Journal of Fish Biology (2009) 75, 2571–2584

doi:10.1111/j.1095-8649.2009.02452.x, available online at www.interscience.wiley.com

Captive hybridization of two geographically isolated

pygmy angelfish species, Centropyge fisheri and
Centropyge resplendens

F. U. Baensch*† and C. S. Tamaru‡

*Reef Culture Technologies LLC, 4079 Koko Drive, Honolulu, HI 96816, U.S.A. and
‡Department of Molecular Biosciences and Bioengineering, University of Hawaii-Manoa,
1955 East-West Road, Ag. Science 218, Honolulu, HI 96822, U.S.A.

(Received 6 November 2008, Accepted 21 September 2009)

This study documents the rearing of two pygmy angelfish species, Centropyge fisheri and Centropyge
resplendens, and the early life history and reproduction of their hybrid offspring. A C. fisheri female,
collected from Hawaii, and a C. resplendens male, captive-bred from parental stock collected from
Ascension Island, were maintained at the hatchery facility for 7 months. Continuous spawning was
achieved at a photoperiod cycle of 14L:10D and a water temperature of 26·5◦ C, range ±1◦ C. Over
the 110 day period, the C. fisheri female spawned 102 times, 57% of which resulted in embryos
(fertilized eggs). The mean ± s.d. fecundity per spawn was 730 ± 459 eggs (range 52–1967).
Fertility (% eggs that developed into embryos) of all eggs that were preserved was 22·4 ± 25·6%.
A total of 235 hybrid juveniles were raised through metamorphosis with an average larval survival
of 16·4%. Eight F1 hybrid juveniles isolated for further study began to display signs of reproductive
behaviour c. 300 days post-hatch (dph). Spawn resulting in non-fertile eggs were first obtained
319 dph, and fertilized eggs developing into embryos were obtained after 411 dph from at least
two female individuals. While no attempt was made at rearing the F2 larvae, embryo and larval
development were normal up to 8 dph. Reproduction and development observed for all hybrid
generations in this study were normal, similar to other Centropyge species and indicates a very close
phylogenetic relationship between what are currently considered distinct species, e.g. C. fisheri and
C. resplendens. © 2009 The Authors
Journal compilation © 2009 The Fisheries Society of the British Isles

Key words: fish larval culture; larval development; phylogenetic relationship; species concept;
taxonomy.

INTRODUCTION
The pygmy angelfishes (genus: Centropyge) are well known for their bright coloura-
tion, small size and captive hardiness. The genus is the largest within the marine
angelfish family (Pomacanthidae), comprising a total of 34 described species (Robert-
son & Allen, 2002; Pyle, 2003). Many species of the Pomacanthidae are highly prized
among aquarist and therefore heavily collected, accounting for c. 8% of the total trade
in marine aquarium fishes (Green, 2003). Within the family, the pygmy angelfishes
account for the majority of exports from the insular Pacific, which is considered

†Author to whom correspondence should be addressed. Tel./fax: +1 808 735 4405; email: rct@hawaii.rr.com

2571
© 2009 The Authors
Journal compilation © 2009 The Fisheries Society of the British Isles
2572 F. U . B A E N S C H A N D C . S . TA M A R U

the most heavily collected region in the world (Pyle, 2003). Reproductive biology is
similar among the species within the genus (Lobel, 1978; Moyer & Nakazono, 1978;
Bauer & Bauer, 1981; Thresher, 1982; Moyer et al., 1983; Moyer & Zaiser, 1984;
Aldenhoven, 1986; Hioki & Suzuki, 1987; Lutnesky, 1988, 1989, 1991; Hioki et al.,
1990). Reports by the various investigators also indicate similarities on their repro-
ductive strategy, behavioural patterns, timing and frequency of spawning, embryonic
and early larval development, exhibition of sexual dimorphism and dichromatism,
and a defined social structure that is similar for each species.
The fishes belonging to the genus Centropyge, like many reef fishes that produce
small pelagic eggs, have delicate and complicated larval stages. Despite numerous
endeavours over the past 30 years, the artificial propagation of this prized group has
only recently been successful, and the long sought prospect of commercial farming
enterprises is still in its infancy (Holt, 2003). The rearing technology, however,
continues to mature and has allowed for the production of cultured products to reach
the market (Baensch, 2002, 2003). Such advances offer some level of optimism that
a means to reduce the collection pressure of wild-caught individuals may be close at
hand. In addition, when available, artificially propagated individuals have advanced
the knowledge base as they provide invaluable and possibly the only source of
information regarding the early life history of this genus, especially for rare species.
The Pomacanthid family contains a relatively large number of reported natu-
ral hybrids for marine species, similar to the butterflyfishes (Chaetodonidae) (Pyle
& Randall, 1994). In that reference, eight presumed or probable naturally occur-
ring hybrids between various Centropyge species are reported. The various hybrids
and their probable parentage were: C. flavissimus (Cuvier) × C. vrolikii (Bleeker);
C. eibli Klausewitz × C. flavissimus; C. eibli × C. vrolikii ; C. loricula (Günther)
× C. potteri (Jordan & Metz); C. loriculus × C. shepardi Randall & Yasuda;
C. bispinosa (Günther) × C. shepardi ; C. heraldi Woods & Schulz × C. bicolor
(Bloch) and C. venusta (Yasuda & Tominaga) × C. multifasciata (Smith & Rad-
cliffe). All of these hybrids were located within the same geographic location as
their parent species. Hybrids between species inhabiting different oceans, however,
have yet to be reported due to obvious geographic barriers.
Defining a species in a way that can be applied to naturally occurring organisms
has proven to be challenging, to say the least, and has occupied scientists for over
a century. A universal species definition remains elusive and with the advent of
molecular biology, another dimension in assessing this most basic unit in biology or
taxonomic rank will undoubtedly keep scientists busy for centuries to come. In the
study of sexually reproducing organisms and in the current investigation, however,
the widespread use of the biological species concept, which defines a species as
members of populations that actually or potentially interbreed in nature and produce
fertile offspring, is referred to.
Currently, the taxonomic status and phylogenetic relationships of the genus
Centropyge revolve around morphology, cytogentic, biochemical and molecular data.
(Chung & Woo, 1998; Pyle, 2003; Bellwood et al., 2004; Schindler & Schneidewind,
2004; Bowen et al., 2006). While naturally occurring hybrids between several species
of Centropyge have been reported in the wild, at best, they remain only presumed or
probable hybrids as none have been tested using the latest in molecular techniques
(Pyle & Randall, 1994). A more direct approach, albeit artificial, is to create hybrids
by introducing sexually mature individuals that are normally geographically isolated

© 2009 The Authors

Journal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 2571–2584
 H Y B R I D I Z AT I O N O F C E N T R O P Y G E F I S H E R I A N D C . R E S P L E N D E N S 2573

from one another. This addresses the core issue of defining a species according to
the biological species concept.
From a practical standpoint, when one of the species is extremely rare and inher-
ently more valuable, it is a common practice to investigate the production of the
rarer individual by crossing it with a more common relative. This particular situa-
tion provides opportunities from both an applied and basic standpoint and presented
itself with the more common Hawaiian C. fisheri and rare C. resplendens forming
the basis of this study.

MATERIALS AND METHODS

Two adult fish were used during this 7 month study. The male individual was a captive-
bred C. resplendens (7·6 cm total length, LT ) and the female was a wild-caught C. fisheri
(6·2 cm LT ). The male C. resplendens was captive-bred at the hatchery (approximate hatch
date was May 2004) from wild-caught parents that were donated by a private individual from
Japan and originally collected from Ascension Island. The female C. fisheri was collected in
April 2006 in local waters around the island of Oahu, Hawaii.
On 10 May 2006, the two fish were placed in a 215 l barrel (92 × 60 cm; black) system.
The photoperiod was set to 14L:10D and water temperature was controlled at 26·5◦ C, range
±1·0◦ C throughout the period. The system design was similar to that used for condition-
ing Centropyge debelius Pyle (Baensch & Tamaru, 2009). Water quality was maintained by
siphoning the tank bottom two to three times a week and a 50% water change every month.
The fishes received three daily feedings of a homemade gel diet, flaked food (Ocean Nutri-
tion Formula 2; www.oceannutrition.com), small pellets (Spectrum, New Life International;
http://nlsfishfood.com) and frozen Mysis sp. (Hikari; www.hikari.info).
The pair was periodically observed for signs of reproductive behaviour. At the time this
was noted, egg collectors were placed within the tank’s outflow. Eggs were collected in
a 10 l planter pot modified with 160 μm screen on the bottom. Collected eggs were pre-
served for each of the 110 days the pair had spawned and was terminated due to the sale
of the broodstock. Embryos were counted and removed for larval rearing trials during the
first 42 days of the study. Both embryos and infertile eggs were preserved for the remain-
ing spawn. Egg development was documented on three separate occasions from fertilization
and through hatching at an incubation temperature of 26·5◦ C, range ±1·0◦ C. Methods and
equipment for egg separation, preservation and documentation are the same as those described
in Baensch & Tamaru (2009). Initially, the larvae were transferred every 7–10 days to new
40 l tanks at a maximum stocking density of two larvae l−1 . At 4–5 mm (LT ), larvae were
transferred to 80 l tanks at a maximum stocking density of one larva l−1 . Water temperature
was controlled between 26 and 27◦ C, range ±1◦ C throughout the larval period. For max-
imizing rearing space and live food resources, two larval rearing trials were conducted in
which the hybrid larvae were mixed with a limited number of C. debelius larvae. The hybrid
larvae could easily be distinguished from the C. debelius larvae due to their difference in size.
One rearing trial was conducted with only hybrid larvae. In all rearing trials, larvae were fed
size-strained wild copepods collected from local waters (Kaneohe Bay, Oahu, Hawaii) up to
75 days post-hatching (dph) and newly hatched Artemia sp. nauplii from day 35 to day 90
(Table I). The nutritional profile of the strains of Artemia sp. nauplii used in the rearing trials
(Great Salt Lake strain + Vietnam strain) was not modified using any commercial enrichment
media. Between 4 dph (early pre-flexion) and 50 dph (late post-flexion), the microalgae Nan-
nochloropsis occulata were maintained as a background algae at a density ranging between
1 × 105 and 1 × 106 cells ml−1 in the rearing tank as described by Tamaru et al. (1994). The
feeding regime and the design and maintenance of the rearing system are similar to that used
in a previously reported investigation rearing C. debelius larvae (Baensch & Tamaru, 2009).
Larvae were photographed (Nikon, SMZ 1500 and Nikon, Coolpix 995, micrometer;
www.nikoninstruments.com) and preserved (7 days in 5% buffered formalin followed by 70%

© 2009 The Authors

Journal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 2571–2584
2574 F. U . B A E N S C H A N D C . S . TA M A R U

Table I. Centropyge resplendens male × Centropyge fisheri female larvae and juvenile
feeding regime

Food type Food size range (μm) Age fed (day) LT (mm)
No food — 0–3 1·6–2·6
Copepod nauplii 50–90 4–15 2·6–4·2
Mostly copepodites 100–200 10–45 3·6–11
Adult copepods 300–500 40–75 9–25
Artemia sp.∗ 300–500 35–90 8·1–29
Grow-out diet∗∗ 700–1500 65– 15–

LT , total length.
∗ Newly hatched Artemia sp. (small and large strains).
∗∗ Grated gel diet, cyclopeeze (Argent Chemical Laboratories; www.cyclop-eeze.com), flakes, pellets.

ethanol) throughout the rearing period to document growth and development. All measure-
ments were made using computer software (CardioCalipers, Iconico, Version 3.3; http://www.
iconico.com/CardioCaliper/) on digital photographs.
Juvenile hybrids were grown out in three multi-tank 100% recirculating systems, each
with a total volume of 410 l. Each 75 l tank was stocked with c. 20 juveniles. System water
temperature was maintained at 26◦ C, range ±1◦ C and provided with 14L:10D photoperiod.
The system design, maintenance and the feeding regime were similar to those used for growing
out C. debelius juveniles (Baensch & Tamaru, 2009).
Eight juveniles were separated for future study and placed into a 440 l aquarium. The design
of 660 l recirculation system and tank maintenance and feeding regime was similar to that
used to spawn the parent adults. A water temperature of 25◦ C, range ±1◦ C and photoperiod
of 13L:11D were maintained for the remainder of the study. The harem was periodically
observed for signs of reproductive behaviour. When this was noted, an egg collector was
placed within the broodstock tanks outflow c. 1 h before lights out. Eggs were collected and
preserved during three periods (each c. 10 days) to determine fertility and fecundity. Fertile
eggs were separated and allowed to develop until starvation to examine early development
on two separate occasions. No attempts were made to rear the progeny (with food) resulting
from spawning of the hybrids.

RESULTS

S PAW N I N G
Spawning behaviour was first observed within 2 days of the fishes being placed
together and was typical of that observed for other Centropyge species (Bauer &
Bauer, 1981). Spawning behaviour commenced up to 90 min before lights out,
and spawning was observed from 10 to 30 min before lights out. The paired male
C. resplendens and female C. fisheri first spawned on 16 June 2006, which was
35 days after being introduced to each other. The first spawning event resulted
in the production of 52 eggs with an observed 12 embryos (e.g. 23·1% devel-
oping). Spawning of eggs was obtained at a water temperature of 26·5◦ C, range
±1·0◦ C and a 14L:10D photoperiod. No additional light and temperature cycling
was provided for this pair. The pair produced eggs almost every night from 16 June
to 14 November 2006. On at least five occasions, embryos were obtained when
spawning had not taken place before lights out. The spawned eggs mean diameter

© 2009 The Authors

Journal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 2571–2584
 H Y B R I D I Z AT I O N O F C E N T R O P Y G E F I S H E R I A N D C . R E S P L E N D E N S 2575

2000

1600

Number of eggs spawned

1200

800

400

0
01/07/06 01/08/06 01/09/06 01/10/06
Dates

Fig. 1. Daily changes in the number of eggs produced by the Centropyge resplendens male and Centropyge
fisheri female pair [Fertile ( ) and infertile ( ) eggs].

± s.d. 0·69 ± 0·02 mm (n = 14). Like other Centropyge species, hybrid embryos
are positively buoyant, spherical and transparent, and contained a single oil globule.
Embryo development was found to be the same as that of other previously docu-
mented Centropyge species (Bauer & Bauer, 1981; Hioki & Suzuki, 1987; Hioki
et al., 1990).
Fecundity and the number of embryos and infertile eggs varied noticeably through-
out the spawning period (Fig. 1). Details of the egg production observed for the
mating pair are summarized in Table II. The pair spawned a total of 102 times over
the course of the 110 days that made up the reporting period and produced mean
± s.d. of 730 ± 459 eggs per spawning. The estimated total number of eggs pro-
duced was 74 500 and ranged between 52 and 1967 eggs per spawning event. Not
all spawn were fertile with 58 (e.g. 57%) of the observed 102 spawnings containing
embryos. The production of embryos mean ± s.d. 300 ± 301 eggs per spawning
was also observed to be highly variable, ranging between 12 and 1597 embryos per
spawning event.

L A RVA L R E A R I N G
For this study, three larval rearing trials were initiated between 16 June and
28 June 2006 utilizing an estimated 1436 embryos from the C. resplendens male
× C. fisheri female cross. The end result of these efforts was a total of 235 hybrid
juveniles (e.g. 73 dph), which gives a 16·4% overall survival. Survival of the three
trials was 16·0, 16·3 and 17·1%, respectively.
A photograph of a 31 dph hybrid post-flexion larva 6·7 mm LT is given in Fig. 2.
For comparison, photographs of similar age or stage representatives of C. fisheri and
C. resplendens larvae obtained from previous work (Baensch, 2002; unpubl. data) are
also included in the figure. At this age and size, the caudal fin rays of the hybrid larvae

© 2009 The Authors

Journal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 2571–2584
2576 F. U . B A E N S C H A N D C . S . TA M A R U

Table II. Summary of egg production data for Centropyge resplendens male and Centropyge
fisheri female from 16 June to 3 October 2006

Mean temperature ± range (◦ C) 26·5 ± 1·0

Photoperiod 14L:10D
Spawing period (110 days) 16/6/06 to 3/10/06
Number of spawning events 102
Total number of eggs produced 74 500
Range of number of eggs spawned per spawning 52–1967
Mean ± s.d. number of eggs per spawning 730 ± 459
Total embryos produced 18 571
Total number of fertile spawning events 58
Per cent fertile spawning events 57
Mean ± s.d. number of embryos per fertile spawning 300 ± 301
Number of embryos (range) 12–1597
Mean ± s.d. fertility 22·4 ± 25·6%

(a) (b) (c)

Day 32 Day 31 Day 30

Day 47 Day 52 Day 55

Day 64 Day 73 Day 66

Fig. 2. Summary of post-flexion and juvenile developmental stages (a) Centropyge fisheri, (b) the hybrid
C. fisheri × Centropyge resplendens of the current study (trial 1) and (c) C. resplendens. (a), (c), docu-
mented in past rearing studies. Scale indicates 1·0 mm.

were fully developed. The pelvic fins were now obvious. Dark pigmentation had
developed on the dorsal region of its body with the overall larval body being mostly
silver. Anal, dorsal and pectoral fin elements were fully formed. The larvae were
virtually indistinguishable from similar sized C. fisheri and C. resplendens larvae.
A 52 dph hybrid post-flexion larval measuring 18·2 mm in LT is also presented
in Fig. 2. At this stage, the hybrid larvae exhibited a noticeable change in behaviour,
becoming more stationary and swimming less erratically. They oriented themselves
more towards the bottom of the rearing tanks, hiding in poly vinyl chloride pipes,

© 2009 The Authors

Journal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 2571–2584
 H Y B R I D I Z AT I O N O F C E N T R O P Y G E F I S H E R I A N D C . R E S P L E N D E N S 2577

which marked the start of settlement. The hybrid larva exhibited a yellow and pur-
ple body, a clear caudal fin and black anal and dorsal fins. A comparable stage
C. fisheri larva had a brown body, a clear caudal fin and black dorsal and anal fins.
A similar staged C. resplendens larva had a yellow body, a yellowing caudal fin
and yellow–black dorsal and anal fins. Eyes were yellow and brown, and pelvic fins
were bright blue in all three varieties.
A 72 dph hybrid juvenile measuring 25·1 mm LT also is shown in Fig. 2. At this
stage of development, full juvenile colouration was exhibited. Swimming motion
was more juvenile-like (slower, less erradic and less directional). Most hybrid larvae
became juveniles by 65 dph. C. resplendens and most C. fisheri larvae completed
metamorphosis 5 and 10 days earlier respectively. At the completion of settlement,
the three varieties were near 20 mm LT . They were nearly identical with regard
to their overall morphology but differed in colouration and colour patterns. The
head and dorsal region of the hybrid was orange, with a central body that was
purple, an orange and purple anal fin and a tail that was clear with some orange
pigmentation. Centropyge fisheri had a predominantly brown body and a clear caudal
fin. Centropyge resplendens had a bright blue body, blue anal fin and orange dorsal
and caudal fin.

G R O W- O U T
All juveniles obtained from the three rearing trials were reared to adults. Juvenile
mortality was very low, with only seven animals expiring over the course of the
240 day rearing period to adults. Data for growth were subjected to regression anal-
ysis (SYSTAT Inc, 1985; www.systat.com) and used to construct a statistical model
for growth of the hybrid offspring from hatching to 307 dph (Fig. 3).
Each data point up to day 60 dph represents the mean LT of one to three fish
sampled randomly out of the rearing tank. Growth was steady throughout the rearing
period. A quadratic equation was found to provide the best fit of the observed data
and explained 94·7% (r 2 = 0·947, P < 0·001) of the variation in growth obtained
over the course of the 307 days in culture (Fig. 3).

60
L T (mm)

40

20

0
0 100 200 300
Days post-hatch

Fig. 3. Summary of observed growth in total length (LT ) of the hybrid ( ) over the course of 307 days in
culture. The curve was fitted by: y = 0·384x − 0·000747x 2 − 1·907.

© 2009 The Authors

Journal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 2571–2584
2578 F. U . B A E N S C H A N D C . S . TA M A R U

H Y B R I D S PAW N I N G
Eight F1 hybrid juveniles isolated for further study began to display signs of
courtship behaviour after c. 300 dph. The largest fish, presumably a dominant male,
would exhibit spawning behaviour with two of the smaller fishes, presumably the
dominant females. No obvious dichromaticism was observed for any of the hybrid
individuals. The remaining fishes were smaller than the two largest females and did
not engage in the mating ritual when approached by the male.
Spawned eggs from the two female F1 hybrid progeny were first detected 319 dph.
Initially no fertility was observed. As the spawnings progressed, however, embryos
were detected in the spawn produced when their hybrid progeny was between 411
and 421 dph. The larvae developed normally up to first feeding (4 dph) after which
mortality increased. Mortality was complete after 8 dph, the same life span of other
Centropyge species reared without food (unpubl. data).

DISCUSSION
The C. fisheri and C. resplendens individuals were initially obtained from
populations that have a distinct colouration and are located near opposite sides of
the globe. Centropyge fisheri is described primarily as brown in colour and endemic
to the Hawaiian Islands and Johnston Atoll, whereas C. resplendens is blue with a
yellow dorsal fin and tail, and endemic to Ascension Island in the central Atlantic
Ocean (Debelius et al., 2003; Pyle, 2003).
Several geographic barriers temporally and spatially separate the two popula-
tions. The most recent direct connection between the tropical Atlantic and east
Pacific Ocean ended with the rise of the Isthmus of Panama, c. 3·1–3·5 BP (Coates
& Obando, 1996). An older connection between the tropical Atlantic and Indian
Oceans closed with the collision of Africa and Eurasia, c. 15–20 BP (Hallam, 1994).
Expectations based on the geographical isolation would designate C. fisheri and
C. resplendens as two distinct species separated by at least several millions of years
of evolution. Recent evidence (Bowen et al., 2006), including the results of this study,
however, indicates that the two species may be more directly linked (phylogeneti-
cally) than indicated by their current taxonomic status and geographic distribution.
It has been reported that C. fisheri and C. resplendens are closely related to
C. nigriocellus Woods & Schultz, C. flavicauda (Fraser & Brunner), C. acanthops
(Norman), C. argi Woods & Kanazawa, and C. aurantonotus Burgess (Debelius
et al., 2003; Pyle, 2003; Bowen et al., 2006). All six members have similar teeth,
mouth, gill rakers and body shape. They also share identical social behaviours
and occupy the same ecological niche (Schneidewind, 2004). Centropyge flavicauda
varies from green to blue to brown and is widespread in the Indo-Pacific;
C. nigriocellus is yellow with a black ocellus, inhabiting various island groups in
the central Pacific; C. acanthops is blue with an orange dorsal, occupying the east
coast of Africa; C. argi is blue with a yellow head, from Florida, the Bahamas and
Caribbean; C. aurantonotus is blue with an orange head, from the Lesser Antilles to
Brazil. Modern fish descriptions list the seven congeners as separate species due to
their different colour patterns (Allen et al., 1998; Debelius et al., 2003; Michael,
2004; Endoh, 2007). Recent studies, however, raise questions as to their exact
taxonomic relationship.

© 2009 The Authors

Journal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 2571–2584
 H Y B R I D I Z AT I O N O F C E N T R O P Y G E F I S H E R I A N D C . R E S P L E N D E N S 2579

In the only complete examination of pygmy angelfishes to date, the genus

Centropyge was divided up into several species complexes based on morpholog-
ical (e.g. meristic and morphometric) characteristics (Pyle, 2003). In accordance
with Schindler (2003), C. fisheri and C. resplendens were placed together with
C. nigriocellus, C. acanthops, C. argi and C. aurantonotus in the monophyletic sub-
genus Xiphypops within Centropyge.
Pyle (2003) discovered that C. flavicauda had been morphologically misidentified
by Smith (1955) as different from C. fisheri. He reported that, based on collec-
tions and field observations, numerous individuals in Hawaiian and Indo-Pacific
populations are morphologically indistinguishable, concluding the two previously
recognized species to be synonymous.
Schindler & Schneidewind (2004) postulated the African C. acanthops to be syn-
onymous with C. fisheri in addition to the Indo-Pacific C. flavicauda. The study
argues that, based on geographical proximity, morphological similarity and the wide
range of colour pattern inherent in C. flavicauda, the two forms could not be con-
sidered two separate species. The interpretation that C. acanthops is only a morpho-
logical variant of C. fisheri is supported by the field observations of a population
from the Maldives with an intermediate colour pattern that appears to represent a
self-sustaining population rather than hybrids (Schneidewind, 2004). The Maldives
are located at the geographic border of the known distribution of C. acanthops and
C. flavicauda. There appears to be an east to west variation in colouration rang-
ing from more dark brown populations to brownish-orange and yellowish ones,
and within selected populations there is even much more variation in this char-
acter than previously thought. Schindler & Schneidewind (2004) concluded that the
three congeners represent one evolutionary species, named C. fisheri after the oldest
taxonomic classification. Further reaffirming this argument is that the interbreeding
between the three possible morphological variants is well documented under captive
conditions (Frische, 1999, pers. data).
Theoretically, the coldwater upwelling of the Benguela Current has separated
tropical reef fauna in the Atlantic and Indian Oceans since its origin >2 million years
ago (Shannon, 1985; Marlow et al., 2000; Rocha & Bowen, 2008). Nevertheless,
the close relationship between many tropical organisms in the two oceans indicates
more recent larval dispersal in the order of only 150 000–500 000 years (Vermeij
& Rosenberg, 1993; Bowen et al., 1998, 2006; Rocha et al., 2005a). Short periods
of warming at the end of each ice age have caused temporary disruptions in the
coldwater Benguela Current upwelling (200 000, 300 000 and 400 000 BP) (Chang
et al., 1999; Peeters et al., 2004; Rocha et al., 2005a), thereby allowing occasional
crossings of planktonic larvae into the Atlantic Ocean. Such dispersal is especially
possible for species with the ability to delay metamorphosis when conditions are
unsuitable for settlement. This phenomenon was discovered in all Centropyge species
reared at the present facility, with larvae able to delay settlement up to 50 days during
unsuitable conditions (Baensch & Tamaru, 2009; unpubl. data).
A comparison study of the 454 base pair (bp) segment of the mitochondrial DNA
(mtDNA) control region of the five Xiphypops members by Bowen et al. (2006)
indicates a recent colonization pathway of C. acanthops from the Indian Ocean
directly to the west Atlantic Ocean in the order of 250 000 years evolving into C.
argi and C. aurantonotus followed by at least two waves of dispersal from the west
Atlantic Ocean to the Mid-Atlantic Ridge evolving into C. resplendens.

© 2009 The Authors

Journal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 2571–2584
2580 F. U . B A E N S C H A N D C . S . TA M A R U

The DNA genealogy of the three Atlantic species was found to be statistically
identical and did not exhibit previously isolated lineages, indicating that there was
insufficient time for mutations to accumulate in the mtDNA markers surveyed, a
pattern also observed in the Pacific pygmy angelfish, C. loricula (Schultz et al.,
2007). Dissimilarities between genetic separation and colouration have recently been
reported in a host of reef fishes including wrasses, hamlets, damselfishes and but-
terflyfishes (McMillan et al., 1999; McCartney et al., 2003). Bowen et al. (2006)
findings indicate that colour pattern alone is a poor base for taxonomic distinctions in
the absence of support from ecology, reproductive biology, genetics or morphology.
The lack of characters beyond colouration that distinguish C. argi, C. aurantonotus
and C. resplendens led Bowen et al. (2006) to suggest that they are morphological
variants of the same species. They caution against revising taxonomic nomencla-
ture based solely on mtDNA data, recommending that further evidence should be
obtained through morphological, captive breeding and field studies.
The above studies indicate that the Xiphypops subgenus contains six closely related
members that fall into two species complexes: the younger Atlantic complex that
evolved from C. acanthops c. 0·25 million BP and the older Pacific complex that
evolved from C. fisheri or C. flavicauda, also relatively recently. Centropyge resplen-
dens and the Hawaiian C. fisheri are the most distant relatives within this group,
based on estimates of genetic divergence (Bowen et al., 2006) and the more obvious
colour pattern and distribution differences.
The results of this study demonstrate that the biological relationship between
C. fisheri and C. resplendens is consistent with them being part of a closely related
group of clades possibly representing emerging biological species. The data sup-
porting this premise can be further summarized as follows: (1) A pair bond was
formed immediately and reproductive behaviour observed within 2 days after the
two fish were placed together. Members of Xiphypops are considered to be very
aggressive towards conspecifics and their close relatives (Michael, 2004; pers. obs.).
Pair bonding between species would be unlikely, unless closely related. (2) The pair
spawned without manipulation shortly (35 days) after being paired. This indicates
that both fish require similar environmental conditioning to spawn (unpubl. data),
which further supports a close biological relationship. (3) The reproductive strategy
of Centropyge is to produce a moderate number of eggs, daily, over a long reproduc-
tive life span (Bauer & Bauer, 1981). Spawning was nearly continuous, and batch
size and egg fertility averaged 730 eggs and 24%, respectively, consistent with the
reproductive performance of a mating pair of the same species. Spawning for other
Xiphypops species was found to occur almost daily, with batch size and average
fertility ranging from 93 to 977 eggs and 3 to 87%, respectively (Bauer & Bauer,
1981). The characteristics of the C. fisheri × C. resplendens spawning obtained in
the current investigation fall within the range of what has been documented for true
Centropyge species in captivity and are indicative of the close biological relation-
ship shared by the two species (unpubl. data). (4) The development of the larvae and
juveniles was observed to be normal and similar to that of other species, raised using
similar rearing protocols (Baensch 2002, 2003; Baensch & Tamaru, 2009; unpubl.
data). The larvae are almost indistinguishable from C. fisheri and C. resplendens
larvae up to settlement (near day 40) based on photomicrographs. After this time,
colouration appears to be the only distinguishing feature. No deformities or other
developmental flaws were observed. (5) Survival of F1 larvae was 16%. A total of

© 2009 The Authors

Journal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 2571–2584
 H Y B R I D I Z AT I O N O F C E N T R O P Y G E F I S H E R I A N D C . R E S P L E N D E N S 2581

10 Centropyge species were reared through metamorphosis to date using a similar

technique to the one described in Baensch & Tamaru (2009). Survival of all pre-
vious Centropyge rearing trials was lower than that of the hybrid, averaging 3·8%
(unpubl. data). Hybrid vigour may have contributed to the increase in survival and is
certainly an area for future investigations. (6) F1 offspring were healthy and behaved
normally. A total of 235 juveniles were reared to 2·5 cm (i.e. market size) or larger.
Growth, behaviour and development appeared normal, similar to the other species
reared (unpubl. data). (7) F1 offspring were able to produce fertile eggs with hatched
larvae that developed normally. Eight offspring that were retained for further study
formed a haremic social unit, typical of Xiphypops species. (8) The F1 offspring
display variations of an intermediate colour form of C. resplendens and C. fisheri,
closely resembling C. acanthops. Particularly interesting is the colour variation of
the caudal fin, which was clear in some individuals, as in C. fisheri, and coloured
in others, as in C. resplendens. The colour of the caudal fin is considered the main
colour difference that separates the three Atlantic Xiphypops species from the three
Indo-Pacific forms (Schindler & Schneidewind, 2004).
This study has demonstrated that the female C. fisheri and the male C. resplendens
reproduce normally, similar to other Centropyge species, and that the offspring are
healthy and capable of producing normal embryos. These results indicate an excep-
tionally close biological relationship between the two species. The hybrid’s colour
pattern supports Bowen et al. (2006) findings that C. acanthops is the closest, most
intermediate relative to both C. resplendens and C. fisheri, which strengthens the
theory that C. resplendens recently invaded the Atlantic Ocean via southern Africa.
How long have C. fisheri and C. resplendens been separated? The best quantita-
tive estimates are again obtained from genetic analysis. Comparisons of the 454 bp
segment of the mitochondrial DNA (mtDNA) control region estimates C. fisheri and
C. resplendens divergence to range between 630 000 and 950 000 years (6·3–9·5%)
based on a mutation rate of 10% per million years (Bowen et al., 2006).
Cytochrome b analysis of C. fisheri and C. resplendens indicates a divergence time
of c. 150 000–400 000 years (0·3–0·9%) based on a mutation rate estimate of
2% per million years in reef fishes (J. Schulz, unpubl. data). These time estimates
are regarded as short intervals to evolve reproductive isolation, achieve speciation
(Bowen et al., 2006) and assist in the explanation of the results obtained during
the current investigation. Both studies found the divergence times between the four
remaining Xiphypops species (C. nigriocellus was not analysed) to be shorter than
between C. fisheri and C. resplendens, suggesting a closer biological relationship.
The close biological relationship between the six Xiphypops species and many
other spatial-temporally separated marine organisms indicates that applying allopatric
speciation to marine organisms may not always be appropriate. First, the oceans are
a circum-global aquatic medium with few permanent physical barriers. Second, the
long pelagic stage of many marine organisms allows for dispersal over vast dis-
tances. The astounding diversity of marine life on coral reefs may in part be due to
ecologically based selection (ecological speciation), in addition to reproductive isola-
tion mechanisms (Rocha et al., 2005b). Thereby, spatial-temporally separated marine
organisms with minimal ecologically based divergent selection pressure would be
slower to diverge. All Xiphypops members occupy the same niche (Pyle, 2003;
Schindler & Schneidewind, 2004), which could explain their persistent close biolog-
ical relationship.

© 2009 The Authors

Journal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 2571–2584
2582 F. U . B A E N S C H A N D C . S . TA M A R U

In summary, the results of the current investigation indicate a very close phy-
logenetic relationship between the two current taxonomically discrete species, e.g.
C. fisheri and C. resplendens, and support in synonymizing the more closely related
Xiphypops species. Further investigations may shed light on the existence of multi-
ple species complexes each with its array of varying morphological variants for the
pygmy angelfishes.

Funding for the research that forms the basis for this report was obtained from the ongo-
ing sales of cultured products from Reef Culture Technologies LLC. The preparation and
revision of this manuscript was funded in part by a grant–cooperative agreement from the
National Oceanic and Atmospheric Administration, Project # A/AS-1, which is sponsored by
the University of Hawaii Sea Grant College Program, SOEST, under Institutional Grant No.
NA05OAR4171048 from NOAA Office of Sea Grant, Department of Commerce. The views
expressed herein are those of the authors and do not necessarily reflect the views of NOAA
or any of its sub-agencies. UNIHI-SEAGRANT-JC-06-23.

References
Aldenhoven, J. M. (1986). Local variation in mortality rates and life-expectancy estimate of
coral-reef fish Centropyge bicolor. Marine Biology 92, 237–244.
Allen, G. R., Steene R. & Allen, M. (1998). Guide to Angelfishes and Butterflyfishes. Perth,
Australia: Tropical Reef Research.
Baensch, F. (2002). The culture and larval development of three pygmy angelfish species.
Freshwater and Marine Aquarium Magazine 25, 4–12.
Baensch, F. (2003). Marine copepods and the culture of two new pygmy angelfish species.
Freshwater and Marine Aquarium Magazine 26, 156–162.
Baensch, F. & Tamaru, C. S. (2009). Spawning and development of larvae and juveniles of
the rare blue Mauritius angelfish, Centropyge debelius (1988), in the Hatchery. Journal
of the World Aquaculture Society. 40, 425–439.
Bauer, J. A. Jr. & Bauer, S. E. (1981). Reproductive biology of pygmy angelfishes of the
genus Centropyge (Pomacanthidae). Bulletin of Marine Sciences 31, 494–513.
Bellwood, D. R., Van Herwerden, L. & Konow, N. (2004). Evolution and biogeography of
marine angelfishes. Molecular Phylogenetics and Evolution 33, 140–155.
Bowen, B. W., Clark, A. M., Abreu-Grobois, F. A., Chavez, A., Reichart, H. & Ferl, R. J.
(1998). Global phylogeography of the Ridley sea turtles (Lepidochelys spp.) inferred
from mitochondrial DNA sequences. Genetica 101, 179–189.
Bowen, B. W., Muss, A., Rocha, L. A. & Grant, W. S. (2006). Shallow mtDNA coalescence
in Atlantic pygmy angelfishes (genus Centropyge) indicates a recent invasion from the
Indian Ocean. Journal of Heredity 97, 1–12.
Chang, Y. P., Chang, C. C., Wang, L. W., Chen, M. T., Wang, C. H. & Yu, E. F. (1999).
Planktonic foraminiferal sea surface temperature variation in the Southeast Atlantic
Ocean: a high-resolution record MD962085 of the past 400,000 years from the
IMAGES II-NAUSICAA cruise. Terrestrial Atmospheric and Oceanic Sciences 10,
185–200.
Chung, K. & Woo, N. (1998). Phylogenetic relationships of the Pomacanthidae (Pisces:
Teleostei) inferred from allozyme variation. Journal of Zoology 246, 215–231.
Coates, A. G. & Obando, J. A. (1996). The geologic evolution of the Central American
Isthmus. In Evolution and Environment in Tropical America (Jackson, J. B. C., Coates,
A. G. & Budd, A., eds), pp. 21–56. Chicago, IL: University of Chicago Press.
Debelius, H., Hiroyuko, T. & Kuiter, R. H. (2003). Angelfishes: A Comprehensive Guide to
Pomacanthidae. Chorleywood: TMC Publishing.
Endoh, K. (2007). Angelfishes of the World (Oceanographic Series). Miami, FL: Ricordea
Publishing.
Frische, J. (1999). Erfolgreiche nachzuchten im Meerwasseraquarium. Ruhmannsfelden:
Bede-Verlag.

© 2009 The Authors

Journal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 2571–2584
 H Y B R I D I Z AT I O N O F C E N T R O P Y G E F I S H E R I A N D C . R E S P L E N D E N S 2583

Green, E. (2003). International trade in marine aquarium species: using the global marine
aquarium database. In Marine Ornamental Species: Collection, Culture and Conserva-
tion (Cato, J. & Brown, C., eds), pp. 31–48. Ames, IA: Iowa State Press.
Hallam, A. (1994). An Outline of Phanerozoic Biogeography. Oxford: Oxford University
Press.
Hioki, S. & Suzuki, K. (1987). Reproduction and early development of the angelfish,
Centropyge interruptus, in an aquarium. Journal of the Faculty of Marine Science and
Technology, Tokai University 24, 133–140.
Hioki, S., Suzuki, K. & Tanaka, Y. (1990). Development of eggs and larvae of the angelfish,
Centropyge ferrugatus. Journal of the Faculty of Marine Science and Technology, Tokai
University 37, 34–38.
Holt, J. (2003). Research on culturing the early life stages of marine ornamental fish. In
Marine Ornamental Species: Collection, Culture and Conservation (Cato, J. &
Brown, C., eds), pp. 252–254. Ames, IA: Iowa State Press.
Lobel, P. S. (1978). Diel, lunar, and seasonal periodicity in the reproductive behaviour of the
pomacanthid fish, Centropyge potteri, and some other reef fishes in Hawaii. Pacific
Science 32, 193–205.
Lutnesky, M. (1988). Sexual dimorphism, dichromatism, and protogynous hermaphroditism
in the pomacanthid angelfish, Centropyge potteri. Pacific Science 42, 46.
Lutnesky, M. (1989). Stimulation, inhibition and induction of “early” sex change in the
pomacanthid angelfish, Centropyge potteri. Pacific Science 43, 196–197.
Lutnesky, M. (1991). A temporal-polygyny-threshold model and spawning in the pomacanthid
angelfish, Centropyge potteri. Pacific Science 45, 97–98.
Marlow, J. R., Lange, C. B., Wefer, G. & Rosell-Mele, A. (2000). Upwelling intensification
as part of the Pliocene-Pleistocene climate transition. Science 290, 2288–2291.
McCartney, M. A., Acevedo, J., Heredia, C., Rico, C., Quenoville, B., Bermingham, E. &
McMillan, W. O. (2003). Genetic mosaic in a marine species flock. Molecular Ecology
1, 2963–2973.
McMillan, W. O., Weigt, L. A. & Palumbi, S. R. (1999). Colour patterns, assortative
mating, and genetic differentiation in brightly coloured butterflyfishes (Chaetodon-
tidae). Evolution 53, 247–260.
Michael, S. (2004). Angelfishes and Butterflyfishes. Charlotte, VT: T.F.H. Publications.
Moyer, J. T. & Nakazono, A. (1978). Population structure, reproductive behaviour and pro-
togynous hermaphroditism in the angelfish, Centropyge interruptus, at Miyake-jima.
Journal of the Faculty of Marine Science and Technology, Tokai University 25, 25–39.
Moyer, J. T. & Zaiser, M. J. (1984). Early sex change: a possible mating strategy of
Centropyge angelfishes (Pisces: Pomcanthidae). Journal of Ethology 2, 63–67.
Moyer, R., Thresher, R. & Colin, P. (1983). Courtship, spawning and inferrel social organi-
zation of American angelfish (Genera: Pomacanthus, Holocanthus, and Centropyge;
Pomcanthidae). Environmental Biology of Fishes 9, 25–39.
Peeters, F. C., Acheson, R., Brummer, G. J. A., de Ruijter, W. P. M., Schneider, R. R.,
Ganssen, G. M., Ufkes, E. & Kroon, D. (2004). Vigorous exchange between the Indian
and Atlantic oceans at the end of the past five glacial periods. Nature 430, 661–665.
Pyle, R. L. (2003). A systematic treatment of the reef-fish family Pomacanthidae (Pisces:
Perciformes). PhD Thesis, Department of Zoology, University of Hawai‘i, Honolulu,
HI, USA.
Pyle, R. L. & Randall, J. E. (1994). A review of hybridization in marine angelfishes
(Perciformes: Pomacanthidae). Environmental Biology of Fishes 41, 127–145.
Robertson, D. R. & Allen, G. R. (2002). Shorefishes of the Tropical Eastern Pacific: An Infor-
mation System. CD-ROM. Balboa, Panama: Smithsonian Tropical Research Institute.
Rocha, L. A. & Bowen, B. W. (2008). Speciation in coral reef fishes. Journal of Fish Biology
72, 1101–1121.
Rocha, L. A., Robertson, D. R., Rocha, C. R., Van Tassell, J. L., Craig, M. T. & Bowen,
B. W. (2005a). Recent invasion of the tropical Atlantic by an Indo-Pacific coral reef
fish. Molecular Ecology 14, 3921–3928.
Rocha, L. A., Robertson, D. R., Roman, J. & Bowen, B. W. (2005b). Ecological speciation
in tropical reef fishes. Proceedings of the Royal Society B 272, 573–579.

© 2009 The Authors

Journal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 2571–2584
2584 F. U . B A E N S C H A N D C . S . TA M A R U

Schindler, I. (2003). Analysis of body shape in species of centropyge (Teleostei: Peraiformes:

Pomacanthidae). Zeitschrift für FischKunde 6, 37–47.
Schindler, I. & Schneidewind, F. (2004). Revision of Centropyge fisheri (Snyder, 1904)
(Teleostei, Pomacanthidae). Zeitschrift für Fischkunde 7, 31–42.
Schneidewind, F. (2004). Noch einmal Centropyge sp. rotkopf von den Malediven. Die
Aquarien- und Terrarienzeitschrift (DATZ) 57, 22–27.
Schultz, J. K., Pyle, R. L., DeMartini, E. & Bowen, B. W. (2007). Genetic connectivity
among colour morphs and Pacific archipelagos for the flame angelfish, Centropyge
loriculus. Marine Biology 151, 167–175.
Shannon, L. V. (1985). The Benguela ecosystem. Part I. Evolution of the Benguela phys-
ical features and processes. Oceanography and Marine Biology Annual Reviews 23,
105–182.
Smith, J. L. B. (1955). The fishes of the family Pomacanthidae in the western Indian Ocean.
Annals and Magazine of Natural History 8, 377–384.
Tamaru, C. S., Murashige, R. & Lee, C. S. (1994). The paradox of using background
phytoplankton during the culture of striped mullet, Mugil cephalus L. Aquaculture
119, 167–174.
Thresher, R. E. (1982). Courtship and spawning in the emperor angelfish, Pomacanthus
emperator, with comments on the reproduction of other pomcanthid fishes. Marine
Biology 70, 149–156.
Vermeij, G. J. & Rosenberg, G. (1993). Giving and receiving: the tropical Atlantic as donor
and recipient region for invading species. American Malacological Bulletin 10,
181–194.

© 2009 The Authors

Journal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 2571–2584