BULLETIN OF MARINE SCIENCE, 65(2): 337–379, 1999 REVISION, PHYLOGENY AND COMMENTS ON BIOGEOGRAPHY OF SOAPFISHES OF THE GENUS RYPTICUS (TELEOSTEI: SERRANIDAE) Ricardo Zaluar Passos Guimarães ABSTRACT Soapfishes of the Atlanto-East Pacific genus Rypticus are reviewed. Rypticus macrostigmus Courtenay and R. brachyrhinus Courtenay are junior synonyms of R. bornoi Beebe and Tee- Van and R. randalli Courtenay respectively. A phylogenetic analysis of the genus based on outgroup comparison with other members of the tribe Grammistini indicates Rypticus is a monophyletic lineage supported by six modifications in axial-skeleton structure and reduc- tion in the number of infraorbital bones. A single, fully resolved tree illustrates the hypoth- esized relationships among species. Some of the cladogenetic events hypothesized based on morphology are corroborated by vicariant biogeography. The tropical to warm-temperate marine percoid genus Rypticus Cuvier comprises the Atlanto-East Pacific serranids commonly referred to as soapfishes, alluding to their ca- pacity to produce large amounts of skin mucous when stressed. This mucous contains a strong toxin, grammistin, which is repulsive for potential predators (Randall et al., 1971). Soapfishes are small to medium-sized cryptic inhabitants of reefs, where they feed pri- marily at night on small crustaceans, mollusks and fishes (Graham, 1972; Hobson, 1965; present study). They are protogynous hermaphrodites (Smith, 1965) and have planktonic eggs and larvae (Houde, 1982; Kendall, 1984). Rypticus was originally described more than 160 yrs ago and its position among other percoids varied considerably in proposed schemes (compare, for example, Smith and Atz, 1969: fig. 3; Kendall, 1976: fig. 1). Johnson (1983) cladistically defined the Serranidae and included Rypticus, some other soapfish genera (i.e., Grammistes, Pogonoperca and Grammistops) and the pseudogrammid genera (i.e., Pseudogramma, Suttonia and Aporops) in tribe Grammistini of the Epinephelinae. Subsequent studies (e.g., Kendall, 1984; Johnson, 1988) have continued to detail the phylogenetic structure of the Serranidae. Finally, Baldwin and Johnson (1993) considered Rypticus as the immediate sister-group to (Grammistops + (Jeboehlkia + (Aporops + (Pseudogramma + Suttonia)))), this whole group as the sister-group of (Grammistes + Pogonoperca), all in a monophyletic Grammistini. Intrageneric relationships of Rypticus were only previously hypothesized by Courtenay (1965, 1967) and McCarthy (1979), but neither author included all species in their analysis, and these were based on simple observations of shared characters. Eleven species of Rypticus are currently recognized, eight from the Atlantic (Courtenay, 1967) and three from the Eastern Pacific (McCarthy, 1979). Since reexamination of type specimens and/or topotypes indicated that some of these are invalid and because a sig- nificant amount of specimens and information have been collected and published since these last reviews, a complete taxonomic review of the group is warranted. Monophyly of the genus and relationships among species are hypothesized based on a cladistic analysis of characters. Possible vicariant scenarios are discussed for geographi- cally isolated clades with a sister-group relationship. 337 338 BULLETIN OF MARINE SCIENCE, VOL. 65, NO. 2, 1999 METHODS COUNTS AND MEASUREMENTS.—Following Courtenay (1967) and McCarthy (1979), counts and measurements were taken as described by Hubbs and Lagler (1949), with the tip of the upper jaw as the anterior landmark for snout, head, predorsal and standard lengths. Counts of vertebrae, supraneurals, total number of caudal rays and number of co-ossified proximal and middle radials in dorsal and anal fins are from radiographs or cleared and stained specimens. Counts of dorsal and anal-fin rays were taken from radiographs, cleared and stained or dissected alcohol specimens. In the descriptions, each meristic range is followed by the modal value in parentheses. Information on the number of gill rakers on the first arch and pectoral-fin rays are based on Courtenay (1967) and McCarthy (1979). In the descriptions and tables 1–2, 4–6, data for number of dorsal-fin spines, dorsal-fin rays, caudal vertebrae, anal-fin rays, predorsal distance, head length, snout length and upper jaw length of Atlantic species are a combination of the present data and those of Courtenay (1967). OSTEOLOGY.—Osteological terminology mostly follows that of Johnson (1983, 1984) and Baldwin and Johnson (1993). Osteological studies were performed primarily on specimens cleared and coun- terstained (cs) for bone and cartilage (Pothoff, 1984). Additionally, radiographs (r) and dried skel- etons (ds) were also examined. TYPES.—Among nominal species with extant type material, those of R. maculatus Holbrook, R. microps Castelnau, R. coriaceus Cope and R. bicolor Valenciennes were not examined in the present study, but examination of their original descriptions and/or topotypes indicates the synonymies of Courtenay (1967) and McCarthy (1979) should be followed. MAPS.—Locality records plotted on maps are those listed in the “Appendix: Material Examined” section and those cited in the following publications: Bullock and Smith (1991), Caldwell (1963), Courtenay (1967; 1981), Eskinazi and Lima (1968), Hetzel and Castro (1995), Hoese and Moore (1977), Koike and Guedes (1981), Mago-Leccia (1965), Martins-Juras et al. (1987), Maugé (1990), McCarthy (1979), Oliveira (1974; 1979), Rodriguez (1973), Rosa (1980), Roux (1973), Smith (1976) and Starks (1913). Records of R. saponaceus from Arraial do Cabo, state of Rio de Janeiro and Guarapari, state of Espírito Santo and that of R. subbifrenatus from the city of Rio de Janeiro are based on personal field observations. ABBREVIATIONS.—Institutional abbreviations are as listed in Leviton et al. (1985), except for UFRJ which refers to the fish collection at the Laboratório de Ictiologia Geral e Aplicada, Universidade Federal do Rio de Janeiro. PHYLOGENETIC ANALYSIS.—The character analysis conducted in this study follows the basic pre- mises of phylogenetic systematics (e.g., Hennig, 1966; Wiley, 1981), i.e., taxa are grouped based on shared apomorphies. Polarity of character states was determined by outgroup comparison (see Nixon and Carpenter, 1993). This was primarily accomplished via direct observation of other Grammistini available for examination. For those characters with more than one state in the outgroups (all Grammistini genera except Rypticus), its plesiomorphic condition in the ingroup (Rypticus) was determined with the aid of Maddison and co-workers’ (1984) algorithm applied upon Baldwin and Johnson’s (1993) phylogeny of the Epinephelinae. Because expansion of branches can alter topology of a previous tree, a simultaneous analysis of the entire Grammistini plus all species of Rypticus was conducted. This was done by adding to the matrix (Table 11) all those characters that contribute directly to the proposed Grammistini phylogeny, i.e., characters 16, 21, 32, 33 and 36 to 52 of Baldwin and Johnson (1993). Regarding these characters, the present work either supports or does not have any evidence to consider their states incorrectly assigned to the outgroups by those authors, except as noted in the text. A parsimony analysis in which reversals and convergences are permitted and counted equally as a single step was conducted by means of the “ie” algorithm of Hennig86, which certainly generates the tree with a minimum length (Farris, 1988). This was done with the aid of the Tree Gardener matrix and tree editor (Ramos, 1996). GUIMARÃES: REVISION OF RYPTICUS 339 Table 1. Frequency distribution of number of dorsal spines (A) and dorsal rays (B) in Rypticus. AB III IVI I021222324252627282 R. bicolor -31 - ----2---- R. bistrispinus 57 -- ----12 320211 R. bornoi 22-5 ----18012- R. courtenayi 12 -- ----26--- R. maculatus 37 4- --1-8181- 7 R. nigripinnis 11-8 -----23-- R. randalli 977 - ---10 927--- R. saponaceus -79 - ---16 948--- R. subbifrenatus -1113 1141444461--- Table 2. Frequency distribution of total number of vertebrae (A), number of anal rays (B) and total number of caudal rays (C) in Rypticus. AB C 254 23141516171425262 R. bicolor 3- ----2 -2- R. bistrispinus -63--71421 34- R. bornoi 1-5 -13216 32- R. courtenayi 8- - -36- 134 R. maculatus 318 114291 --1 R. nigripinnis 5- --121 -2- R. randalli 4-6 -2127- -53 R. saponaceus 7-6 -145442 -1- R. subbifrenatus 1-1206 158 7- 562 Table 3. Frequency distribution of number of pyloric caeca (A) and number of preopercular spines (B) in Rypticus. AB 345 12345 R. bicolor -31 12 3 3-1 R. bistrispinus 4-- --321 7 R. bornoi 31- -912- 3 R. courtenayi -11 21 2 --- R. maculatus --1 771 --- R. nigripinnis 4-- 83 0 --- R. randalli 7-- 288-- 0 R. saponaceus -3- -44-- 4 R. subbifrenatus 3--19 237-- POLYMORPHISM.—Characters 13, 14, 15 and 18 have one or more taxa with more than one state. Because HENNIG86 does not support a “01” entry, each polymorphism was treated differently in order to avoid loss of information (e.g., Nixon and Davis, 1991). For example, three species pre- sented polymorphism for character 15 (i.e., have two or three opercular spines), but each has a different frequency distribution of states. While R. randalli typically has two spines, R. nigripinnis and R. maculatus typically have three. Because reduction in number of opercular spines is a novelty within the Serranidae, it was assumed that presence of three spines (state 0, coded “0” in the ma- 340 BULLETIN OF MARINE SCIENCE, VOL. 65, NO. 2, 1999 5 52 51 40 49 48 ----1-- ------- 47 1 5221-1 46 131----- 22 45 112------ 29 . s u c i t p 44 1 532- - -1- - 13 14 y R n i h 43 1461------- 15 t g n e l d 42 18 r a d n a t 193---------- 41 13 s f o t n -1-1-11 -2-2221 e 15 30 c r e p n 14698 0 i 10 39 e c n a t s 38 i d l a s r 37 o d e r p f 36 o n o i t 4 u -111 2 ---112-15-1-------- ----- -----1493 ---- --------1241 ------5387 ----- 1--145 0 35 b i r t s i d y c n s e u u s t s s i q u a u i s y e e n n n u a r i i t c e n l n p r F i r a l i a f l s e o p .