Phylogenetic relationships and taxonomic revision of the Blenniid Fish genera Cirripectes and Scartichthys


Material Information

Phylogenetic relationships and taxonomic revision of the Blenniid Fish genera Cirripectes and Scartichthys
Physical Description:
vii, 314 leaves : ill. ; 28 cm.
Williams, Jeffrey Taylor, 1953-
Publication Date:


Subjects / Keywords:
Blenniidae   ( lcsh )
Fishes -- Classification   ( lcsh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1986.
Includes bibliographical references (leaves 307-313).
Statement of Responsibility:
by Jeffrey Taylor Williams.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000880497
notis - AEH8316
oclc - 14952835
System ID:

This item is only available as the following downloads:

Full Text

Cirripectes AND Scartichthys







During my work on this project, many people have assisted me in

various ways. I thank all of them. In particular, I thank W. F.

Smith-Vaniz, E. Bohlke, W. Saul, and staff (Academy of Natural Sciences

of Philadelphia); P. C. Heemstra (J. L. B. Smith Institute of

Ichthyology); T. Iwamoto, W. N. Eschmeyer, L. Dempster, and staff

(California Academy of Sciences); R. Winterbottom and staff (Royal

Ontario Museum); C. R. Gilbert, G. Burgess and J. B. Miller (Florida

State Museum); K. Hartel (Museum of Comparative Zoology, Harvard); M.

L. Bauchot and M. Desoutter (Museum National d'Histoire Naturelle); A.

G. Gerberich, K. A. Bruwelheide, L. P. Norrod, S. L. Jewett, J. R.

Gomon, and Victor G. Springer (Smithsonian Institution), and the staff

at all other museums that provided specimens, radiographs, and/or other

technical assistance. I thank V. G. Springer, W. F. Smith-Vaniz, and

H. Bath for their gracious hospitality while I was visiting museums.

For providing color transparencies of specimens, I thank J. E. Randall,

V. G. Springer, R. Winterbottom, and B. Carlson.

C. R. Gilbert and V. G. Springer generously gave their time to

discuss my project. I gratefully acknowledge their help.

Financial support for this study was provided in part by the

Department of Zoology and Florida State Museum, University of Florida;

the California Academy of Sciences; and the National Science Foundation

(Predoctoral Improvement Grant DEB 8207313). I thank my parents for

their support throughout the study.

Finally, I thank my wife, Karen, for her support, editorial

comments, and seemingly limitless understanding.



ACKNOWLEDGEMENTS ........................................ ii

ABSTRACT .......................................................... vi

INTRODUCTION ............................................ ......... 1

HISTORICAL TREATMENT OF THE BLENNIIDAE..................... ...... 6

METHODS........... ...... ...** .......*** ... ..* ............00.. 9
Taxonomic Methods....................................... ... .... 9
Methods of Phylogenetic Analysis................................ 10

PHYLOGENETIC ANALYSIS............................................. 12
Family Blenniidae............................ .................. 15
Salariini-"Blenniini" Group ...................................... 18
"Blenniini"..................................................... 19
Salariini ........................................ ................ 20
Altius Group........................................**** ******** 28

Node I Group...........................................**********. 29
Genus Istiblennius..................................... ..***... 32
Node II Group ........................... ......................* 32
Genus Entomacrodus...................... .... ................** 36
Node III Group................................. ...............** 36
Genus Stanulus ........ ..... ...... 000 ...... .. ....**** ....************. 37
Node IV Group.......... ..............**..... ... ..... .... 38
Genus Pereulixia..................... .......... ...*.......**** 45
Node V Group. ......... .............................. ............. 49
Genus Exallias.................. ..... ....... .........***** ****** 54
Node VI Group.... ............................................... 54
Genus Cirripectes.......................................... ..... 57
Node VII Group......... ....... ...... ....... ... ................ 61
Genus Ophioblennius.................................**.*********. 61
Genus Scartichthys................................... 62


Node VIII Group.................... .......................********** 68
Node IX Group.........................................................*************... 73
Node X Group. .................. ........................***************** 73

Node XI Group ............................... ...................... 76
Node XII Group.. ..... ......... ................... .... ...... .... 76
Node XIII Group....... ..... ..................................... ... 77
Node XV Group................................................... 80
Node XVI Group.................................................. 81
Node XVII Group.................................................. 82

Cirripectes SWAINSON............................ .................. 87
Key to the Species of Cirripectes................. ............ 109
Cirripectes alboapicalis (Ogilby)................................. 143
Cirripectes auritus Carlson.................................... 149
Cirripectes castaneus (Valenciennes)............................. 155
Cirripectes chelomatus Williams and Mauge......................... 165
Cirripectes filamentosus (Alleyne and Macleay).................. 170
Cirripectes fuscoguttatus Strasburg and Schultz.................. 180
Cirripectes gilbert new species................................. 184
Cirripectes hutchinsi new spepies............................... 189
Cirripectes imitator Williams................................... 195
Cirripectes jenningsi Schultz.......... ....................... 199
Cirripectes kuwamurai Fukao...................................... 203
Cirripectes obscurus (Borodin)................................. 205
Cirripectes perustus Smith....................................... 209
Cirripectes polyzona (Bleeker)................................. 213
Cirripectes quagga (Fowler and Ba)................... ............ 219
Cirripectes randalli new species................................ 226
Cirripectes springer new species ............................... 230
Cirripectes stigmaticus Strasburg and Schultz.................... 234
Cirripectes vanderbilti (Fowler)................................ 241
Cirripectes variolosus (Valenciennes) ....,............ ......... 246

Scartichthys JORDAN AND EVERMANN.................................. 254
Key to the Species of Scartichthys.............................. 260
Scartichthys crapulatus new species............................. 265
Scartichthys gigas (Steindachner)................................ 267
Scartichthys variolatus (Valenciennes) ......................... 272
Scartichthys viridis (Valenciennes).............................. 276

Exallias JORDAN AND EVERMANN...................................... 282
Exallias brevis (Kner)........................................... 283

HISTORICAL ZOOGEOGRAPHY........................................... 291

APPENDIX................ ................................ .......... 302

LITERATURE CITED. ................................ ................. 307

BIOGRAPHICAL SKETCH................................................ 314

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

Cirripectes AND Scartichthys


Jeffrey Taylor Williams

May 1986

Chairman: Carter R. Gilbert
Major Department: Zoology

The blenniid fish tribe Salariini comprises a group of marine shore

fishes found circumglobally in tropical and subtropical seas. Members

of the tribe are quite similar ecologically, with most being herbivores

or deposit feeders and inhabiting rocky or coralline areas.

The salariin genera Cirripectes, Scartichthys, Exallias,

Ophioblennius, and Pereulixia form a monophyletic group closely related

to the genera Stanulus, Entomacrodus, and Istiblennius. A phylogenetic

analysis supports the hypothesis that each of the following genera is

the sister genus to a combination of all preceding genera:

Scartichthys, Ophioblennius, Cirripectes, Exallias, Pereulixia,

Stanulus, Entomacrodus, and Istiblennius (including at least the I.

gibbifrons species complex). A hypothesis of the relationships among

the species of Cirripectes is illustrated in a cladogram.

The genera Cirripectes and Scartichthys are taxonomically revised.

Cirripectes occurs throughout the Indo-Pacific region and comprises the

following 20 species: C. alboapicalis (Australia to Easter Island), C.

auritus (western Indian Ocean to Line Islands), C. castaneus (Indian

Ocean and Red Sea to western Pacific Ocean), C. chelomatus (Coral Sea

to Tonga Islands), C. filamentosus (western Indian Ocean to western

Pacific Ocean), C. fuscoguttatus (Pacific Plate), C. gilbert new

species (Indian Ocean), C. hutchinsi new species (western Australia),

C. imitator (Taiwan to Japan and Ogasawara [Bonin] Islands), C.

jenningsi (Gilbert Islands to Tuamotu Archipelago), C. kuwamurai

(Japan), C. obscurus (Hawaiian Islands), C. perustus (western Indian

Ocean to Gilbert Islands), C. polyzona (western Indian Ocean to Line

Islands), C. quagga (western Indian Ocean to Pitcairn Island Group), C.

randalli new species (Mauritius and Cargados Carajos Shoals), C.

springeri new species (Philippines to Solomon Islands), C. stigmaticus

(western Indian Ocean to Samoa), C. vanderbilti (Hawaiian Islands and

Johnston Island), and C. variolosus (Pacific Plate).

Scartichthys is restricted to the eastern Pacific Ocean and

comprises four species: S. crapulatus new species (northern Chile and

vicinity of Valparaiso), S. gigas (Panama to northern Chile), S.

variolatus (San Felix, San Ambrosio, and Juan Fernandez Islands), and

S. viridis (vicinity of Valparaiso to southern Peru).

Area cladograms derived from the phylogenetic analyses are

analyzed. Eleven geographic patterns indicated by the cladogram are



The blenniid fish tribe Salariini comprises a group of marine shore

fishes found circumglobally in tropical and subtropical seas. Members

of the tribe are quite similar ecologically; most are herbivores or

deposit feeders and inhabit rocky or coralline areas.

The objectives of the present study are to 1) test the hypotheses

that the genera Cirripectes, Scartichthys, Ophioblennius, Pereulixia

and Exallias comprise a monophyletic group and are each monophyletic;

2) if the first hypotheses are confirmed, form phylogenetic hypotheses

of interrelationships of the species of Cirripectes; 3) taxonomically

revise Cirripectes and Scartichthys, and provide complete species

descriptions of and distribution maps for each of the included species;

and 4) analyze the distribution patterns of these genera and their

included species.

Smith-Vaniz and Springer (1971) gave a synopsis of the Salariini to

provide a foundation for future work on the tribe. They pointed out

several genera that needed additional study and provided a preliminary

hypothesis of the relationships of the 24 genera they recognized in the

tribe. Their method was not cladistic, but they did establish the

monophyletic nature of the tribe Salariini. Although they did not

cladistically analyze the members of the tribe, they provided a

dendrogram (Smith-Vaniz and Springer, 1971: fig. 51) of the genera

showing their interrelationships. This dendrogram suggests that the

genera Cirripectes, Scartichthys, Exallias, Pereulixia, and

Ophioblennius are more closely related to each other than to other

members of the tribe.

The genus Cirripectes Swainson is widespread throughout the

Indo-Pacific. It is a common inhabitant of coral and rocky reefs and

is usually common in heavy surge areas. Specimens of Cirripectes are

common in museum collections, but complex nomenclatural problems and

the lack of adequate keys to the species have resulted in numerous

misidentifications. Smith-Vaniz and Springer (1971) listed 24 nominal

species of Cirripectes, and four additional species have since been

described (Table 1). There has been no comprehensive revision of the

genus and all published keys to species have been of a regional nature

(Chapman, 1951, for the Indo-Australian Archipelago; Fukao, 1984, for

Japan; Schultz and Chapman, 1960, for the Marshall and Mariana Islands;

Smith, 1959, for the western Indian Ocean). As a result, numerous

populations from different geographic areas have been described as new

species, even though these populations are actually members of the same

widespread species. There is also sexual polychromatism in some

species and, in many instances, each color morph has been named as a

distinct species. Like most genera of marine shore fishes, there has

been no phylogenetic analysis of the interrelationships of the

Cirripectes species.

Table 1. List of nominal species of Cirripectes with correct
identifications. Arrangement is alphabetical by species name.

Species, Author, Publication Date

Correct Identification

Salarias alboapicalis Ogilby, 1899
Cirripectes auritus Carlson, 1981
Blennius canescens Garman, 1903
Ophioblennius capillus Reid, 1943
Salarias castaneus Valenciennes in Cuv. and
Val., 1836
Cirripectes chelomatus Williams and Mauge, 1983
Ophioblennius clarki Reid, 1943
Salarias cruentipinnis Day, 1888
Cirripectes cruentus J. L. B. Smith, 1959
Salarias filamentosus Alleyne and Macleay, 1877
Cirripectes fuscoguttatus Strasburg and Schultz, 1953
Cirripectus gibbifrons J. L. B. Smith, 1947
Cirripectes imitator Williams, 1985
Cirripectes indrambaryae H. M. Smith, 1934
Cirripectes jenningsi Schultz, 1943
Cirripectes kuwamurai Fukao, 1985
Cirripectes lineopunctatus Strasburg, 1956
Salarias nigripes Seale, 1901
Exallias obscurus Borodin, 1927
Cirripectes perustus J. L. B. Smith, 1959
Salarias (Cirripectes) polyzona Bleeker, 1868
Rupiscartes quagga Fowler and Ball, 1924
Cirripectes reticulatus Fowler, 1946
Salarias sebae Valenciennes in Cuv. and
Val., 1836
Cirripectes stigmaticus Strasburg and Shultz, 1953
Ophioblennius vanderbilti Fowler, 1938
Salarias variolosus Valenciennes in Cuv. and
Val., 1836
Cirripectus variolosus patuki De Buen, 1963





Cohen (1956) briefly reviewed the eastern Pacific genus

Scartichthys Jordan and Evermann. He examined specimens of several

nominal species, provided a partial synonymy tor the genus, and

recognized S. gigas and S. viridis as valid species. He did not

attempt to treat all of the nominal species, thus leaving the status of

many species in doubt. Smith-Vaniz and Springer (1971) listed thirteen

nominal species (Table 2) and estimated that there were only two valid

species of Scartichthys.

Ophioblennius Gill (reviewed by Springer, 1962), Pereulixia Smith,

and Exallias Jordan and Evermann, and their individual species, are

relatively well-defined, but their interrelationships within the

Salariini have not been formally hypothesized.

Concerning generic relationships, Smith-Vaniz and Springer

(1971:20) suggested that Scartichthys might be congeneric with

Cirripectes, but deferred combining them until more detailed studies

were made. Their statement (p. 20) about the relationship between

these two genera is inconsistent with statements made elsewhere in

their paper. They suggested (p. 24) that Exallias appeared to be most

closely related to Cirripectes and (p. 36) that Ophioblennius appeared

to be most closely related to Scartichthys. Later in the paper (their

Figure 51), they suggested that Cirripectes and Scartichthys are sister

groups, which are in turn the sister group of Exallias, and that these

three genera together form the sister group of Pereulixia.

Ophioblennius is then shown as the sister group of all of these

genera. This confusion concerning the relationships among these genera

illustrates the need for a complete and thorough analysis of their


Table 2. List of nominal species of Scartichthys with correct
identifications. Arrangement is alphabetical by species name.

Species, Author, Publication Date

Correct Identification

Salarias concolor Philippi, 1896
Salarias cuvieri Gunther, 1861
Salarias eques Steindachner, 1898
Ophioblennius fernandezensis Clark, 1938
Salarias gigas Steindachner, 1876
Ophioblennius mazorkae Hildebrand, 1946
Salarias modestus Philippi, 1896
Salarias petersoni Fowler, 1940
Salarias rubropunctatus Valenciennes in Cuv. and
Val., 1836
Blennophis semifasciatus Kner and Steindachner, 1866
Salarias variolatus Valenciennes in Cuv. and
Val., 1836
Salarias viridis Valenciennes in Cuv. and Val., 1836
Ophioblennius xiphiodon Clark, 1938





Springer (1968b) reviewed the literature dealing with

classification of the Blenniidae. He cited numerous studies of higher

level (family or above) relationships, but noted that Norman (1943)

provided the only comprehensive intrafamilial classification. Norman

(1943) recognized three blenniid subfamilies: Ophioblenniinae,

Blenniinae, and Salariinae. Springer (1968b) split Norman's Blenaiinae

into the subfamily Blenniinae, comprised of the tribes Salariini.

Blenniini, and Omobranchini, and the subfamily Nemophidinae and noted

that Norman's Ophioblenniinae was based on larval stage individuals

belonging to several different genera, all but one of which (Somersia,

Labrisomidae) he placed in the tribe Salariini. Springer and

Smith-Vaniz (1972) chose not to recognize blenniid subfamilies,

preferring instead to recognize five tribes in the family Blenniidae

(Blenniini, Salariini, Omobranchini, Nemophini, and Phenablenniini).

Springer (1968b) proposed that Blennius was most closely related to the

Omobranchini and Nemophini line. Smith-Vaniz (1976), noting that the

Blenniini of Springer and Smith-Vaniz (1972) was polyphyletic, removed

Blennius (which would be the true Blenniini) from the Blenniini, and

hypothesized a sister group relationship between Blennius and a group

comprised of the tribes Owobranchini, Phenablenniini, and Nemophini.

Smith-Vaniz (1976) did not propose a hypothesis for the relationship of

his monophyletic group to the remaining members of the Blenniidae,

leaving the Salariini, "Blenniini," and Blennius in an unresolved

trichotomy (his Figure 88).

In his revision of the tribe Blenniini, Bath (1977) commented on

the sutured dentaries of Blennius but retained it in the tribe

Blenniini. Bath stated that the tribe was defined by a group of

unspecialized characters, thus leaving open the possibility that some

of the genera would ultimately be removed.

Springer (1968b) stated that the Salariini were either offshoots of

the Blenniini or shared a common ancestor with them. Smith-Vaniz and

Springer (1971) and Springer and Smith-Vaniz (1972) supported a close

relationship between the tribes Salariini and Blenniini, but proposed

no distinctive characters to support this hypothesized relationship.

Smith-Vaniz and Springer (1971) provided data supporting the

monophyletic nature of the tribe Salariini and discussed in detail the

nature of the dentition and jaws. Although the lateral portions of the

premaxillary capsule are enclosed in certain genera, they found that at

least the mesial portion of each premaxilla was an open capsule (i.e.

no anteroventral wall of bone) in all genera of the Salariini.

Smith-Vaniz and Springer (1971) diagnosed all genera they recognized in

the Salariini and presented their partially intuitive concept of the


relationships among the genera in a dendrogram (Smith-Vaniz and

Springer, 1971:Figure 51).


Taxonomic Methods

Counts and measurements follow Smith-Vaniz and Springer (1971) and

Williams (1985). Specimens used for osteological examination were

prepared by clearing and counter-staining with Alizarin-red S and

Alcian blue following the method of Dingerkus and Uhler (1977). Counts

rf vertical fin rays and characters associated with the vertebral

column (i.e. pleural ribs, epipleural ribs, etc.) were taken from

radiographs. When the last anal-fin ray was split through base, it was

counted as one.

Species accounts are arranged alphabetically by species. Only

primary synonyms are listed. In the material examined section for each

species, the presence of a standard length measurement associated with

a catalog number indicates that counts and measurements were taken from

that specimen(s). All other material listed was examined to confirm

identification and for use in establishing geographic ranges.

Counts of dentary and premaxillary incisors are approximate due to

the difficulty of counting these tiny teeth.

Institutional abbreviations used in listings of material examined

are from Leviton et al. (1985). Other abbreviations used are

SL-standard length, HL--head length, LL-lateral line, AN--anterior

nostril; EIP-extra interorbital sensory pore position;

IFO-infraorbital sensory pore series; MD-mandibular sensory pore

series; MSP--mid-snout pores; PBN-pore positions behind nuchal flap;

PN-posterior nostril; POP-preopercular sensory pore series;

SO-supraorbital sensory pore series; ST-supratemporal sensory pore


Methods of Phylogenetic Analysis

Procedures used to determine relationships between taxa follow the

general principles of Hennig (1966), as discussed by Vari (1978) and

Parent (1981). Only monophyletic groups (i.e. a group that includes

all descendants and only descendants of a hypothetical common ancestor)

are recognized as valid taxonomic units. Each monophyletic group is

defined by shared derived characters (synapomorphies). Polarity of

each character (i.e. the character state toward which the character is

assumed to have evolved in the taxa being analyzed) is determined by

outgroup analysis (Maddison et al., 1984). The principle of parsimony

is used to resolve cases where derived (apomorphic) characters support

conflicting hypotheses of relationship (i.e. homoplasiously distributed

characters). Although evolution may not proceed in accordance with the

principle of parsimony, this approach minimizes the number of

assumptions that must be made to explain character transformations.


The result of a phylogenetic analysis is a hierarchical arrangement

of increasingly inclusive monophyletic groups. This arrangement,

usually illustrated as a branching diagram (cladogram), provides a

hypothesis of the phylogenetic relationships of the taxa analyzed. The

phylogenetic method is the preferred technique for this study as it

provides a clear hypothesis of relationships that can be tested further

by examining other characters.


In this section, I review the synapomorphies of the Blenniidae and

other taxonomic groupings at decreasing levels of universality that

lead to Cirripectes and its allies. Although much work remains to be

done before the relationships of blenniid genera and species are fully

resolved, this discussion will provide data necessary to support my use

of certain taxa as outgroups for determining character polarities.

Characters used to define groupings at higher taxonomic levels (i.e.

family, tribe) are based on a general survey of selected species within

each group. The major emphasis of my study is on Cirripectes and its

allies in the Salariini. The relationships among the genera closely

related to Cirripectes are depicted in the cladogram of Figure 1.

Groupings of genera on this cladogram will be referred to in the text

by the node number. Generic groupings outside the Cirripectes

assemblage of Figure 1 are not shown on a cladogram because the

relationships proposed herein for these taxa are tentative and require

additional study.

The goal of this study is to provide a working phylogenetic

hypothesis of blenniid relationships and should be considered a first

step in the dynamic process of phylogenetic analysis.

0 C3
So a)
I *141 M 'Y *y-i .0
a) 4 W. -r
c Ca I -i w-4 0

-4 *Mi-l 0 B C t -
*-1C I 4-i 5* *-40 0 -*. 4.)

CU4- O 11.0 0 C B
'-4 r a i% &4.J 41 -rH 4-14 0W-4 C
cd uC *cc -4 ,Q ,C C= 0 -4cU.
M -4 44 3t G a 04-JI I
S0 0-41 0 a I 0 00 O -

l e '0 0 ,C 4, 0 .4 3 N CO

(- U *ril CO 4 1 0 C r1 4 -4 C*U m
= 0 I ( > I CO IN Cq 0 C | 1
40 -- C *M W W (3 U 0VC4

" -I CC o C a C S O = 4J4-O

4* -a U CO W C4 .- CC > 0 4 C

<4i C 0 -4 Cr 5- 0 60 4OJ
,i oo 4) 44 Wu WQ 4 M 0 i M -4 'O

CU > (l 3 Wol 0> cc 0i o-
4J iW W 0C 00 4- W4
x0 U -4 0 3 OC 4- CO co O )
0o4 4 O 4e 0 1 0

ca wo t) CL Lw w 0 c Go 0 0 la.e a
w* A0 0 c00 a 4 CU A4 034 CU r
0-> w3 0B 1 a 0 1 > -H 00 C

(a r40 *C C 3"4 0 O -4 W 40 0
a) c a 0 C0 W -4w 0- O

(4 CU I I i WQ O V i O C P 0 "
4- ,4 w C I 4 0 0 U X1 4 0 4
)0 04.O W Xa0 W 4 0 O
a-. (13 4 a) ) 44 -Z 0 w : wC -

*U OM 0 4l B I l *i 0 *il am
.O4 0 0o t O>I 03 r. "-l
-C0. 0 0- CO 0 i 0 0

*r- 04e 4-) w cd ca -44 .O C I >4 I 3 0- -
) U I M L W C 3 = C 0 r 4 0

m m -r4 :r C 0 A e 0 0 C4 C4 0 0 C
144 a $g wu u I 0 U
O 5. 4i 4 c 1 4 a) .. .i
-r40-4 CC a U 41 0 m C cc -k c

0w4 00 WC 14 Cd Ul (v w co 0
ai CU a 0 4 i C CB W UU CB -40 0
W r Iw 0 M 1 3 t c O j 0

0 c 0 0 w w c m w I z

,a4 CU 4J *4 0 C 4J 9-I % W *T C-m 0
4J l IB 03 = C w l 0 C4 0 B

c 01 Cd cCO c0 uc I -a I

"4J S )004 0.2 ) r- 4 0.1 00
I Mm o 1 -i 4U -" C U W i.-

Ca c 3 O. Ca

-4 0 r. *(a m I.0 X C : 0 ca
-r *-4 C -4 O > ca u aW iC-ai

4 =) a) a) w 41 S 4J B w (0 1w
4 00 0 -4 co a) CC C- U (U
W M 0 0 41 0 Au 0) $ U 0 -t COC: >
ao a h 1 -14r r. g Q O I o S ca 0

'VC 0 I CO C (0 0 ( U I CU 04 -
c 4) 1 rn o m Im. S D cs I w I w

CUO t 4J 0 () r M I 0 0 "o0 4 M I 0
4- u 6 S > --i O i 4 Oi

U B ca c4 & 4 1 B1 W 01 1. a .
0 c-a 0 -4N 0 -4 u i C ^ a;; i I-N

G caU at- 0 00 = CX o4 m4j
c c m I I-a C U D ="0-occ0 I
w u Z 4-i >, a r- 0 o u1 4 U r= 0 1

wor< -4 CLC = 1 4 o 0 W I .- w= N

O O I h 0 I u a > M( o a C (O N
2 2 0 = w I r 6 CU ao 0 a 4) a. o N

00- C I -0-0w W-W-4 (1) 0 A U',4 CU W
O O r- l.m 0 'c i t -o
S CU Q. LU O O Cl w -i (a 0 3 Q)

94 =a CL~ 4i 1J $W 14W U '0 4 "0 U 3 CL > -"

0 CJ

SO a
CoJ Cj




o j



I'o A4 :

Family Blenniidae

Springer (1968b) provided an extensive osteological description of

the family Blenniidae, based primarily on Entomacrodus nigricans. He

also discussed three characters that distinguish the Blenniidae from

all other blennioids: (1) coracoid reduced and fused to cleithrum; (2)

distinctive relationship of the interopercle with the epihyal

(=posterior ceratohyal) and interhyal; and (3) distinctive premaxillary

and dentary dentition of adults.

In addition to these three characters, monophyly of the Blennlidae

is supported by two other shared derived characters. The first is a

distinctive relationship of the urohyal with the dorsal and ventral

hypohyals, a condition not found in any other blennioids. The blenniid

urohyal has two lateral projections (one dorsal and one ventral) on

each side (Figure 2), each with a strong ligamentous attachment to its

respective hypohyal(s). There is variation in the positioning of the

lateral projections among the blenniid tribes, but all members have

this distinctive urohyal-hypohyal association.

The other derived character shared by the Blenniidae is the

presence of fleshy rugosities on the anal-fin spines of adult males.

As anal-fin spine rugosities are not found in other blennioid fishes

(sensu George and Springer, 1981), or other Perciform fishes, the

Figure 2. Urohyal bone. (Anterior
lateral views; B and D-dorsal views.)
processes (Acanthemblemaria aspera). C
processes (Chasmodes saburrae).

toward left; A and C-left
A and B-no developed lateral
and D-well-developed lateral




g **
Q /y^~-*y *

Presence of these fleshy rugosities is the apomorphic condition. The

Omobranchini-Phenablenniini group of Smith-Vaniz (1976) and certain

genera of the Salariini are the only Blenniidae not having anal-fin

spine rugosities. The most parsimonious explanation (based on

Smith-Vaniz's [1976] cladogram as modified in my discussion of the

Salariini-"Blenniini" group) of this character's distribution among

blenniid taxa is that the absence of anal-fin spine rugosities

represent two independent reversals (losses), one in the

Omobranchini-Phenablenniini group and one in the ancestral line giving

rise to the Salariini, with a secondary derivation of the rugosities in

one lineage within the Salariini.

Salariini-"Blenniini" Group

Springer (1968b) and Smith-Vaniz and Springer (1971) proposed a

close relationship between the Salariini and Blenniini, but did not

provide a character to support this relationship. I follow Smith-Vaniz

(1976) in separating Blennius, which is now the only member of the true

Blenniini, from the remaining genera of the tribe, which I will refer

to as the "Blenniini." A thorough study of the "Blenniini" is needed

to determine whether it is monophyletic.

The hypothesized sister-group relationship between the "Blenniini"

and Salariini is supported by two synapomorphies: (1) premaxillary

ascending process only weakly attached to main body of bone, and (2)

2-0-2 pectoral radial formula. The ascending process of the

premaxillary bone has an ossified connection to the main body of the

bone only at the anteriormost and posteriormost edges, while the middle

part is loosely attached by connective tissue. Springer (1968b) stated

that the blenniid ascending process was weakly joined (ossified

connection) at two points, anteriorly and posteriorly, to the ventral

portion of the premaxilla. Although this is true of the

Salariini-"Blenniini" line, other blenniids and blennioids have the

entire base of the ascending process weakly or strongly connected by

bone to the main body of the premaxilla. In addition, the ascending

process is an independent ossification in Pereulixia and Scartichthys,

where the anterior and posterior connections to the main body of the

premaxilla have been replaced by connective tissue. Alticus has an

intermediate condition in that a bony anterior connection has been

lost, but the posterior connection has been retained.

Springer (1968b) proposed that the most primitive pectoral radial

formula for blenniids is 2-1-1. Therefore, the 2-0-2 formula is

considered to be a derived condition. Some members of both tribes have

the 2-1-1 formula, which I interpret as a secondary reversion to the

2-1-1 condition.


As discussed previously, the "Blenniini" referred to here comprises

those genera included in the tribe by Bath (1977), but excluding


Monophyly of the "Blenniini" is supported by two synapomorphies.

The first character is the first basibranchial shaped like a broad,

shallow U, curving downward between the posterior end of the basihyal

and the anteroventral edge of the second basibranchial (Figure 3C). I

have not found a similarly shaped basibranchial in any other blennioid

examined and consider this condition as derived compared with the

straight basibranchial of most other blennid genera or a basibranchial

with the anterior end slightly upturned (Istiblennius, Antennablennius,

Atrosalarias, Ecsenius, Entomacrodus, and Litobranchus).

The second character suiting the "Blenniini" is their distinctively

shaped fourth epibranchial bones (Figure 4B). The "blenniin" fourth

epibranchial is a relatively smooth bone with a flattened longitudinal

bony flange projecting dorsally from the lateral half of the shaft of

the bone. All other blennioids examined have a fourth epibranchial

bone with one or two dorsally directed bony projections situated

dorsally on the shaft slightly before the articulation of the

epibranchial shaft with the infrapharyngobranchial plate. This type of

projection is lacking in the "Blenniini."

These characters have been examined in only a small percentage of

"blenniin" species. Thus, this hypothesis of monophyly for the tribe

needs to be further tested.


Two characters support monophyly of the Salariini: (1)

premaxillary bone an open capsule and (2) distinctively shaped

Figure 3. Left lateral view of basibranchial bone. (Anterior
toward left.) A-plesiomorphic rod-shaped structure (Cirripectes
filamentosus). B-slightly upturned anterior end (Atrosalarias
fuscus). C-derived shallow-U shape of "Blenniini" (Chasmodes



Figure 4. Fourth epibranchial bone. Left dorsolateral view.
(Anterior toward left.) A-single, broad, mid-shaft process
(Atrosalarias fuscus). B-no mid-shaft process (Chasmodes saburrae).
C-two mid-shaft processes positioned at mid-shaft (Cirripectes
filamentosus). Abbreviation: PR-mid-shaft process.


infrapharyngobranchial plate (Figure 5F-K). The first character was

discussed in detail by Smith-Vaniz and Springer (1971). The

premaxillary bone of the Salariini has a band of connective tissue in

place of the anteroventral bony wall present in all other blennioid


The infrapharyngobranchial plate of the Salariini has a broad,

flattened bony flange projecting medially from the tooth-bearing

portion of the bone. This flange extends anteriorly to articulate with

the medial tip of the second epibranchial. Further elaborations of the

infrapharyngobranchial flange occurring in Cirripectes and its allies

are discussed in the account for the node III group. In the

Tripterygiidae, "Blenniini" (the hypothesized sister group of the

Salariini), and all other groups in the-family Blenniidae, the

infrapharyngobranchial plate is composed primarily of the tooth-bearing

portion with a narrow medial flange of bone that has an anterior

rod-shaped protuberance articulating with the second epibranchial. The

families Dactyloscopidae, Chaenopsidae, Clinidae, and Labrisomidae also

have an infrapharyngobranchial plate with a large bony flange

projecting anteromedially from the tooth-bearing portion, but the

lateral edge of the flange is developed as a rod-shaped cartilaginous

swelling. The expanded bony flange appears to have developed

independently in the two groups and, thus, is considered a synapomorphy

of the Salariini.

I recognize two major groups within the Salariini, which I refer to

as the Alticus group and the Rhabdoblennius group.

Figure 5. Dorsolateral views of left infrapharyngobranchial
plates. (Anterior toward left.) A-Enneanectes altivelis.
B-Acanthemblemaria aspera. C-Chasmodes saburrae. D-Omobranchus
punctatus. E-Meiacanthus grammistes. F-Atrosalarias fuscus.
G-Stanulus seychellensis. H-Ophioblennius atlanticus.
I-Cirripectes perustus. J-Cirripectes polyzona. K-Cirripectes
quagga. Dashed line indicates anteriormost limit of tooth-bearing
portion of bone. Abbreviations: F-medial flange; RC-rod-shaped







Alticus Group

The Alticus group comprises the genera Ecsenius, Andamia, Alticus,

Dodekablennos, Praealticus, Istiblennius, Ophioblennius, Pereulixia,

Exallias, Cirripectes, Scartichthys, Stanulus, Entomacrodus,

Nannosalarias, Crossalarias, Atrosalarias, Salarias, and Glyptoparus.

These genera share a distinctive dentary bone that forms an open

capsule as a result of the loss of the anterodorsal bony wall. The

anterodorsal bony wall of the dentary is present in all other blennioid

fishes, but is only poorly developed in the salariin genera Hirculops

and Rhabdoblennius. With the exception of one of about 45 species (V.

G. Springer, 1971 and pers. comm.) of Ecsenius, members of the Alticus

group have a very high number of premaxillary teeth (55 to almost 400)

and dentary teeth (39 to about 300) as adults. The remaining salariin

genera have 17-50 premaxillary teeth and 16-38 dentary teeth. All

other blenniids typically have tooth counts less than or in the lower

end of the range found in the Salariini.

The relationships of the remaining salariin genera (the

Rhabdoblennius group), Cirrisalarias, Medusablennius, Mimoblennius,

Litobranchus, Antennablennius, Alloblennius, Hirculops and

Rhabdoblennius, are unresolved. The occurrence of lower tooth counts

in these genera than in the Alticus group cannot be used to infer

relationships, as this is the plesiomorphic condition. It is not known

whether they form a monophyletic group. A comprehensive analysis of

the interrelationships of all salariin genera is needed.

Node I Group

Although generic interrelationships within the Alticus group are

unresolved, the genera referred to here as the node I group appear to

represent a monophyletic unit. Springer (1968b) suggested these genera

were related because they all have the terminal anal pterygiophore

supporting two rays. Relationships within the node I group are

presented in the cladogram in Figure 1. Monophyly for this group,

comprising the genera Istiblennius, Entomacrodus, Stanulus, Pereulixia,

Scartichthys, Ophioblennius, Exallias, and Cirripectes, is supported by

three derived characters at node I: (1) last pleural ribs borne on

first or second caudal vertebral centrum (this is usually centrum 11 or

12); (2) terminal anal pterygiophore bearing two rays; (3) each dentary

of ophioblennius stage larvae with a large recurved canine positioned

midlaterally on each dentary.

The node I group, with the exception of Stanulus, consistently has

the last pleural ribs borne on the first or second caudal vertebra.

Stanulus (some specimens rarely have them on the first caudal

vertebra), the "Blenniini," and other blenniids typically have the last

pleural ribs borne on the last precaudal vertebra. Among the salariin

genera, only Alticus, Atrosalarias, and Salarias contain one or more

species that usually have the last pleural ribs borne on a caudal

vertebra, but each of these genera also has members with the last

pleural ribs borne on a precaudal vertebra. Dodekablennos also has the

last pleural ribs borne on a caudal vertebra and may be the sister-

group to the node I group, but the unavailability of cleared and

stained material leaves its status uncertain. The condition of the

last pleural ribs borne on a caudal vertebra is considered apomorphic,

while the last pleural rios borne on a precaudal vertebra is

plesiomorphic. Within the node I group, only Stanulus typically has

the last pleural ribs on the last precaudal vertebra (Springer,1968a).

Although this character argues for the exclusion of Stanulus from this

clade, additional characters at node II support its inclusion on the

basis of parsimony. Some species of the genus Istiblennius have the

last pleural ribs borne on a precaudal vertebra. This and other

characters suggest that Istiblennius, as presently defined, may be

polyphyletic. A complete revision of Istiblennius is needed to define

its limits.

The second character supporting monophyly for the node I group is

the presence of two rays on the terminal anal pterygiophore. Although

Smith-Vaniz and Springer (1971) stated that Praealticus frequently has

the last anal-fin ray divided through the base, I found this to be an

extremely rare and probably exceptional condition in Praealticus.

Istiblennius shows variation in this character that appears to be

species specific. Many Istiblennius species have the last two anal

rays on one pterygiophore, whereas other species have the last two rays

borne on separate pterygiophores. Most genera in the node I group show

some variation in this character, as discussed by Springer (1967), but

they usually have the last anal pterygiophore bearing two distinct

rays. Other blenniids and many blennioids (including the Chaenopsidae,

Labrisomidae, and Ophiclinini) have the last anal pterygiophore bearing

a single ray. The presence of two rays on the last anal

pterygiophore is considered the derived condition for blenniid fishes,

and supports monophyly for the node I group. I should note that the

last anal ray split to base is pleisomorphic for the Perciformes.

Among blennioids, at least the Blenniidae (except the node I group) and

Chaenopsidae do not have the last ray split. The simplest explanation

for the presence of the split ray in the node I group is that it is a

reversal and, thus, is apomorphic at this node.

The third character supporting monophyly for the node I group is a

characteristic of the ophioblennius stage larvae. The ophioblennius

stage larvae of each of the genera in the node I group, except

Pereulixia, for which a larval stage has not been found, have two

posterolaterally directed, large, recurved canines located about

midlaterally on each dentary. Istiblennius is problematical in that an

ophioblennius larval stage is known only for members of the I.

gibbifrons species complex (Smith-Vaniz and Springer, 1971). If the

absence of this larval stage in other Istiblennius species is real and

not a collecting artifact, then the genus as presently defined may be

polyphyletic. No other blennioid larvae possess these large hooked

canines midlaterally on each dentary. Larval Nemophini have large

canines in the jaws, but these teeth are not hooked. The presence of

large hooked teeth is considered a derived condition and is evidence

for monophyly of the node I group Although larval Pereulixia are not

known, the presence of several additional characters, discussed later

in this paper, places this genus in the node I group on the basis of


Genus Istiblennius

The genus Istiblennius as defined by Smith-Vaniz and Springer

(1971) (or some part of the genus including the Istiblennius gibbifrons

species complex) appears to be the sister group of the remaining

members of the node I group. As indicated above, I know of no

character that supports a hypothesis of monophyly for the species

included in Istiblennius. The genus is presently being revised.

Node II Group

This group of genera (Entomacrodus, Stanulus, Pereulixia,

Scartichthys, Ophioblennius, Cirripectes, and Exallias) shares two

derived characters: (1) each anal-fin spine of adult males enveloped

in a fleshy rugosity and (2) posteriormost branchiostegal ray with an

elongated proximal portion extending well onto the posterior ceratohyal

and passing dorsad of the proximal end of the preceding branchiostegal

ray (Figure 6C).

In addition to the node II group, adult males of Nannosalarias

typically have fleshy rugosities enveloping each anal-fin spine. As

discussed in the Blenniidae section, this condition is a synapomorphy

of the Blenniidae. Based on my hypothesis of relationships within the

Salaiiini, the occurrence of anal-fin spine rugosities in the node II

group must be a secondary derivation because the rugosities have been

lost at the hypothesized node at which the Alticus and Rhabdoblennius

groups diverge. Inasmuch as the genera of the node II group and

Figure 6. Left lateral view of branchiostegals and associated
bones. (Anterior toward left.) A-fifth and sixth branchiostegal rays
extending about equally onto ceratohyals (Crossosalarias macrospilus).
B-sixth branchiostegal ray extending slightly dorsad of fifth
(Stanulus seychellensis). C-sixth branchiostegal extending dorsad of
and anterior to fifth (Cirripectes perustus). Abbreviations:
AC-anterior ceratohyal; DR-dorsal hypohyal; PC-posterior ceratohyal;
VH-ventral hypohyal.


Nannosalarias do not share other characters defining the "Blenniini,"

the most parsimonous explanation is that this apomorphic character is

either independently derived or convergent. Within the Alticus group,

the rugosities appear to have been independently derived in

Nannosalarias because it does not share the other characters defining

the node I group. I interpret the presence of fleshy anal-fin spine

rugosities as a synapomorphy of the node II group at this level of


The sixth (=posteriormost) branchiostegal ray, although variously

shaped among blenniid genera, has two distinctive conditions associated

with the development of the proximal portion (herein referred to as the

head) that overlaps the posterior ceratohyal: (1) head of

posteriormost and preceding branchiostegal rays extending equally far

onto'the posterior ceratohyal (Figure 6A) or (2) head of posteriormost

ray extending as much as three to four times as far onto posterior

ceratohyal as preceding ray (Figure 6C). Condition 1 is the typical

condition found in other salariin and blennioid fishes and is

considered the plesiomorphic state. The node II genera, except

Pereulixia, which has condition 1, share condition 2. The most

parsimonious explanation for the occurrence of condition 1 in

Pereulixia is that it is a reversal to the plesiomorphic condition.

In condition 2, the elongated head of the sixth branchiostegal ray

has an anterodorsally directed component projecting over the dorsal tip

of the preceding ray(s) and a posteriorly directed spur sometimes

extending along a slight groove in the posterior ceratohyal. There is

a ligamentous connection of the elongated portion of the branchiostegal

head to the posterior ceratohyal. Condition 2 does not occur in other

blennioid fishes and, thus, is a synapomorphy of the node II genera.

The head of the sixth branchiostegal ray in Stanulus is not as

extensively developed as in the node II genera. In Stanulus, the head

is only slightly elongated in an anterodorsal direction, reaching a

point just above the preceding ray (Figure 6B). Stanulus is a genus of

small species and the condition might be neotenic. This condition is

interpreted as being a modification of condition 2.

Genus Entomacrodus

I know of no uniquely derived character that will define the genus

Entomacrodus as a monophyletic taxon. Entomacrodus has vomerine teeth,

but vomerine teeth are also present in Pereulixia and some Stanulus.

Smith-Vaniz and Springer (1971) and Springer (1967, 1968a) stated

that Entomacrodus is most closely related to Stanulus, based on their

similarity in overall appearance. I have found no synapomorphy to

support this hypothesis.

Node III Group

This group comprises the genera Stanulus, Pereulixia, Scartichthys,

Ophioblennius, Cirripectes, and Exallias. I have found only one

character that defines this assemblage. These genera, except

Pereulixia and Scartichthys, possess a distinctively shaped

infrapharyngobranchial bone (Figure 5G-K). The anteromedially

positioned flange of the infrapharyngobranchial plate (either a fused

complex of one or more of infrapharyngobranchial bones 2, 3, and 4, or

a loss of one or two of these bones) has one or two large foramina,

giving the bone a doughnut-shaped appearance (Figure 5). This

condition appears to be unique and defines these genera as a

monophyletic group.

The infrapharyngobranchial plate of Pereulixia and Scartichthys is

typical of other salariin genera (Figure 5F; see previous discussion

for the Salariini) and does not show the modifications present in the

other members of this group. This condition is interpreted as a

reversal based on parsimony.

Genus Stanulus

Stanulus, the hypothesized sister group of the other node III

genera, has a distinctive ophioblennius stage larva that is much

smaller than the larvae (at an equivalent stage of development) of

other genera in the node I group. Similarly, adult Stanulus are

smaller than other members of this group (except Entomacrodus

thalassinus and E. strasburgi). In addition, Stanulus has scalelike

flaps along the anterior portion of the lateral line. Pereulixia and

several species of Cirripectes have similar scalelike flaps associated

with the anterior lateral line pores, but these genera possess

additional characters that argue for an independent derivation of the

flaps in each genus.

Node IV Group

The node IV group comprises Pereulixia, Scartichthys,

Ophioblennius, Cirripectes, and Exallias. Three derived characters

define this group: (1) transverse row of cirri extend across nape

(Figure 7); (2) symphyseal portion of the premaxillary bone connected

to remainder of premaxillary by a slender bridge of bone beneath the

anterior edge of the base of the ascending process, symphyseal portion

is free posteriorly (Figure 8B); (3) lacrimal with an elongate

anterorentral extension that reaches ventrally as far as the lower edge

of the maxillary bone and has a sensory canal opening at or near its

ventral edge (Figure 9B).

A distinctive transverse row of cirri across the nape in the node

IV genera manifests itself variously in the different genera. Although

some genera of the Labrisomidae (sensu George and Springer, 1981) also

have a transverse row of nuchal cirri, this character is considered to

be convergent, as these groups are clearly not closely related. The

different patterns of this transverse row of cirri will be discussed

later for each genus or group of genera.

The distinctive shape of the body of the premaxillary bone in the

node IV group appears to have resulted from a loss of bone in the

concavity beneath the base of the ascending process. The posterior tip

of the free symphyseal portion is tightly connected by ligaments to the

posterior edge of the base of the premaxillary ascending process and

the anterior edge of the maxillary bone. The base of the ascending

process straddles the unossified concavity in the body of the

Figure 7. Dorsal views of nuchal cirri. (Anterior toward top.)
A-type A, Cirripectes kuwamurai. B-type B, C. variolosus. C-type
C, C. gilbert. D-type D, C. perustus. E-type E, C. alboapicalis.
F-type F, C. jenningsi. G-type G, C. vanderbilti. H-Pereulixia
kosiensis (dashed line indicates cirri not illustrated). I-Exallias
brevis. J-Scartichthys viridis.

B \ 1 K>7
p 11 ,raI 8

C -11
`5 :'--,'

- Al I ''I

ii. i

G ~` I~ 'b''

rI! II

I' ,I N7 ) rl 7

~a/i /n

Figure 8. Dorsal views of left premaxillary bones (ascending
process removed from each bone). (Anterior toward top.) A-Alticus
saliens. B-Scartichthys gigas.


A 0 h t

B:f w
~- O~

Figure 9. Left lateral views of left lacrymal bones. (Anterior
toward left.) A-pores exit near main canal (Alticus saliens).
B-branch of canal almost reaches ventral margin of bone (Scartichthys







premaxillary bone and appears to serve as a point of attachment for

connective tissue holding the teeth in place. Other genera of the

Blenniidae either do not have an open gap in the body of the

premaxillary bone, or, if there is an opening, there is an anterior and

posterior bony connection to the main body of the premaxilla (Figure

8A). The loss of the posterior bony connection between the symphyseal

portion and the main body of the premaxillary is a synapomorphy of the

node IV group.

The distinctively shaped lacrimal (Figure 9B) of the node IV group

is not found elsewhere in the Blenniidae. In most other salar;in

genera, the lacrimal is much shorter and broader than it is in the node

IV group, and the ventrally directed sensory canal(s) exits from the

bone at a point relatively close to the main infraorbital canal. The

slender, elongate, anteroventrally directed extension of the lacrimal

is a synapomorphy of the node IV genera. Ecsenius is exceptional in

having an elongate lacrimal (Springer, 1968b:pl.9), but the ventral

sensory canal exits relatively close to the main infraorbital canal.

Genus Pereulixia

The genus Pereulixia forms the sister group of the node V group and

is distinctive in having a nuchal cirri arrangement consisting of four

slightly overlapping groups of cirri. The cirri are frequently

complex, with multiple branches arising from a single base. The skull

Figure 10. Left lateral view of supraoccipital region of skull of
Pereulixia kosiensis illustrating well-developed crest.


of Pereulixia has a high bony crest extending anteriorly from the

dorsalmost edge of the supraoccipital bone to a point above the

posterior edge of the orbits (Figure 10). The crest is formed from

dorsal expansions of the frontal and parietal bones, each of which

comprises about half of the crest and are firmly sutured to each

other. Other Salariini may have short bony crests on the skull, but

the crest does not reach the height of the dorsalmost tip of the

supraoccipital bone. Even salariin taxa with well-developed fleshy

crests have only low bony crests on the skull. Both of these

characters are considered derived. In other blenniid tribes, a large

bony crest is present in large Hypleurochilus and Omobranchus

aurosplendidus. These occurrences are considered independent


Pereulixia'has larger (thus fewer) dentary incisors than

premaxillary incisors. Exallias has a similar dentition, but has more

incisors in both jaws. Ecsenius has dentary incisors larger than

premaxillary incisors, but there are fewer incisors. This character of

dentary incisors larger than premaxillary incisors is considered to be

a derived condition. Based on the characters defining the node III and

V groups, I hypothesize independent derivations in these genera.

The relationship of Pereulixia to other salariin genera is somewhat

tentative. Although it shares several characters with the node IV

group, its placement here is only slightly more parsimonious (two fewer

steps) than placing it as the sister group to Istiblennius and the node

II group.

Node V Group

The node V group includes Scartichthys, Ophioblennius, Cirripectes,

and Exallias. This assemblage is defined by the following five

synapomorphies: (1) fourth epibranchial with two distinct processes

located on the dorsal side near the midpoint of the shaft of the bone,

one process directed anterodorsally and the other posterodorsally

(Figure 4); (2) fifth ceratobranchial with an expanded flange directed

ventrally and posteriorly from the main tooth-bearing portion of the

bone, flange with a straight vLetical margin posteriorly that forms a

sharp right angle where it meets the relatively straight ventral margin

of the flange (Figure 11); (3) each premaxillary bone of ophioblennius

stage larvae with one or two large recurved canines near symphysis of

premaxillaries; (4) each dentary bone of ophioblennius stage larvae

with one or two large recurved canines at symphysis of dentaries; and

(5) dorsal-fin membrane attached to caudal fin. The above characters 1

through 4 are present in (and exclusive to) all node V genera and are

synapomorphies of the group. Character 5 is a derived character but

occurs homoplasiously.

No other blennioids have two distinct processes on the fourth

epibranchial bone. The plesiomorphic condition is a single broad

dorsal projection in this position on the fourth epibranchial and

occurs in all blennioids except the "Blenniini" and node V group.

There appear to be two derived conditions for the fourth epibranchial:

two projections or no projections. The former condition is a

synapomorphy of the node V group and the latter a synapomorphy of the

tribe "Blenniini."

Figure 11. Left lateral views of left fifth ceratobranchial
bones. (Anterior toward left.) A-Acanthemblemaria aspera.
B-Chasmodes saburrae. C-Mimoblennius atrocinctus. D-Hirculops
cornifer. E-Alticus saliens. F-Scartichthys gigas. Abbreviations:
CN-constriction; PF-posterior flange; UF-unconstricted flange;
VF-ventral flange.


A 1*
/ -"*-,.u ~^ -

- UF

In the node V group, the fifth ceratobranchial has a distinctive,

unconstricted, ventroposteriorly directed flange (Figure 11F) that

extends posteriorly to a point beyond the main body (tooth-bearing

portion) of the bone. Alloblennius, Mimoblennius, Hirculops, and

Litobranchus also have a posteriorly directed flange, but, in these

genera, the flange is quite distinctive in having a rounded extension

of the flange at the posterior edge and a constriction separating it

from an angular process on the ventral edge (Figure 11C-D). Other

salariin genera have a reduced posterior flange that does not project

posteriorly beyond the ddrsal limb of the fifth ceratobranchial (Figure

11E). Genera of the "Blenniini" have a rounded flange of bone on the

posterior edge of the dorsal limb of the fifth ceratobranchial, another

flange located ventrally below the tooth-bearing portion of the bone,

and a distinct constriction between the flanges at the dorsoposterior

edge of the tooth-bearing portion (Figure 11B). All other blennioids

have a slender fifth ceratobranchial bone with a very narrow flange of

bone on the ventromedial edge of the tooth-bearing portion (Figure

h1A). The slender fifth ceratobranchial with a very narrow flange of

bone is considered the plesiomorphic condition. The unconstricted,

extensively developed flange of bone on the fifth ceratobranchial bone

is a synapomorphy of the node V group.

The third and fourth synapomorphies for the node V group are

dentitional characters of the ophioblennius stage larvae. The first of

these is the presence of large, hooked canines at the symphysis of the

premaxillary bones. The second character is the presence of large,

hooked canines at the symphysis of the dentary bones. Watson (in Leis

and Rennis, 1983) distinguished the genera Exallias, Entomacrodus,

Cirripectes, and Stanulus from the Nemophini by their large, hooked

canines in the jaws. Although Watson suggested that the hooked canines

serve to identify salariin larvae, these dentitional characters define

only the node V genera. The genera of the node V and I groups share in

having a large hooked canine midlaterally on each dentary, but, unlike

the node V group, other members of the node I group do not have hooked

cymphyseal canines. These distinctive hooked symphyseal canines also

are lacking in all other blennioid larvae; thus, I consider them to

represent a derived condition.

Most members of the node V group have the dorsal-fin membrane

attached to the caudal fin at a point above or posterior to the

caudal-fin base as adults. Most salariin genera have the dorsal-fin

membrane attached to the caudal peduncle anterior to the caudal-fin

base. Although the dorsal fin is attached to the caudal fin in many

other blennioids, this attachment is considered to be apomorphic at

this level of analysis, inasmuch as the dorsal fin is not attached to

the caudal fin in other node I genera (except some species of

Istiblennius). Although the dorsal-fin membrane is not attached to the

caudal fin in Scartichthys and a few species of Cirripectes, I consider

this condition to be either a reversal, or a secondary derivation,

based on parsimony.

Genus Exallias

Exallias, a monotypic genus, is the sister group of the node VI

group and is distinctive among all blennioids in having the lateralmost

of the nuchal cirri very short, with cirri increasing in length toward

the nape.

Both sexes of Exallias lack dentary canines. Although some species

of Istiblennius also lack dentary canines, all other genera in the node

I group have them. The most parsimonious explanation is that the lack

of dentary canines in Exallias is an apomorphy at this level of


Exallias differs from other members of the node V group in having

enlarged dentary incisors. Exallias usually has about three or four

times as many premaxillary incisors as dentary incisors. Pereulixia

has a similar dentition, but has fewer incisors on both jaws

(Smith-Vaniz and Springer, 1971: table 2). Other members of the

Salariini (see discussion of Ecsenius in Pereulixia section) have

similar sized incisors in each jaw with up to twice as many

premaxillary incisors as dentary incisors. I consider the possession

of enlarged dentary incisors an apomorphic condition that has been

independently derived in Exallias and Pereulixia.

Node VI Group

This group comprises Cirripectes, Scartichthys, and Ophioblennius

and is defined by one synapomorphy, a large prominence at about the

Figure 12. Posterior view of left maxillary bones. A-Alticus
saliens. B-Scartichthys gigas. Abbreviation: MP-maxillary


midpoint of the posteromedial edge of the maxillary bone (serves as the

point of insertion for the primordial ligament).

The well-developed maxillary prominence (Figure 12B) is not present

in other blennioid fishes, but, occasionally a slender flange in

approximately the same position, or slightly anterior to this position,

is present in other salariin genera. The presence of a well-developed

maxillary prominence is a synapomorphy of the node VI group.

Genus Cirripectes

Cirripectes, sister group of the node VII group, contains a diverse

assemblage of species. I have found only two synapomorphies that

define Cirripectes: (1) distinctive nuchal cirri arrangement (Figure

7) and (2) adult male urogenital papilla with one or two tapering

filaments associated with gonopore (Figure 13).

All Cirripectes have nuchal cirri that are about equal in length,

typically simple, and arranged in a single row that may be divided into

as many as four groups. When there are two groups on one side of the

nape, the two groups do not overlap where their bases meet. In all but

two species, there is a small break in the row of nuchal cirri at the

nape. This condition differs from that of all other node IV genera.

Figure 13. Ventral views of male urogenital papillae. (Anterior
toward top.) A-type I, Cirripectes perustus. B-type II, C.
vanderbilti. C-type III with posteriorly positioned gonopore, C.
alboapicalis. D-type III with distally positioned gonopore.
E-Scartichthys viridis.



Exallias has the cirri increasing in length toward the nape and

typically in a single continuous row across the nape. Scartichthys and

Ophioblennius have a relatively short row of cirri on either side of

the nape, with a wide gap separating the two rows. Pereulixia has four

groups of nuchal cirri with overlapping bases. Large Pereulixia

specimens usually have the cirri extensively branched, sometimes

appearing as a series of palmate cirri. Because the nature and

arrangement of the nuchal cirri in Cirripectes is unique, I hypothesize

that the Cirripectes condition is a synapomorphy for the species of the

genus. No other apomorphic characters are shared by all -.mbers of the


Adult males have one or two slender filaments associated with the

gonopore (Figure 13A-C). These filaments are variously modified in

several species groups within Cirripectes. In all Cirripectes species

groups, except one, the filaments extend beyond the gonopore and taper

to a point. The relatively long tapering filaments contrast with the

short blunt-tipped filament(s) found in Exallias and the node VII

genera. Other blennioids have a gonopore at the tip of a short

papilla. I consider each of the following as a synapomorphy for its

group: the single short blunt protuberance in Exallias, the two blunt

protuberances in the node VII group, and the one or two slender pointed

filaments in Cirripectes.

Node VII Group

Scartichthys and Ophioblennius are included in the node VII group.

Three synapomorphies define this group: (1) male genital papilla

flattened anteroposteriorly, with two short blunt projections on either

side of the gonopore (Figure 13); (2) transverse row of nuchal cirri in

two widely separated groups (Figure 7); and (3) high number of caudal

vertebrae (22-26). The distinctive shape of the male genital papilla

does not occur in other blennioids and I consider it to be a derived


The nuchal cirri arrangement in these two genera varies among

species, but always consists of a group of cirri (as few as two in some

Ophioblennius) on each side of the nape with a broad gap between the

groups. This condition is hypothesized to be a secondary modification

of the condition characterizing the node IV group and, as such, is a

synapomorphy for the node VII group.

Although other blenniids have a high number of caudal vertebrae,

all other genera included in the node IV group have 19 to 22 caudal

vertebrae (only a few species of Cirripectes have as many as 22). At

this level of analysis, I consider the high number of caudal vertebrae

(22-26) to be a synapomorphy for the node VII group.

Genus Ophioblennius

Ophioblennius is distinctive among blenniids in having disconnected

overlapping portions of the lateral line. The anterior section of the

lateral line ends at a point above the posterior section of the lateral

line near the middle of the body. The posterior section of the lateral

line begins slightly anterior to and beneath the end of the anterior

section of the lateral line and continues to the caudal-fin base

(frequently as a series of short tubes).

As discussed by Smith-Vaniz and Springer (1971), adults of this

genus have dentary canines that are as long as or longer than the pupil

diameter. Other members of the node I group have much shorter canines

(or none at all) as adults. These two characters are synapomorphies of

the species of Ophioblennius.

Genus Scartichthys

The genus Scartichthys is distinctive in that the posterior part of

the lateral line consists of about 18-21 (typically 21) short tubes,

each with a pore at either end. Other members of the node IV group

usually have from none to 10 lateral line tubes. The high number of

lateral line tubes is a synapomorphy of this genus.


Although the results of my study are in many cases tentative, I

have attempted to present all supportive data clearly and concisely.

As this study is the first attempt to provide a phylogenetic analysis

of any group within the Salariini, it is possible that my hypothesis of

relationships will be altered as the taxa are analyzed in greater


The only previous attempt to provide a hypothesis of relationships

within the Salariini was a dendrogram based on overall similarity

(Smith-Vaniz and Springer, 1971; reproduced here as Figure 14). My

hypothesis of phylogenetic relationships (Figure 1) differs from

Smith-Vaniz and Springer's (1971) dendrogram in the relationships of

several genera. In my analysis, Istiblennius (based primarily on the

I. gibbifrons species complex) is included in the node I group, whereas

Smith-Vaniz and Springer (1971) placed it with Praealticus, Alticus,

and Andamia. It seems likely that this difference results from the

particular species of Istiblennius used in each analysis. I agree with

Smith-Vaniz and Springer (1971:50) who state that Istiblennius might be

polyphyletic and, thus, both placements could be correct for a specific

subset of Istiblennius species. A revision of Istiblennius is needed

before its relationships can be assessed adequately.



















"4 I




c o,,




Smith-Vaniz and Springer (1971) considered Nannosalarias to be the

sister group of my node II group. As I did not treat the

interrelationships of the members of my Alticus group, outside of my

node I group, it is possible that Nannosalarias is the sister group of

the node I group. Further study of the Salariini is needed to

determine the interrelationships of the remaining genera in the Alticus


Smith-Vaniz and Springer (1971) considered Stanulus and

Entomacrodus to be sister genera that together form the sister group of

my node IV group. They grouped Entomacrodus and Stanulus on the basis

of overall similarity, but did not provide a specific character to

support the hypothesized sister group relationship. In my analysis,

Entomacrodus is the sister group of the node III group, which includes

Stanulus and is defined by the presence of a large foramen in the

infrapharyngobranchial plate. Thus, Stanulus appears to be more

closely related to the node IV group than to Entomacrodus.

Smith-Vaniz and Springer (1971) also regarded Ophioblennius as the

sister group of the other genera in the node IV group. They placed it

in this position because it possesses a combination of unique and

distinctive characters that they felt isolated it from the other

genera. In a cladistic analysis, unique characters cannot be used to

infer relationship or lack of relationship with taxa not possessing

these characters. In my analysis, Ophioblennius and Scartichthys are

sister genera, which together are the sister group of Cirripectes. My

determination of the relationship of Ophioblennius to the other genera

in the node IV group is based on a suite of characters at nodes V, VI,

and VII.

Although many of the characters I used to determine relationships

among the genera in my node I group are homoplasiously distributed, a

cladistic analysis is the preferred method because it specifies how a

character is distributed and clearly illustrates weak and strong

components of the hypothesis. A cladistic hypothesis is defined by a

specific distribution of characters among taxa and can be tested by

adding additional characters or reassessing the original characters.

As a result of the relative ease of testing a cladistic hypothesis, it

is possible that any given hypothesis of relationships will change when

additional characters are added to the analysis. Because of the

potential for a given cladogram to change when additional-characters

are added, formal taxonomic names should be applied with care in order

to minimize future nomenclatural confusion.


I present a hypothesis of relationships among species of

Cirripectes in the cladogram in Figure 15. To avoid confusion I began

numbering the nodes on this cladogram with the next available node

number (VIII) from the generic cladogram (Figure 1). All references to

nodes I through VII refer to groups on the cladogram in Figure 1. The

characters defining Cirripectes (the basal node in Figure 15) as a

monophyletic genus were given in the earlier account of Cirripectes and

will not be repeated here. Characters defining the terminal taxa

(species) are given in the species accounts in the systematic section

of this paper.

Morphological characters vary little among species of Cirripectes.

Most species are differentiated from each other by subtle differences

in color pattern or slight differences in meristic or other

characters. Because of the morphological similarity of the species,

few characters are available for use in a phylogenetic analysis. The

phylogenetic relationships I hypothesize herein should be considered a

working hypothesis of interspecific relationships.

Node VIII Group

This group is hypothesized to be the sister group of the remaining

Cirripectes species (node X group). The node VIII group comprises C.

4O 0) U
c 41 '-
m I a-). 4 -4 -
-l I ( 3 Z u
ac C -I 0. -4 u s I
a c Q) co c w- I
0 4 ***- 0 0 I
S 3 -I-, L I -400
Sto C U "'4 0 -
'-4 a r-4 I. I-4
a ca a) -O 0 w
C 0 cc w o 0
a) 1 ,, *1 .d ,S Q
0 -' 0 :t o U 0. I
,- 00 c, -H I 0 c

S. 1 M C O 01

H: Ow *t 4 0
)-4 0 3 4-4 c 0a M a)
S" 0 0 S0 4 =

WIV.,- 4 00, -4 0
o, C -4 0- 0 1 M 3r
m0 0 ca I 0 I c*

J I .m c I I(U

-4 -4 0. SI I I
r0 ,~ 4 0 W,
r. c .Cu L 0 : C
4U (0 U 6 4 CO 0 0 *
wW X c00I W -A0-I <

bo 0 m4I 0 0

Sc- 0 0 4 o e-i 0 0
0 4-3 1-. 0 I I '*

o (Ue a 0.m U -0 4 *
O.4 0.1-4 3 B) 00-I 0
0 0 0 1 00 0

i -I I 0.C I-
14 a 0a o a4 -4 aj i ai
W S -40 1 0 4J W0
0C I O 1 0 '4-4 O 0
ou a a i ai 1 1 a

Q-4 0 U 0 0 0 0
*ai-i j 01 0

O i ic a r e o IN
S0 0 U 4 0 r I
.43- 0C-e W r-4 Q)

Q.41 4L C u L C m 0
w0 Cc -A U W I
W O W L 0 r 0

0 w- 0 0* 4m

c *0 0 4 -4 0 w rI
1 3t I z-- I-I =.
a 01 0 m o u

C: 8 I C C1

*0 *w I 0 4-i -

r 0 0go 0 I wo
9 a 4) *t o ca
*ir 02 U0 I i- I

*4 0 4 *W- I :: *
0 )4O 0 3 1 c4
S0 00 .- *1 **0 .(

- a 0 0 V C 1 0 a)
o WD I Ou w I =
W4 (U W CO C5 CC 04 |

*'.'40 0 010
mr 0 0 z I- 0 e 0 I

C0 0wI I 0 4(1 I I
-4m o a W C--
0) ) t-i m u g *
v i i- C C Q ) o o

* OB M ep 0 0 B I B)
cO O) v 0 0 1 r
-4 L M44 4 4 4



nco o >
I x



X xx
tr m~


sn.4 trr



P7-7 -p

quagga, C. jenningsi, C. alboapicalis, and C. obscurus. These four

species are associated in an unresolved trichotomy between C. quagga,

C. jenningsi, and the node IX group comprising C. alboapicalis and C.

obscurus. Four synapomorphies define the node VIII group: (1) lower

lip crenulate; (2) adult male genital papilla with a single slender

filament, gonopore located at distal tip of filament or on posterior

side near base of filament (Figure 13); (3) posteriormost pleural rib

borne on second caudal vertebra; and (4) foramen in

infrapharyngobranchial plate very large, reflecting a narrowing of the

posterior part of the bony ring encircling the foramen (Figure 5).

Members of the node VIII group have distinct crenulae on the

anteromedial edge of the lower lip. In other Cirripectes species, the

medial portion of the lower lip has a smooth edge. Among the node I

genera, a crenulate lower lip occurs only in certain species of

Istiblennius and Cirripectes and, on the basis of parsimony, I consider

these occurrences to be independent derivations of lower lip crenulae.

Exallias has a deeply plicate lower lip, but does not have distinct

crenulae on the anteromedial edge. The presence of distinct crenulae

on the lower lip is a synapomorphy of the node VIII group.

The shape of the male genital papilla is apparently unique to the

node VIII group and is considered a derived condition. The existence

of several different adult male genital papilla morphologies in the

node V group of genera makes it difficult tG determine derived and

plesiomorphic conditions. Three distinctive papilla morphologies are

found among Cirripectes species. As none of these papilla morphologies

are found in other blennioids, I interpret each of all three conditions

as derived. Because each papilla morphology occurs only in

Cirripectes, it is not possible to use outgroup analysis to polarize

the sequence of derivation of the three morphologies. Polarizing the

character based on the most parsimonious distribution of derived

characters for Cirripectes species results in a transformation series

beginning with the short blunt papilla (plesiomorphic) leading to the

following successively derived conditions: single slender filament,

two widely separated filaments, and two short, closely appressed

filaments. Although this technique can be considered circular

reasoning, I tentatively accept this sequence, while realizing that an

alteration of my cladogram could alter the polarity of this character.

The presence of the posteriormost pleural ribs on the second caudal

vertebra is considered to be a derived condition that has arisen

independently in some Scartichthys species and the node VIII group. I

interpret the occurrence of pleural ribs on the second caudal vertebra

as a synapomorphy of the node VIII group at this level of analysis.

In other members of the node V group with a foramen in the

infrapharyngobranchial bony plate, the posterior section of the ring of

bone encircling the foramen is relatively thick. As the bony portion

extends as far posteriorly in the node VIII group as in other members

of the node V group, it seems likely that the narrowing of the

posterior section of the bony ring in members of the node VIII group

resulted from a loss of bone from the interior edge of the posterior

section of the ring. I consider the reduction in thickness of the

posterior section of the ring of bone encircling the foramen to be a

synapomorphy of the node VIII group.

Node IX Group

Cirripectes obscurus and C. alboapicalis have scalelike flaps

anteriorly on the lateral line. Stanulus and Pereulixia also have

scalelike flaps. As discussed in the accounts of Stanulus and

Pereulixia, this is an apmorphic character that has been independently

derived in three different lines. I consider the presence of

lateral-line flaps to be a synapomorphy of the node IX group at this

level of analysis.

Node X Group

The node X group comprises the remaining Cirripectes species, which

are in the node XI and XII groups, The node X group is defined by two

apomorphic characters: (1) adult male genital papilla with two widely

separated filaments (Figure 13) and (2) adult males with bulb-shaped

testes (Figure 16).

Within blennioids, an adult male genital papilla with two slender

filaments, one on either side of the gonopore, is unique to the node X

group. This condition is modified in two groups of Cirripectes in that

the filaments are reduced in length and closely appressed, but there

are still two distinct filaments. The presence of two distinct

filaments on the male genital papilla is a synapomorphy of the node X


Figure 16. Ventral view of left testis of two Cirripectes
species. (Anterior toward left.) A-elongate testis (Cirripectes
obscurus). B-bulb-shaped testis (C. vanderbilti).





Adult males of the node X group have bulb-shaped testes. All other

blenniids examined have slender elongate testes. Bulb-shaped testes

are a synapomorphy of the node X group.

Node XI Group

Cirripectes imitator, C. fuscoguttatus, and C. gilbert are

associated in a trichotomy at node XI. These three species share a

distinctively shaped row of nuchal cirri (Figure 7). The nuchal cirri

are arranged in four groups of cirri with the cirri in the ventralmost

group on each side borne on a greatly enlarged flap. The node XI group

also has a higher number of cirri than any other Cirripectes. The

distinctive morphology of the row of nuchal cirri is a synapomorphy of

the node XI group.

Node XII Group

The node XII group is associated in an unresolved trichotomy

comprising Cirripectes variolosus, C. vanderbilti, and the node XIII

group. Members of the node XII group share a nuchal cirri morphology

in which there is, on either side of the nape, a slightly expanded

basal flap bearing the ventralmost cirri (Figure 7). The node XII

species usually have four groups of nuchal cirri. This arrangement has

been modified in C. vanderbilti, where the two groups of cirri on

either side of the nape are fused at their bases, resulting in two

continuous rows of cirri with a short gap between the rows at the

mid-point of the nape (Figure 7G). In C. perustus, there are

frequently three groups of nuchal cirri, one group forming a continuous

row across the nape and a separate row ventral to the medial row on

each side (Figure 7D). The node XV group has a fourth condition, in

which the ventralmost group on either side is a greatly expanded flap

bearing rudimentary cirri, with the cirri dorsal to this flap not

connected basally (Figure 7A). Although there are four different

modifications of the arrangement of nuchal cirri in the node XII group,

all, except the highly modified condition in the node XV group, have a

slightly expanded flap bearing the ventralmost nuchal cirri. I

hypothesize that the general morphology (and secondary modifications) of

the nuchal cirri in the node XII group is a synapomorphy of the group.

Node XIII Group

The node XIII group is an unresolved trichotomy comprising

Cirripectes hutchinsi, C. randalli, and the node XIV group. The node

XIII group lacks an interorbital pore (EIP) that is present in all

other Cirripectes (Figure 17), Exallias, Ophioblennius, and Stanulus.

The lack of an extra interorbital pore position is a synapomorphy of

the node XIII group.

Figure 17. Cephalic sensory pore patterns of Cirripectes. A and
B-complex pattern, dorsal and lateral views of Cirripectes imitator.
C-simple pattern, lateral view of C. perustus. Dashed lines indicate
bases of cirri not illustrated. Abbreviations: AN--anterior nostril;
EIP-extra interorbital sensory pore position; IFO-infraorbital
sensory pore series; MD-mandibular sensory pore series; MSP-mid-snout
pores; PBN-pore positions behind nuchal flap; PN-posterior nostril;
POP-preopercular sensory pore series; SO-supraorbital sensory pore
series; ST-supratemporal sensory pore series. Lines indicate first and
last pores of each series.






Node XIV Group

The node XIV group comprises the node XV, XVI, and XVII groups that

are associated in an unresolved trichotomy. Members of the node XIV

group, except Cirripectes polyzona, share a relatively simple cephalic

sensory pore system, in which most pore positions have one or two pores

(Figure 17C). Cirripectes polyzona and other members of Cirripectes

have a complex cephalic sensory pore system, in which there are usually

six or more pores at most positions (Figure 17A-B). Although a simple

pore pattern is found in many other blennioids, Exallias and most

Scartichthys have a complex pore pattern. The simple pore pattern is a

synapomorphy of the node XIV group at this level of analysis. The

complex pore pattern of C. polyzona may be independently derived, as it

shares another derived character with the node XVI group. An

independent derivation of the complex pore pattern is the most

parsimonious explanation. If the simple pore pattern is not considered

an apomorphic character, the node XIV group collapses to form a

polychotomy with the node XIII group.

Node XV Group

The node XV group comprises Cirripectes auritus and C. kuwamurai,

which share two apomorphic characters associated with the nuchal cirri

(Figure 7): (1) large flap bearing rudimentary cirri present on each

side of nape and (2) medial cirri (those located dorsal to nuchal flap)

arising directly from skin, bases of cirri not connected by a low


The distinctive nuchal flaps of Cirripectes auritus and C.

kuwamurai are unique among blennioids. These flaps are a synapomorphy

of the node XV group.

In the node XV group and many specimens of Cirripectes filamentosus

and C. chelomatus, the dorsalmost nuchal cirri arise directly from the

skin. Other Cirripectes and those related genera with a transverse row

of nuchal cirri have the dorsalmost cirri on either side of the nape

connected to each other by a low basal membrane. The most parsimonious

explanation for the distribution of this character is that it has been

derived independently in Cirripectes filamentosus and the node XV

group. I consider the presence of independently based cirri to be an

apomorphic character that provides additional evidence for the

monophyly of the node XV group.

Node XVI Group

The members of the node XVI group are associated in an unresolved

trichotomy comprising Cirripectes polyzona, C. castaneus, and C.

stigmaticus. The node XVI group shares in lacking a pore position

behind the ventral edge of the row of nuchal cirri (Figure 17). Two

pore positions are present in Exallias, Ophioblennius, and

Scartichthys. I interpret the loss of the pore position as a

synapomorphy of this group.

Node XVII Group

The node XVII group comprises Cirripectes filamentosus, C.

chelomatus, C. springer, and C. perustus. Members of the node XVII

group have a smaller diameter foramen in the infrapharyngobranchial

plate (Figure 5) than other members of the node V group. I interpret

this decrease in size of the foramen in the infrapharyngobranchial

plate as a secondary modification and, as such, a synapomorphy of the

node XVII group.

Within the node XVII group, a sister group relationship between

Cirripectes chelomatus and C. filamentosus is supported by two

apomorphic characters: (1) adult male genital papilla with two short,

closely appressed, slender filaments, with one of the filaments located

on either side of the gonopore (Figure 13) and (2) dorsalmost nuchal

cirri arising directly from the skin and not connected basally by a low

membrane (Figure 7).

The male genital papilla of C. vanderbilti, C. chelomatus, and C.

filamentosus has been secondarily modified so that the filaments are

shorter and more closely appressed than are the two long, widely

separated, slender filaments in other members of the node X group. I

consider the closely appressed filaments in C. vanderbilti to be an

independent modification because this species does not share the

characters defining the node XIII and XIV groups. The presence of twu

short, closely appressed, slender filaments on the adult male genital

papilla is interpreted as a synapomorphy of C. chelomatus and C.


As discussed in the account of the node XV group, the dorsalmost

nuchal cirri arising directly from the skin is an apomorphic condition,

while the dorsalmost cirri connected basally by a low membrane is

plesiomorphic. The nuchal cirri morphology in the node XV group is

unique among blennioids and appears to have evolved separately from the

independently based nuchal cirri in C. chelomatus and C. filamentosus.

For this reason, I consider the independently based nuchal cirri in C.

chelomatus and C. filamentosus as evidence of monophyly for these two


Other relationships within the node XVII group are unresolved. The

three possible phylogenies for the node XVII group are shown in Figure

18. Three characters are involved in the different placements of taxa

in the cladograms in Figure 18: (1) loss of the mid-snout sensory pore

positions (Figure 17),. (2) incomplete lateral line, and (3) a

dorsal-fin pterygiophore formula (representing the number of proximal

pterygiophores inserting into each of the first four interneural spaces

with the first value being the number of proximal pterygiophores

inserting anterior to the first neural arch) of 1-1-0-1.

Cirripectes filamentosus, C. chelomatus, and C. springer lack

mid-snout pore positions. All other members of the node III group have

one or more pores in this position. The absence of mid-snout pores is

a synapomorphy of these three species.

An incomplete lateral line that extends posteriorly to about the

middle of the body is shared by Cirripectes filamentosus, C.

chelomatus, and C. perustus. All other members of the node IV group





1-4 4


4) 41

4-4 .

0 $



. C4 C
u O

0 m%

0 .


V-4 6O


*1-4 -
r. 0

0 W-

.,4 ".4
- 4

".4 -4

0 0


0vo v


S,,.1 ~


~ ~kv

have a complete lateral liie (reaches at least to caudal peduncle). An

incomplete lateral line is a synapomorphy of these three species.

A dorsal-fin pterygiophore formula of 1-1-0-1 is shared only by

Cirripectes stigmaticus, C. perustus, C. springeri, and Ophioblennius.

The dorsal-fin pterygiophore formula is typically 0-2-0-1 in the

Salariini. Although the dorsal-fin pterygiophore formula varies among

blennioid fishes, I have not found the 1-1-0-1 formula in other

blennioids. The 1-1-0-1 formula is apomorphic, but occurs

homoplasiously and, based on a parsimony argument, has been derived

independently in Ophioblennius and twice within Cirripectes (in the

node XVI and XVII groups).

Because the mid-snout pore, lateral-line, and dorsal-fin

pterygiophore formula characters each support a different pattern of

phylogenetic relationships, the relationships among Cirripectes

springeri, C. perustus, and the C. filamentosus-C. chelomatus group

must remain unresolved until one or more additional characters are

discovered that support one of the three hypotheses of relationship.

Cirripectes SWAINSON

Cirripectes Swainson, 1839:182, 275 (type species Salarias variolosus

Valenciennes in Cuvier and Valenciennes, 1836, by original

designation). Spelled Cirripectus on pp. 79, 80.

Diagnosis. Cirripectes is a salariin blenniid with: (1) a

transverse row of uniform-length, simple, nuchal cirri that may be in a

co.Linuous row or in up to four groups, but, when in more than one

group, with little or no gap at the dorsalmost point on the nape; and

(2) an adult male urogenital papilla with one or two tapering filaments

extending beyond the gonopore. Among blenniids, these conditions are

unique to Cirripectes.

Cirripectes may also be distinguished from other salariin genera

with a transverse row of nuchal cirri by the following combination of

characters: no vomerine teeth, large canine located posteriorly on

each dentary bone, 13-17 (rarely 13) segmented dorsal-fin rays, and

14-18 segmented anal-fin rays. Selected counts for the species of

Cirripectes are provided in Table 3.

Distribution. Cirripectes is widely distributed throughout the

Indo-Pacific region (Figures 19-26).

Nomenclatural discussion. Swainson (1839) used two spellings in

the publication describing this genus. Early in his paper, he


- .4




0 -4
.0 -v4-



J-j QQ

4-4 s$4


o cc

w w

r74 GOd

0' u

".4 Wi~


. -^ 0 -

.A I I ,-I
N ;" "
zZ U
f r i... P
.., d ,.

'. 0 .
- i 0, -

z .
a ^ ^ s '

0* 3
^V h- 1 il ,'

o. tD

.I0*' 1
I- I i ^ [

V^N?^ f Ir 4~/' ; T
o a ~

cc .^^ --'

00~~~ i8 '. c;Do2

\ s qz I *
~~J>S s=| t

E5 1 -.. S
i^ < S ( o

.0 -
z _J 3
( 0 8i 3
A? C,~




- I


a -






"-4 4)






".4 (13




o 0 0 o

if I u "

o e ?

. ? L 0 a

0 -

f o r

/ l 4
5o -~ L A


C 0 4^ Oa ULA/ < <

gF 8
I ? S
*~ ~ t .~ {,d<
*;'* ^c L '*

.i- I 0
2 a"-. -a .- o
04~c i'4A
v3 a

0 2 A'A4 ^ *
-22^- 4, r
4 I.
. a: a .;- ^ i>

4 0 O5 A
V "I ./ *IP O i

5, 02 s' ,
t~ **f i S a ; *
P~ ,

--6 S5 *' ^ *' 2 .4 ~^
<~~ -.< -*-iX f /

Li '0 i Cs

2:S S z I5 j
r^ + u o0 0o 4 ^ s.

a -.. -" 4

cZ- 2

4.443 4'-
L, 010
ir8 L '. "
.. cCO

A~ A
4i ~ -















o 0

0 i

a I
-1 2, 0 S

. 8 I'

1 44
, 01 s5


- ID

S 5
*? r,

SI C .

-. 1C

,-/ /' ds .-"_

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EV0KL4NTP_2QP25W INGEST_TIME 2012-02-17T17:03:49Z PACKAGE AA00003809_00001