, geographically, are much more comprehensive than those for most of the fishes, show ongoing large-scale connections across the EPB in two genera represented in the TEP only by transpacifics (Lessios et al. 1998, 2003). In two genera, isthmian-relict species co-occur with resident eastward-migrant transpacific species in the TEP (Lessios et al. 1996, 2001). In one other genus the TEP fauna consists of a pair of endemics derived from an isthmian relict (Lessios et al. 1999). Thus the range of rela- tionships in this small set of taxa parallels that among the shore fishes. Conclusions: The Level and Pattern of Migration across the EPB The extreme width of the EPB and the slow- ness of currents that traverse it have led to the conclusion that migration is limited to the few species with the longest larval durations (Scheltema 1988). However, additional infor- mation reinforces previous conclusions that many central Pacific shore-fish taxa that have pelagic larval durations that are adequate for an EPB transit and larval characteristics likely to facilitate long-distance dispersal are not represented among successful eastward mi- grants. Although data on pelagic larval dura- tions are still lacking for many central Pacific taxa, there is enough to indicate that variation in that measure of dispersal capability does not explain some conspicuous absences of speciose taxa in the transpacific fauna. Ab- sences of certain higher taxa whose members likely have adequate pelagic larval durations probably are due to phylogenetic variation in some other aspects of larval life histories, but chance may well aifect which species in a particular taxon make their way across the EPB in sufficient numbers to establish suc- cessful immigrant populations. Transit times for westward migration are likely to be on the order of twice as long as for eastward migra- tion; hence the pool of capable species must be much smaller for westward than for east- ward migration. Flotsam migration across the EPB, which has been recognized for corals (Jokiel 1984, Grigg and Hey 1992), seems more likely to occur in a westward direction because the eastern source of flotsam (a con- tinental shore spanning ~20? of latitude) is huge in comparison with the western source (small islands scattered over a vast area of ocean). Modeling of flotsam movements in the TEP indicate that most westward flotsam dispersal would occur on the NEC (Garcia et al. 1999), hence to the Hawaiian Islands (in- cluding Johnston Island). However, few spe- cies whose adults are demersal are known to associate with flotsam, which severely limits the potential importance of such dispersal. Elotsam dispersal may be important for pe- lagic shore fishes because many westward migrants have pelagic adults that are known to associate with flotsam. Large diflFerences in the availability of comparable habitat on the two sides of the EPB may be a major factor limiting overall levels of migration?the only shore-fish hab- 552 PACIFIC SCIENCE ? October 2004 itat present at the head of the eastbound currents on the western side of the barrier (tropical oceanic reefs) is rare in the TEP. The scarcity of that habitat in the TEP and the small fauna (number of species and in- dividuals) it supports there also reduces the potential for westward migration. Conditions on each side of the EPB evidendy limit the establishment of some immigrant shore fishes, which have appeared repeatedly as vagrants in a region or undergo large fluctu- ations in abundance there. However, other than efl^ects of the scarcity of corals in the TEP on a few reef fishes it is unclear how such limitations might operate. Distributional and existing genetic data support the prevailing view that eastward mi- gration across the EPB is the predominant mode but indicate that westward migration has been substantially underestimated, with the ratio of eastward to westward move- ments probably being on the order of 3:1. A bias toward eastward migration previously has been attributed to: (1) eastward currents being stronger than westward flows; (2) the TEP having a depauperate fauna with few species capable of long-distance dispersal (the reverse of the situation in the Indo-central Pacific); and (3) adults of Indo-central Pacific species having broader ecological capabilities than adults of TEP endemics (summarized in Grigg and Hey [1992]). Although eastward migrants exhibit a range of dispersal charac- teristics, westward migrants comprise elas- mobranchs, species with pelagic adults, and species with specialized larvae capable of extended dispersal. Such a difference in the composition of the eastward and westward migrant faunas is consistent with the idea that westward migration is more diflicult than eastward migration. However, our analyses indicate that effective differences between eastbound and westbound currents may not be as great as previously thought. The TEP shore-fish fauna is only about one-third the size of that in the Indo-Malayan center of Indo-Pacific diversity and is substantially smaller than the shore-fish fauna of the entire central Pacific. However, because the reef- associated fauna of the TEP (~550 species [Robertson and Allen 2002]) is about the same size as that of any of the three main groups of islands on the western side of the EPB, there is no real support for the notion that a bias toward eastward migration reflects emigration from a richer to a poorer fauna. There are no indications that interactions with the indigenous TEP fauna have shaped taxonomic or ?cologie biases in the structure of the eastward-immigrant fauna. Rather the ?cologie structure of that fauna simply re- flects the structure of the donor fauna, tem- pered by effects of taxonomic variation in dispersal ability. Although long-distance dis- persal capabilities may be important to de- termining the taxonomic structure of the central Pacific fauna (e.g., Mora et al. 2003), currendy there are no indications that TEP fishes have lower dispersal capabilities than central Pacific members of the same family. The TEP has one of the most dynamic envi- ronments of any tropical region in the world, a much more dynamic environment than that of its sister region, the tropical Northwest Adantic. The native TEP coral fauna was largely wiped out by environmental stresses of various types, especially those associated with El Ni?os, leading to large differences between the coral faunas and abundance of coral reefs in those two regions. However, differences between the shore-fish faunas of those two regions are much smaller, and there is no evidence of a strong, widespread impact of those stresses on the TEP's shore- fish fauna. Hence, there is no reason to as- sume that limited ecological capabilities of TEP species might limit westward migration of fishes even if it does so for invertebrates. Many more westward migrants are known from the Hawaiian Islands than from the Marquesas; in fact all 23 supposed westward migrants are found at Hawai'i and only five at the Marquesas. Several factors likely con- tribute to this bias. First, the Marquesas rep- resent a much smaller target than the Hawaiian Islands. Second, even though west- ward flow toward the Marquesas is faster than that toward Hawai'i, the Marquesas may ef- fectively be more isolated than Hawai'i by an eastward current that intermittentiy impinges on them. Third, although the westward cur- rent flowing toward Hawai'i has a tropical Tropical Transpacific Shore Fishes ? Robertson et al. 553 origin, the current carrying any propagules to the Marquesas originates in a cool part of the eastern Pacific with a largely temperate fauna. Fourth, flotsam migration seems more likely to occur toward Hawai'i (and Johnston Island) than to the Marquesas. However, a sampling bias may also contribute to this dif- ference: with a long history of ichthyological studies and the large population of fishers and scuba divers likely to notice unusual fishes, the Hawaiian fauna is very well known. In contrast the Marquesas are isolated, with a small human population including very few people equipped to identify transpacific spe- cies, little recreational scuba activity, and a very short history of research on its fish fauna with substantial parts of the archipelago yet to be visited by professional ichthyologists. Previous analyses relied on distributional data to indicate the likely direction of migra- tion across the EPB by transpacific fishes. Recent genetic studies have revealed part of the history of range expansion among wide- spread transpacific taxa and indicate substan- tial variation in how the EPB has aifected the composition of the TEP shore-fish fauna. However, most such studies lack sufficient geographic scope and analytical sensitivity to provide a general picture of the level of bias in the direction of historical and current mi- gration across the EPB. ACKNOWLEDGMENTS We are indebted to Gerald R. Allen, Richard L. Pyle, John E. Randall, John L. Earle, Bruce C. Mundy, and Bruce B. Collette for sharing their information on the fish fauna of various parts of Oceania, including the Ha- waiian, Line, and Marquesas Islands. Robert J. Lavenberg contributed to an early draft of this paper. Thanks to Giacomo Bernardi, Brian W. Bowen, John L. Earle, Haris A. Lessios, Bruce C. Mundy, and John E. Ran- dall for reviewing the manuscript; to Bruce B. Collette, John E. Olney, and William F. Smith-Vaniz for comments on the status of several transpacific taxa; and to J. Howard Choat for information on pelagic larval dura- tions. Ernesto Pe?a prepared the data for Table 5. Literature Cited Allen, G. R. 1985. An annotated and illus- trated catalogue of the lutjanid species known to date. EAO Eish. Synop. No. 125, Vol. 6. Allen, G. R., andj. E. Randall. 1977. Review of the sharpnose puifers (subfamily Can- thigasterinae) in the Indo-Pacific. Rec. Aust. Mus. 30:475-517. Allen, G. R., and D. R. Robertson. 1994. Eishes of the tropical eastern Pacific. Uni- versity of Hawai'i Press, Honolulu. . 1996. An annotated checklist of the fishes of Clipperton Atoll, tropical eastern Pacific. Rev. Biol. Trop. 45:813-843. -. 2002. Halkhoeres salmofasciatus, a new species of wrasse (Labridae) from Isla del Coco, tropical eastern Pacific. Aqua 5:65- 72. Arai, T., M. Marui, M. J. Miller, and K. Tsukamoto. 2002. Growth history and inshore migration of the tropical eel An- guilla marmorata in the western Pacific. Mar. Biol. (Berl.) 140:309-316. Bagarinao, T. 1994. Systematics, distribution, genetics and life history of milkfish, Chanos chanos. Environ. Biol. Eishes 39:23-41. Bearez, P. 1996. Lista de los peces marinos del Ecuador continental. Rev. Biol. Trop. 44:731-741. Benoit, H. P., P. P?pin, and J. A. Brown. 2000. Patterns of metamorphic age and length in marine fishes, from individuals to taxa. Can. J. Eish. Aquat. Sei. 57:856-869. Bermingham, E., S. McCaiferty, and A. P. Martin. 1997. Eish biogeography and mo- lecular clocks: Perspectives from the Pan- amanian isthmus. Pages 113-128 in T. D. Kocher and C. A. Stepien, eds. Molecular systematics of fishes. Academic Press, New York. Bernardi, G., G. Bucciarelli, D. Costagliola, D. R. Robertson, and J. B. Heiser. 2003. Ecology and evolution of the coral reef fish genus Thalassoma (Labridae): 1. Molecular phylogeography and biogeography. Mar. Biol. (Berl.) 144:369-375. Bloom, S. A. 1981. Similarity indices in community studies: Potential pitfalls. Mar. Ecol. Prog. Ser. 5:125-128. 554 PACIFIC SCIENCE ? October 2004 B?hlke, E. B., ed. 1989. Fishes of the western North Atlantic: Leptocephali. Mem. Sears Found. Mar. Res. No. 1, Pt. 9, Vol. 2. Yale University Press, New Haven, Connecti- cut. B?hlke, E. B., and J. E. McCosker. 2001. The moray eels of Australia and New Zealand, with the description of two new species (Anguilliformes: Muraenidae). Rec. Aust. Mus. 53:71-102. B?hlke, E. B., and J. F. Randall. 2000. A re- view of the moray eels (Anguilliformes: Muraenidae) of the Hawaiian Islands, with descriptions of two new species. Proc. Acad. Nat. Sei. Phila. 150:203-278. Bowen, B. W., and W. S. Grant. 1997. Phy- logeography of the sardines (Sardinops spp.): Assessing biogeographic models and population histories in temperate upwell- ing zones. Evolution 51:1601-1610. Bowen, B. W., A. L. Bass, L. A. Rocha, W. S. Grant, and D. R. Robertson. 2001. Phylo- geography of the trumpetfishes {Aulostomus spp.): Ring species complex on a global scale. Evolution 55:1029-1039. Briggs, J. C. 1960. Fishes of worldwide (circumtropical) distribution. Copeia 1960:171-180. . 1961. The eastern Pacific barrier and the distribution of marine shore fishes. Evolution 15:545-554. . 1964. Additional transpacific shore fishes. Copeia 1964:706-708. . 1974. Marine zoogeography. McGraw-Hill, New York. . 1995. Global biogeography. Elsevier, New York. 1999?. Coincident biogeographic patterns: Indo-West Pacific Ocean. Evo- lution 53:326-335. -. 1999^". Modes of speciation: Marine Indo-West Pacific. Bull. Mar. Sei. 65:645- 656. . 2003. Marine centers of origin as evolutionary engines. J. Biogeogr. 30:1-18. Brogan, M. W. 1994. Distribution and re- tention of larval fishes near reefs in the Gulf of California. Mar. Ecol. Prog. Ser. 115:1-13. Chambers, R. C. 1997. Environmental influ- ences on egg and propagule sizes in marine fishes. Pages 63-102 in R. C. Chambers and E. A. Trippell, eds. Early life history and recruitment offish populations. Chap- man and Hall, London. Chang, C. W., W. N. Tzeng, and Y. C. Lee. 2000. Recruitment and hatching dates of grey mullet {Mugil cephalm L.) juveniles in the Tanshui estuary of Northwest Taiwan. Zool. Stud. 39:99-106. Chavez, F. P., and R. L. Smitii. 1995. Bio- logical and chemical consequences of open ocean upwelling. Pages 149-170 in C. P. Summerhays, K. C. Emeis, M. V. Angel, R. L. Smith, and B. Zeitzschel, eds. Up- welling in the ocean: Modern processes and ancient records. Wiley, New York. Chavez, E. P., P. G. Strutton, G. E. Friede- rich, R. A. Feely, G. C. Eeldman, D. G. Foley, and M. J. McPhaden. 1999. Bio- logical and chemical response of the equatorial Pacific Ocean to the 1997-98 El Ni?o. Science (Washington, D.C.) 286:2126-2131. Chavez, F. P., J. Ryan, S. E. Lluch-Cota, and M. C. Niquen. 2003. From anchovies to sardines and back: Multidecadal change in the Pacific Ocean. Science (Washington, D.C.) 299:217-221. Chen, L. 1999. Ontogenetic development in post-setdement scarids (Pisces: Scaridae): Transition in feeding modes. Ph.D. thesis, James Cook University, Townsville, Aus- tralia. Chow, S., and H. Takeyama. 2000. Nuclear and mitochondrial DNA analyses reveal four genetically separated breeding imits of tiie swordfish. J. Fish Biol. 56:1087- 1098. Clarke, T. A. 1995. Larvae of nearshore fishes in oceanic waters of the central equatorial Pacific. Pac. Sei. 49:134-142. Coates, A. G., and J. A. Obando. 1996. The geologic evolution of the Central Ameri- can isthmus. Pages 21-56 in']. B. C.Jack- son, A. E. Budd, and A. G. Coates, eds. Evolution and environments in tropical America. University of Chicago Press, Chicago. Colborn, J., R. E. Crabtree, J. B. Shaklee, E. Pfeiler, and B. W. Bowen. 2001. The evolutionary enigma of bonefishes {Albula Tropical Transpacific Shore Fishes ? Robertson et al. sss spp.): Cryptic species and ancient separa- tions in a globally distributed shorefish. Evolution 55:807-820. Collette, B. B. 1995. Belonidae. Pages 919- 926 in W. Fischer, F. Krupp, W. Schneider, C. Sommer, K. F. Carpenter, and V. H. Niem, eds. Guia FAO para la identificaci?n de especies para los fines de la pesca. Pacifico Centro-Oriental. Vol. 2. FAO, Rome. Collette, B. B., and C. R. Aadland. 1996. Re- vision of the fi-igate tunas (Scombridae, Auxis), with descriptions of two new sub- species from the eastern Pacific. Fish. Bull. 94:423-441. Collette, B. B., and H. Banford. 2001. Status of the eastern Pacific aguj?n needlefish Ty- losurus pacificus (Steindachner, 1876). Rev. Biol. Trop. Suppl. 1:51-57. Collette, B. B., and C. F. Nauen. 1983. FAO species catalogue. Vol. 2. Scombrids of the world. FAO, Rome. Collette, B. B., C. Reeb, and B. A. Block. 2001. Systematics of the tunas and mack- erels. Fish Physiol. (Tuna) 19:1-33. Compagno, L. J. V. 1984?. FAO species catalogue. Vol. 4. Part I. Sharks of the world: Hexanchiformes to Lamniformes. FAO, Rome. . 1984?. FAO species catalogue. Vol. 4. Part II. Sharks of the world: Carcharhi- niformes. FAO, Rome. Compagno, L. J. V., and S. F. Cook. 1995. The exploitation and conservation of fresh- water elasmobranchs: Status of taxa and prospects for the future. J. Aquatic. Aquat. Sei. 7:62-90. Correial, A. T., C. Antunes, E. J. Isidro, and J. Coimbra. 2003. Changes in otolith mi- crostructure and microchemistry during larval development of the European con- ger eel {Conger conger). Mar. Biol. (Berl.) 140:165-173. Crosetti, D., W. S. Neilson, and J. C. Avise. 1994. Pronounced genetic structure of mi- tochondrial DNA among populations of the circumglobally distributed grey mullet {Mugil cephalus). J. Fish Biol. 44:47-58. Darwin, C. 1872. The origin of species by means of natural selection. 6th ed. Ran- dom House, New York. Dawson, C. F. 1985. Indo-Pacific pipefishes (Red Sea to the Americas). Gulf Coast Re- search Laboratory, Ocean Springs, Missis- sippi. De Sylva, D. P., and F. Williams. 1986. Sphyraenidae. Pages 721-726 in M. M. Smith and P. C. Heemstra, eds. Smith's sea fishes. Springer, New York. Doherty, P. J., S. Planes, and P. Mather. 1995. Gene flow and larval duration in seven species of fish from the Great Bar- rier Reef Ecology 76:2373-2391. Ekman, S. 1953. Zoogeography of the sea. Sidgwick and Jackson, London. Eldin, G. 1983. Eastward flows of the south equatorial central Pacific. J. Phys. Ocean- ogr. 13:1461-1467. Emerson, W. K. 1994. A Zoogeographie summary of the marine molluscs of Clip- perton Island (tropical eastern Pacific Ocean). Festivus 26:62-71. Eschmeyer, W. N., F. S. Herald, and H. Hammann. 1983. A field guide to Pa- cific coast fishes of North America, from the Gulf of Alaska to Baja California. Houghton and Mifflin, Boston. Fauchald, K. 1977. Polychaetes from inter- tidal areas in Panama, with a review of previous shallow-water records. Smithson. Contrib.Zool. 221:1-75. Fedoryako, B. I. 1979. Triggerfishes of the genera Canthidermis and Xanthichthys (Ba- listidae, Tetraodontiformes) from the Indian and Pacific Oceans. J. Ichthyol. 19:19-30. Firing, F., and R. Lukas. 1983. El Ni?o at the equator and at 159? W. Trop. Ocean- Atmos. Newsl. 21:9-11. Firing, F., R. Lukas, J. Sadler, and K. Wyrtki. 1983. Equatorial undercurrent disappears during 1982-83 El Ni?o. Science (Wash- ington, D.C.) 222:1121-1123. Fischer, W., F. Krupp, W. Schneider, C. Sommer, K. F. Carpenter, and V. H. Niem, eds. 1995. Guia FAO para la iden- tificaci?n de especies para los fines de la pesca. Pacifico Centro-Oriental. Vols 1-3. FAO, Rome. Fowler, A. J. 1989. Description, interpre- tation and use of the microstructure of otoliths from juvenile butterflyfishes (fam- 556 PACIFIC SCIENCE ? October 2004 ily Chaetodontidae). Mar. Biol. (Berl.) 102:167-181. Fowler, H. W. 1928. The fishes of Oceania. Mem. Bernice P. Bishop Mus. 10:1-540. Pritsche, R. A. 1976. A review of the cornet- fishes, genus Fistuhria (Fistulariidae), with a discussion of intrageneric relationships and zoogeography. Bull. Mar. Sei. 26:196- 204. Garcia, M., M. Hall, A. Pares-Sierra, and P. Arenas. 1999. Simulated trajectories of floating objects entering the eastern tropi- cal Pacific Ocean. Inter-Am. Trop. Tuna Comm. Spec. Rep. 11:346-395. Garrison, G. 2000. Peces de la Isla del Coco (Isla del Coco Fishes). InBio, Herederia, Costa Rica. Garth, J. S. 1974. On the occurrence in the eastern tropical Pacific of Indo-West Pa- cific decapod crustacean commensal with reef-building corals. Proc. 2nd Int. Coral ReefSymp. 1:397-404. Glynn, P. W., and J. S. Auk. 2000. A bio- geographic analysis and review of the far eastern Pacific coral reef region. Coral Reefs 19:1-23. Glynn, P. W., and G. R. Wellington. 1983. Corals and coral reefs of the Gal?pagos Islands. University of California Press, Berkeley. Glynn, P. W., R. H. Stewart, and J. F. McCosker. 1972. Pacific coral reefs of Panama: Structure, distribution, and pred- ators. Geol. Rundschau 61:483-519. Gooding, R. M., and J. J. Magnuson. 1967. Fcological significance of a drifting object to pelagic fishes. Pac. Sei. 21:486-497. Grant, W. S., and B. W. Bowen. 1998. Shal- low population histories in deep evolu- tionary lineages of marine fishes: Insights from the sardines and anchovies and lessons for conservation. J. Hered. 89:415-426. Graves, J. F. 1998. Molecular insights into the population structures of cosmopolitan marine fishes. J. Hered. 89:427-437. Graves, J. E., and A. F. Dizon. 1989. Mi- tochondrial DNA sequence similarity of Atlantic and Pacific albacore tuna {Thun- nus alalungd). Can. J. Fish. Aquat. Sei. 46:870-873. Graves, J. E., and J. R. McDowell. 1995. Inter-ocean genetic divergence of istio- phorid billfishes. Mar. Biol. (Berl.) 122:193-203. Grigg, R. W., and R. Hey. 1992. Paleo- ceanography of the tropical eastern Pa- cific Ocean. Science (Washington, D.C.) 255:172-178. Grove, J. S. 1984. Fl Ni?o 1982-1983 and new records of Indo-West Pacific fishes at the Galapagos. West. Soc. Malacol. Annu. Rep. 21:5. . 1986. Some evolutionary implica- tions of protracted Fl Ni?o events in the colonization and ecological balance offish populations in the Gal?pagos Archipelago. UNESCO Intergovernmental Oc?ano- graphie Commission, Workshop Report 49:77. Grove, J. S., and R. E. Lavenberg. 1997. Fishes of the Galapagos Islands. Stanford University Press, Stanford, California. Hastings, P. A. 2000. Biogeography of the tropical eastern Pacific: Distribution and phylogeny of chaenopsid fishes. Zool. J. Linn. Soc. 128:319-335. Heemstra, P. C. 1995. Lobotidae. Page 1226 in W. Fischer, F. Krupp, W. Schneider, C. Sommer, K. F. Carpenter, and V. H. Niem, eds. Guia FAO para la identi- ficaci?n de especies para los fines de la pesca. Pacifico Centro-Oriental. Vol. 3. FAO, Rome. Heist, F. J., J. A. Musick, and J. F. Graves. 1996. Genetic population structure of the shortfin mako ilsurus oxyrinchus) inferred from restriction fragment length poly- morphism analysis of mitochondrial DNA. Can. J. Fish. Aquat. Sei. 53:583-588. Hobson, E. S., and V. Walters. 1968. First eastern Pacific record of Canthigaster am- boinensis, an Indo-West Pacific puiferfish. Copeia 1968:861-862. Hubbs, C. L., and R. H. Rosenblatt. 1961. Fifects of the equatorial currents of the Pacific on the distribution of fishes and other animals. Pages 340-341 in Proc. 10th Pacif Sei. Congr. (abstract). Hughes, T. P., D. R. Bellwood, and S. R. Connolly. 2002. Biodiversity hotspots, cen- tres of endemicity and the conservation of coral reefs. Ecol. Lett. 5:775-784. Tropical Transpacific Shore Fishes ? Robertson et al. SSI Humann, P., and N. Deloach. 2003. Reef fish identification: Galapagos. 2nd ed. New World Publications, Jacksonville, Florida. Hunter, J. R., and C. T. Mitchell. 1966. As- sociation of fishes with flotsam in the oifshore waters of Central America. Fish. Bull. 66:13-29. Jokiel, P. L. 1984. Long-distance dispersal of reef corals by rafting. Coral Reefs 3:113- 116. Jones, G. P., M. L. Caley, and P. L. Mundy. 2002. Rarity in coral reef fish communi- ties. Pages 81-102 in P. F. Sale, ed. Coral reef fishes: Dynamics and diversity in a complex ecosystem. Academic Press, New York. Jordan, D. S. 1908. The law of geminate species. Am. Nat. 62:73-80. Kawahara, S., Y. Uozumi, and H. Yamada. 1988. First record of a carangid fish, Tra- churus murphyi from New Zealand. Jpn. J. Ichthyol. 35:212-214. Kokita, T. 2003. Potential latitudinal varia- tion in egg size and number of a geo- graphically widespread reef fish, revealed by common-environment experiments. Mar. Biol. (Berl.) 143:593-601. Lavenberg, R. J. 1992. A new moray eel (Muraenidae: Gymnothorax) from Oceanic islands of the South Pacific. Pac. Sei. 46:58-67. Leis, J. M. 1984. Larval dispersal and the Fast Pacific Barrier. Oceanogr. Trop. 19:181- 192. . 1986. Fcological requirements of Indo-Pacific larval fishes: A neglected Zoogeographie factor. Pages 759-766 in Proc. 2nd Int. Conf Indo-Pacific Fish. Ichthyological Society of Japan, Tokyo. -. 1989. Larval biology of butterfly- fishes (Pisces, Chaetodontidae): What do we really know. Environ. Biol. Fishes 25:87-100. -. 1991. The pelagic stage of reef fishes: The larval biology of coral reef fishes. Pa- ges 183-230 in P. F. Sale, ed. The ecology of coral reef fishes. Academic Press, New York. Leis, J. M., and B. M. Carson-Fwart, eds. 2000. The larvae of Indo-Pacific coastal fishes: An identification guide to marine fish larvae. Fauna Malesiana Handbooks. 2. Brill, Leiden. Leis, J. M., and M. McCormick. 2002. The biology, behavior and ecology of the pe- lagic, larval stage of coral reef fishes. Pages 171-200 in P. F. Sale, ed. Coral reef fishes: Dynamics and diversity in a complex eco- system. Academic Press, New York. Lessios, H. A., B. D. Kessing, G. M. Well- ington, and A. Graybeal. 1996. Indo- Pacific echinoids in the tropical eastern Pacific. Coral Reefs 15:133-142. Lessios, H. A., B. D. Kessing, and D. R. Robertson. 1998. Massive geneflow across the world's most potent marine biogeo- graphic barrier. Proc. R. Soc. Lond. B Biol. Sei. 265:583-588. Lessios, H. A., B. D. Kessing, D. R. Robert- son, and G. Paulay. 1999. Phylogeog- raphy of the pantropical sea urchin Eucidaris in relation to land barriers and ocean currents. Evolution 53:806-817. Lessios, H. A., B. D. Kessing, and J. S. Pearse. 2001. Population structure and speciation in tropical seas: Global phylo- geography of the sea urchin Diadema. Evolution 55:955-975. Lessios, H. A., J. Kane, and D. R. Robert- son. 2003. Phylogeography of the pan- tropical sea urchin Tripneustes: Contrasting patterns of population structure between oceans. Evolution 57:2026-2036. Longhurst, A. 1998. Fcological geography of the sea. Academic Press, New York. Markle, D. F., and J. E. Olney. 1990. Sys- tematics of the pearlfishes (Pisces: Cara- pidae). Bull. Mar. Sei. 47:269-410. Mayr, E. 1954. Geographic speciation in tropical echinoids. Evolution 8:1-18. McCormick, M. I. 1994. Variability in age and size at setdement of the tropical goatfish Upeneus tragula (MuUidae) in the northern Great Barrier Reef lagoon. Mar. Ecol. Prog. Ser. 103:1-15. . 1999. Delayed metamorphosis of a tropical reef fish {Acanthurus triostegus): A field experiment. Mar. Ecol. Prog. Ser. 176:25-38. McCormick, M. I., and B. W. Molony. 1995. Influence of water temperature during the larval stage on size, age and body condition 558 PACIFIC SCIENCE ? October 2004 of a tropical reef fish at settlement. Mar. Ecol. Prog. Ser. 118:59-68. McCosker, J. E. 1987. The fishes of the Gala- pagos Islands. Oceanus 30:28-32. . 1998. (Review of) Fishes of the Gal?pagos Islands, by J. S. Grove and R. J. Lavenberg, 1997, Stanford University Press. Copeia 1998:809-812. . 2002. Notes on Hawaiian snake eels (Pisces: Ophichthidae), with comments on Ophkhthus bonaparti. Pac. Sei. 56:23-34. McCosker, J. E., and P. H. Humann. 1996. New records of Galapagos fishes. Not. Galapagos 56:18-22. McCosker, J. E., and D. R. Robertson. 2001. Aplatophis zorro, a new species of eastern Pacific snake-eel, with comments on New World ophichthid distributions (Anguilli- formes: Ophichthidae). Rev. Biol. Trop. Suppl. 49:13-19. McCosker, J. E., and R. H. Rosenblatt. 1975. The moray eels (Pisces: Muraenidae) of the Galapagos Islands, with new records and synonomies of extralimital species. Proc. Calif Acad. Sei. Ser. 4, 40:417-427. . 1993. A revision of the snake eel ge- nus Myrichthys (Anguilliformes: Ophich- thidae) with the description of a new eastern Pacific species. Proc. Calif. Acad. Sei. 48:153-169. McCosker, J. E., L. R. Taylor, and R. R. Warner. 1978. Ichthyological studies at Gal?pagos. Not. Galapagos 17:13-15. McCosker, J. E., R. H. Bustamante, and G. M. Wellington. 2003. Ereshwater eels. Anguilla mamiorata, discovered at Gal?- pagos. Not. Galapagos 62:2-6. Meekan, M. G., J. H. Carleton, A. D. McKinnon, K. Elynn, and M. Eurnas. 2003. What determines the growth of tropical reef fish larvae in the plankton: Eood or temperature? Mar. Ecol. Prog. Ser. 256:193-204. Merlen, G. 1988. Two new butterflyfish rec- ords from Wenmen Island. Not. Gala- pagos 46:8-10. Mora, C, and A. E. Ospina. 2001. Tolerance to high temperatures and potential impact of sea warming on reef fishes of Gorgona Island (tropical eastern Pacific). Mar. Biol. (Berl.) 139:765-769. Mora, C, V. Francisco, and E. A. Zapata. 2001. Dispersal of juvenile and adult reef fishes associated with floating objects and their recruitment into Gorgona Island reefs, Colombia. Bull. Mar. Sei. 68:557- 561. Mora, C, P. M. Chittaro, P. E. Sale, J. P. Kritzer, and S. A. Ludsin. 2003. Patterns and processes in reef fish diversity. Nature (Lond.) 421:933-936. Mundy, B. C. In press. A checklist of the fishes of the Hawaiian Archipelago. Bishop Mus. Bull. Zool. Myers, R. E. 1999. Alicronesian reef fishes. 3rd ed. Coral Graphics, Guam. Nakamura, I. 1985. EAO species catalogue. Vol. 5. Billfishes of the world. EAO, Rome. Nilsson, G. E., and S. Ostlund-Nilsson. 2004. Hypoxia in paradise: Widespread hypoxia tolerance in coral reef fishes. Biol. Lett. 271:S30-S33. Oda, D. K., and J. D. Parrish. 1981. Ecology of commercial snappers and groupers in- troduced to Hawaiian reefs. Proc. 4th Int. Coral Reef Symp., Manila 1:59-67. Paulay, G. 1997. Diversity and distribution of reef organisms. Pages 298-353 in C. Birkeland, ed. The life and death of coral reefs. Chapman and Hall, New York. Pfeiler, E., M. A. Mendoza, and E. A. Man- rique. 1988. Premetamorphic bonefish (Al- bula sp.) leptocephaH from the Gulf of California with comments on life history. Environ. Biol. Eishes 21:241-249. Pietsch, T. W., and D. B. Grobeker. 1987. Erogfishes of the world: Systematics, zoo- geography and behavioral ecology. Stan- ford University Press, Stanford, California. Planes, S. 1993. Genetic diiferentiation in re- lation to restricted larval dispersal of the convict surgeonfish Acanthurus triostegus in French Polynesia. Mar. Ecol. Prog. Ser. 98:237-246. . 1998. Genetic diversity and dispersal capabilities in marine fish. Evol. Biol. 30:253-298. . 2002. Biogeography and larval dis- persal inferred from population genetic analysis. Pages 201-220 in P. E. Sale, ed. Coral reef fishes: Dynamics and diversity Tropical Transpacific Shore Fishes ? Robertson et al. 559 in a complex ecosystem. Academic Press, New York. Planes, S., and C. Fauvelot. 2002. Isolation by distance and vicariance drive genetic structure of a coral reef fish in the Pacific Ocean. Evolution 56:378-399. Planes, S., M. Parroni, and C. Chauvet. 1998. Evidence of limited gene flow in three species of coral reef fishes in the lagoon of New Caledonia. Mar. Biol. (Berl.) 130:361-368. Planes, S., G. Lecaillon, P. Lenfant, and M. Meekan. 2002. Genetic and demographic variation in new recruits of Naso unicomis. J.Eish Biol. 61:1033-1049. Pyle, R. L. 1999. Patterns of Pacific reef and shore fish biodiversity. Pages 157-175 in L. G. Eldredge, J. E. Maragos, P. E. Holthus, and H. E. Takeuchi, eds. Marine and coastal biodiversity in the tropical island Pacific region. Vol. 2. Population, development, and conservation priorities. Pacific Science Association, Honolulu. Randall, J. E. 1956. A revision of the surgeon- fish genus Acanthurus. Pac. Sei. 10:159- 235. . 1963. Review of the hawkfishes (family Cirrhitidae). Proc. U.S. Nati. Mus. 3472:389-451. . 1972. The Hawaiian trunkfishes of the genus Ostracion. Copeia 1972:767-768. . 1985. Eishes. Pages 462-481 in B. Delesalle, R. Galzin, and B. Salvat, eds. French Polynesian coral reefs. Vol. 1. Proc. 5th Int. Coral Reef Congress, Tahiti, 1985. Antenne Museum-EPHE, Moorea. . 1997. The hawkfish Cirritichthys ser- ratus Randall, a junior synonym of C. falco. Micronesica 30:199-203. . 2000. Revision of the Indo-Pacific labrid fishes of the genus StethojuUs, with descriptions of two new species. Indo-Pac. Fishes 31:1-42. Randall, J. E., and M. L. Bauchot. 1999. Clarification of the two Indo-Pacific spe- cies of bonefishes, Albula glossodonta and A forsten. Cybium 23:79-83. Randall, J. E., and B. A. Carlson. 1999. Ca- ranx caballus, a new immigrant carangid fish to the Hawaiian Islands from the tropical eastern Pacific. Pac. Sei. 53:357-360. Randall, J. E., and J. L. Earle. 2000. Anno- tated checklist of the shore fishes of the Marquesas Islands. Bishop Mus. Occas. Pap. 66:1-39. Randall,]. E., and J. E. McCosker. 1975. The eels of Easter Island with a description of a new moray. Contrib. Sei. (Los Angel.) 264:1-32. Randall,;. E., and B. C. Mundy. 1998. Batistes polylepis and Xanthichthys caeruleolineatus, two large triggerfishes (Tetraodonti- formes: Balistidae) from the Hawaiian Is- lands, with a key to the Hawaiian species. Pac. Sei. 52:322-333. Randall, J. E., and H. A. Randall. 2001. Re- view of the fishes of the genus Kuhlia (Perciformes: Kuhhidae) of the central Pacific. Pac. Sei. 55:227-256. Randall, J. E., K. Matsuura, and A. Zama. 1978. A revision of the triggerfish genus Xanthichthys, with a description of a new species. Bull. Mar. Sei. 28:688-706. Randall,;. E., G. R. Allen, and D. R. Robert- son. 2003. Myripristis earlei, a new soldier- fish (Beryciformes: Holocentridae) from the Marquesas and Phoenix Islands. Zool. Stud. 42:405-410. Richmond, R. H. 1990. The effects of El Ni?o/Southern Oscillation on the dis- persal of corals and other marine organ- isms. Pages 153-190 in P. W. G. Glynn, ed. Global ecological consequences of the 1982-83 El Ni?o Soudiern Oscillation. Elsevier, Amsterdam. Riginos, C, and B. C. Victor. 2001. Larval spatial distributions and other early life- history characteristics predict genetic dif- ferentiation in eastern Pacific blennioid fishes. Proc. R. Soc. Lond. B Biol. Sei. 268:1-6. Robertson, D. R. 1998. Do coral-reef fish faunas have a distinctive taxonomic struc- ture? Coral Reefs 17:179-186. Robertson, D. R., and G. R. Allen. 1996. Zoogeography of the shorefish fauna of Chpperton AtolL Coral Reefs 15:121-131. . 2002. Shorefishes of the tropical eastern Pacific: An information system. CD-ROM, Smithsonian Tropical Re- search Institute, Balboa. Robinet, T., R. Lecomte-Einiger, K. Escou- 560 PACIFIC SCIENCE ? October 2004 beyrou, and E. Feunteun. 2003. Tropical eels Anguilla spp. recruiting to R?union Island in the Indian Ocean: Taxonomy, patterns of recruitment and early life his- tories. Mar. Ecol. Prog. Ser. 259:263-272. Rosenblatt, R. H. 1967. The Zoogeographie relationships of the marine shore fishes of tropical America. Stud. Trop. Oceanogr. (Miami) 5:579-592. Rosenblatt, R. H., and B. W. Walker. 1963. The marine shore fishes of the Gala- pagos Islands. Occas. Pap. Calif. Acad. Sei. 44:97-106. Rosenblatt, R. H., and R. S. Waples. 1986. A genetic comparison of allopatric popula- tions of shore fish species from the eastern and central Pacific Ocean: Dispersal or vi- cariance? Copeia 1986:275-284. Rosenblatt, R. H., J. E. McCosker, and I. Rubinoif. 1972. Indo-West Pacific fishes in the Gulf of Chiriqui, Panama. Contrib. Sei. (Los Angel.) 234:1-18. Rubio, E. A., A. Suarez, E. Estupian, W. Henao, and B. Vargas. 1992. Los recursos ictiol?gicos de la Isla Malpelo (Colombia). I. Una revision de su conocimiento y nue- vos reportes para la ictiofauna de la Isla. Pages 642-658 in VIII Seminario Nacional de las Ciencia y las Tecnolog?as del Mar y Congreso Centroamericano y del Caribe en Ciencias del Mar. Memorias Tomo 2. Ruttenberg, B. I. 2000. An unusual pulse of recruitment of two coral reef fishes in the Galapagos Islands coincident with the 1997-98 El Ni?o. Bull. Mar. Sei. 67:869- 874. Scheltema, E. S. 1988. Initial evidence for the transport of teleplanic larvae of benthic invertebrates across the East Pacific bar- rier. Biol. Bull. (Woods Hole) 174:145- 152. Scoles, D. R., B. B. Collette, and J. E. Graves. 1998. Global phylogeography of the mackerels of the genus Scomber. Eish. Bulk 96:823-842. Shasky, D. R. 1985. Update on moUusks with Indo-Pacific faimal affinities in the tropical East Pacific. West. Soc. Malacol. Annu. Rep. 18:27. Smith, C. L. 1997. National Audubon guide to the tropical marine fishes of the Carib- bean, the Gulf of Mexico, Elorida, the Ba- hamas, and Bermuda. Knopf, New York. Smith-Vaniz, W. 1995. Carangidae. Pages 940-986 in W. Eischer, E. Krupp, W. Schneider, C. Sommer, K. E. Carpenter, and V. H. Niem, eds. Guia EAO para la identificaci?n de especies para los fines de la pesca. Pacifico Centro-Oriental. Vol. 2. EAO, Rome. Sponaugle, S., and R. K. Cowen. 1994. Larval duration and recruitment patterns of two Caribbean gobies (Gobiidae): Contrasting early life histories in demersal spawners. Mar. Biol. (Berk) 120:133-143. Springer, V. G. 1982. Pacific Plate biogeog- raphy with special reference to the shore fishes. Smithson. Contrib. Zool. 367:1- 182. Starnes, W. C. 1988. Revision, phylogeny and biogeographic comments on the cir- cumtropical marine percoid fish family Priacantiiidae. Bull. Mar. Sei. 43:117-203. Stepien, C. A., and R. H. Rosenblatt. 1996. Genetic divergence in antitropical marine fishes (Trachurus, Merluccius, and Scomber) between North and South America. Cope- ia 1996:586-598. Stepien, C. A., J. E. Randall, and R. H. Rosenblatt. 1994. Genetic and morpho- logical divergence of a circumtropical complex of goatfishes: Mulloidichthys vani- coknsis, M. dentatus, and M. martinicus. Pac. Sei. 48:44-56. Streelman, J. T., M. Alfaro, M. W. Wesmeat, D. R. Bellwood, and S. A. Karl. 2002. Evolutionary history of the parrotfishes: Biogeography, ecomorphology, and com- parative diversity. Evolution 56:961-971. Stroud, G., B. Goldman, and W. Gladstone. 1989. Larval development, growth and age determination in the sharpnose puffer Canthigaster valentini (Teleostei: Tetraodontidae). Jpn. J. Ichthyol. 36:327- 337. Teske, P. R., M. I. Cherry, and C. A. Matthee. 2003. The evolutionary history of seahorses (Syngnathidae: Hippocampus): Molecular data suggest a West Pacific origin and two invasions of the Atiantic Ocean. Mol. Phylogenet. Evol. In Press. Thompson, D. A., L. T. Eindley, and A. N. Tropical Transpacific Shore Fishes ? Robertson et al. 561 Kerstitch. 1979. Reef fishes of the Sea of Cortez. Wiley, New York. Thomson, J. M. 1997. The Mugilidae of the world. Mem. Queensl. Mus. 41:457-562. Thresher, R. E. 1991. Geographic variation in the ecology of coral reef fishes: Evi- dence, evolution and possible implica- tions. Pages 401-436 in P. F. Sale, ed. The ecology of coral reef fishes. Academic Press, New York. Thresher, R. E., and E. B. Brothers. 1985. Reproductive ecology and biogeography of Indo-West Pacific angelfishes (Pisces: Pomacanthidae). Evolution 39:878-887. . 1989. Evidence of intra- and inter- oceanic regional dilferences in the early life history of reef-associated fishes. Mar. Ecol. Prog. Ser. 57:187-205. Thresher, R. E., P. L. Colin, and L. J. Bell. 1989. Planktonic duration, distribution and population structure of western and cen- tral Pacific damselfishes (Pomacentridae). Copeia 1989:420-434. Vermeij, G. J. 1978. Biogeography and adap- tation: Patterns of marine life. Harvard University Press, Cambridge, Massachu- setts. . 1987. The dispersal barrier in the tropical Pacific: Implication for moUus- can speciation and extinction. Evolution 41:1046-1058. . 1991. When biotas meet: Under- standing biotic interchange. Science (Washington, D.C.) 253:1099-1104. Vermeij, G. J., and G. Rosenberg. 1993. Giving and receiving: The tropical Atiantic as donor and recipient region for invad- ing species. Am. Malacol. Bull. 10:181- 194. Victor, B. C. 1986. Duration of the plank- tonic larval stage of one himdred species of Pacific and Atiantic wrasses (Eamily Lab- ridae). Mar. Biol. (Berl.) 90:317-326. . 1987. Growth, dispersal, and identi- fication of planktonic labrid and pomacen- trid reef-fish larvae in the eastern Pacific Ocean. Mar. BioL (Berl.) 95:145-152. Victor, B. C, and G. M. Wellington. 2000. Endemism and the pelagic larval duration of reef fishes in the eastern Pacific Ocean. Mar. Ecol. Prog. Ser. 205:241-248. Victor, B. C, G. M. Wellington, D. R. Robertson, and B. Ruttenberg. 2001. The effects of the El Nino-Southern Oscilla- tion event on the distribution of reef asso- ciated labrid fishes in the eastern Pacific Ocean. Bull. Mar. Sei. 69:279-288. Ward, R. D., N. G. Elliott, B. H. Innes, A. J. Smolenski, and P. M. Grewe. 1997. Global population structure of yellowfin tuna, Thunnus albacares, inferred from al- lozyme and mitochondrial DNA variation. Eish. Bulk 95:566-575. Wellington, G. M., and D. R. Robertson. 2001. Variation in larval life-history traits among reef fishes across the Isthmus of Panama. Mar. Biok (Berl.) 138:11-22. WelHngton, G. M., and B. C. Victor. 1989. Planktonic larval duration of one hundred species of Pacific and Atiantic damsel- fishes (Pomacentridae). Mar. Biol. (Berl.) 101:557-567. . 1992. Regional dilferences in dura- tion of the planktonic larval stage of reef fishes in the eastern Pacific Ocean. Mar. Biol. (Berl.) 113:491-498. Whitehead, P. J. P. 1985. Clupeoid fishes of the world: An annotated and illus- trated catalogue of the herrings, sardines, pilchards, sprats, anchovies and wolf- herrings. EAO species catalogue. Vol. 7. Pts 1 and 2. EAO, Rome. Whitehead, P. J. P., and R. Rodriguez- Sanchez. 1995. Albulidae. Pages 851-853 in W. Fischer, E. Krupp, W. Schneider, C. Sommer, K. E. Carpenter, and V. H. Niem, eds. Guia EAO para la identi- ficaci?n de especies para los fines de la pesca. Pacifico Centro-Oriental. Vol. 2. EAO, Rome. Williams, D. McB. 1991. Patterns and pro- cesses in the distribution of coral reef fishes. Pages 437-474 in P. E. Sale, ed. The ecology of coral reef fishes. Academic Press, New York. Wilson, D. T., and M. I. McCormick. 1999. Microstructure of settiement-marks in the otoliths of tropical reef fishes. Mar. Biol. (Berl.) 134:29-41. Wilson, D. T., and M. G. Meekan. 2002. Growth-related advantages for survival to the point of replenishment in the coral 562 PACIFIC SCIENCE ? October 2004 reef fish Stegastes partitus (Pomacentridae). Mar. Ecol. Prog. Ser. 231:247-260. Wilson, G. 2001. Early post-settlement life history in Lethrinus species (Pisces: Leth- rinidae): Systematics, distribution and de- velopment. Ph.D. thesis, James Cook University, Townsville, Australia. Wyrtki, K. 1965. Surface currents of the eastern tropical Pacific Ocean. Inter-Am. Trop. Tuna Comm. Bull. 9:296-304. . 1967. Circulation and water masses in the eastern equatorial Pacific Ocean. Int. J. Oceanol. Limnol. 1:117-147. 1985. Sea level fluctuations in the Pacific during the 1982-83 El Ni?o. Geo- phys. Res. Lett. 12:125-128. Wyrtki, K., E. Firing, D. Halpern, R. Knox, G. J. McNally, W. C. Patzert, E. D. Stroup, B. A. Taft, and R. Williams. 1981. The Hawaii to Tahiti shuttle experiment. Science (Washington, D.C.) 211:22-28. Zapata, E. A., and P. A. Herron. 2002. Pelagic larval duration and geographic distribu- tion of tropical eastern Pacific snappers (Pisces: Lutjanidae). Mar. Ecol. Prog. Ser. 230:295-300. Appendix 1 Additional Information on Transpacific Fishes Listed in Table 1 Notorynchus cepedianus: occurs on the northwestern and southwestern edges of the Pacific but not the Pa- cific Plate. Carcharhinus alhiniarginatus: occurs in the northwestern Pacific and scattered central Pacific sites, including to the west of the Line Islands; possibly vagrant in Hawai'i. Carcharhinus brachyuriis: known from the northern and southern temperate parts of the eastern Pacific (extending into the fringes of the TEP) and the continental northwestern and southwestern Pacific. Carcharhinus galapagensis: found at scattered locations in the central (but not western) Pacific, including the Phoenix Islands to the west of the Line Islands, and southern French Polynesia. Carcharhinus lencas: occurs along the western edge of the Pacific, known in the South-central Pacific from a single record at the Tuamotus. Carcharhinus limbatus: found along the western edge of the Pacific but at few locations in the central Pacific, not including the Line Islands. Carcharhinus obscuras: unconfirmed re- ports from Hawai'i; no records from elsewhere on the Pacific Plate. Triaenodon obesus: at Cocos Island in the TEP it is ubiquitous and can readily be observed in aggregations of dozens of individuals; also reaches moderate densities on the coast of Panama and occurs as far north as the tip of Baja California and at the Revillagigedos Islands (E. Ochoa, pers. comin., 2003). Sphyma lewini: occurs along the western edge of the Pacific, in Oceania restricted to the easternmost is- lands: Hawai'i (as a resident), the Line Islands, and French Polynesia. Sphyma mokarran: occurs in the western Pacific, but in Oceania is known only from one sight record at Hawai'i by R. Pyle (pers. comm., 2000). Sphyma zygaena: occurs in the northwestern and southwestern corners of the Pacific; on the Pa- cific Plate only at the Hawaiian Islands. Odontaspis ferox: in the northwestern and southwestern Pacific; known in the central Pacific only from Hawai'i, where it probably is a resident (B. C. Mundy, pers. comm., 2003). Dasyatis dipterura: known outside the TEP from a single juvenile specimen collected in Hawai'i (J. E. Randall, pers. comm., 2002). Taeniura meyeni: in the TEP common at Cocos Island (Garrison 2000) and the Gal?pagos (Grove and Lavenberg 1997; J. Earle, pers. comm., 2003). Anguilla marmorata: McCosker et al. (2003) discussed a specimen of what likely is this species collected in the Gal?pagos in 1997, although there are indications of previous ar- rivals. McCosker et al. (2003) considered these to be vagrants that crossed the EPB. No other anguillids are known from the TEP. From a reproductive per- spective this species may effectively be disjunct be- cause the only known spawning area in the Pacific is in the western Pacific. Enchelychore lichenosa: known only from Japan/Taiwan and the Gal?pagos; two specimens collected at the Gal?pagos in 1964 and two in 1966 (McCosker and Rosenblatt 1975), three dif- ferent individuals photographed between 1990 and 1992 (Robertson and Allen 2002, Humann and Deloach 2003); not known elsewhere despite exten- sive collecting at intervening locations. Enchelynassa canina: in the TEP known only from Clipperton, where it is not uncommon, and western Panama. Gymnothorax buroensis: not uncommon in rotenone collections at Clipperton and at islands in the Gulf of Chiriqui, Panama; known in Hawai'i (where it is replaced by G. eurostus, see B?hlke and Randall [2000]) from a single specimen, perhaps of TEP ori- gin? Gymnothorax flavimarginatus: moderately com- mon at the oifshore islands of the TEP and in western Panama. Gymnothorax javanicus: one large adult photographed by D.R.R. at Cocos Island in 1997 (Robertson and Allen 2002, and see Garrison 2000), one by P. Humann in the Gal?pagos after 2000 (J.E.M., unpubl data), and two 1.5- to 2-m individ- uals observed by D.R.R. (and photographed by J. Earle) at Coiba Island, Panama, in May 2003. Gym.- nothorax meleagris: photographed by Humann and Deloach (2003) at the Gal?pagos; one adult observed at Cocos Island in 1997 byj. Earle (pers. comm., 1997). Gymnothorax pictus: not uncommon in the intertidal reef flat (its typical habitat) at CKpperton. Gymno- thorax undulatus: first recorded in western Panama in the early 1970s (Rosenblatt et al. 1972); often seen by D.R.R. diere in 1999 and 2003 and in the Pearl Is- lands (Gulf of Panama) in 2002; one individual seen in the Revillagigedos by A. Kerstitch (pers. comm., 1995); not currendy known from the Marquesas but Tropical Transpacific Shore Fishes ? Robertson et al. 563 present in the Society and Tuamotu Islands. Uropte- ryghis mpraforatus: several adults of various sizes col- lected by D.R.R. with rotenone at 5 m depth in the groove-and-spur zone of the Clipperton reef in April 1998. Etmmeus teres: in the Pacific there are dis- junct populations in the eastern Pacific (including the TEP), Hawai'i, Japan, and southeastern Australia. Whitehead (1985) treated these as a single species but noted that they could be separate. Encheliophis dubius: outside the TEP known only from Hawai'i, from a single confirmed record (B. C. Mundy and J. E. Randall, pers. comms., 2003). Encheliophis vermicularis: only found at the western and eastern fringes of the Pacific. Randall's (1985) record from the Society Islands appears to be an error (J. E. Randall, pers. comm., 2000). Brotula multiharhata: in the TEP known from two small juveniles (identified by J. Neilsen) collected by D.R.R. in a rotenone station at 40 m depth at Clipperton in April 1998. Antennarius coccineus: replaced in the Hawaiian Islands by its sister species, A. drombiis. Antennarius commerson: scattered records at widely distributed TEP sites over a quarter century (Pietsch and Grobeker 1987, Garrison 2000, Robertson and Allen 2002, Humann and Deloach 2003; A. Kerstitch, pers. comm., 1995); this large, cryptic, typically uncommon species occurs to depths of 70 m, and its apparent absence from the Line and Marquesas Islands may reflect inadequate sampling. Myripristis bemdti: abundant at the ofishore islands of the TEP and moderately common at islands in the Gulf of Chiriqui, western Panama. Pkctijpops lima: its absence from the Line Islands may be a sampling artifact because P. lima is a small, highly cryptic fish that typically is collected by poison from deep caves. Doryrhamphiis excisus excisus: two subspecies occur in the TEP, D. excisus paulus (endemic to the Revillagi- gedos Islands) and D. e. excisus, which occurs through- out the remainder of the TEP (Dawson 1985), where it is common (D.R.R., pers. obs.). Aulostomus chinensis: common at the oceanic islands of the TEP but rare on the mainland. Taenianotus triacanthus: known in the TEP from one specimen photographed in the Galapagos (McCosker 1987). Cookeolus japoniais: the lack of a Marquesan record may reflect a sampling bias, because it usually occurs in deep water (to 400 m). Priacanthus alalaua: known only from Hawai'i and, in the TEP, at the Revillagigedos Islands and Baja California. Carangoides caballus: abundant on the mainland of the TEP but uncommon at the oceanic islands. Caranx lugubris: in the TEP most common around the offshore islands. Caranx melam- pygus: in the TEP most common at the oceanic is- lands; also common at islands in western Panama. Caranx sexfasciatus: in the TEP is most common around the offshore islands and islands of western Panama. Decapterus macrosoma: not known from the Marquesas, but occurs in the Society Islands and westward from there. Decapterus muroadsi: besides the TEP, populations occur in the northwestern and southwestern Pacific and in the central Pacific only at Hawai'i. Elegatis bipinnulata: often found associated with flotsam far offshore. Gnathanodon speciosus: com- mon on most of the mainland of the TEP and found at all the offshore islands (although only as juveniles at the Gal?pagos [J.E.M., unpubl. data]) except Clip- perton. Seri?la lalandi: antitropical, rare in the Ha- waiian Islands; not known from Micronesia, the Line Islands, and Marquesas Islands; occurs in the southwestern to southeastern Pacific and Japan. Se- ri?la rivoliana: absent from Erench Polynesia and the Line Islands; present throughout Micronesia and the West Pacific; rare in Hawai'i. Trachurus murphyi: in the eastern Pacific extends from the southern fringe of the TEP to Chile, and between there and New Zealand (Kawahara et al. 1988). Uraspis helvola: tax- onomy of this genus needs revision; we follow Smith- Vaniz (1995), use the older of the two available names, and assume there is one circumtropical spe- cies; in the central Pacific known only from Hawai'i and the Marquesas Islands; not common in Hawai'i and records are sporadic (B. C. Mundy, pers. comm., 2001) and may be a vagrant there. Aphareus furca: first collected at Cocos Island in 1925 (AMNH lot 8554); adults of various sizes not uncommonly observed by D.R.R. there in 1997 (see also Garrison 2000); re- sembles and shares habitat with Paranthias colonus, which is abundant at Cocos; hence it could easily be missed in dense aggregations of that serranid and be common at Cocos. Pristipomoides zonatus: this wide- spread Indo-central Pacific species is known from a single specimen collected in the Gal?pagos (J.E.M., unpubl. data). Mulloidicthys vanicolensis: part of a pan- tropical species complex with four members that oc- curs throughout the Indo-central Pacific; genetic data (see section on Genetic Studies in Discussion) show that it co-occurs with the TEP endemic M. dentatus at Clipperton, Baja CaKfornia, and Panama (H. A. Les- sios, D.R.R., and B. Kessing, unpubl. data). Chaetodon auriga: the only TEP records are several specimens noted in the Gal?pagos between 1984 and 1986 (Grove 1984, 1986, Merlen 1988). Chaetodon kleinii: in the TEP known from a single individual photo- graphed in the Gal?pagos (Grove and Lavenberg 1997); does not occur in Erench Polynesia (where it is replaced by its sister species, C. tiHchrous). Chaetodon l?nula: known in the TEP from one fish observed at Cocos Island (Garrison 2000) and another at the Gal?pagos (Merlen 1988). Chaetodon meyeri: in the TEP single individuals have been seen at the Re- villagigedos (A. Kerstitch, pers. comm., 1995) and the Gal?pagos. Chaetodon unimaculatiis: the only TEP record is one individual photographed in the Gal?- pagos (Humann and Deloach 2003). Forcipiger fla- vissimus: notably common at Clipperton in the TEP, uncommon on most of the mainland (Allen and Robertson 1996; D.R.R., pers. obs.), except common at the tip of Baja California (J. Earle, pers. comm., 2003). Sectator ocyurus: common throughout the TEP; isolated individuals at Moorea, Hawai'i (repeatedly), and Japan; recently at Baker Island, ~500 km west of the Line Islands (E. DeMartini, pers. comm., 2000); common at the Marquesas (Randall and Earle 2000; D.R.R., pers. obs.); small groups of adults observed at Rangiroa in the Tuamotus in 2004 (J. Earle, pers. 564 PACIFIC SCIENCE ? October 2004 comm.). Kiihlia mugil: absent from the Hawaiian and Marquesas groups, each of which has an endemic congener (Randall and Randall 2001). Cin'hitichthys oxycephalus: abundant throughout the TEP. Oxy- cirrhitus typiis: no published records from French Polynesia but recendy photographed in the Mar- quesas (J. Earle, pers. comm., 2003). Mugil cephahis: circumglobal, subtropical to warm temperate, with populations in the western Pacific, the TEP, and, in the central Pacific, only at Hawai'i; known from Rapa in southeastern Polynesia (Randall 1985); several TEP populations synonymized with M. cephalus by Thom- son (1997). Iniistms pavo: common at western Panama and Cocos Island (D.R.R., pers. obs.); not known from the Line Islands but is widespread in the central Pacific and seems likely to be there. Novamlichthys taeniourus: common throughout the TEP (D.R.R., pers. obs.). Stethojulis bandanensis: common at several offshore islands in the TEP; in 1999 either arrived at or increased in abundance at the Galapagos (Victor et al. 2001); replaced in the Hawaiian and Marquesas Islands by endemic congeners (Randall 2000). Tha- lassoma purpureum: common at Clipperton in 1994 (Robertson and Allen 1996), less so elsewhere in the TEP; observed by D.R.R. at die Gal?pagos (1990) (and see also Humann and Deloach 2003), Cocos Is- land (1992 and 1997), and Isla Montuosa (western Panama) in 1998, where spawning was seen. Calo- tomus carolimis: common in the Revillagigedos Islands (D.R.R., pers. obs., 1992, 1994); not recorded in the Gal?pagos before 1977 (McCosker et al. 1978). Scarus ghobban: common on much of the mainland of the TEP. Scams nibroviolaceus: common throughout much of the TEP, on the mainland and offshore is- lands. Acanthiinis achules: known in the TEP from two pelagic juveniles, one collected near Clipperton and the other near the tip of Baja California. Acanthiirus nigricans: abundant at Clipperton, Cocos Island, and the Gal?pagos, less common on mainland reefs; rela- tively uncommon in the Hawaiian Islands (most common on the island of Hawai'i [D.R.R., pers. obs., 2003]). Acanthurus triostegus triostegus: there are three subspecies: A. t. sandvicensis from Hawai'i, A. t. mar- quesensis from the Marquesas, and A. t. triostegus from the remainder of its range, including the Line Islands and the TEP (Randall 1956); in die TEP is particu- larly abundant at the offshore islands and in western Panama. Acanthurus xanthopterus: common on the TEP mainland and some of the oceanic islands (e.g.. Cocos Island). Ctenochaetus niarginatus: abundant at several offshore islands in the TEP (D.R.R., pers. obs.). Naso annulatus: only two records for the TEP: Allen and Robertson (1996) observed one adult at Clipperton in 1994, as did D.R.R. at Cocos Island in November 1997. Naso brevirostris: McCosker and Humann (1996) recorded one individual in the Gal?pagos (see also Humann and Deloach 2003). Naso hexacanthus: Robertson and Allen (1996) observed a pair of adults at Clipperton in 1994, and D.R.R. ob- served a group of three same-sized adults at Cocos Island in 1997. Naso lituratus: known in the TEP only from about a dozen same-sized adults observed at Clipperton in 1994 (Robertson and Allen 1996) and 1998 (by D.R.R.). Naso vlamingii: McCosker and Humann (1996) recorded one individual at the northern Gal?pagos (see also Humann and Deloach 2003); one subadult was photographed in the Gal?- pagos in 1990 by R. Steene (see Robertson and Allen 2002). Sphyraena barracuda: known in the TEP from one specimen caught in western Panama and two caught in the Gal?pagos (see Robertson and Allen 2002; J.E.M., unpubl. data). We classify this as cir- cumtropical because the taxonomic stams of Indo- Pacific and Adantic populations (which differ in color and behavior [J. E. Randall, pers. comm., 2003]) has not been formally resolved. Sphyraena qenie: recorded from Mexico by De Sylva and Williams (1986) and observed at Isla Isabela, SE Gulf of Cahfornia by DRR in October 2003; common at the islands of western Panama between 1977 and 2003, and at Clipperton in 1998 (D.R.R., pers. obs.). Trichiurus lepturus: common in most of the TEP; absent from the Pacific Plate, present in the western Pacific. Scomber australasicus: known in the TEP only from the Revillagigedo Islands, where it is a resident (see section on Genetic Studies in Discussion). Scomber jap?nicas: no confirmed records anywhere in the cen- tral Pacific. Balistes polylepis: low numbers observed and collected in the main Hawaiian Islands over the past 20 yr; D.R.R. and John Earle found this species to be not uncommon at sites scattered along much of the western coast of Hawai'i Island in July 2003; hence is a resident at Hawai'i (see also Randall and Mundy 1998); known from one individual in the Marquesas (Randall and Earle 2000). Canthidermis maculatiis: the taxonomy of this species is problematic, with between one and five species, including a TEP endemic (Fedoryako 1979); treated here as a single circumtropical species; not recorded from the Line or Marquesas Islands but likely occurs there because juveniles and adults are semioceanic. Melichthys niger: known from all the offshore islands in the TEP, particularly abundant at Clipperton (Robertson and Allen 1996); uncommon on the mainland. Melichthys vidiia: not uncommon at Cocos Island (D.R.R., pers. obs., 1997, and see Garrison 2000); D.R.R. observed one adult at Clipperton in 1998 and various adults around Isla Montuosa in western Panama in the same year; rare at the Gal?pagos (Humann and Deloach 2003). Xanthichthys caeruleolineatus: known only from Cocos Island and the Gal?pagos in the TEP; rela- tively deep-living (to 200 m) and uncommon within usual scuba limits (Randall et al. 1978), hence its range and population stams within the TEP are un- clear. Xanthichthys mento: in the TEP known mainly from the offshore islands; antitropical in the central and West Pacific; occurs at Japan; common at Easter Island (D.R.R., pers. obs., 1997). Aluterus monocerus: contrary to assertions by Grove and Lavenberg (1997), this species is widely distributed, although not common, in the TEP (see Robertson and Allen 2002). Cantherines dumerilii: notably abundant at Clipperton and Cocos Island in the TEP. Ostracion meleagi'is me- Randall (1972) classified the Hawaiian popu- Tropical Transpacific Shore Fishes ? Robertson et al. ses lat?on of O. mekagris as an endemic subspecies, O. m. camurum, and suggested (J. E. Randall, pers. comm., 2003) diat the TEP form might also belong to that subspecies. Until this matter has been formally re- solved we treat the TEP form as O. mekagiHs meka- gris. Lactaria diaphana: only pelagic-phase individuals found in TEP, in oflfshore waters; these may represent an undescribed oceanic species (J. Leis, pers. comm., 2000); hence we class TEP individuals as oceanic. Canthigaster amhoinemis: a single record from the TEP, in the northern Gal?pagos in 1967 (Hobson and Walters 1968). Canthigaster janthinoptera: known in the TEP from one specimen at the G?al?pagos and another in western Panama (see Robertson and Allen 2002); replaced in Hawai'i by an endemic sister spe- cies. Canthigaster vakntini: in the TEP known from two individuals collected at the Gal?pagos; absent from western edge of the EPB and the Phoenix Is- lands to the southwest of the Line Islands (G. R. Al- len, pers. comm., 2002); occurs in the Tuamotus (~500 km southwest of the Marquesas) and the Car- oline Islands (~1,000 km west of the Marquesas); ab- sent from the Gilbert Islands (west of the Line Islands [G. R. Allen, pers. comm., 2002]); present at the southern Marshall Islands between 5 and 10? N (Allen and Randall 1977, Randall 1985, Myers 1999, Mundy in press), ~2,000 km west of the Line Islands, which are at 2-7? N. Chilomycterus reticulatus: common at the Revillagigedo Islands in 1994 (D.R.R., pers. obs.); in the central Pacific resident only at Hawai'i and perhaps Easter Island; pelagic juveniles collected at the Tuamotus, the Line Islands, and near the Mar- quesas; pelagic juveniles are large (to 20 cm SL [J. Leis, pers. comm., 2000]) and probably have great powers of dispersal. Cyclichthys spilostyhis: only two specimens known from the TEP, which were photo- graphed at diiferent sites and 16 yr apart in the Gal- ?pagos (Humann and Deloach 2003); elsewhere in the Pacific not known eastward of New Caledonia (> 10,000 km in a direct line from the Gal?pagos); not known to have a large, probably long-lived pelagic juvenile (J. Leis, pers. comm., 2000). Diodon hohcanthiis: largely absent from the central Paci- fic (common only at Hawai'i); a record from the Society Islands (Randall 1985) likely is of D. litiirosiis (J. Leis, pers. comm., 2000); common throughout the TEP but antiequatorial in the central and western Pacific.