The utility of the cephalic lateralis system in Apogonid taxonomy
Laura M. Rodman

 Introduction:
            The Apogonidae, comprised of more than 250 species in 21 genera, is the eighth most speciose family in the Perciformes (Allen and Morrison, 1996; Allen and Robertson, 1994; Randall, Allen, and Steene, 1990).  Apogonids are relatively small fishes, reaching maximum lengths near 20 centimeters.  Apogonids are characterized as having two separate dorsal fins, the first with 6-8 spines and the second with one spine and 8-14 rays; an anal fin with two spines; scales that are ctenoid, cycloid, or absent; seven branchiostegal rays; and usually 24 (10 + 14) vertebrae (Nelson, 1996).  They are distributed circumglobally in tropical and subtropical waters, with their greatest diversity occurring on coral reefs in the Indo-Pacific.  Apogonids inhabit a wide range of habitats, from the deep ocean to the estuaries and freshwater streams of numerous Pacific islands (Springer, 1982).
            The systematic relationships of the species in the Apogonidae are poorly understood.  Fraser (1972) completed what is to date, probably the most comprehensive osteological study of the family.  Fraser anticipated that a greater understanding of the phylogenetic relationships within the Apogonidae would result from detailed study of proposed natural groupings.  He suggested that osteological characters provide a natural source of phyletic information.  This assumption provided justification to the means by which Fraser examined the fishes' osteology, and through comparison of apogonid, lower percoid, and beryciform osteological units hypothesized apogonid evolutionary history. 
        Fraser's study preceded the widespread acceptance and utilization of cladistic methods to elucidate phylogenies.  Although his work to establish the evolutionary history of osteological units via comparison with an "outgroup" appears cladistic in nature, Fraser did not apply cladistic methods to analyze his data.  In terms of inter-familial relationships, Fraser ascertained differences between apogonid and beryciform osteology and correctly hypothesized that the beryciform character states were the more ancestral forms. 
      Fraser did not always, however, accurately determine the intra-familial polarity of evolutionary change when more than one state existed for a given character.  Although he used beryciform osteological data and meristics to define primitive character states, he did not always adhere to the suggested trends when providing direction to character evolution.  For example, if the serrated condition of the opercle and preopercle in berycoids is assumed to be primitive, a smooth opercle and preopercle should be considered more advanced.  Fraser assigned the correct polarity to the loss of opercular spination but, incorrectly assigned the polarity of preopercular serration.  He designated contradictory conditions, the loss of serrations on the preopercular edge and the gain of serrations along the preopercular ridge, as advanced (Gon, 1996).
        Instead of using cladistic methodology to estimate the degree of relatedness between genera, Fraser based his hypothesized inter-generic relationships upon the degree of physical similarity between taxa.  He divided the apogonid genera into two phenetic groups upon the basis of the following characteristics: the retention or loss of the basisphenoid, the presence or absence of the supramaxilla, the morphology of the caudal fin, and the degree of cryptic coloration.  Within these groups, Fraser sequenced the genera in the direction of one of the two extremes; placing them between the more "advanced" and "primitive" genera with which they shared the largest number of characters.  He then collated these sequences into a phenogram, of sorts.  Interpretational difficulties, such as character reversions, parallel evolution of hypothetically advanced characters, and independent losses of characters within multiple lineages, plagued Fraser's resultant classification.  He found that the evolutionary path was more obvious for some character units than for others; and often, character sets suggested conflicting evolutionary histories.
        Fraser's neglect to define taxa upon the basis of synapomorphic characters lies at the root of his problematic analysis.  This is particularly apparent within the large and diverse genus Apogon.  Fraser defined the genus upon the following characters: five free hypurals, three epurals, a well-developed basisphenoid, a serrated preopercular posterior-edge, no supramaxilla, ctenoid scales, a shelf on the third infraorbital, a complete pored lateral line, nine segmented second-dorsal fin rays, and eight segmented anal fin rays.  These characters are symplesiomorphic in nature (Gon, 1995).  The first eight character states are shared with the berycoids and primitive percoids and Fraser was admittedly unsure of the degree of primitiveness of the latter two characters. 
        The probability of creating unnatural, i.e. non-monophyletic, taxa and taxonomic-hierarchies increases when plesiomorphic characters are used to define them.  In such a classification, when a character undergoes a change in multiple taxa it often appears anomalous in the evolutionary scheme, and a mosaic distribution of character states results.  Such is the case for the genus Apogon.  Four subgenera do not conform to the distinguishing characters of the genus: Yarica species retain the supramaxilla, Paroncheilus species have cycloid scales, Zoramia species have lost the infraorbital shelf, and species in Brephamia have incomplete lateral lines.  Although Fraser did not address this weakness of his defining character set, he recognized that even after his revisionary work Apogon may still represent an unnatural taxon.
        Fraser determined the subgeneric classification of Apogon using the same characters he used to distinguish the apogonid genera.  He divided the genus into 10 subgenera: Apogon, Brephamia, Lepidamia, Nectamia, Paroncheilus, Pristiapogon, Pristicon, Yarica, Zapogon, and Zoramia.  When evaluated within the context of the genus, as described by Fraser, eight of the subgenera may be defined by one unique character.  Apogon is the only subgenus with 0 uroneurals.  Brephamia is the only subgenus with an incomplete lateral line.  Lepidamia has much smaller scales, and thus a significantly higher number of them in the lateral line, than do other Apogon subgenera.  Paroncheilus is characterized by the presence of cycloid, versus ctenoid, scales.  Pristiapogon is the only subgenus characterized by having the intercalary facet barely on the otic bulla.  The presence of a supramaxilla distinguishes the subgenus Yarica.  Zapogon possesses a smooth ceratohyal.  Zoramia is distinguishable by the significantly long lengths of the first-dorsal fin spines.  Pristicon and Nectamia are the only two subgenera not defined by one unique (within Apogon) character.  The two subgenera are distinct from other Apogon subgenera because they lack the previously mentioned unique characters, and possess 2 or 3 predorsals and one pair of uroneurals.  The number of epipleurals distinguishes them from one another.  Pristicon has epipleurals on all but the last pair of ribs whereas, Nectamia has epipleurals on all but the last two pairs of ribs. 
        Theoretically, the characters used to define a genus should be different from those used to discern the subgenera and the states of the generic characters should be consistent across the subgenera.  Thus, the taxonomic utility of the generic characters is lessened at the subgeneric level.  So although it is possible to distinguish between the proposed Apogon subgenera, it is not possible to say that they are all groups of the same taxonomic rank.  Nor is it possible, because of the weak definition of Apogon itself, to inconclusively state that they are all subgenera of this genus. 
        Regardless of these problems, Fraser's work remains heavily relied upon as a source of information for defining and distinguishing apogonid genera and subgenera.  The taxa have remained, as defined by Fraser, with little modification through the years.  In order to expedite species identification, authors have created diagnostic keys utilizing external characters to distinguish the genera and subgenera, always citing Fraser's work as the source of osteological and internal character data (Fraser and Lachner, 1985; Fraser and Struhsaker, 1991). 
        Much of the work after Fraser's, has been committed to revising the taxonomy of the species in the family's genera and subgenera.  Nomenclatural difficulties plague the Apogonidae.  Many of the species and genera were described in the 1800s and either descriptions or illustrations were inaccurate or type material has either been lost or is in poor condition, thus making consequential identification difficult, if not impossible.  Fraser acknowledged the necessity of further taxonomic analysis at the specific level, stating that supra-specific relationships will not be understood until the taxonomy and systematics of the species are better known.
        Some taxonomic difficulties have been resolved with resulting changes to Fraser's classification of Apogon.  Randall, et al. (1990) removed Apogon aureus from synonymy with Apogon fleurieu and recognized them both as valid species.  Previously, upon the basis of an inaccurate specific description, A. fleurieu had been placed in the Oplegnathidae (Whitley, 1959).  Gon (1987) recognized the species' illustration as an apogonid and placed A. fleurieu in Fraser's Nectamia, synonymizing it with A. aureus.  The recognition of Apogon fleurieu as a valid taxon resulted in Ostorhinchus Lacepéde replacing Nectamia Jordan as a subgenus of the genus Apogon. 
        Gon (1995) reevaluated the subgenus Lepidamia and found it to be composed of four species: Apogon kalosoma, Apogon multitaeniatus, Apogon natalensis, and Apogon omanensis.  Gon did not add any new characters to the subgenera's description and referenced Fraser (1972) for osteological and swimbladder data.  He did, however, attempt to utilize cladistic methods in his analysis.  He argued that small scale size represents a derived character state in the Apogonidae, and is thus a synapomorphy which separates Lepidamia from the remaining Apogon subgenera.  Gon utilized meristic data and color characters to create a cladogram depicting the relationships among the Lepidamia species.  Gon stated that his phylogenetic hypothesis is supported by the generalization that rare (in frequency of occurrence) characters and small geographic distributions are equatable to derived characters and taxa.
        Gon (1996) has further revised the genus Apogon by removing Jaydia from its synonymy and elevating the taxon's status to the subgeneric level.  Jaydia was originally described by Smith (1961) to accommodate two species he believed did not belong in the genus Apogon.  Fraser (1972) did not believe sufficient differences existed to distinguish Jaydia from Apogon, or at least warrant its generic recognition, and synonymized the two genera.  Gon recognized six characters supporting the monophyly and recognition of Jaydia as a subgenus, four of which represent unique characters or character states within the genus Apogon.  Jaydia species are characterized by one of three preopercular serration patterns: a weakly serrate preopercular edge and a smooth to weakly serrate ridge; a serrate preopercular edge and ridge; or a smooth preopercular edge and ridge.  The latter two patterns are unique (within Apogon) to this subgenus.  A second distinguishing characteristic is the presence of a rounded caudal fin in adult Jaydia.  In all other Apogon subgenera, adults retain the emarginately shaped caudal fin from their earlier life history stages.  Gon considered the rounded caudal fin shape to be derived; as did Fraser (1972) who sited it as a diagnostic character for the hypothetically advanced genera Astrapogon, Pseudamia, and Pseudamiops.  The position of the longest dorsal-fin spine is also unique in Jaydia species.  In Jaydia, the longest dorsal-fin spine rests on the third pterygiophore; in all other apogonids, it rests on the second pterygiophore.  Gon considered this a synapomorphic character for the group.  Also, all Jaydia species possess or have secondarily lost (according to Gon's phylogenetic analysis) light organs associated with the intestine.  The organs produce light via a luciferin-luciferase reaction.  Jaydia species are not the only apogonids to possess light organs.  They are, however, the only apogonid species to employ the luciferin-luciferase mechanism for light production.  Gon recognized this character as a fourth synapomorphy uniting the group.  Gon validated the elevation of Jaydia to subgeneric status upon the basis of these derived characters. 
        Antithetically, Gon described an additional character which suggests that defining Jaydia as a subgenus of Apogon results in an unnatural, i.e. polyphyletic, generic taxon.  Apogonids show the complete range of reduction and loss of the eighth dorsal spine and its associated elements.  In this process, the loss of the spine is succeeded by the fusion of the eighth distal and ninth basal pterygiophores.  All Jaydia species have lost the eighth first dorsal-fin spine but, only one exhibits fusion of the eighth and ninth pterygiophores.  Examination of species from other Apogon subgenera (Nectamia, Apogon, and Pristiapogon) indicate that the pterygiophore elements are at least partially fused in other Apogon species.  Because the fusion of these two elements is a final step in the loss of the eighth first-dorsal spine, the fused condition is considered derived.  Gon suggested that the prevalence of the primitive, unfused, state in Jaydia implies that the subgenus cannot be nested within the genus Apogon.  Gon stated that further phylogenetic analysis of Jaydia may lead to the reinstitution of the group's generic status.
        Additional work had led to the refinement of Apogon subgeneric characters and consequently, to the removal of species from the subgenera.  Fraser (1972) had previously placed Apogon unicolor in the subgenus Pristicon.  This species he later removed on the basis of the presence of a ventral fleshy, unossified flap on the preopercle; small body scales; and a black stomach and intestine (Fraser, 1998).  Fraser stated that he believes Apogon unicolor may represent another, yet undescribed, lineage within the genus Apogon.
        Fraser (1998) augmented his previous study by publishing notes on the distribution of a number of character states across the Apogon subgenera.  Fraser divided the subgenera into phenetic groups upon the basis of the number of first-dorsal fin spines.  Species of Apogon, Zapogon, Paroncheilus, Pristicon, Yarica, Brephamia, and Zoramia have six first-dorsal spines.  Lepidamia and Pristiapogon species have seven first-dorsal spines.  Species of Ostorhinchus may have six, seven, or eight first-dorsal spines.  Fraser examined the distribution of the degree of preopercular-margin ossification and serration across the subgenera.  The six-spined species can be divided into three phenetic groups.  Apogon and Zapogon species have a preopercle with an unossified ventral margin and a serrate vertical edge.  The preopercle of Paroncheilus species have a partially ossified ventral edge and a serrate vertical edge.  The remaining six-spined Apogon species have preopercles with ossified, serrate ventral and vertical edges.  Fraser's data in combination with Gon's (1996) analysis of preopercular ridge and edge serration has led to additional distinguishing characteristics for the Apogon subgenera.  Pristicon may be further distinguished from Ostorhinchus on the basis of the degree of preopercular serration.  Pristicon species have a completely serrate preopercle, including a serrate ridge; whereas, Ostorhinchus species have a preopercle with a serrate edge and a smooth ridge.  Although detailed character analyses and examination of character-state distributions has led to an increase in the amount of knowledge available, thereby facilitating species identification; little systematic headway has been made from the additional information as cladistic methods have not applied to data analysis. 
        Taxonomic revision has proven to be a daunting task, as a lack of definable, non-variable characteristics precludes accurate species identification.  Apogonid species are often distinguished by coloration.  Although characteristic color markings may identify species relatively easily when the collection is fresh, they usually fade rapidly and leave the taxonomist with few clues to the fishes' identity (Smith, 1961; Greenfield, in press).  Also, the meristic characters are often stable and do not differ among closely related species.  One example of just such a complicated species identification is found within Pristiapogon, a subgenus of the family's largest genus Apogon. In this subgenus, four morphologically similar species exist: Apogon menesemus, Apogon taeniopterus, Apogon exostigma, and Apogon abrogramma.  A. menesemus and A. taeniopterus form one sympatric species pair and A. exostigma and A. abrogramma form another sympatric species pair.  The two allopatric species pairs differ only in their color patterns; meristic and morphometric measurements do not differ significantly.  Sympatric species differ in their color patterns and in at least one meristic or morphometric characteristic; however, overlap in frequency of  measurement occurs, inhibiting conclusive species identification of faded and preserved specimens (Fraser and Lachner, 1985). 
        Another group of species whose identities have been frequently confused is: Apogon coccineus, Apogon crassiceps, Apogon doryssa, and Apogon erythrinus (Smith, 1961; Shen and Lam, 1977; Shao and Chen, 1986; Greenfield, in press).  These four species superficially resemble one another in small size, shape, and lack of color pattern and pigmentation on the scales.  Apogon erythrinus has alternately been synonymized with Apogon crassiceps and Apogon coccineus (Smith, 1961; Shen and Lam, 1977; Randall, et al., 1990; Greenfield, in press).  Until recently the species have been distinguished upon the basis of differences in the number of gill rakers and proportional length measurements of the body and fins.  However, these characteristics were not unambiguous for intra-specific variation was known to occur within the geographic range of several of the species.  For example, it was possible to distinguish Apogon crassiceps and Apogon doryssa on the basis of morphometric variation.  But, Apogon erythrinus exhibited great enough morphological variation within its described geographic range to cover all these differences (Shen and Lam, 1977).  It was thought that specimens of A. erythrinus from the Red Sea attained a larger average size, a greater number of gill rakers, and had shorter fin spines than did specimens taken east of the Philippine Islands (Lachner, 1953; Smith, 1961).  Unequivocal identification of species from this group was considered difficult, if not impossible. 
        Recent work has distinguished these species and validated each of their existences (Greenfield, in press).  Apogon doryssa may be distinguished from the other species by its significantly long second-dorsal fin spine, triangular-shaped head, unique nasal configuration, and slender caudal peduncle (Masuda, et al., 1984; Greenfield, in press).  Greenfield recognized that the remaining species fall into two phenetic groups.  These groups he referred to as the Apogon erythrinus complex and the Apogon crassiceps complex.  The members of these complexes share the following characteristics: six first-dorsal fin spines, one second-dorsal fin spine and 8-9 rays, two spines and 7-8 rays in the anal fin, 16 or fewer total gill rakers on the first gill arch, 13-14 pectoral-fin rays, and two predorsals (Greenfield, in press).
        Greenfield (1994) revised the definition of Apogon erythrinus, listing it as a species endemic to the Hawaiian Islands and Johnston Atoll.  Up to this point, the species had been considered to have a large geographic distribution and a relatively broad character definition.  Detailed morphological examination has shown that these characteristics do not define only Apogon erythrinus but rather, define a group of closely related species, Greenfield's Apogon erythrinus complex.  As defined by Greenfield, this complex encompasses four species: Apogon erythrinus, Apogon indicus, Apogon marquesensis, and Apogon susanae.  These species share two characteristics.  The skin at the end of the snout, near the anterior nostril, is not formed into a flap; it is smooth.  All four species possess two full scales between the lateral line and the second and third spines of the first-dorsal fin (Greenfield, in press).  In contrast, species in the crassiceps complex have a free-edged flap of skin near the anterior nostril and a single, large, full scale between the lateral line and first dorsal fin (Greenfield, in press).  The crassiceps complex includes not only Apogon crassiceps and Apogon coccineus but also, a number of other Apogon species historically confused with Apogon erythrinus. 
        At the time of Fraser's 1972 publication the Apogonidae was thought to be composed of three subfamilies: the Epigoninae, the Pseudaminae, and the Apogoninae.  Johnson (1984) provided sufficient evidence to separate and elevate the Epigoninae to family status.  Several years later, Johnson (1993) defined three apomorphic characters shared between the Pseudaminae and the Apogoninae.  The first, is a short distal radial associated with the last spine of the first dorsal fin., so short, that there exists near contact between the proximal-middle element and the base of the spine.  Second, the configuration of the dorsal gill-arch elements of apogonids is distinctive: no articulation exists between the second epibranchial and second pharyngobranchial, the third epibranchial is expanded and the fourth is narrow, and the fourth upper pharyngeal toothplate lacks a fourth pharyngobranchial cartilage.  The last synapomorphy is the presence of horizontal and vertical rows of sensory papillae on the fishes' head and body.  Previously, these papillae had been thought to characterize only the pseudamine apogonids.  Johnson recognized their existence in most apogonine genera, and submitted it as a unifying character for the family. 
        The utilization of cutaneous sensory papillae, and other characters of the cephalic lateralis system, in taxonomic designation is not a new concept in ichthyology.  The systematic utility of these characters has been known and utilized, at least for gobiids, since Sanzo (1911) published his paper touting the importance of the sensory papillae patterns in goby systematics.  Within the past two decades cheek papillae patterns have been used within the Gobiidae to distinguish species, define groups of species, and define groups of genera.
        Miller (1998) created a key distinguishing three species of West African Eleotris upon the basis of a set of characters from the lateral line system.  Initially, identification was based upon the number of lateral-line scales only.  Although a practical method of identification, in terms of field utility, it did not always yield inconclusive species identification as scale count overlap did occur.  When the lateral-line scale count was employed in conjunction with characters from the head lateral line system, species identification was unambiguous.  Winterbottom and Burridge (1992, 1993a, 1993b) used cheek sensory papillae patterns to divide the goby genus Priolepis into two species-groups: those with a transverse pattern of cheek papillae and those with a reduced transverse pattern of cheek papillae.  McKay and Miller (1997) used characteristics of the cephalic lateralis canals and sensory papillae patterns to define a group of genera, the European sand gobies.
        The characters of the cephalic lateralis system have not gone unmentioned in apogonid taxonomy.  Lachner (1953) described Gymnapogon japonicus as resembling gobiids in general body shape and in having papillae arranged in a definite pattern over the head and body.  In this same publication, Lachner included illustrations and descriptions of papillae patterns in other species of Gymnapogon.  Mention of the presence of papillae and pores leading to subsurface channels has been made for many of the species and genera of the Pseudaminae (Smith, 1961; Randall, Lachner, and Fraser, 1985; Gon, 1986; Baldwin and Johnson, 1999).  Other authors have noted the existence of similar structures in Apogonine species (Randall, Fraser, and Lachner, 1990; Randall and Hayashi, 1990; Allen, 1995; Allen and Morrison, 1996).

Research Justification:
        Although some authors have noted the existence of cephalic canals, pores, and papillae, no one has attempted to describe or compare patterns among apogonid species or supra-specific taxa.  Published descriptions are general and brief, failing to do little more than state that such structures exist.  No one has attempted to define a methodology for naming the canals and pores of apogonid fishes, nor has anyone attempted to quantify and name the papillae patterns.  The potential taxonomic utility of the character set appears to have been nearly overlooked.  A few authors have suggested that information useful in elucidating interfamilial relationships may lie in these characters, for few families within the Perciformes possess cephalic sensory papillae associated with the lateralis system (Johnson, 1993; Baldwin and Johnson, 1999).  Even though parallels have been drawn between the cephalic lateralis systems of apogonids and gobiids, and the characters have proven themselves to be highly useful in goby taxonomy, no one has yet investigated their utility in apogonid species identification and taxonomy.
        Miller and Wongrat (1979) suggest that for highly diverse taxa their evolutionary history and taxonomy can best be understood by conjoining the analysis of complex characters sets, such as the body and cephalic lateral line system, with the analysis of simpler characters of reduction provided by the skeleton.  It is my hope that the cephalic lateralis characters, when added to Fraser's preexisting osteological data, will shed light on the identification and classification of the subgenera and species of the genus Apogon.

Questions to answer:
·      Do cephalic canal, pore, and sensory papillae patterns differ among apogonid species and subgenera?
·      Can the differences be used to create diagnoses for identifying species, species groups, or subgenera?
·      Are the differences consistent with previously establishes taxa?
·      Do differences in pore and canal configuration delineate groups at a higher taxonomic ranks than do difference in papillae patterns?
·      Can the nomenclature used for the goby cephalic lateralis system be applied and/or adapted for use with the apogonids?
·      Can cephalic lateralis characters differentiate otherwise nearly identical species?
·      Can the characters be used to identify long-preserved specimens in which other identifying characters have deteriorated over time?
·      Can the cephalic lateralis characters be polarized and used in cladistic analysis?

Research Plan:
            The primary objective of this study is to analyze the components of the cephalic lateralis system of the species and subgenera in the genus Apogon.
            Representative species of each subgenus will be studied and the pore and papillae patterns drawn.  Visualization of papillae patterns will be facilitated by staining the specimens with Cyanine Blue following the procedure published by Saruwatari, et al. (1997).  The pores and stained papillae will be observed with the aid of a dissecting microscope. 
            Previously established pore and papillae pattern nomenclature, used in gobioid taxonomy, will be modified and a standardized nomenclature will be instituted for the apogonid cephalic lateralis system. 
            The figures and specimens shall be compared in order to discern differences which may exist among species and subgenera.  The differences will be analyzed for congruence with established taxa.  Diagnostic patterns will be described for previously established Apogon species-groups and subgenera.  The differences shall be compared among the established taxonomic groupings in order to ascertain if patterns indicative of particular taxonomic levels or of particular taxa exist. 
            Ultimately, Fraser's osteological data will be analyzed using cladistic methodologies.  His characters will be subjected to a phylogenetic analysis program (PAUP).  The genus Holapogon will be employed as the outgroup.  Fraser (1973) described the genus citing it as the most primitive genus in the Apogoninae.  In comparison to other apogonid genera, Holapogon species retain the most complete array of primitive osteological character states, including: a high number of dorsal-fin spines, a perforated ceratohyal, and a complex caudal skeleton.
            The resultant cladograms will give insight into the monophyly of Apogon.  And by optimizing the cephalic lateralis system characters onto the most parsimonious tree, the taxonomic utility of the characters will be revealed.

Literature Cited:
Allen, G. 1995.  A new species of cardinalfish (Apogon: Apogonidae) from northwestern Australia.  Records of the Western Australian Museum.  17: 177-180.

 Allen, G. and S. Morrison.  1996.  A new species of cardinalfishes (Apogonidae) from northern Australia.  Records of the Western Australian Museum.  17: 439-442. 

Allen, G. and D. Robertson.  1994.  Apogonidae.  In: Fishes of the Tropical Eastern Pacific.  University of Hawaii Press: Honolulu.  121-123.

Baldwin, C. and G. Johnson.  1999.  Paxton concilians: A new genus and species of Pseudamine Apogonid (Teleostei: Percoidei) from northwestern Australia: The sister group of the enigmatic Gymnapogon.  Copeia.  4: 1050-1071.

Fraser, T. 1972.  Comparative osteology of the shallow water cardinal fishes (Perciformes: Apogonidae) with reference to the systematics and evolution of the family.  Ichthyological Bulletin of the J. L. B. Smith Institute of Ichthyology Rhodes University, Grahamstown.  34: 1-105.

Fraser, T.  1973.  Evolutionary significance of Holapogon, a new genus of cardinal fishes (Apogonidae), with a redescription of its type-species, Apogon maximus.  J. L. B. Smith Institute of Ichthyology Special Publication.  10: 1-7.

 Fraser, T.  1998.  A new species of cardinalfish (Apogonidae) from the Philippines, with comments on species of Apogon with six first dorsal spines.  Proceedings of the Biological Society of Washington.  111(4): 986-991.

Fraser, T. and E. Lachner.  1985.  A revision of the cardinalfish subgenera Pristiapogon and Zoramia (Genus Apogon) of the Indo-Pacific Region (Teleostei: Apogonidae).  Smithsonian Contributions to Zoology.  412: 1-47.

 Fraser, T. and P. Struhsaker.  1991.  A new genus and species of cardinalfish (Apogonidae) from the Indo-West Pacific, with a key to apogonine genera.  Copeia.  3: 718-722.

Gon, O.  1986.  Apogonidae.  In: Smith's Sea Fishes.  Smith, M. M. and M. M. Heemstra eds. Macmillan, Johannesburg, pp. 546-561.

Gon, O.  1987.  Redescription of Apogon (Ostorhinchus) fleurieu (Lacepéde, 1802) with notes on its synonymy.  Japanese Journal of Ichthyology.  4(2): 138-145.

Gon, O.  1995.  Revision of the cardinalfish subgenus Lepidamia (Perciformes, Apogonidae, Apogon).  Israel Journal of Zoology.  41: 1-22.

Gon, O.  1996.  Revision of the cardinalfish subgenus Jaydia (Perciformes, Apogonidae, Apogon).  Transactions of the Royal Society of South Africa.  51: 147-194. 

Greenfield, D.  1994.  Assemblage structure of small, cryptic, coral-reef fishes: an Atlantic-Pacific comparison.  Proceedings of the Fourth Indo-Pacific Fish Conference.  Bangkok, Thailand.  pp. 338-339.

Greenfield, D.  In press.  A revision of the Apogon erythrinus complex (Teleostei: Apogonidae).  Journal?

Johnson, G.  1984.  Percoidei: Development and Relationships.  In: Ontogeny and Systematics of fishes. H. G. Moser, et. al. eds.  Special Publication No. 1.  American Society of Ichthyologists and Herpetologists.  464-498.

Johnson, G.  1993.  Percomorph Phylogeny: Progress and Problems.  Bulletin of Marine Science.  52(1): 3-28.

 Lachner, E.  1953.  Family Apogonidae: Cardinalfishes.  In: Fishes of the Marshall and Marianas Islands.  Schultz, L. P., Herald, A. D., Lachner, E. A., Welander, A. D., and Woods, L. P.  Smithsonian Institution, United States National Museum. 1 (202): 412-498.

Masuda, H., K. Amaoka, C. Araga, T. Uyeno, and T. Yoshino, eds.  1984.  The Fishes of the Japanese Archipelago.  Tokai University Press: Japan.  pp. 143-151.

McKay, S. and P. Miller.  1997.  The affinities of European sand gobies (Teleostei: Gobiidae).  Journal of Natural History.  31: 1457-1482.

 Miller, P.  1998.  The West African species of Eleotris and their systematic affinities (Teleostei: Gobioidei).  Journal of Natural History.  32: 273-296.

 Miller, P. and P. Wongrat.  1979. A new goby (Teleostei: Gobiidae) from the south china Sea and its significance for gobioid classification.  Zoological Journal of the Linnean Society, 67: 239-257.

Nelson, J. 1996.  Fishes of the World, 3^rd ed.  John Wiley and Sons, Inc.: New York.  600pps.  

Randall, J., G. Allen, and R. Steene.  1990.  Cardinalfishes: Family Apogonidae.  In: Fishes of the Great Barrier Reef and Coral Sea.  University of Hawaii Press, Honolulu.  137-153.

Randall, J., T. Fraser, and E. Lachner.  1990.  On the validity of the Indo-Pacific cardinalfishes Apogon aureus (Lacepéde) and A. fleurieu (Lacepéde), with description of a related new species from the Red Sea.  Proceedings of the Biological Society of Washington.  103 (1): 39-62.

Randall, J. and M. Hayashi.  1990.  Apogon selas, a new cardinalfish from the western Pacific.  Japanese Journal of Ichthyology.  36 (4): 399-403.

Randall, J., E. Lachner, and T. Fraser.  1985.  A revision of the Indo-Pacific apogonid fish genus Pseudamia, with descriptions of three new species.  In: Indo-Pacific Fishes.  Eds.  J. E. Randall, H. A. Randall, and H. Bennett.  Bishop Museum Press, Honolulu.

Sanzo, L.  1911.  Distribuzione della papille cutanee (organi ciatiforme) e suo valore sistematico nei gobi.  Mitteilungen aus der Zoologischen Station zu Neapel.  20: 249-328. 

Saruwatari, T., J. Lopez, and T. Pietsch.  1997.  Cyanine Blue: A versatile and harmless stain for specimen observation.  Copeia.  4: 840-841.

Shao, K. and J. Chen.  1986.  Ten new records of cardinalfishes from Taiwan, with a synopsis of the family Apogonidae.  Journal of the Taiwan Museum.  39 (2): 61-104.

Shen, S. and C. Lam.  1977.  A review of the cardinal fishes (Family Apogonidae) from Taiwan.  Acta Oceanographica Taiwanica.  Science Reports of the National Taiwan University.  7: 154-192.

Smith, J. 1961.  Fishes of the family Apogonidae of the Western Indian Ocean and Red Sea.  Ichthyological Bulletin, Rhodes University 22: 373-419. 

 Springer, V. 1982.  Apogonidae.  In: Pacific plate biogeography, with special reference to shorefishes.  Smithsonian Contributions to Zoology.  367: 14.

 Whitley, G.  1959.  Ichthyological snippets.  The Australian Zoologist.  12(4): 310-323.

Winterbottom, R. and M. Burridge.  1992.  Revision of Egglestonichthys and of Priolepis species possessing a transverse pattern of cheek papillae (Teleostei; Gobiidae), with a discussion of relationships.  Canadian Journal of Zoology.  70: 1934-1946.

Winterbottom, R. and M. Burridge.  1993a.  Revision of Indo-Pacific Priolepis possessing a reduced transverse pattern of cheek papillae, and predorsal scales (Teleostei; Gobiidae).  Canadian Journal of Zoology.  71: 2056-2076.

Winterbottom, R. and M. Burridge.  1993b.  Revision of the species of Priolepis possessing a reduced transverse pattern of cheek papillae and no predorsal scales (Teleostei; Gobiidae).  Canadian Journal of Zoology.  71: 494-514.