foxandhound
Transkrypt
foxandhound
C1adistir.s ( 1992) 8:295-318 THE PHYLOGENETIC RELATIONSHIPS AMONG REQUIEMAND HAMMERHEAD SHARK% INFERRING PHYLOGENY WHEN THOUSANDS OF EQUALLY MOST PARSIMONIOUS TREES RESULT Gavin J. P. Naylor’ ’ Department of Biology University of Michigan, Ann Arbor, Michigan 48109-1048, U.S.A. Rectivedfwpublication 20 August 1992; accepted 6 October1 992 Abstmct-Protein variation among 37 species of carcharhiniform sharks was examined at 17 presumed loci. Evohttionary trees were inferred from these data using both cladistic character and a distance Wagner analysis. Initial cladistic character analysis resulted in more than 30 000 equally parsimonious tree arrangements. Randomization tests designed to evaluate the phylogenetic information content of the data suggest the data are highly significantly different from random in spite of the large number of parsimonious trees produced. Different starting seed trees were found to influence the kind of tree topologies discovered by the heuristic branch swapping algorithm used. The trees generated during the early phases of branch swapping on a single seed tree were found to be topologically similar to those generated throughout the course of branch swapping. Successive weighting increased the frequency and the consistency with which certain clades were found during the course of branch swapping, causing the semi-strict consensus to be more resolved. Successive weighting also appeared resilient to the bias associated with the choice of initial seed tree causing analyses seeded with different trees to converge on identical final character weights and the same semi-strict consensus tree. The summary cladistic character analysis and the distance Wagner analysis both support the monophyly of two major clades, the genus Rhizopionodon and the genus Sphyma.. The distance Wagner analysis also supports the monophyly of the genus Gzrcharhinus. However, the cladistic analysis suggests that Grrcharhinus is a paraphyletic group that includes the blue shark Rionnc~ +Lcn. Introduction Carcharhiniform sharks comprise about 55% of the approximately 350 living shark species. The Carcharhiniformes are currently divided into eight families: The cat sharks (Sciliorhinidae), the finback catsharks (Proscillidae), the false catsharks (Pseudotriakidae), the barbelled houndsharks (Leptocharidae), houndsharks (Triakidae) , the weasel and snaggletoothed sharks (Hemigalidae) , the requiem sharks (Carcharhinidae) and the hammerhead sharks (Sphyrnidae) (Compagno, 1988). Of these eight families, the requiem sharks and the hammerhead sharks are generally better known due to the large size of several of their constituent species, their occasional implication in shark attacks and their increasing commercial importance to fisheries. Cawharhinus contains more species than any other genus within the family Carcharhinidae. It comprises 30 species whose collective ranges spread over tropical, sub-tropical and temperate regions (Garrick, 1982; Bigelow and Schroeder, 1948; Compagno, 1984). Many species within the genus are phenotypically similar and are often taxonomically confused. The confusion is frequently compounded by extensive intraspecific variation among geographically isolated populations. As a result, a number of synonyms have been used to describe species Oi$N-3007~y”/040?9.5+24$08.00/0 0 1992 The Willi Hennig Sorietv within the genus. Recent revisions by Garrick (out of more than 100) of these nominal The marked most species heterogeneity 1985), within species indicate and morphological of Cu~&~rhi~~us has allowed intra-generic from morphological 1982), (1982, that only 30 species are valid. data for only the most obvious uniftirmity relationships sibling among to be proposed species pairs (Garrick, i.e. C. v&-C. dussumni, C. ~PUCU--C.nmboinunsis, C. limhntus-C:. nmh&v-hynchoidps and C. ntnhlyrh?rn~hos-C1 whdk. The establishment of phylogenetic relationships beyond Baranes these pairs had not been rigorously and Shahrabany-Baranes, metric data) (Lavery, mltil 1992). genetic Lavery surveyed The present relationships within taxa using allozyme included attempted (with the exception 1986, who used phenetic allozyme variation study represents the genus data derived clustering in 21 carcharhiniform an effort to investigate Cmrharhiwus, and among from 37 (:archarhiniform of of morphotaxa the phylo- its closely related taxa, 1 1 of which were in the survey by Lavery (1992). Materials and Methods Sharks were collected and December America, 1987 Belize, Philippines in the field and from fish markets from south the Bahamas, and Trinidad, the northwestern taken by me with the exception Suiso of the Sea Grant of the Virginia possible different carcharhiniform Localities and samples Samples of heart, shark in the field stored at -75°C. equal volume for Islands. of Marine of Curcharhinus nitrogen and skeletal 2 grams until Approximately Supernatants which were collected were frozen the were by Mark C. altimus collected by A total of 21 (out of a 16 species from eight were other surveyed. 1. muscle were removed placed from in separate to the laboratory where each cryotubes they were half of each tissue sample was homogenized water and centrifuged 1984 of North of specimens to Carcharhinus, of each), transported February coasts in the Red Sea, Samples Science. and related sizes are given in Appendix liver, kidney Gulf Hawaii, and one that are closely of distilled 10 minutes. Institute between and the gulf of Aqaba in Oahu (approximately and kept in liquid east of two C. gdupagensis species genera the Hawaiian Program Dr J. Musick 30) Australia, in an at 0°C and 12000 g (10 000 rpm) at -75°C until required for electro- phoresis. Supernatant Connaught fractions Starch seven different sumed were run in horizontal and 7% sucrose, buffers as outlined loci were initially screened, scored. A list of optimum presented in Table 1. The buffers relative by weight. by Selander starch gels made up with 12.5% Gels were prepared et al. (1971). More using one of than of which 17 showed activity clearly enough and tissues mobility directly from primary gels or from secondary Richardson et al. (1986) when electromorphs used for the of electromorphs “line up” were seen 17 presumed was scored 35 preto be loci is either comparison gels sense to have closely similar mobilities. PHYL~(;ENUK ANAIXIS Data collected in the present study were subjected to both cladistic character and distance Wagner analysis (Farris, 1972) using Cavalli-Sforza and Edwards (1967) PHYLOGENYOF Presumed 2% REQUIEM AND HAMMERHEAD SHARKS Table 1 tissues used and electrophoretic loci scored, Protein Locus acronym” Aspartate amino transferase Aspartate amino transferase Carbonate dehydratase Dipeptidasel Esterase Esterase Isocitrate dehydrogenase Glutamate dehydrogenase Glycerol-Sphosphate dehydrogenase Leucine amino peptidase Lactate dehydrogenase Lactoyl glutathione lyase Malic enzyme Malic enxyme Malate dehvdrogenase Superoxide dismutase Superoxide dismutase GOT-l GOT-2 CA PEP EST-l EST-02 IDH GDH a<;PD IAP LDH CL0 ME-l ME-2 MDH SOD-l SOD-2 conditions Enzyme Tissue commission No.* type’ 2.6.1.1 2.6.1.1 4.2.1.1 3.4.13.1 3.1.1.1 3.1.1.1 1.1.1.42 1.3.1.3 1.1.1.8 3.4.1 I 1.1.1.27 4.4.1.5 1.1.1.40 1.1.1.4f~ I. 1.37 1.15.1.1 1.15.1.1 H H R H H H H K M H H I< M M H H H 1 Stain rrfrrrnc-C’ Electrophoretic conditions” I I I I .5 6 2 2 5 5 2 3 ,F i i i I :1 I I ‘, 1 I 7 ? ? 3 1 -I L, ‘> ; ‘1 i, ’ I.ocus acronyms follow Richardson et al. (19%). ‘Nomenclature Committee of the International Union of Biochemistry (1978). ’ H, heart muscle; K, kidney; M, skeletal muscle. ’ 1. Tris-ritrate EDTA, 250 V, 5 h; 2, Tris-versene-borate. 200 V. 6 h; 3, Tris-citrate pH 6.7. 150 c’, 5.j 11: 4. Poulik, 2.50 V, 5-6 h; 5. Tris-citrate pH 8.0, 130 V, 5 h: 6. I.ithium hydroxide, 350 V, 6-7 h: 7, TrisHCl, 250 V. 5 h. ’ 1, Selander et al. (1971); 2, Sicilian0 and Shaw (1976): 3. Shaw and Prasad (1970); 4, Hal-r-is and Hopkinson (1976); 5, Recipes routinely used in the laborator-\ of R. Highton and outlined in H~tlgt~ (1986). ’ CJycyl lrurine used as substrate for Dipeptidasr. chord distance. the combined group in both to the carcharhinids characters (Compagno, A cladistic was treated analyses. and sphyrnids macrostwma were Hemigaleus Hemigaleids based are regarded on a number used analysis 3.0k (Swofford, as a character. was carried 1990) out using on a Macintosh Electromorphs character states. In the few cases where common electromorph of morphological was entered within the computer program IIci. Each presumptive loci were treated heterozygotes were encountered. as the character state. locus as unordered the most As apparent ht*tcro- zygosity over all presumed loci was low (less than 20/u), and most samples represented by five specimens, this method of coding proved satisfactorv but two cases. These specimen which cases specimens with other were both cases where was heterozygous were heterozygous (seed “A”) ment branch the taxon was represented for the presumed for one unique taxa. They were thus scored A heuristic wise addition as as the sister 1988: 255-256). parsimony PAUP, version microstoma and Hemigaleus outgroup as possessing locus under electromorph were ~OI- all by’a single study. both In and one shared the shared electromorph. procedure was used to search for most parsimonius trees. The step(simple) option of PAUP was used to generate the initial setd tree while the tree bisection swapping. subsequent branch survey and because reconnection Only tree topologies swapping. Because electrophoretic (TBR) of minimum a large number mobilitv option was used to implr- length were retained of taxa were included characters were coded f’ol in the withotlt XI\ 298 G. J. I’. NAYLOR presupposed transformation sequence (unordered), most parsimonious trees (MPTs) were generated. a large number (37267) of EVALUATION OF MULTIPLE MOST PARSIMONKWS TREES The phylogenetic infiation content of the data The phylogenetic information content of the data set was evaluated using the randomization tests of Archie (1989). Character state assignments were randomly permuted within each character of the original data matrix to remove phylogenetic information (Archie, 1989). Fifty different randomized matrices were generated, each one reflecting the character state distribution of the original data. Each matrix was subsequently subjected to the heuristic (TBR) tree searching procedure of PAUP 3.0k. Searches were left to run to exhaustion or were stopped once 10 000 MPTs had been found. The tree length derived from the original data matrix was contrasted with the histogram of tree lengths derived from the 50 randomized matrices (Fig. 2) and subjected to a one-sample t-test. Seed tree bias The effect of different starting seed trees on the outcome was investigated by contrasting results using stepwise addition (seed “A”, set “A”) with those obtained from a second seed tree, “B” constructed using random addition (set “B”). This second tree “B” was subjected to TBR branch swapping until 20 000 MPTs had been generated. A majority rule consensus of the 20000 seed “B” trees was constructed and contrasted with the majority rule consensus of the seed “A” trees. Signal consistency during branch swapping S&samples of trees produced by branch swapping were taken to test whether the same “kinds” of trees were produced throughout the course of a heuristic search. Each sub-sample comprised a block of 5000 trees produced sequentially by the branch swapping algorithm (i.e. tree Nos l-5000, 5001-10000, lOOOl15 000, 15 001-20 000). Majority rule consensus trees were constructed for each block of 5000 trees and compared with one another and with the majority rule consensus tree for all 20 000 trees using partition distances (Penny and Hendy, 1985). This procedure was carried out separately for trees in set A and for trees in set B. The consistency of the hierarchical signal among s&samples was also investigated graphically sense Naylor (1992). Each sub-sample of 5000 trees was subjected to the plot group frequency option of PAUP. This option orders “groups found” (clades) by their frequency of occurrence in a sample of MPTs. The “groups found” in subsamples were re-ordered to coincide with the order found for the entire 20 000 tree set. The frequency of occurrence of “groups found” was plotted for each of the four subsamples and compared. This procedure was carried out separately for both tree sets A and B. Successive weighting Successive weighting (Farris, 1969, 1989; Carpenter, 1988) was adopted in an attempt to reduce the number of most parsimonious trees produced. Autapomorphic character states (electromorphs) were removed from the data set and PHYLOGENYOF REQUIEM AND HAMMERHEAD SHARKS 299 recoded as missing values to ensure that the resealed consistency indices used in successive weighting were not artificially inflated (see Farris, 1989). Characters (presumed loci) were weighted according to the best fits of their resealed consistency indices to the initial 20 000 trees. TBR branch swapping was carried out on the initial seed tree with the new character weights enforced. After 20 000 new trees had been generated, the characters were re-weighted according to their new best fits. This protocol was re-iterated until character weights stabilized, at which point the semi-strict consensus of the resultant set of trees was computed. The procedure was applied to tree sets A and B, separately. The two final sets of successively weighted trees (one from set A, the other from set B) were sub-sampled as outlined above. Hierarchical signal constancy among subsamples within tree sets, and between tree sets, was contrasted with its counterpart in the unweighted analysis. The topology of the distance Wagner tree was entered into PAUP to determine its length. Results and Discussion A total of 201 electromorphs was detected over the 17 presumed loci surveyed. Of these 116 (58%) were found to be unique (autapomorphic) to individual taxa. The electromorphs detected were alphabetically coded in the character state matrix to reflect their relative electrophoretic mobilities and are presented in Appendix 2a. Those most anodal are denoted by “A” while the most cathodal are denoted by “Z”. (Because 27 alleles were found at the PEP locus the most cathodal allele was scored using the integer symbol “l”.) The rooted distance Wagner network (Fig. 1) had a cophenetic correlation coefficient of 0.929, an Fvalue of 7.1 (Prager and Wilson, 1976) and a length of 92 steps when fitted to the character matrix. 32 767 most parsimonious trees (MPTs) of length 86 steps (computed excluding autapomorphies) and retention index (RI) 0.892 (Farris, 1989) were generated by branch swapping (TBR) on seed tree “A”. This number represents a ceiling imposed by computer memory rather than a complete set of MPTs. 20 000 MPTs of length 86 steps and Rl 0.892 were generated by branch swapping on the seed tree “B” at which point TBR branch swapping was deliberately interrupted. Results of the randomization tests indicate that the MPT length of 86 steps is 31.44 standard deviations shorter than the mean MPT length of 159.18 found for the 50 randomized data sets (Fig. 2), suggesting that the data matrix has a high overall phylogenetic content even though it yields a large number of MPTs. The majority rule consensus tree (Margush and McMorris, 1981) for set A was resolved into 27 component forks [Fig. 3(a)]. The majority rule consensus tree for set B was resolved into (29) component forks [Fig. 4(a)] which differed in topo logical structure from that of set A. In the unweighted implementation, majority rule trees for sub-samples within a tree set yielded closely similar topologies to the majority rule tree for all trees in a set. This similarity in “hierarchical signal” throughout the course of branch 300 (;..I. P. NAM.OR C.sorrah Caltimus - C.plumbeus I Cleucas Cmelanopterus lmacrorhinus T.obesus N.brevirostris P.glauca S.lewini Stiburo S.tudes S.mockarran - Szygaena G.cuvieri Racutus R.lalandii Rporosus R.terranovae H.macrostoma H.microstoma , 0.0 I 0.1 1 I I I 0.2 0.3 0.4 0.5 I 0.6 I 0.7 DISTANCE FROM ROOT Fig. 1. Distance Wagner tree (Farris, 1972) of 37 species of carcharhiniform sharks. The tree was constructed using Cavalli-Sforra and Edwards (1967) chord distance and was rooted with the two species of Hemigakus as an outgroup. The tree topology had a length of 92 steps when fitted to the original character state matrix. PHYLOGENYOF 301 REQUIEM AND HAMMERHEAD SHARK!3 lo- 0 ( I 90 80 I 110 I 100 ” I 120 ---r---- -4 140 130 16” 150 MPT length Fig. 2. Histogram plot of most parsimonious tree lengths derived from 50 rdrldom permutation\ ot the data matrix (Archie, 1989). The original data matrix yields an MPT length of86 steps which is :%I.+I standard deviations shorter than the mean of 159.18 steps found for the permuted data sets, wggcw~ng that the original data matrix has latent hierarchical structure. swapping on frequency plots a single starting tree (Fig. 5). The minor reflected in the among sub-samples contrasts with the variation partition distance distances to be about five times smaller than among corresponds The [Fig. 6(a)]. to different topological search The “islands similarity tree set, suggests heuristic between is also variation partition of trees” that only a portion indices clearly trees among of the total number the majority searches on different rule consensus set also shows that majority a single branch in majorit): rule consensus signal produced less resolved swapping; rare topologies, as more by contrast, reflects a f~-orrl ;I over all tht to use heuristic srarch- of executing incomplett to conclusion on a singlt rule trees among more throughout seed tree than does the semi-strict becomes within advised to use the several seeds option. similarity of the hierarchical one search set (1991). suhsamples of trees resulting the alternatives seed trees or executing seed tree, one is better The topological between tree This probahh as put forward by Maddison rule to estimate and has to choose group tree sets show within tree set distances. trees in a tree set. It follows from this that, if one is obliged ing methods within tree sets A and B in the pair-wise matrix of tree in majority are required sub-sample “rare” accurately the course consensus. groups the majority the predominant sub-samples reflects of branch (The art- found rule the constant\ swapping semi-strict during consensus, hierarchical within ;I tic’<’ being signal on consensus the courst’ of’ unaffectrd hi throughout branc.11 swapping). Successive weighting resulted in the stabilization ‘The final weights for tree set A were identical are presented topology in Appendix for both less sensitive to bias arising sets will be required 2b). The tree sets, suggesting semi-strict consensus that the successive from different to test the generality of weights after three iterations. to those of tree set B (final weights seed trees. tree weighting (Further had the same procedure work on other i\ data of this supposition). Sub-sampling of the final set of successively weighted trees from both tree stats revealed a different hierarchical signal profile during branch swapping than was observed for the unweighted implementation as described by Naylor (1992). Frequency distributions of groups found (Fig. .5) had broad plateau regions at the top of curves corresponding to higher numbers of groups found in 100% of tree4 _ Lmacrorhinus C.acronotus C.albimarginaius C.attimus -[ C.plumbeus C.fafciformis Cgalapagensis Cbngimanus - Cobscurus -- P.gfauca - Cperezi Camblymynchos Cwheeleri C.brachyurus Cbrevipinna c.leucas C.limbatus Cmacbti Cmelancpterus C.porOSuS Csealei Csorrah ____ - - H.macrostoma H.microstoma C.acronotus C.albimarginatus C.attimus C.plumbeus C.amblyrhynchos C.wheeleri C.brachyurus C.brevipinna C.falciformts Cgalapagensis Cobscurus C&don C.leucas C.limbatus C.bngimanus Cmacbtt Cmelanopterus Cparezi C.porosus Csealei Csorrah Gcuvieri L.macrorhinus N.brevirostris P.gtauca R.acutus Rtalandei Rporosus Rterranovae S.lewinr Smockarran S.ttburo Studes Szygaena T obesus -4 r ---I - - - S.lewini Smockarran Szygaena Stiburo S.tudes N.brevirostris Tobesus L.macrorhinus G.cuvierr Racutus R.lalandei R porosus R.terranovae Cbrachyurus C.brevipinna Csorrah C.leucas C.limbatus C.melanopterus -----I C.a”imus Columbeus PHYLOGENYOF REQUIEM AND HAMMERHEAD SHARKS 303 as seen in the unweighted analysis. However, while the within-tree set distances are comparable to those seen for the unweighted case, the among-tree set distances are roughly half those seen for the unweighted case. Invoking Maddison’s (1991) “islands of trees” analogy once again, it would appear (for this data set at any rate) that successive weighting has drawn islands in the “archipelago” closer together without affecting their size. (a) Majority rule consensus - No weighting I I I 1 (W Semi-strict consensus - No weighting (cl Semi-strict consensus - Successive welghting Fig. 4. (a) Majority rule consensus tree of the first 20 000 MPTs generated by branch swapping on seed B. (b) Semi-strict consensus tree of the first 20 000 MPTs generated by branch swapping on seed B. (c) Semi-stl-ict consensus tree of successively weighted trees generated by branch swapping on seed B. Final (asymptotic) weights used are given in Appendix 2b. Note that the topology for Fig. 5(c) is the same as that of Fig. 4(c), reflecting the convergence of tree topologies toward the same solution. G..J. P. NAYI.OK set A 0 100 group categories set 0 200 0 100 200 group categorias Fig. 5. Hierarchical signal consistency among subsamples of trees. Each plot depicts the frequencies of corresponding groups (clades) found in each of four sub-samples. Each subsample comprised 5000 trees (l-5000,5001-10 000, 10 001-15 000,15 001-20 000). Note that the cuxves produced by successive weighting have characteristically extended plateau regions and truncated tails, reflecting, respectively, an increased number of clades found in 100% of the trees and a decreased number of rare clades found among all the trees. Both factors cause the strict consensus tree derived from successively weighted trees to be more resolved. The semi-strict consensus of the successively weighted analysis was chosen as the most reliable summary of the phylogenetic signal in the data on the grounds that it was: (a) uninfluenced by seed tree bias; and (b) the most resolved tree that was entirely consistent across all s&samples (the majority rule trees varied across subsamples for some of the groupings found). VARLATION BETWEEN ANALV~ES While the retention indices for the MPTs generated in the parsimony analysis are high (0.892) and the cophenetic correlation coefficient is high for the distance Wagner tree, the conclusions concerning relationships vary between analyses. The semi-strict consensus tree derived from the successively weighted character analysis is much less resolved than is the distance Wagner tree. This is largely because the semi-strict consensus tree is a particularly stringent summary of the MPTs found. As a result, there are clades found in the distance Wagner tree that appear unresolved in the semi-strict consensus tree, but that are supported by a large percentage of PHYLOGENYOF the MPTs used to generate conservative, only successively weighted The genera genera Rhizopionodon, (the than sharks (Sphyma) topological resolution reef white-tip weighted is not character seen congruence However, Relationships within Carcharchinus species pairs are defined The among of the blue then the genus However, when shark P. gkzuca is included, the genera analysis. Interestingly, Wagner clade using the constraints clade (0.79% increase loci scored the genus option in 9.5% of Somts plumbru.s- [C. abollt rrla- it has experienced outside weighted implementation, based on parsimony that independent accelerated teeth referable 1905), (l.andini, to whereas referable 19’7’7). If indeed but rather the sister group must be explained. (One explanation membc~- of cxvolrltion rclativc, at so many loci and gi\rell extra to make steps in the utl- in tree length systematic How&r, support it is notcworth\~ Compagno (Jarrharhinus appear in the to Prionace do not is of course, that did not develop ( l!)Xti) of‘ sharks. Middle appear Ptionacv is riot a derived a clade containing in the, statemcnls the idea that thr bIu~- Eoctsnc, until member- I hc~ \vithitl (2~rcharhinrr.c. I.O.WJC~O~C, Tn’aenodon, Negaption, as suggested by the distance Wagner analysis, spicuous absence of its fossil teeth during the Eocene. Oligocene, descendant of an old lineage until the Pliocene.) tht. of the presunlecl a derived molecular position. first of’ fi-onI to 62 I.53 link I’. g/ouca to the obscures group Carcharhinus teeth derivrd 61 664 at eight 1% increase and fossil evidence teeth from cloes not require less than species within the genus serrated (‘. in the number of P. glauca from length that, if it is indeed as to its true phylogenetic that its broad 0/~sr11171.~, from the (,‘awhwhilir~u~ the exclusion it may be premature shark might be a derived C. If the weights in tree Carcharhinus and causes morphological containing is no increase P. glauco is autapomorphic Carcharhinus, If I? &IILC~~is group in both analyses. analysis but as the sister taxon to tht, there then 2a). This suggests (2rrrharhinu.r tree. analyses. 7dm!4~i]. is problematic. implementation). in an increase in length). implementation, Pliocene in all Loxodon, Negaption and Trkzenodon in thta of PAUP, (Appendix that its placement (Stromer, was found if P. glawa is excluded are implemented, results weighted Similarly shark I Wagner are disagreements to its sister taxa. Given that P. glaum is autapomorphic suggests shark), lemon of the successiveh between it falls in a clade Carcharhinus, (in the unweighted weighting (Jarcharhinus Relation- tiger tree. 1’tionace &zutn distance successive (Rhirtr Carcharhinus tree forms a monophyletic Carrharhinus clade containing steps required outside consensus But there and C. longimanus in the character galapagensk analyses. Negaprion (the considerably brarhyww]. sharks these sub-groups. placement omitted, between in both trees, i.e. [C. amb@-hynrhos-4. [C. brev ip’znna-C. C. altimus], tionships sharpnose this basal placement vq to be consistent basally on the distance the consensus too of thr reliable. tend Galpocerdo (the are placed semi-strict the MPTs that were used to generate tree clades in all trees and havtt concordance. shark), in the analyses. and (the slit-eye shark), Loxodon consensus Carcharhinus monophyletic trees. Rhizojwionodon and Gakocedro, are placed This At the risk of seeming semi-strict here as phylogenetically other also show some Sphyma, Triaenodon consensus. in the fall into distinct of within-clade among found Hammerhead for example, a high degree ships within analyses. pionodon), the semi-strict groups tree sets are regarded relationships between and those 30.5 REQUIEM AND HAMMERHEAD SHARKS that I? ,&ucn its unique then thy CX)I~and hlioc.t~~~c~ is the moclt~r~1 tooth morphologic C;..]. P. NAMX)K (a) PARTITION MAJORITY (DERIVED a b METRIC DISTANCES RULE CONSENSUS FROM UNWEIGHTED c d BETWEEN TREES DATA SET) e f g h t I I :22 I 23 25 :23 I I 24 26 \\ Consensus Tree a b : :! g h \\ __ -4 I_*_. :_23 _ _ _ _2_4_ _ _2P_ KEY: \. MPTs Used l-5000 5001-10000 10001-15000 15001-20000 l-5000 50001-10000 10001-15000 15001-20000 \ TREE SET A TREE SET B > Fig. 6. (a) Partition metric distances (Penny and Hendy, 198.5) among majority rule trees constructed from s&samples of trees generated by branch swapping on unweighted data sets. Trees a to d are derived from branch swapping on seed tree “A”, while trees e to h are derived from branch swapping on seed tree “B”. The partition distances show that groups of trees derived by branch swapping on a given seed tree are more similar to one another (triangular outlines) than they are to trees generated from a different seed (rectangular outline). This observation is concordant with Maddison’s (1991) finding of islands of trees. CORRESPONDENCETO THE STRATIGRAPHK: SEQIJENCE The remainder of the discussion addresses the tree topology semi-strict consensus tree of successively weighted trees [Figs Because carcharhiniform 1984) the first recorded depicted in the 3(c) and 4(c)]. sharks have such an excellent fossil tooth record (Maisey, stratigraphic appearance of each taxon is presented to PHYLOGENYOF M REQUIEM AND HAMMERHEAD SHARKS 307 PARTITION METRIC DISTANCES BETWEEN MAJORITY RULE CONSENSUS TREES (DERIVED FROM SUCCESSIVELY WEIGHTED b a c d e DATA SET) f g h j&, 0 ‘. 1 \ \ 0 2“. '5 . , \ \ 0 \ I 15 4 z“, I________-----> :Is_ _ _;3_ _ - _,;_ - - - - 11: I I y. 13; I I I4“, , I , I ;16 11 '14 I 9 11 11 13: :_w ___l_J__ _13_- ._JJ1 KEY: Consensus Tree a b C d e f 9 h :6 I I I I I I :_4___ . \ . \ 4“. . \ . \ _4____4_1‘> MPTs Used l-5000 5001-10000 10001-15000 15001-20000 \ l-5000 50001-10000 10001-15000 \ I 5001-20000 TREE SET A TREE SET B J Fig. 6. (b) Partition metric distance (Penny and Hendy, 1985) among majority rule trees constructed from sub-samples of trees generated by branch swapping after asymptotic successive weighting. Trees a to d are derived from branch swapping on seed tree “A “, while trees e to h are derived from branch swapping on seed tree “B”. The partition distances show groups of trees derived by branch swapping on a given seed tree to be more similar to one another than to trees generated from a different seed. Note, however, than the partition distances between different tree sets are considerably smaller for the successively weighted case than they are for the unweighted case depicted in Fig. 6(a). determine whether the cladogram branching appearance times. The tiger shark Caleocerdo cuvieri emerges most basal to the selected outgroup sequence is concordant as one of seven polytomous Hemigakus. As mentioned previously, with first branches Galeomdo and Rhiwprionodon form the two most basal branches in 95% of the successively weighted MPTs, perhaps suggesting a more basal position for these two taxa than SO8 actually depicted first recorded Capetta, in the semi-strict from The four species cluster 1988). basal placement Eocene, of sharpnose traits (Compagno, with clade species similarities hammerhead from that incorrectly the same similar than makes shapes. seven the on the teeth argument between snout occurring Within 1988; pers. The fossil from fossil the origin tooth ancestor. represent and Rhiqtrrionodon would that the the lateral vestige and origin the two diminutive Loxodon are so similar appear (Capetta, in the Middle The genus Fossil hierarchical of Rhizoprionodon 1987). 1905). this period Teeth teeth The are species of Rhiroptionodon of Nigeria first most commonly within the genus or Scoliodon. Triaenodon at the and Sphyrna. Teeth in size have not yet been Teeth small and to my from Negaprion first 1955). from the recorded C. egertoni (Capetta, RELATIONSHIPSWITHINTHECLADECONTAINING 1987; of sphyrnids and Scoliodon, both a large clade that appears recorded a shape independently Triaenodon are quite (White, tooth shark, in the fossil record from onward is the C. leucaslike Relationships white-tip in the fossil record. Carcharhinus comprises Ca&arhinus (Stromer, to those from trowel-like Sphyma tibwo and S. tudes level as Galeocerdo, RhiqtAonodon reported Eocene reef be (1988) of an incipient with the origin species the macrorhinus, first appearances have not been been (Capetta, Scoliodon shark, Negapn’on breuirosttis, all branch knowledge of might Compagno with the very similar be consistent to the of sphyrnids obsesus and the Lemon established have of the hammerheads He suggests in conjunction subsequent that reliable group teeth It is possible, horizons evidence. Loxodon and shape pair distinct fossil 1987). form a sub-clade. The slit-eye shark, from This monophyletic earliest (Capetta, at earlier This the hammerheads, same polytomous R. @rosus and R. counts. R. acutus. for the development obs.) at any time (Lower sequence are markedly a clearly Miocene of Scoliodon might hammerheads Compagno, form found ridges taxa. The branching by morphology. polytomy. Lower supposed cephalofoil. The only by vertebra1 branch pre-anal as Physogaleus, Scoliodon or Rhizqiwionodon due to their Scoliodon or Rhizo@ionodon-like flanges and with the early appearance in the fossil record. If this is the case, then is currently hammerhead mono- two primitive in the more derived and relatively primitive either a convincing a distinctly furrows is corroborated sphyrnid identified strikingly 1952: labial is consistent (Sphyrnidae) from small of (GnLocPrcto are (Arambourg, occurs with R. Zalandii while all three sharks Sphyrna are recorded however, teeth Rhi.zojwionodon shares distinguishable from the largely circumtropical branching Fossil of Morocco Rhizofwionodon form Galporerdo: long of its teeth the Rhizoptionodon terranovae are sibling The sharks None of these characters 1987) have many cranial tree. Eocene) from the polytomy. of Rhiqbrionodon Capetta, consensus (Lower 1987). branching morphological within the Ypresian 198 1; Capetta, phyletic older P. N.Z\TL.OR (;.,I. to include Middle Eocene P. gluuca. of Egypt in the fossil record from 1987). CARCHARHINUSAND PRIONACE Carcharhinus are only partially resolved. C. UCTP notus is placed as the sister taxon to all other members of the genus. C. isodon branches next, after which there is then a polytomy containing 11 branches. Eight of these 11 branches give rise directly to terminal taxa. Two of the branches give PHYLOGENYOF rise to the species pairs wheeler-i). One branch REQUIEM A.ND HAMMERHFm taxa. All eight sharks in this resolved length. All but one of the eight tudinally branch between containing C. galapagensis, PREVI~LXYPR~POSED Springer (1950, suggested pointed ridge-backed and that members pectoral treatment characters that of a group group, include large ridge-backed tips assessment (1982) rounded except The of the genus group Garrick (1982) sub-groupings (1982) forms. rounded tips with (Garrick, that both his of specirs too 1982). be regarded Garrick as the c-entral C. perezi and (:. long- obscun~s group, th us excluded was erected C. son-ah, C. falr@rnis, Eulamia group but included Compagno Hr with (1988) to teeth (.A rtlb- C. longimmnrts, concurred with (;. perezi from the group and included here suggest that the large ridge-C. long-imc~nuv, (.‘. /al& (.‘. ppfpzi) ;irc’ all members of a monophvlcti( P. glauca. and Compagno (1988) seal& group. character (1988) have postulated based on morphological suggest several in the semi-strict analysis. Compagno’s tentati\.c Both (&rick he closely is supported co~~sc~xas inclusion other similarity. that C. po)us~~.sshould This close relationship tree but does not appear sively weighted be divided into those the inclusion might to as the data presented within Carcharhinus the C. dussumm’-C. Wagner those (C. ol~~~rus, (A galapagtxsis, C. altimus and Compagno and and C. altimus, (I plumhew, fin morpholoE. allozyme tha: also includes be divided with each other that he removed C. albimarginatus. ,formis, C. plumbpus, clad< sharks that had, for the most part, broad triangular backed members should necessitated he referred Garrick in spite of its distinct sensu lato might as it was considered C. galapagensis marginatus and C. sealei from Springer’s oarrick’s off next, polytomous CAK~IIMHIN~V (Eulamia) groups included which (C. pmzi excepted). forms to be aligned which This to an apical Carrharhinus was abandoned C. obscures and manus. longi- C. altimus and C. plumbewc (Carcharhin2c.s sense stricto) and ridge-backed and smooth-backed in other proposed than 2 m total ridge running C. faltiformis, branches the sister taxon suggested forms dorsal to more (1. eight C. obscurus, C. longimanus and P. glauca. 1951) Pterolumniops. This diverse forms in turn that the ridge-backed first the clade. INTERRELATIONSHIPSWTHIN into smooth-backed clade containing have an interdorsal pair. The silky shark, by C. perezi, which (C. amblyrhynchos, well-resolved clade are large, growing (I? gh,cu) the two dorsal fins. Within off basally as a species followed and C. bruchyurus) (C. breuipinna, gives rise to an internally 309 SHARKS linked to by the distance tree or the succc~- of‘ (I mdoti and (:. wwrclh in this group is not supported. Garrick and suggested (A ZI&JP~P~~ on characteristics. This that the C. albimar@natus basis suggested tree but does not appear of overall was closely morphology, relationship in the semi-strict related tooth is supported co~~sc~~s~~s to CL ambly~-h~rld/o.~ _ shape and vcr-tehal by the distance tree. Compagno’s L$‘agnt.rtreatment of (:. perezi as the sister group to C. clmb1vrhvnchn.sand (.L 7rhwlti is not supported. The suggestion by both authors that- the sibling species C. 1irnbatu.c and C.‘. atnhi~rhvnrhoidus are most closely related beuipinnn f-rills closest to CAbryuipirzna is not supported. to C. brachyurus. The striking similarity In this study. (1. between cl. litd/atw and C. brvc~ipinna in body shape, color pattern, tooth shape and behavior appears to be the result of convergent evolution rather than of rlose rrlatedness. Support for G.,J. P. NAM.C)K 310 this supposition is seen in the disparate to color pattern. tips. These diffusely C. &&ztusjuveniles markings pigmented unpigmented The present recent (pers. but develop with age (Garrick, been dipped diminish 182), ontogenies in intensity obs.). of the two species with respect are born with distinctive black pigmented fin with age so that large adults appear onl) C,‘. brpvipinnn juveniles are born In contrast, an increasingly intensive so that large adults often black appear pigment on the fin tips as though their fins had in black ink (pers. obs.). allozyme survey by Lavery study. In his study, Lavery tunately, the two data evidence for the sets cannot phylogeny. scored (1992) is closely 38 allozyme comparable loci for 21 taxa. to the Unfor- be readily combined to evaluate the total are thus restricted to topological Comparisons C.acronotus Csealei Cdussumeri C.porosus C.titzoyensis Calbimarginatus Camblyrhynchos C.wheeleri C.altimus Cperezi C.plumbeus Cgalapagensis C.obscurus C.longimanus C.brachyurus C.brevipinna Climbatus Camblyrhynchoides Cfalciformis C.sorrah C.leucas Camboinensis C.melanopterus Ccautus Cbornensis Rhizoprionodon Fig. 7. Phylogenetic hypothesis for sharks of the genus Carcharhinus after Garrick (1982). Topology deduced from Garrick’s text. PHYLOGENYOF REQUIEM AND HAMMERHEAD SHARK!S 311 similarities and differences among inferred trees. In Lavery’s study, one most parsimonious tree resulted with a length of 215 steps and a retention index of 0.545. The monophyly of the genus Curcharhinus was not supported in his study. Instead, the genus was seen to be paraphyletic including members of two other genera (Negapn’on amtudiem and Galeocerdo cwieri) and one member of an entirely different family (Hemipistis ehgutus) . A re-analysis of Lavery’s data shows that if the genus Curcharhinus is constrained to be monophyletic, five most parsimonious trees of length 219 steps result. This represents a four-step increase over his presented most parsimonious tree. (Lavery reported that “placing these three species in their recognized positions outside the genus Curcharhinus results in a tree C.acronotus C.brachyurus C.falciformis C.melanopterus C.cautus Csignatus Nasolamia C.albimarginatus Caftimus C.galapagensis C.longimanus C.obscurus C.plumbeus Cleucas Camboinensis Glyphis Lamiopsis Negaprion Prionace Camblyrhynchos C.perezi C.wheeleri C.brevipinna Climbatus Camblyrhynchoides C.leidon Cisodon C.porosus C.sealei Csorrah C.fitzroyensis E Cdussumeri C.bomensis Cmacloti Chemiodon Triaenodon Fig. 8. Phylogenetic hypothesis for sharks of the genus Carcharhinus Topology deduced from Compagno’s text and presented cladogram. atter Compagno (1988). (i. J. I’. NAYLOK 3 12 almost 10% longer ment] I was unable [either 228 or 229 steps, to reproduce depending this finding). on the precise By reciprocal sion of Neguption breuirosttis, Galeocerdo cuuieri and Hw@zkus the present study results in an increase As pointed out by Lavery as they examine polling different the taxa from phylogenetic structure (1992), portions both in tree length the two studies within Carrharhinus do not overlap phylogeny. is recommended within the genus the incluin of two steps. of the Currhnrhinus studies, arrange- contrast, to more extensively A larger survey, confidently assert Curchnrhinus. CIASSIFICATION Results from both the the Carcharhinidae, hammerhead cladistic the herein (genus of the Carcharhinidae, are in agreement, Proposals suggested to the (1934, 1939, that “the prime reason without sufficient forms] and Springer for these and Eulumia dependent accelerated more, (1982). to, molecular the shows that reflects evolution hypothesized a or largely supported of cladogram a formal re-classification relationships being collected (1982) established in intra-generic of s.1. the data. within available. be If a of Further- the genus monophyletic several monotypic Curchurhinus data. ad hoc incidences strictly put are strongly by the available any known affinities proposed to explain becomes other Curcharhinus is excepted.) of relationships tree, that concerning (1853)) Garrick from, cladogram, must be invoked ommended evidence forms] the presented unresolved 1951). distinction the tentative of the genus would involve naming are currently re-classification s.1. into two genera, classification further the data presented to become some of the groupings and are only tenuously is adopted because Curchurhinus characters, within on the basis of only one species for [ridge-backed However, (1950, taxa failing of Curchurhinus Results of this study partially corroborate forward by Garrick classification the a tentative Sphyrnini cladistic that includes have been made by Owen Curcharhinus 1943) reference (Sp rm g er’s division [smooth-backed analysis tribe that Compagno’s the genus Whitley and species.” which proposed data and the molecular usage is that for the most part they were erected each group (1988) indicate based on morphological assigned I suggest analysis be adopted. to subdivide Gill (1962), were character Compagno As both the morphological hammerheads and is a paraphyletic Sphyna). hammerheads Carcharinidae. Wagner as it now stands, sharks reclassification in which distance re- taxa. It is rec- postponed until DNA sequence data for this purpose. Acknowledgments Several people assisted particularly like to thank allowed me to accompany in the collection the commercial them fishing of material for this study. I would fishermen and scientists who either or actually They are: from the east coast of North America: N. Sanders, J. Musick and J. Colvocoresis; collected specimens for me. D. De Maria, R. Bakey, E. Sanders, from Hawaii: E. Schellenburger, S. Raiser and M. Suiso; from the Red Sea: A. Sulliman and “Muhammed”. Others who generously provided either direct assistance or advice were R. Apilado, “Mungpert” and D. Cadra (Philippines), J. Parrish and Sea Life Park (Hawaii), P. Coutin PHYLOGENYOF REQUIEM AND HAMMERHEAD SHARKS 3 I3 and P. Showers (Sierra Leone), R. Fox, D. Doubilet and H. Fraise (Australia), M. Doty, L. Cohen, T. Bowman (Belize), S. Blum, J. Casey, R. Shipp, J. Dendo, G. Nelson, L. Abele, V. Springer, G. Burgess, J. Castro, J. Ott, E. Silverfine, D. Carr, J. Jibrin, S. Wilkes, S. Fitzgerald, D. Chipman and E. Knurek (U.S.A.). I thank Carla Hass for patiently teaching me electrophoretic techniques and E. Clark, M. Doty, R. Highton, C. Miter, G. Vermeij, R. Voss, G. Wilkinson and four anonymous reviewers for their helpful suggestions for the improvement of the manuscript. I would like to give special thanks to R. Highton for the use of his electrophoresis laboratory, to G. Vermeij for his persistent encouragement and to E. Clark for first suggesting I pursue this project. Financial support was provided by The WIN Foundation, The University of Maryland Foundation, The University of Maryland Computer Science Center, D. Shen and the National Science Foundation (grant BSR 87-08121 to G. J. Vermeij and grant BSR 83-07115 to R. Highton). REFERENCES ARAMBOUR~;,C. 1952. Les vertebres fossiles des gisements de phosphates (Manx-AlgrrirTunisie). Serv. Geol. Maroc, Notes et Mem. 92: l-372. ARCHIE,J. W. 1989. A randomization test for phylogenetic information in systematic data,. Syst. Zool. 38: 239-252. BARANES, A. AND 0. SHAHRABANYBARANES.1986. A numerical taxonomic study of the northern Red Sea carcharhinids. Proc. 2nd Int. Conf. Indo-Pacific Fishes, pp. 246255. BIGFLOW,H. B. AND W. C. SCHROEDER. 1948. Fishes of the western North Atlantic. Part I Mem. Sears Found. Mar. Res. Yale Univ. Press, New Haven, Connecticut. CAPETTA,H. 1981. Additions a la faune de selaciens fossiles du Maroc. 1: sur la presence dtss genres Heptrunchias, Alo@ et Cldonlor!+s dans L’Ypresien des Ouled .Abdoun. CAobios 14: 563-575. CAPE?TA, H. 1987. Chondrichthes II. Mesozoic and (:enzoic Elasmobranchii. Vol. Jb. tiandbook of Paleoichthyology. Gustav Fischer Verlag, Stuttgart, West Germany. C~ENTER, J. M. 1988. Choosing among multiple equally parsimonious cladograms. <:ladistics 4: 291-296. CAVAI.I.I-SFOR%A, I.. 1,. ANDW. F. EDWARDS.1967. Phylogenrtic analysis: Models and estimation procedures. Evolution 21: 550-570. COMPA(;NO,I,. J. V. 1984. FAO species catalogue. Vol. 4. Sharks of the world. An annotated and illustrated catalogue of shark species known to <late. Part 2. Oarcharhiniformcs. F.40 Fish. Synop. 125: 251-655. COMPA(:NO, L. J. V. 1988. Sharks of the Order Carcharhiniformrs. Princrton Ilni\. Press. Princeton, New Jersey. FARRIS,J. S. 1969. A successive approximation approach to character weighdng. Svst. ZooI. 18: 256268. Farris, J. S. 1972. Estimating phylogenedc trees from distance ma1rict.s. ,4m. Nat. IO(i: 645-668. FARRIS,J. S. 1989. The retention index and the rcscaled consistrncv index. (:ladistics I,: 417-419. GARRICK,,J. A. F. 1982. Sharks of the genus Carch~mhinrr.s.NOAA Trrh. Rep., NMFS. (1it.c.. 445. GARRKK,J. ,4. F. 1985. Additions to a revision of the shark genus Carcharhirms: Synotlyrny OI ,-\ptionodon and Hypoprion, and description of a new species of C’wchnrhirru~. N( )A..1 Tech. Rep., NMFS. Circ. 34. Gr1.1..T. 1862. Analytical synopsis of the order of squali: and revision of the nomenclature OI the genera. Ann. Lyceum Nat. Hist. 7: 367-408. HARRIS, H. AND D. A. HOPKINSON. 1976. Handbook of Enzyme Electrophoresis in fluman (Genetics. North Holland Publishing, Amsterdam, The Netherlands. 314 G. J. P. NAYLOR HEDGES,S. B. 1986. An electrophoretic analysis of holarctic hylid frog evolution. Syst. Zool. 35: l-21. LANDINI,W. 1977. Revizione degli ittiodontoliti pliocenici della collezione Lawley. Palaeontographia Italica. 70: 92-134. LIVERY, S. 1992. Electrophoretic analysis of phylogenetic relationships among australian carcharhinid sharks. Aust. J. Mar. Freshwater Res. 43: 97-108. MADDISON, D. R. P1991. The discovery and importance of multiple islands of most-parsimonious trees. Syst. Zool. 40: 315-328. MASSEY,J. G. 1984. Higher elasmobranch phylogeny and biostratigraphy. Zool. J. Sot. 82: 33-54. MARGUSH,T. ANDF. R. MCMORRIS.1981. Consensus n-trees. Bull. Math. Biol. 43: 239-244. NAYLOR,G. J. P. 1992. Plotting frequency distibutions of Phylogenetic groupings found among sets of most parsimonious trees. Cladistics 8: 161-164. NOMENCLATURE COMMITTEEOF THE INTERNATIONAL UNION OF BIOCHEMISTRY.ed. Enzyme Nomenclature 1978. Academic Press, New York. OWEN,SIR RICHARD.1853. Descriptive Catalogue of the Osteological Series Contained in the Museum of the Royal College of Surgeons of England, Vol. 1. Taylor and Francis, London PENNY,D. ANDM. D. HENDY.1985. The use of tree comparison metrics. Syst. Zool. 34: 75-82. PRAGER,E. M. ANDA. C. WILSON. 1976. Congruency of phylogenies derived from different proteins. A molecular analysis of the phylogenetic position of cracid birds. J. Mol. Evol. 9: 45-57. RICHARDSON, B. J., P. R. BAVERSTOCKAND M. ADAMS. 1986. Allozyme Electrophoresis. A Handbook for Animal Systematics and Population Studies. Academic Press, Orlando, Florida. SELANDER,R. K., M. H. SMITH,S. Y. YANG,W. E. JOHNSON,ANDJ. B. GENTRY.1971. Biochemical polymorphism and systematics in the genus Peromyscus. I. Variation in the old-field mouse (Peromyscus polionotus). Stud. Genet. VI. Univ. Texas Publ., 7103: 49-90. SHAW,C. R. ANDR. PRASAD.1970. Starch gel electrophoresis of enzymes-a compilation of recipes. Biochem. Genet. 4: 297-320. SICILIANO,M. J. ANDC. R. SHAW.1976. Separation and visualization of enzymes on gels. In I. Smith (ed.). Chromatographic and Electrophoretic Techniques, Vol. 2, Fourth edn. Heinemann, London, pp. 185-209. SPRINGER,S. 1950. A revision of North American sharks allied to the genus Curcharhinus. Amer. Mus. Nov., 1451; 1-13. SPRINGER,S. 1951. Correction for “A revision of North American sharks allied to the genus Curcharhinus”. Copeia 3: 244. STROMER,E. 1905. Die fischreste des mittleren und oberen eocans von Agypten. Beitr. Pal. Oesterr. 18: 37-58, 163-92. SWOFFORD, D. L. 1990. Phylogenetic Analysis Using Parsimony (PAUP), version 3.0k. Computer software. Illinois Natural History Survey, Champaign. WHITE, E. I. 1955. On Lyamna eulybuthrodon. Blake. Ann. Mag. Nat. Hist. 8: 191-193. WHITLEY,G. P. 1934. Notes on some Australian sharks. Mem. Queensl. Mus. 10: 180-200. WHITLEY,G. P. 1939. Taxonomic notes on sharks and rays. Aust. J. Zool. 9: 227-262. WHITLEY,G. P. 1943. Ichthyological descriptions and notes. Proc. Linn. Sot. N.S.W., Sydney. PHYLOGENYOF REQUIEM AND HAMMERHEAD SHARKS Appendix 315 1. Material Examined Curcharhinus acronotus: 454791, 60516.3 (LORAN), Charleston, S. Carolina, N. America (372, 391); Bulls Bay, Nr Charleston, S. Carolina, N. America (373); Dauphin Is., Alabama, N. America (163); Mayport, Jacksonville, Florida, N. America (442). Curcharhinus dbimurginatus: Mercedes, Camarines Norte, Philippines (287) Curc~urhinus ultimus: Norfolk Canyon, 60 miles east of Virginia Beach, N. America (110, 111, 116, 122,569). Curcharhinus umblyrhynchos; Mercedes, Camarines Norte, Philippines (288, 289); Nihoa Island, (N. W. Hawaiian Islands), Hawaii (545); Tumalutab Is., Zamboanga City, Mindanao, Philippines (327,328). Curcharhinus bruchyurus; Saint Kilda, Vincent Gulf, S. Australia (003). Curcharhinus bre+innu; Mayport, Jacksonville, Florida, N. America (433) ; Ponce Inlet, Daytona, Florida, N. America (445, 450, 452, 461). Carchurhinus fulc+rmis: Moro Gulf: S. of Mindanao (via Arena Blanco, Zamboanga), Philippines (329, 330, 332, 333, 339); Port of Spain fish mkt. Trinidad (474,506,506,508,509). Curcharhinus gulupugensis: Raena Point, Oahu, Hawaii (128, 130, 132) ; Nihoa Island, (N. W. Hawaiian Islands), Hawaii (544); seaward side of Rabbit Island, Oahu, Hawaii (572). Curchurhinus is&n: Bulls Bay, Nr Charleston, S. Carolina, N. America (409, 410, 411,41la, 411b). Carchurhinus Zeucus; Dauphin Is., Alabama, N. America (100); Mayport, Jacksonville, Florida, N. America (436); Ponce Inlet, Daytona, Florida, N. America (457,458); sand shoal inlet, eastern shore, Virginia, N. America (370). Curcharhinus limbatus: Dauphin Is., Alabama, N. America (154, 156); Ponce Inlet, Daytona, Florida, N. America (444, 446); sand shoal inlet, eastern shore, Virginia, N. America (368). Curcharhinus longimanw: 19 10 N, 160 51 W (off Hawaii) (574); 19 15 N, 160 49 W (off Hawaii) (541, 542, 543); off Nihoa Island, (N. W. Hawaiian Islands), Hawaii. (573). Carchurhinus mdoti: Turtle Islands, Sulu Sea. (Navotas mkt, Manila), Philippines (284). Curcharhinus melunoperus: Anda, Pangasinan. (Alaminos mkt), Philippines (261, 262); Straights of Tiran, Gulf of Aqaba, Red Sea (005,006). Curchurhinus obscurus: Dauphin Is., Alabama, N. America (95,99); Norfolk Canyon, off Virginia Beach, Virginia, N. America (567, 568); Ocean City, Maryland, N. America (246). Curcharhinusperezi: Deep Water Caye, Grand Bahama Is., Bahamas (233, 234, 235); Glover’s Reef, Dangriga, Belize (179,180). Curchurhinus plumbeus: Makapu Point, Oahu, Hawaii (555, 556); Maunalua Bay, S. Oahu, Hawaii (570,571); Mayport, Jacksonville, Florida, N. America (437). Curcharhinus pmosus: Manzanilla Bay, eastern Trinidad (495,496,497, 498,499). Curchurhinus seuki: Turtle Islands, Sulu Sea. (Navotas mkt., Manila), Philippines (273,274,275, 276,277). Curchurhinus sorruh: Navotas mkt, Manila, Philippines (origin probably Palawan) (259, 358); S. China Sea coast of Palawan (Navotas mkt Manila), Philippines (296,297) ; Turtle Islands, Sulu Sea. (Navotas mkt, Manila), Philippines (300). 316 G.J. P. NAYLOR Curcharhinus whpelen‘: El Tur, South Sinai (in Straits of Tiran). side of Tiran Is. at “Melba” Egypt (008). Galpocerdo ruvieri: Dauphin Is., Alabama, Hawaii (552,553,554); Hemigalew Ocean mucrostoma: Turtle N. America City, Maryland, Islands, (248); E<gypt (004): Makapu N. America South Point, Oahu, (239). Sulu Sea. (Navotas mkt. Manila), Philippines (314, 315, 316, 317, 318). microstoma: Visayan Sea. (Cebu City mkt), Philippines Hemigahs Loxodon mucrmhinus: Navotas Fish landing, Sulu Sea. (Navotas mkt, Manila), Philippines Neguption breuirosttis: off Cosgrove, Panama City, Florida, Manila, Florida N. America NewJersey, Rhiqbrionodon 217); Turtle N. America acutus: Freetown, Islands, Sulu (324, 325). (258) ; Turtle Islands, (285, 285). Keys, Florida, N. America (558, 560); (75). Prionace gluucu: 19 15 N, 160 49 W. (200 Pleasant, Philippines m SSW of Oahu, Hawaii (540) ; Point (416). (Sierra Sea, Fisheries (Navotas mkt, by catch), Manila), Sierra Leone Philippines (215. (272. 312, 313). Rhiwpionodon lalundii: Port of Spain fish mkt. Trinidad (485, 486, 487, 488, 489) Rhizoptionodon porosus: Port of Spain fish mkt, Trinidad Rhizoptionodon N. America terrunouae: 454’791. (37.5, 376, 390); Sphyma huini: American Is., Alabama, Panama Shoals, N. America City, Florida, $Lh~rnu mockawan: Jacksonville, Is., Alabama, Florida (257); Dauphin Is., N. America City. Charleston, N. America Keys, Florida, Ocean N. America Florida, (LORAN), 60516.3 Dauphin (511). N. America Maryland, S. Carolilla, (164). (539); Dauphin N. America (24.5): (08 1) . Alabama, (443); N. America Panama (101, City, Florida, 256): Mayport. N. America (089, 090). ~phymu tihro: Manzanilla L$dqwzu hdes: Manzanilla mkt, Trinidad Bay, eastern Trinidad (500, 50 1) . Bay, eastern Trinidad (502, Sphyrna zygaenu: Ocean City, N. America Triaenodon obesus: Between Island, 503, 504); Port of’ Spain fish (49 1) . Tiran (N. W. Hawaiian (244). and Sinafir Islands), Hawaii Is., Gulf of Aqaba, (546, 551); South Egypt (010); Nihoa side of Tiran Egypt (009). (Three-digit numbers in parentheses correspond to assigned field numbers) Is., PHYLOGENY OF REQUIEM AND HAMMERHEAD SHARK!3 Appendix 2a. Character Matrix Based on Electromorph Autapomorphic Presumed locus” 123456789 C. acronotus C. albimarginalus 6 altimus EGKFKI EFMKENBDJ M E I E C J E F Q E F H E E M,J ELMJ E G M C H M E H D C. amblyrhynrhos C. brachyuruh C. bmtipinna C faln@rnu C. galapagpnsis Cz’.isodon C. leuras C. limbatus C. longimanus (.‘. marloll (: melanoptenl, (.A obscuruc (.‘. p flezi C plumba,Puc C. porosus C. st7ab-i C: .sorrah (.Y VJhrPlen C. cuvien FL macrostoma H mirrostomn I.. macrorhinus N. brpoiro.~tns I’. glauca R. acufu.\ R. lalandii R. porosus R. twranovar .s. kwini S. mockarran .S. tibum $. tudrr S. zygaeno 7: obessu.r E.J D E E A H 12 MJ F K F F E H F M E I E B N B C 0 E H M E C J F ? A HMBFAF H N B F E L. F ERKMI E I. M F D C A D C A D C A D C G H G G H G C K G G K G G H G F E E 12 13 14 I5 (; 0 A D B If C K I N N K S I‘ Ii K (: (: I’ 0 1: I’ V \’ X LV H \I’ 1 % % I. 1: bl M R 1 (: F F F J F H F C (; F F F F F F E’ F E G’ B B D D D D c; B G (; K B E E E C C (: C: c (: (: <: F I I I I I I I I H ” (; G (: D G (: (: (: B 0 B B L hl E J B I B H A <; I. c: I. F 0 B P B P B P B A N (1 P A S A S (: H X4 E C (: (: B C C (: (1 (: (: A .A <: c: I F F F F G c: E H G C: I I I (; I ; (; (; J ‘ii I <; I I I E E E (: I 1 D D D D B F .4 B F I (; B (k A D D (i <. B C: B B B B B B B B B B B B B B B B B D <; I A B B E E F E H H H H H B F D D D F F D D D F D D F F F D D D D G A A H Z F E E E E <: B (: (: B F J D D D D D D D D D D 1 D I> J ‘D 1) I) D K H H (: E A B B B B (; (; (; (; G F F F F F K I N N N N N D K f: N B F F (: B K H H H H B D C C A F J G G G C: H H H H H E .J II I I> K I K K F F K D H I E P K MJ 10 C N N ? E E N K N J 1 Including D K D K K I I K K F G F F Differences, States N H <: I (: I D D (: (: G F (; 1; D (; “ (I) EST-I; (2) EST-2; (3) GOT-I: (4) GOT-2; (5) LDH-H: (6) ME-l: (7) ME-2 (8) SOD-I: (10) (YGPD; (11) MDH; (12) ID”: (13) CA; (14) PEP; (15) GDH; (16) GI.0: (17) IAP. ! b 1) (: (: (: ; b: 16 I7 :i c: (1 H (: D (: (; ”A .\ A 11 A .\ A .1 :i (: E ” .I .i .\ 1 .\ ‘(: (: (’ \ \ -\ ( \ (’ ”\ c. \ \ 1 ) \ J F \ \ ( ,\ (: .\ (. \ ,\ (. (: (: A K .A :I A I :\ .\ :\ (: ( 1\ 1% \ .\ \ \ ,i 4 \ ( \ I; \ (91 %)1)-L’: G. J. P. NAYLOR 318 Appendix 2b. Character Matrix Numerically Coded with Autamorphic States Removed [Final, asymptotic character weights used for successive weighting (Fan-is, 1969) are @en for each locus] Presumed locusa 1 c. aLronotus C. albimqinatus C. altimus C. amblyrhynchos C. brachyurus C. breoipkzna C. /al@rmis C. galupagensis C. isodon c. leuca.5 C. limbatus C. kmgimanus C. ma&i C. melunoptews C. obscwus C. perezi C. plumbeus c. porosus C. se&i C. sorrah C. wheeleri G. cwieri H. macrostoma H. microstoma L. mucwrhinus N. brevirostris P. giuuca R acutus R lalandii R porosus R terranovae s. lewini S. mockarran s. t&u70 S. tudes S. zygakma T. obesus 2 5 ? 2 7 4 ? 2 5 ? 142323 2455261?? 3 1 4 275276325? 165?33 2145161253? 4 2 4 ? 2 7 ? 3 2 5 3 142??? 2 4 ? 2733253142? 235463? 2 5 2 1 6 2954661253162433 2 5 5 ? 752253142? ? 6527622531426?3 2 6 ? 2 3 ? 2 3 ? 3142633 5 4 6 6 1 2 ? 253162433 ? ? ? ? ? 6125,l 4 2 6 ? 5 7 6 ? 25314?? 295466 1 2 5 3 1 6 2 6 ? 4761??? 162433 2752763253165133 2 ? ? 2 3 ? ? 2 5 3 1 4 ? 1 ? 2352253142?35 265273? 2 5 ? 1 4 214516 1 2 5 3 1 4 3???5?? 23,?????1 5?12??11 3 2 ? 1 5 ? 12???1 1 3 2 ? 33?2???2?3???? 3 3 ? 5 ? 4 ? 2531’???3 2955????5?1 6 32234?1 322251? 1 223471322251? 1 223471321?51821 1223471 1251821 3 2 463?5144 13333?13 4 6 3 ? 323??3 5 2 ? 4 4 ? 4 8 3 1 5 1 4 ? ? 3333512 4 8 3 1 5 1 4 ? 13333512 4 6 3 ? 442323?13 5 2 ? 3 3 ? 2 2 ? 225316???4 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 2 ? 3 3 2 2 3 3 2 1 3 ? 3 ? 2 ? ? 5 3 2 ? 2 4 3 3 3 2 3 3 3 2 2 ? 2 3 3 3 3 4 4 7 7 2 2 2 6 6 4 ? ? ? 2 2 3 1 1 a (1) EST-l, wt = 1000; (2) EST-2, wt = 489; (3) GOT-l, wt = 1000: (4) GOT-2, wt = 564; (5) LDH-H, wt = 533; (6) ME-l, wt = 762; (7) ME-2, wt = 467; (8) SOD-l, wt = 643; (9) SOD-2, wt = 1000; (10) aGPD, wt = 333; (11) MDH, wt = 1000; (12) IDH, wt = 778; (13) CA, wt = 1000; (14) PEP, wt = 1000: (15) GDH, wt = 1000; (16) GLO, wt = 729.