U·M·I - ScholarSpace [PDF]

Sep 4, 1989 - 73 - 78. Schmale, M. 1981. Sexual selection and reproductive success in males of the bicolor damselfish, E

3 downloads 4 Views 8MB Size

Recommend Stories


umi-umd-2275.pdf (10.51Mb)
Never let your sense of morals prevent you from doing what is right. Isaac Asimov

UMI WeedCat
Raise your words, not voice. It is rain that grows flowers, not thunder. Rumi

Umi Purwandari
Life is not meant to be easy, my child; but take courage: it can be delightful. George Bernard Shaw

Umi Mardiyati & Khusfatun Khasanah
You're not going to master the rest of your life in one day. Just relax. Master the day. Than just keep

UMI Guide to Copyright
Pretending to not be afraid is as good as actually not being afraid. David Letterman

UMi Plus Extreme Origami
Ask yourself: Am I happy with my career? If not, what could I change about my job to be happier and

note to users umi
Ask yourself: How much of your time during an average week is spent doing things you dislike or that

registered nurses' use of personal communication ... - ScholarSpace [PDF]
Keywords: distraction; cellular phone; Internet; nurses ... frequency of personal communication device (basic cell phones, enhanced ... device had negatively affected their performance as a nurse (7.4%), that they ...... Table 4.23 Positive Versus Ne

registered nurses' use of personal communication ... - ScholarSpace [PDF]
Keywords: distraction; cellular phone; Internet; nurses ... frequency of personal communication device (basic cell phones, enhanced ... device had negatively affected their performance as a nurse (7.4%), that they ...... Table 4.23 Positive Versus Ne

umi-umd-5897 (2.260Mb)
At the end of your life, you will never regret not having passed one more test, not winning one more

Idea Transcript


INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

U·M·I Uruversity Microfrlms International A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor. M148106-1346 USA 313/761-4700 800/521-0600

Order Number 9300316

The life history and reproductive success of the coral blenny, Ezallias brevis (Kner, 1868) Carlson, Bruce Allan, Ph.D. University of Hawaii, 1992

Copyright @1992 by Carlson, Bruce Allan. All rights reserved.

V·M·I

30() N. Zech Rd.

Ann Arbor. MI 4X106

THE LIFE HISTORY AND REPRODUCTIVE SUCCESS OF THE CORAL BLENNY, Exallias brevis (Kner, 1868)

A DISSERTATION SUBMITTED TO THE GRADUATE DIVISION OF THE UNIVERSITY OF HAWAII IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN ZOOLOGY AUGUST 1992

BY BRUCE A. CARLSON

Dissertation Committee: Leighton R. Taylor, Chairman Ernst S. Reese George S. Losey E. Alison Kay Richard W. Grigg

c

copyright by Bruce A. Carlson 1992 All Rights Reserved

iii

This dissertation is dedicated to my parents, Richard and Maxine; their love and encouragement made this work possible.

iv

ACKNOWLEDGEMENTS

I wish to thank the members of my dissertation committee for their understanding, advice, and patience during the many years this work was in progress.

I would also like to thank

Marjorie Awai who accompanied me on dozens of dives to Hanauma Bay and Kahe Point at Waianae.

Without her constant

encouragement to continue, this work may never have been completed.

I also wish to extend special thanks to Thomas

Hourigan who joined me on many dives, and whose insight and comrr.ents were particularly helpful.

I also want to

acknowledge the contribution of the late Michael Weekley who volunteered to continue making my daily observations in Hanauma Bay during May 1981, while I was off-island. Finally, I would like to extend my appreciation to Tim Tricas for his advice and encouragement, to John E. Randall for the use of his photographs, to Susie and Thomas Kelly for their assistance with photos and artwork, to the staff of the Waikiki Aquarium

an~

particularly Daryl Imose for taking on

many of my duties at the Waikiki Aquarium while I was in the field and on sabbatical completing this dissertation.

And,

finally, to Lu Eldredge who graciously offered to proof-read a draft of this dissertation and made many helpful suggestions for improvements.

v

ABSTRACT

This work was undertaken to describe aspects of the life history of the coral blenny, Exallias brevis (Perciforrnes: Blenniidae), and patterns of individual reproductive success for an entire population of this fish. Field observations were conducted using scuba on a shallow reef at Hanauma Bay on the island of Oahu, Hawaii. Sixty-three fish at Hanauma Bay were fin-clipped and tagged with colored glass beads for identification, measured, and released.

During the period from November 1980, through

October 1981, a total of 240 dives was made to provide a near-daily record of spawning activity of these individuals. or "shelter"). six of the 11 focal-males did not have eggs during observations, but these males did spend time apparently preparing nest sites.

During 20 hours of observations of

these males, I occasionally observed them concentrating "nests bites" on rock patches. 66

Sometimes they grabbed small

pieces of algae and ripped them off the rock. material was spit out.

All this

Three males made a total of 93 such

bites on rocks during a period of 40 minutes. Early morning and late afternoon activity patterns During early morning and late afternoon observations, only three individuals, including two females and one male, were observed.

These observations were made on individuals

at Hanauma Bay and the results are listed in Table 2.3. A female at location (0.0,14.0) was observed on August 10, 1988 from 0645 hrs to 0945 hrs (temperature 27.S 0C). 0645 hrs she was already out and feeding.

At

However, she did

not venture very far when foraging, swimming less than one meter during each foray.

Also, she made few feeding forays

and only 15 coral-bites during this first hour.

However,

during the second and third hours of observations she spent progressively more time swimming and feeding.

Furthermore,

some of her feeding forays were more than six meters from the coral head where

I

first observed her at 0645.

A second female at location (A.0,13.0) was also observed for three hours but starting later in the morning, 0745 hrs, on August 8, 1988 (temperature 27.0 0C).

During these three

hours the percentage of time she spent swimming increased, as for the first female but the number of coral-bites increased only during the second hour then decreased during the third hour.

There were no differences in the distances covered

67

Table 2.3. -- Activity budgets of individual Exallias brevia during early mornings and late afternoons. Column headings: Female (or Male) indicates the sex and map coodinates of focal-individuals at the Coral Gardens study site in Hanauma Bay; %Swim is the percentage time spent swimming; %R,S is the percentage time spent either resting, or in shelter; %Int. is the percentage time spent interacting with conspecifics or extraspecifics; CBt is the total number of coral-bites; and CBt is the approximate percentage of the total observation time required to make the coral-bites. ===========================================================

Female Date D,14 8-10-88 II II

" "

Time 0645-0745 0745-0845 0845-0945

3.72 4.25

%R,S 97.66 95.63 95.06

%Int. 0.00 0.12 0.00

CBt 15 19 25

CB% 0.42 0.53 0.69

3.28 3.37 3.95

96.42 96.02 95.86

0.08 0.08 0.00

8

19

0.22 0.53 o, 19

%Swim 1. 92

II

II

II

"

0745-0845 0845-0945 0945-1045

8-08-88

0630-0730

remained in shelter hole

7-28-88

0800-0900 0900-1000

3.63 5.58

96.24 93.61

0.08 0.00

1400-1500 1500-1600 1600-1700 1700-1730

5.80

93.45 98.23 98.55 99.74

0.00 0.00 0.00 0.00

27

1. 63 1. 37

3 1

0.75 0.14 0.08 0.03

95.42 97.00 98.33

0.00 0.00

22 9 6

0.61 0.25 0.17

A,13

Male Q,6 II

8-08-88

II

"

II

7-27-88

" " " " " "

" " " 7-29-88 II

"

1430-1530 1530-1630 1630-1730

0.23 3.97 2.75 1.50

68

0.00

7

2

29 5

0.05 0.81

during each of the three hours, with a maximum distance of about seven meters between feeding locations. A male at location (Q.0,6.0) was observed in a shelter hole at 0630 on the morning of August 8, 1988, but he remained in the hole until 0730. at that time.

He was not caring for eggs

Observations were then switched to the female

at (A.0,13.0) beginning at 0745. The same male at (Q.0,6.0) was observed on an earlier date, July 28, 1988, for a period of two hours beginning at 0800 hrs (temperature 26.6 0C).

During the first hour he

spent mvch of his time swimming around his territory occasionally poking his head into holes and sometimes entering holes for brief periods (less than one minute).

He

fed only twice, once at the start of the hour and a second time near the end of the hour.

The greatest distance covered

during this period was less than four meters.

During the

second hour of observations on this male, beginning at 0900 hrs, the number of coral-bites and the swimming time both increased considerably.

Also, the distance between the two

most distant feeding points during this hour was about seven meters. During the afternoons of July 27 and 29, 1988, the male at (Q.0,6.0) was observed for 3.5 hours and 3 hours, respectively (temperatures 27.0 o C, and 27.9 0 C, respectively). During both afternoons, the amount of time he spent swimming and feeding diminished with time, while an increasing amount 69

of time was either spent perching or sheltering.

On July 27,

he moved into his shelter hole at 1702 hrs and did not reemerge by the time observations terminated at 1720 hrs.

On

July 29, he moved into his shelter hole at 1657 hrs and remained there until observations were terminated at 1730 hrs.

Much of Hanauma Bay was in the shadow of Koko Crater at

this time of day, but the sun did not "set" behind the ridge until 1732 hrs on July 29, 1988. Activity bUdget of a non-Hawaiian E. brevis The results of focal-male observations on one individual in Papua New Guinea were similar to those of Hawaiian brevis.

This male was guarding eggs which were located under

a branch of a large Pocillopora sp. head.

~.

(~.

eydouxi ?) coral

The site was near the seaward edge of the drop-off at

a depth of 6.5 m.

The time was 0938 - 1038 hrs.

The

temperature was 27.7 oC at the start of observations and 28.4 oC at the end of the hour. Hawaiian

~.

This male was smaller than

brevis and was estimated to be 85 - 90 rom SL.

During this one hour, he swam 84 seconds (=2.3% of time) he made 13 coral-bites (=0.36% of time) and spent the remaining time (97.34%) either perching in the open or on his nest guarding eggs. his eggs 35 times.

While egg guarding, he mouthed

I estimated the maximum foraging distance

from his nest site to be about three meters.

70

section 5: Foraging and Feeding Ecology

Introduction Most tropical salariin blennies which have been examined have been shown to be herbivores or omnivores, and apparently by extension some popular accounts of

~.

brevis have assumed

that it too is a herbivore (for example, Reader's Digest Book of the Great Barrier Reef, 1984, pp. 288 - 289). Strasburg (1960) reported that

~.

Hiatt and

brevis at Enewetak is a

herbivore, but Hobson (1974) reported that in Hawaii it is a corallivore. This section on the foraging and feeding ecology of ~.

brevis was undertaken to resolve this discrepancy and to

determine if there is a difference between Hawaiian and nonHawaiian

~.

brevis as corallivores or herbivores.

This

section also provides a foundation for the investigation in Chapter IV on filial egg cannibalism by male

~.

brevis.

Methods During focal-individual sampling described in the previous section, observations were made on the number of coral-bites made by individuals and the species of corals preyed upon at both the Hanauma Bay and Kahe Point study sites.

The same data is used in this section but only for

observations made between 0900 hours and 1500 hours: observations totalled 45 hours. 71

An additional 18 hours of

feeding observations were made during which no other activities were recorded. therefore was 63 hours.

The total sampling period No attempt was made to measure the

abundance or areas of each coral species within

~.

brevis

territories for comparison with the proportions of coral species ingested. On November 22, 1985, a weekly photographic record was made of 6 coral-bite scars on Porites lobata resulting from feeding activity of an adult study site.

~.

brevis in the Coral Gardens

Photographs of each scar were taken each week

through December 12, 1985. through January 11, 1986.

Additional observations continued Each scar was relocated by marking

it with a small nail driven into the coral head approximately 1 cm from the scar.

Photographs were taken using an SLR

camera in an underwater housing fitted with a 50 rom macro lens and lighting supplied by an electronic flash. A millimeter scale included in the initial photograph permitted the area of each scar to be measured.

Each

photograph was projected onto a piece of paper and the outline of the scar was then traced.

The area was then

calculated using a polar planimeter.

Coral polyps ingested

with each bite were counted on each initial photograph. Stomach contents of

~.

brevis collected at Kahe Point

were examined under a binocular microscope to see whether or not food items other than coral were present.

72

Additional feeding observations outside of Hawaii were made at Enewetak Atoll in June, 1976, and on the barrier reef near Motupore Island, Papua New Guinea, on November 25, 1984.

Results Hawaiian Exallias brevis feeding observations Exallias brevis at both Oahu study sites preyed upon Porites lobata, Pocillopora meandrina, Montipora spp. verrucosa and

(M.

M. flabellata combined), Leptastrea purpurea,

Pavona varians, and Cyphastrea ocellina.

The number of

coral-bites made on each of these corals is shown in Table 2.4.

Males fed primarily on g. lobata (88% of all coral-

bites) and made an average of 13.94 coral-bites per hour (s.d.

= ±8.17, n = 48 observation hours).

Females also fed

primarily on g. lobata (53.8%), but other corals comprised a larger proportion of their. diet than for males.

Females

also fed at a higher rate than males, averaging 28.40 coralbites per hour (s.d.

= ±11.03, n = 15 observation hours).

One of the six coral-bite scars photographed on November 12, 1985, is illustrated in Figures 2.10a and 2.10b.

Figure

2.10a was taken immediately after the scar was made.

Much of

the surface tissue (coenosarc) was removed, and apparently some polyps were also removed although others were missed, particularly in the lower portion of the scar.

One week

later another photograph of this scar was taken (not illustrated), and the polyps within the scar had regenerated 73

Table 2.4. -- Coral species preyed upon by Exallias brevis at Hanauma Bay and Kahe Point, Oahu. Column headings: CB = coral bites; CB/hr = mean number of coral bites per hour; percent = percentage of each coral species ingested. Observation hours for males totalled 48 hours (n = 15 males); for females the total was 15 hours (n = 5 females). =========================================-==================

CORAL SPECIES Porites lobata Pocillopora meandrina Montipora .§!m. Leptastrea purpurpea Cyphastrea ocellina Pavona varians TOTALS

CB 589 65 10 4 1 __ 0 669

MALES CB/hr Percent 12.27 1.35 0.21 0.08 0.02 0.00 13.94

74

88.0% 9.7% 1. 5% 0.6% 0.1% 0.0% 99.9%

CB 229 76 42 16 0

---&l. 426

FEMALES CB/hr Percent 15.27 53.8% 5.06 17.8% 2.80 9.9% 1.07 3.7% 0.00 0.0% 4.20 14.8% 28.40 100.0%

Figure 2.10.

Regeneration of an Exallias brevis feeding scar on Porites lobata coral. a) Fresh scar photographed on November 22, 1985 immediately after being made by an adult ~. brevis; b) the same scar 20 days later on December 12, 1985. The scale is in millimeters.

75

A

8 76

but their tentacles did not extend beyond the rim of the calyx.

Also, none of the coenosarc tissue had regrown.

The feeding scar remained visible 20 days later (Figure 2.10b) although the polyps had regenerated.

The area of

surface coenosarc remained unpigmented or had not regenerated.

After 50 days the scar had almost completely

healed but remained faintly visible (not illustrated). The mean area of a Porites lobata feeding scar was 2.04 cm2 (s.d. = 0.42, n = 10).

Polyp density for Porites lobata

was calculated from photographs at 71.40 polyps/cm 2 (s.d. ±6.7, n

= 10 samples).

=

Tricas (1986) published a similar

estimate of 76.7 ±5.7 polyps/cm 2 for Porites lobata. A photograph of a Pocillopora meandrina coral immediately after a coral-bite from an adult illustrated in Figure Figure 2.11b.

2.11~,

~.

brevis is

and again one week later in

The tips of each coral branch were much

lighter where feeding had occurred but many of the polyps remained visible and apparently were not removed during feeding. stomach contents of

~.

brevis collected at Kahe Point

revealed little identifiable material.

The bulk of the

contents were gelatinous although some whole polyps, and tentacle rings were present.

Small pieces of calcium

carbonate were also present in some individuals indicating that coral skeletal material was occasionally ingested while

77

Figure 2.11.

Appearance of an Exallias brevis feeding scar on Pocillopora meandrina coral. a) The coral branch immediately below the millimeter scale had just been fed upon by an adult ~. brevis on November 22, 1985; b) the same coral photographed on November 29, 1985.

78

UGJ.

A

B 79

UUJ..

Utf

feeding.

A few stomachs (N = 3) contained fragments of

unidentified algae. Non-Hawaiian Exallias brevis feeding observations One male

E. brevis observed for one hour on the barrier

reef near Motupore Island, Papua New Guinea, fed on Acropora sp., Porites sp., Montipora sp., Pocillopora sp. (£. eydouxi

?), and a faviid coral (Goniastrea ?).

When feeding on

massive corals he made two bites in rapid succession (couplets) but made three bites in rapid succession (triplets) when feeding on Acropora. At Enewetak atoll, I observed a female

~.

brevis perched

on branches of Millepora sp. and also observed numerous feeding scars nearby on this same hydrocoral.

After watching

her for about five minutes she began feeding on the Millepora (Figures 2.12a and 2.12b).

Section 6: Territoriality Introduction By definition, a territory is an area occupied more or less exclusively by means of repulsion of conspecifics or heterospecifics through overt defense or advertisement, whereas a home range is an area that an individual patrols regularly but does not defend (Wilson, 1975).

Territorial

behavior has been documented for many coral reef fishes including the damsel fishes (Pomacentridae), the butterflyfishes (Chaetodontidae), the surgeon fishes 80

Figure 2.12.

Exallias brevis at Enewetak Atoll. a) A female ~. brevis perched among the branches of Millepora sp. hydrocoral; b) ~. brevis feeding scars on the same Millepora colony.

81

82

(Acanthuridae), and others (Reese, 1964; Myrberg and Thresher, 1974; Thresher, 1976; Tricas, 1986).

Territori-

ality has also been reported among blennies (Strasburg, 1953; Gibson, 1969; Stephens, et al., 1970; Nursall, 1977 and 1981).

Nursall's papers are particularly informative in

describing the extent of territoriality in the redlip blenny, Ophioblennius atlanticus.

Focal-individual results, reported

in section 1 and 2 of this chapter, indicate that both male and female

~.

brevis show aggressive behavior towards

conspecifics.

As a consequence, individuals maintain

exclusive areas over extended periods of time.

The intent of

this section is to document the size of these territories.

Methods Individual male and female

~.

brevis were observed for

one hour periods on two or three different days, and their movements were recorded on an underwater slate.

A diagram

of the study site was drawn on the slate and movements of the fish were plotted on this diagram.

The resulting set of

points were later transferred to the study-site map (Figure 1.3).

These polygonal areas were measured using a polar

planimeter.

The term "territory" is used herein to describe

these areas even though defense was not always observed nor tested.

Home ranges may be much larger as some individuals

were occasionally observed far from their usual territories. 83

Results I observed individual male and female

~.

hrevis in their

own exclusive areas, or "territories", day after day. areas were areas not shared with other during spawning.

~.

These

brevis except

Both sexes defended these areas when

conspecifics of the same gender, and occasionally the opposite gender, intruded. On rare occasions,

female~.

brevis chased the blue-eye damselfish, Plectroglyphidodon johnstonianus. These territories varied in size from 6.45 m2 to 32.51 m2.

Table 2.5 lists individual territory sizes for

~.

brevis

at both the Coral Gardens and Kahe Point study sites. Females had larger territories at both sites compared to males.

Among males, those caring for eggs had smaller

territories than males not caring for eggs. Some individuals ranged widely outside their territories.

Male 'WI traversed over 37 m after abandoning one nest

site and establishing another (see Appendix A).

On another

occasion, on January 1, 1981, I observed male '0' 21.25 m away from his territory (see Appendix A).

On March 7, 1984,

I observed an untagged, non-nesting male at at (N.O,10.0), but after he chased away an intruding male, he then proceeded to follow him to (J.O,4.0).

The two did not interact again

but the foray covered a distance of 17.5 m. Females also made forays far outside their territorial boundaries.

For example, during May, 1984, I observed 84

Table 2.5. -- Size of individual territories of male and female E}{allias brevis. Individuals in column headed "IDIi identified with a "k" were observed at the Kahe Point study site; all others were observed at the Hanauma Bay study site. ============================================================

MALES WITH EGGS:

ID 1k 2 11 13 15 16 17 18 ~=

MAL~S

WITH0DT EGGS:

FEMALES:

TERRITORY SIZE 10.84 m2 6.45 m2 10.84 m2 13.67 m2 6.71 m2 6.45 m2 10.06 m2 7.22 m2 9.03 m2 , s.d. = ±2.70

ID TERRITORY SIZE 18.58 m2 1k 4k 26.83 m2 9.03 m2 15 16 9.80 m2 37.41 m2 0.6 ~= 20.33 m2 , s.d. = ±12.00 ID TERRITORY SIZE lk 29.67 m2 3k 32.51 m2 4k 16.77 m2 12.67 m2 4 A,13 25.03 m2 D.14 31.22 m2 ~= 24.64 m2 , s.d. =±8.21

85

female '4' for three hours and measured a territory of 12.64 m2.

But over a period of three years beginning on March, 10,

1981, I observed her. feeding in many locations covering a total area 138.03 m2 . Females also moved outside their territories when spawning.

As discussed in Chapter III, and Appendix B, I

observed female'S' moving at least 39 m from a spawning site to her territory.

On August 8, 1988, I followed an untagged

female after she finished spawning with a male at (F.0,17.0). She headed in a north-westerly direction and after 15 minutes had moved about 38 m before I lost her in the coral.

Section 7: Nest site Characteristics Introduction In his review of the reproductive biology of reef fishes, Thresher (1984) noted that all blennies appear to lay demersal eggs in a cave or shelter-hole guarded by the male. Exallias brevis follows this same pattern. This section describes the specific locations and characteristics of the sites used by individual male

~.

brevis as nests, as a prelude to an investigation of the reproductive success of these individuals in Chapter III.

Methods The location of nest sites at the Coral Gardens study sites was recorded with reference to the site map (Figure 86

1.3).

The orientation of each nest site was determined using

a Morin rangefinding compass ("opticompass") and was defined as the bearing perpendicular to the nest.

Results Locations of nests At Hanauma Bay, male throughout the study area.

~.

brevis established 38 nest sites Two additional nests were located

immediately outside the study area (see Appendix A for male lUI

and male 'WI).

The greatest concentration of nest sites

occurred on the large Porites lobata coral head located within the coordinates (G.O to J.7, and 16.5 to 18.3, Figure 1.3).

This coral head is also clearly visible near the lower

right corner of the bordered area in Figure 1.2 where it appears lighter in color than surrounding coral heads. Eleven different sites were utilized by eight males on this coral head throughout the study period.

The maximum number

of males simultaneously residing on this coral head was seven of which five were guarding nests.

Five additional nest

sites on this coral head have been used by males since I first observed 1976.

~.

brevis at Hanauma Bay beginning in January,

Since 1976, I have not observed any females as

permanent residents anywhere on this coral head. other males established nests at scattered locations throughout the study area.

All of these males were solitary

occupants of the coral heads or rocks where they established 87

nests, with the brief exception of males 'B' and 'I'.

These

two males co-occupied opposite sides of the coral head at (J.5, 8.0) for 27 days before male 'I' moved to the nest site vacated by male 'E'. Characteristics of nest sites Table 2.6 summarizes the characteristics of 38 nests within the study area according to substratum: 1) dead coral rock, 2) freshly cleaned surface of living coral with all polyps removed by the male, or 3) terrigenous boulders; exposure: 1)

nest located within a crevice, 2) in the open

on an exposed rocky surface, or 3) under a ledge; and by the adjacent environment: 1) opening onto a surge channel, or 2) surrounded by living coral.

Nest orientation with respect to

compass heading is presented in Figure 2.16. The great majority of nests were located on dead Porites lobata coral rock.

In all instances, the upper portions of

these coral heads were alive, but the nest sites were located around the perimeter where the coral had died and was covered with algal turf and coralline algae.

In many instances,

males removed some surrounding living coral in the process of preparing these nest sites (Figure 2.13). In two instance, males removed large areas of living tissue from the exposed surfaces of nests (Figure 2.14).

E. lobata while creating

Male 'L' removed approximately 256 cm2

of living coral to create a nest at (P.9, 13.7).

This nest

site, and a similar one created by male '5' at (F.9, 8.9) 88

Table 2.6 Characteristics of 38 nest sites of ~. brevis at Hanauma Bay. Refer to text for discussion of categories. ============================================================

Substratum: Exposure: Environment:

dead coral = 31, live coral = 2, rock = 5 in crevice = 16, exposed = 16, under ledge surge channel = 35, in corals and rock = 3

89

=6

Figure 2.13.

Dead coral substrate used as a nest site by Exallias brevis. The individual in this photo is male 'B', a female is present in the shadow to the left of the nest.

90

~.

:~

,

~

~

~~

,~~~

91

Figure 2.14.

Live coral substrate used as a nest site by Exallias brevis. To create this site, the male cleaned off all coral tissue. This site was expanded after this photo was taken but was subsequently abandoned. The coral in the cleaned area never regenerated.

92

93

were completely exposed, flat, vertical surfaces and clutches of yellow eggs in these nests were visible for a con5iderable distance. A third type of substrate was utilized by males 'C', 'H', 'L' and '4'.

Their nests were located on the surface of

large rocks or boulders which apparently had rolled into the ocean from nearby cliffs (Figure 2.15). Nests could be further described as belonging to one of three types depending on their particular location on a rock or coral head.

sixteen nests were situated in crevices in

the coral or rock structure.

These were the most difficult

nests for observing clutches but were often the easiest for observing adults.

Males and females usually remained

motionless within these crevices during observations which facilitated identification of individuals (see Figure 3.1). sixteen nests were constructed on completely exposed rock or coral surfaces with no protection or cover from surrounding rock.

Clutches of eggs were easily observed on these sites,

but adults usually sought cover in adjacent holes when frightened. six nests were located under rock or coral ledges. Males and females had to turn upside down to deposit and care for eggs laid on the undersides of these ledges.

These nests

were similar to the exposed nests in the ease of observing clutches.

94

Figure 2.15.

A basalt rock used as a nest site by Exallias brevis. The individual in this photo is male lei spawning with female lEI.

95

96

All but three of the 38 nests were located adjacent to surge channels between large coral heads or boulders.

These

channels contained little or no live coral, and the bottom was either sandy, or hard, flat rock.

No nests were found in

the central areas of large fields of live coral, for example within the coordinates (J.O to P.O, and 13.0 to 16.0), despite the presence of ample food resources.

One exception

was the nest of male 'I' located at (J.9, 8.0) which was constructed on the east face of a large coral head and which was surrounded by other coral heads.

This nest was utilized

for only six days from November 11 - 17, 1980 and received only two clutches.

It was abandoned on November 18 when male

'I' moved to the coral head at (M.O, 10.0).

A similar nest

location was used by male 'z' at (Q.1, 6.5) and male 'M' at (J.1,17.5). The compass orientations of nests were grouped into eight equal sectors around the 260 0 bearing representing the direction of prevailing incoming ocean swells.

The

distribution of nest sites within these sectors is shown in Figure 2.16.

The north-facing sector (350 0

-

34 0 )

,

which was

essentially perpendicular to the ocean swell direction had the greatest number of nest sites (16) while the two westfacing sectors (215 0

-

259 0 and 260 0 -304 0 ) had the least

number (1 each).

97

Figure 2.16

Distribution of Exallias brevis nest sites in relation to the predominant direction of ocean swells (260 0 ) . The orientation of each nest site was measured as the compass bearing perpendicular to the nest.

98

2600~--------*,"----------I800

99

Discussion Eggs and embryonic development Thresher (1984) has reported that the average incubation time for demersal spawning fishes is 152.3 hours or 6.35 days (n

= 10); whereas for pelagic spawning fishes it is only 43.1

hours, or

~.80

days (n

= 20).

Exallias brevis is typical of

other demersal spawning fishes with an incubation time ranging from 6 to 9 days depending on water temperature.

The

color changes of eggs that occur during development provide a useful index for aging clutches in the field (see Chapters III and IV).

However because color change is gradual, age

can only be determined with an accuracy of about ±1 day at best.

McDonald (1981) used this technique for aging the eggs

of the blue-eye damselfish, Plectroglyphidodon johnstonianus. The adaptive significance, if any, of the brilliant yellow coloration of newly laid

~.

brevis eggs is not known.

Many demersal spawning fishes produce brightly colored eggs, such as the black-lined blenny, Meiacanthus nigrolineatus, which produces orange eggs (Fishelson, 1976); anemonefish, Amphiprion spp., which produce bright orange eggs (Thresher, 1984); and the Hawaiian sergeant or mamo, Abudefduf abdominalis, which produces red or occasionally yellow colored eggs (Helfrich, 1958).

Bright colors often indicate

an aposematic function, therefore, colorful eggs may be unpalatable or toxic.

Gladstone (1987) has demonstrated that 100

the translucent yellow-colored, demersal eggs of the sharpnose pufferfish (Canthigaster valentini) are rejected by egg predators.

Presumably they contain tetrodotoxin which is

found in the skin, liver, intestines, and ovaries of the adults. The eggs of most damselfishes are readily consumed by egg predators whenever they can elude the defenses of the guardian male.

Similarly, I have observed milletseed

butterflyfish (Chaetodon miliaris), saddle wrasses, (Thalassoma duperrey), and other egg predators

consume~.

brevis eggs when I removed the guardian male from the nest. Apparently, most fish eggs, including those of

~.

brevis, are

not distasteful and therefore their bright coloration is not aposematic. Laboratory research on sticklebacks (Gasterosteus aculeatus) and river bullheads (Cottus gobio) has provided evidence that females of these species prefer males whose nests contain eggs (Ridley and Rechten, 1981; Marconato and Bisazza, 1986).

Sikkel (1988, 1989) has presented field data

which indicate that female garibaldi, Hypsypops rubicundus (a damselfish), preferentially lay eggs in nests which already contain eggs.

Furthermore, nests containing newly laid

bright yellow eggs are most preferred.

Presumably, by laying

eggs among those of other females, the per capita risk to the eggs from both heterospecific and conspecific egg predators is reduced.

This benefit is maximized if new clutches are 101

all laid at about the same time, thus prolonging the amount of time they remain together during development.

Sikkel

(1989, p. 454) notes that "if depositing eggs among those of others is advantageous to female garibaldi, then the highly conspicuous coloration of new eggs may also be advantageous despite the fact that garibaldi nests are typically exposed. The advantage would arise by increasing the detectability of eggs in the nest to other

femal~s".

This remains an

intriguing but essentially untested hypothesis which could be extended to

~.

brevis.

Larvae At the time of hatching, 2.1 rom TL.

~.

brevis larvae measure about

Illustrations of four larvae ranging from 2.2 rom

to 26.0 mm SL are illustrated in Leis and Rennis (1983). Miller, et al. (1979) report that

~.

brevis larvae were taken

at nearly all of their collecting stations during both summer and winter but they were slightly more abundant in winter. Watson and Leis (1974) reported that

~.

brevis was the third

most common blenny larva taken in the Sampan Channel of Kaneohe Bay but was rare in samples of the inner bay. Lobel (personal communication) has brought to my attention that the blenny illustrated on the cover and on page 194 of Newbert (1984) is a larval

~.

brevis.

Newbert

states that he took the photo in the open ocean at a depth of 40 m off Kona, Hawaii.

If this is

~.

brevis, then it is

probably the only photo of a living larva. 102

Collections of

~.

brevis at the Bishop Museum in

Honolulu include three larval specimens (BPBM 9119 and 9157, Nehoa, Hawaii, June 3 & 5, 1967, collected offshore with nightlights).

Two of these specimens, 25.0 and 25.5 rom SL,

had recurved, fang-like teeth typical of many salariin blenny larvae (referred to as ophioblennius larva).

However, the

third specimen, also 25.5 rom SL, had lost its canine teeth and had comb-like teeth typical of adults.

Leis and Rennis

(1983) state that the ophioblennius larvae of salariin blennies may reach 30 rom SL before undergoing metamorphosis. They also illustrate a 26.0 rom

SL~.

brevis larva collected

from a mid-water trawl off Hawaii which appears to be undergoing metamorphosis. I have never observed

~.

brevis juveniles as small as 26

rom on the reef, but given their size, rarity, and secretive behavior, this does not mean that they are not present at that size. juvenile

~.

33.9 rom SL.

As noted in the results, I have rarely observed brevis on the reef, and the smallest of these was It thus appears that

~.

brevis metamorphoses and

settles on the reef at a size between 26 rom and 34 rom SL. Age and growth Tag-recapture growth data can be used to develop an approximate growth curve for a variety of animals, including fishes.

The data presented in Table 2.1 can be used in this

manner but at least one major assumption is required regarding age. 103

The data from Table 2.1 were fitted to a Von Bertalanffy growth curve following methods described in Ricker (1975) and Fabens (1965).

The von Bertalanffy equation is the most

common method used to describe growth in fishes: SLt

=

L~

(l-e- K(t - to»

where 'SLt' is the standard length at any given time It', 'L~

'is the average maximum standard length, 'K' is a

constant describing the rate of increase in growth; and 'to' is the back-extrapolated time when the size of the individual was zero (see Ricker, 1975). The calculation of 'K' and

, are presented in Table

'L~

2.7, and are estimated directly from the tag-recapture data. The growth data for males and females when analyzed separately resulted in similar estimates for 'K' and = .077 for males & .083 for females; 123.59 for females).

L~,

122.84 for males &

The average maximum standard

was estimated to be 123.072 mm.

The value of 123.072 mm for maximum size for

~.

'L~

, is less than the

brevis, but this is not unexpected since

'L~'

is the average maximum size.

63

brevis tagged at Hanauma Bay exceeded this value but

~.

(K

The data were therefore pooled, and 'K'

was estimated to be 0.081. length,

L~=

'L~'

Five individuals out of

only by 1 or 2 mm (Table 3.1, Chapter 3).

One exceptionally

large individual, female 'T', measured 129 rom SL and was the largest female ever encountered.

The largest specimen in

collections at the Bishop Museum was 110.9 mm SL (catalog no. 104

Table 2.7. -- Estimates for the von Bertalanffy growth constant 'K' and the average maximum size 'L=' for Exallias brevis based on tag-recapture data collected at Hanauma Bay. Data for males and females are pooled in final calculations. (Symbols: 'ri' represents the difference between the size at capture, '1, " and the size at recapture, '12" divided by the length of the interval in days, 'd'i and 'sit represents the average standard length during the same interval.) ============================================================

AVERAGE SIZE DURING INTERVAL si = (11 + 12) 12

GROWTH RATE DURING INTERVAL ri = (12 - 11)/d MALES: B (120-119)/357 C (115-111)/466 L (125-114)/373 M (120-117)/358 N (105- 64)/449 T (118-114)/403 X (120-116)/417 6 (104-100)/ 56 FEMALES: F (120-111)/555 G (115-103)/411 S (123-114)/515 V ( 95- 87)/132 V (113- 95)/328 X ( 91- 72)/141 Y (124-116)/414 4 (113- 81)/410 4 (123-113)/765

= = = = = = = =

.

r' .0028 .0086 .0295 .0084 .0913 .0099 .0096 .0714



S'

(120+119)/2 (115+111)/2 (125+114)/2 (120+117)/2 (105+ 64)/2 (118+114)/2 (120+116)/2 (104+100)/2

= = = = = = = =

119.5 113.0 119.5 118.5 84.5 116.0 118.0 102.0



S'

ri

= .0162 (120+111)/2 = 115.5 = .0292 (115+103)/2 = 109.0 = .0175 (123+114)/2 = 118.5 = .0606 ( 95+ 87)/2 = 91.0 = .0549 (113+ 95)/2 = 104.0 = .1348 ( 91+ 72)/2 = 81.5 = .0193 (124+116)/2 = 120.0 = .0780 (113+ 81)/2 = 97.0 = .0157 (125+113)/2 = 119.0 ============================================================

K

L=

= =

(1:ri) (1:Si~ - n 1: (r'si) n 1: si - (1:Si) 2 (1: Si) (1:ri si) - 1: ri r si 2 n1:ri si - 1: riI: si

105

= = =

.002677 "days" .081425 "morrch s " 123.072 mm

BPBM 6111, Tahiti).

Randall (1983) reports examining one

individual 145 rom but this represents the "maximum total length" and is equivalent to about 118 mIn SL. Obtaining an estimate of 'to' is more difficult since data relating size to absolute age must be supplied (Fabens, 1965).

Age determination in fishes is usually made by

counting growth rings on scales or on otoliths.

Salariin

blennies lack scales thereby eliminating this method.

The

otoliths of ten adult g. brevis collected at Kahe Point were examined but no growth rings could be discerned on the otoliths, and therefore no estimates of age could be made using this method either. I therefore "estimated" that the smallest individual that I tagged, male 'N', was 6 months old at a size of 64 rom SL.

This figure may be olf by a month or two but is not an

unreasonable guess.

Exallias brevis larvae probably

metamorphose at about 26 - 34 rom SL, as do some other reef fishes.

For example, the milletseed butterflyfish, Chaetodon

miliaris, settle out at a size of 26 rom SL and an age of about 1.8 months (Ralston, 1975); the convict surgeonfish or manini, Acanthurus triostegus, also settle out at a size of 26 rom SL and an age of 2.5 months (Randall, 1961).

While

there is considerable variation among species in size and age at settlement, an estimated age of about 2 months is not unreasonable "first guess" for a 26 rom g. brevis larva.

106

Given an average growth rate of 0.433 rom/day (=26 rom/60 days), and an estimated age of 2 months for a 26 rom larva, then a 64 rom juvenile would require an additional 2.9 months to grow that large.

Assuming that the growth rate slows

somewhat, then an age of 6 months at 64 rom SL should be a reasonable, and probably conservative, first estimate. The second step in calculating 'to' is estimating the age of other small individuals.

Given an estimate of six

months for male 'N' when he was 64 rom SL, then he was 449 days older when recaptured at a size of 105 mm SL.

During

this period he grew at an average rate of 0.0913 mm/day. Next, considering only those individuals tagged and recaptured at sizes between 64 mm and 105 rom, I estimated their ages at the time of capture by taking the difference between their size and 64 rom, and calculating the time it would have taken them to grow that large if they grew at a rate of 0.0913 rom/day. For example, female 'X' was 72 rom SL when first captured.

By this method she would have been 8.92 months old

(6 months +«72 rom - 64 mm)/(0.0913 rom/day)/30 days/month) 8.92 months).

The initial ages of seven individuals were

estimated using this method and the results are listed in column #1 of Table 2.8.

Column #2 lists the sizes of the

same individuals at the time of recapture with their ages estimated as their initial age plus the length of the interval between capture and recapture. 107

=

The calculation of 'to' is shown in Table 2.8.

Having

derived estimates of 'L oo', 'K', and 'to', the Von Bertalanffy growth equation for

= 123.072(1

SLt

~.

brevis is:

- e-· 08(t + 2.15»

using this equation, a growth curve was plotted and is shown in Figure 2.17.

According to this curve, an average

~.

brevis would reach a size of 83.39 mm SL at an age of 12 months, 107.88 rom SL at 24 months, and 117.26 rom SL at 36 months, and 120.84 mm SL at 48 months. It is not uncommon for small reef fishes to reach maturity at an age of about one year, e.g., Chaetodon multicinctus (Tricas, 1986),

and~.

Interestingly, the smallest mature

miliaris (Ralston, 1975). ~.

brevis that I observed

was female '4' at a size of 81 mm SL. age of 11.3 months on the growth curve.

This corresponds to an Female '4' was

tagged on March 10, 1981 which means that she would have been spawned during March-April of the previous year, which is the peak spawning period

for~.

brevis (see Chapter III).

I recaptured female '4' 13.5 months later (estimated age

= 24.8 months) when she measured 113 rom SL. growth curve, an average mm SL.

~.

According to the

brevis would have been only 108.8

I remeasured female '4' one more time at an estimated

age of 50.0 months

(=

24.8 + 25.1 months).

She measured 125

rom SL at that time compared to the estimated average size of 121.17 rom SL.

Female '4' thus provides a useful test of the

growth equation presented here. 108

The curve may be somewhat

Table 2.8. -- Estimate of 'to' (months) for von Bertalanffy growth curve for Exallias brevis at Hanauma Bay. Estimated ages of seven ~. brevis in column #1 are based on an estimated age of 6 months for a 64 mm SL individual. This individual was measured 449 days later at a size of 105 mm SL which represents an average growth rate of .0913 mm/day during this period. Estimates of initial ages for the six other individuals in column #1 are calculated from this rate and assume that all individuals were 64 mm SL at age 6 months. The second column represents the size of each corresponding individual from column #1 at the time of recapture. The ages in column #2 are the estimated initial ages plus the interval time. (L oo = 123.072 mm, Table 2.7). ============================================================

Males N 6

COLUMN #1 COLUMN #2 SIZE AND ESTIMATED SIZE AND AGE (t2) AGE (t1l AT CAPTURE ~T RECAPTURE S~1 --11- In(Lm -SLt11 SL2_ --t2_ In(Lm -SLt21 64 6.00 4.079 105 20.97 2.894 100 19.14 3.139 104 21.01 2.948

Females SL1 72 X 4 81 87 V V 95 G 103

--118.92 12.21 14.40 17.32 20.24

In(LarS~11

3.933 3.739 3.586 3.335 2.999

SL2 91 113 95 113 115

In (Lm-SLt2L-3.468 2.310 3.335 2.310 2.088

--t2_ 13.62 25.88 18.80 28.25 33.94

============================================================

Regression of 'In(Lm -SLtl' on 't': intercept: 4.6408 slope: -0.0798 to

=

(4.6408 - In(LaJ)/.07981

109

=

K

= -2.1547

months

Figure 2.17.

A von Bertalanffy growth curve for Exallias brevis, derived from capture-recapture growth data. Each data point represents the individuals listed in Table 2.1. Only the recapture size for each individual is plotted as each initial size lies exactly on the growth curve. The four marks identified along the curve represent the average estimated standard length at 1 year, 2 years, 3 years, and 4 years.

110

130

L

YS

-----------------------------.------------------------- --- --------------- ----------- ----- ------ --~------:---~ --- -- -

120

G-

110

117.26

C

-FMX T 120.84

r-"\

E 100 E .c 90

I....J

-g>

I-' I-' I-'

80 83.39

Q)

"0

70

~

~c 60

2

en

50 I

*SL

40

t

=123.07( 1 - e-· 0 8 ( t + 2.15) )

30 20 10

~

0

5

10

25 Months

30

3-5

so

-4 -B

conservative in describing growth of Hanauma Bay

~.

brevis,

assuming female '4' is representative of all individuals in the population, and if her estimated age at first capture is accurate. Following the preparation of this section, a 45.5 rom SL juvenile

E. brevis was collected by Marjorie Awai on

September 4, 1989, and maintained for 35 days in an outdoor 400 liter aquarium with a variety of living corals for food. When recaptured, it measured 51.5 rom SL which represents a growth rate of 0.286 rom/day.

Dr. Ronald Thresher examined

the otoliths of this individual and estimated a planktonic duration of about 28-36 days, a post-settlement age of 79 (± 5) days, and a total age of about 120 days. Bertalanffy equation for

Using the Von

E. brevis, this juvenile would have

been 91 days old when captured and 126 days old when sacrificed.

The close agreement of the two independent age

estimates improves confidence in this growth equation. survivorship and lifespan As reported by McDonald (1981) there is an absence of studies on survivorship and mortality among coral reef fishes.

McDonald listed the median survivorship of five

species of pomacentrids reported in the literature which ranged from 1.0 year to 3.4 years.

Two additional species

studied by McDonald had median survivorships of 2.3 years for the blue-eye damselfish, Plectroglyphidodon jOhnstonianus, and 5.2 years for the Pacific gregory, Stegastes fasciolatus. 112

The median survivorship of 1.7 years for

~.

brevis is

relatively short, but survival apparently is not constant among age classes.

Using the data in Table 3.2 combined with

Appendices A and B, the percentages of individuals in each size class which were observed through the end of the study period were: (I) under 84 rom SL: 14%; (2) 85 - 108 rom SL: 40%; (3) 109 - 117 rom SL: 56%; (4) 118 - 121 rom SL: 60%; (5) over 122 mm SL: 0%.

As noted elsewhere, the smallest

individuals were very difficult to track over extended periods, and therefore the survival estimate of 14% is probably an underestimate.

However, the subsequent three

size classes were easier to follow and a trend for improved survival with size is apparent.

None of the largest, and

presumably oldest individuals, survived through the end of the study. Mortality might be expected to be high among

~.

brevis

due to the bright color of males, the males' necessity to remain close to conspicuous clutches of eggs, and the necessity for females to often travel considerable distances to mate with males thus exposing themselves to predators along the way.

On two occasions I observed males being

stalked by coronetfish, Fistularia commersoni, and frequently observed nesting males with severely torn fins or gashes. These wounds are unlikely to have been inflicted during aggressive encounters with conspecifics as I have never observed any wounds being inflicted after watching many such 113

encounters.

Eels are a likely predator, particularly at

night. The maximum life span of 4.2 years for g. brevis is short compared to estimates of 6 years for P. johnstonianus and 11 years

for~.

fasciolatus reported by McDonald (1981),

and compared with other coral reef fishes.

Estimates of life

span among a variety of coral reef fishes are reported in the literature including: the keyhole angelfish, Centropyge tibicen, 6 years; the cowfish, Lactoria diaphena, 8 years; the zebra lionfish, Dendrochirus zebra, 5 years; the dotted sweetlips, Plectorhynchus pictus, 7 years; Clark's anemonefish, Amphiprion clarkii, 13 years (all of the above Moyer, 1986); the black anemonefish,

A. rnelanopus, 4 years

(Ross, 1978); the yellow damselfish, Eupomacentrus planifrons, 7 years (Pauly & Ingles, 1981); and various butterflyfishes, Chaetodon spp., 3 - 8 years, (summarized in Reese, 1991).

Each of these records must be considered

minimum longevity estimates, because the fish were adults when first observed and in some instances the fish survived beyond the final observation.

By contrast, some temperate

water fishes have considerably longer life spans: the garibaldi damselfish, Hypsypops rUbicunda, may live as long as 17 years (Clarke, 1970); and the yellowtail rockfish, Sebastes flavidus, is reported to have a reproductive life span of 31 - 39 years and a maximum life span of 56 years (Eldridge, et al., 1991). 114

In summary,

~.

brevis has a relatively short life span,

and suffers high annual mortality.

High reproductive effort

is predicted under such conditions, and in Chapter III this will be demonstrated to be true for

~.

brevis.

Activity budget and feeding behavior Exallias brevis can be described as sedentary, or more precisely, as a time-minimizer (sensu Schoener, 1971). Whether they are out in the open on coral heads, or in shelter holes, or on eggs in a nest,

~.

brevis is essentially

inactive except for respiring and some eye and fin movements. Both males and females are similar in this respect and spend over 92% of their time in two sedentary activities:

rest and

shelter. A comparison of activity bUdgets for two other coral reef fishes is presented in Table 2.9.

The first is the red-

lip blenny, Ophioblennius atlanticus, which is a herbivore and a close relative of

~.

brevis.

The second is the

multiband butterflyfish, Chaetodon multicinctus, which is a corallivore and feeds on the same species of corals as brevis.

~.

Nursall (1981) reported that Q. atlanticus spends

60% of its time "resting", and only 8.5% of its time "feeding".

Another tropical blenny, Entomacrodus nigricans

spends up to 84% of its time motionless (Graham et al., 1985) (unfortunately, neither Nursall nor Graham provide data on the standing crop of algae at their respective study sites).

115

Table 2.9. -- Comparison of activity budgets for the red-lip blenny, Ophioblennius atlanticus (OA); the multiband butterflyfish, Chaetodon multicinctus (CM); and the coral blenny, Exallias brevis. Data are based on focal animal observations. =======================================================================

~ ~

0'1

Rest (=resting)4 Shelter (=within crevice) Swim Coral bites (=feeding) Interactions Out of sight 6 1 2 3 4

OAI

CM2

60.0% 5.0% 15.0% 8.5% 2.5% 9.0% 100.0

0.0% 5.0% ? 94.5% 1.8%

---.:l.

101.3%

Exallias brevis 3 males females 50.2% 91.1% 44.4%5 4.4% 2.9% 3.3% 0.5% 0.7% 2.0% 0.5%

100.0%

100.0%

Nursall (1981): n = 4 males, 8 females, 3 unknown; 65 hours. Tricas (1986): n = 5 males, 5 females, 40.25 hours. males, n = 11, 28 hours; females, n = 8, 17 hours. My behavior categories are equated with those used by Nursall (1981) which are listed in parentheses. 5 Includes nest-cleaning time. 6 Nursall (1981) recorded time when focal animal was lost from sight. I did not record this for E. brevis.

Tricas (1986) reported that

~.

multicinctus spends

between 92.8% and 96.2% of its time foraging.

The

explanation advanced by Tricas (1986) is that living coral tissue represents an extremely low energy source and as a result, fish must invest large amounts of time feeding to obtain adequate energy.

But this explanation is inadequate

to explain the sedentary behavior of the same species of corals.

~.

brevis which feeds on

If coral tissue is indeed a low

energy food source, one would expect

~.

brevis to spend more

time feeding than a herbivorous counterpart.

In fact,

~.

brevis spends less time feeding than its herbivorous relative, Q. atlanticus.

This rate is also far less than

another herbivorous blenny, Blennius sanquinolentus, which spends an average 28% of its time feeding (from data in Taborsky and Limberger, 1979, Figure 3). Two major differences between

~

brevis and

~.

multicinctus may account for this apparent discrepancy. First,

~.

multicinctus cannot sink to the bottom and "rest"

as can

~.

brevis.

While~.

multicinctus can stop moving, it

usually hovers in one spot using its pectoral fins and therefore remains somewhat active all the time.

this

activity requires energy and may result in a higher feeding rate for the butterflyfish.

Since~.

brevis lacks a swim

bladder, it can "rest" on the coral without having to constantly swim or hover like

~.

multicinctus.

Resting

undoubtedly requires less energy than swimming, although in 117

strong surge,

~.

brevis must brace itself tightly in the

rocks using its pectoral fins. A more significant difference is the amount of coral tissue consumed per bite.

According to Motta (1988), g.

multicinctus has a small, forceps-like mouth and only one coral polyp is removed from inside each corallite per bite. Hourigan (1987) reports that

~.

multicinctus removes only

portions of one polyp with each bite when feeding on Porites spp.

Mark Roman (personal communication to Ernst Reese)

found that this species removes about 0.7 polyps per bite when feeding on Pocillopora damicornis.

Each adult g.

brevis coral-bite on Porites lobata averages 2.04 ± 0.42 cm2. Given a

E. lobata polyp density of about 71.40 polyps/cm 2,

each bite thus encompasses, on average, 146 polyps.

However,

examination of photographs made of feeding scars indicates that only about 78% of the polyps appear to be removed with each bite, or about 114 polyps per bite.

~.

brevis thus

consumes about 114 times as many polyps per bite as does

~.

multicinctus. Hourigan (1987, Table 5.9, p. 324) reported that a male g. multicinctus feeds at rates ranging from 487-649 coralbites/hour, depending on sex and location.

According to

Tricas (1986, Table XV, p. 175), a male g. multicinctus averages about 48 coral-bites/5 minutes or 572 bites/hour. By comparison, a male

~.

bites/hour (Table 2.4).

brevis averages about 14 coralSince 118

male~.

brevis feed almost

exclusively on Porites lobata this means they consume about 1,596 polyps/hour.

If~.

multicinctus consumes only one

polyp per bite as stated by Motta (1988), then on average they eat only about one third as many polyps per hour as

~.

brevis. ~.

multicinctus may indeed spend a great deal of time

foraging, but the explanation may not be entirely because of the low energy value of coral tissue.

As evidenced by

brevis, an alternative explanation is that

~.

~.

multicinctus

simply does not obtain as many polyps with each bite that it takes.

Hourigan (1987) came to a similar conclusion when

comparing the lower feeding rate of

~.

guadrimaculatus, also

a corallivore, which has a larger mouth than that of

~.

multicinctus. Male and female

~.

brevis are similar in the amount of

time they spend swimming.

For males this averages 1.73

minutes/hour (2.89% of time budget), while for females it averages 2.00 minutes/hour (3.33% of time budget).

However,

females do more feeding during each swimming foray than males and make twice as many coral-bites per hour (13.94 coralbites/hour for males vs. 28.40 coral-bites/hour for females). This is consistent with the hypothesis that females must feed at a higher rate than males to enable them to produce large quantities of energy-rich eggs. III,

~.

As demonstrated in Chapter

brevis females are indeed prodigious egg producers

year-round.

Hourigan (1987) cites four other studies which 119

have described a similar pattern in other fishes and both he and Tricas (1986) have discovered the same relationship in g. multicinctus. Table 2.4 reveals one other difference in feeding behavior between male and female

~.

brevis.

Both males and

females feed predominantly on Porites lobata, but the proportion of coral-bites for males on this coral was much higher than for females.

Females had a higher proportion of

coral-bites directed towards Pocillopora meandrina than did males, and nearly 15% of their coral-bites were made on Pavona varians which males never fed upon.

Further research

is required to determine if these differences are the result of feeding preferences, or differences in coral species distribution within male and female territories.

Male

territories are frequently entirely within the boundaries of massive colonies of Porites lobata and this may limit the variety of corals available as food.

Female territories are

larger than those of males, and they rarely include these massive coral heads, and as a consequence they may encounter a greater variety of corals.

Alternatively, females may

require essential nutrients for egg production which are not available from only one or two species of coral, and therefore must locate and feed on a greater variety of coral species.

Further work needs to be done to distinguish

between these two hypotheses.

120

The functional significance of the feeding pattern of two bites in succession on Porites lobata vs. three on Pocillopora meandrina, is not clear.

Motta (1985) has

described a similar pattern in the ornate butterflyfish, Chaetodon ornatissimus, which in most cases makes one to three bites on a single location on a coral head. feeding scars of

~.

The

ornatissimus are similar to those of

E.

brevis but are smaller (about 1 cm2) and more square in shape, and Motta estimates that they remove about 16 - 50 coral polyps/bite.

Motta does not mention whether or not

there are consistent differences in the pattern of successive bites when this species feeds on different coral species.

In

E. brevis the pattern of "couplets" on Porites lobata and "triplets" on Pocillopora meandrina is virtually invariable. Hobson (1974) was the first to correctly describe Hawaiian

E. brevis as a corallivore, but he could not explain

the different results obtained by Hiatt and Strasburg (1960) who described this species at Enewetak Atoll as a herbivore. Virtually all tropical blennies, including the closely related Cirripectes spp. are herbivores, but my observations at Enewetak Atoll, the Great Barrier Reef, and elsewhere in the Pacific indicate that

E. brevis is indeed a corallivore

at all locations and feeds on a greater variety of corals than in Hawaii.

It is also one of the few species of fish

which feeds on fire coral, Millepora sp.

I have observed

Hawaiian E; brevis apparently feeding on algae on a few 121

occasions and have found bits of algae among its stomach contents.

Perhaps Hiatt and Strasburg (1960) collected an

individual which had fed more extensively on algae or, less likely, they misidentified a specimen of Cirripectes fuscoguttatus from Enewetak Atoll which is similar to

~.

brevis in coloration but is a herbivore. At this time,

~.

the family Blenniidae.

brevis is the only known corallivore in However, detailed observations of

most tropical blennies have not yet been made.

I have

observed Ecsenius yaeyaemensis on the Great Barrier Reef feeding on scleractinian corals, and Losey (1972) has reported observing

~.

bicolor feeding on Acropora corals.

Robert Robertson (1970) reported that Cirripectes sp. and Meiacanthus sp. feed on corals in the Red Sea.

The identity

of the Cirripectes sp. was not given by Robertson nor any -

specific information on the types or amount of corals fed upon.

Robertson attributed these observations to Fishelson.

However, Fishelson (personal communication, 1989) has no knowledge of making such observations and considers Red Sea Cirripectes to be "algivorous". Territoriality Both male and female

~.

brevis occupy exclusive areas

year-round from which they chase away conspecifics of the same or, occasionally, of the opposite sex.

Thus~.

brevis

does maintain territories but both sexes also roam widely across the reef on occasion.

The frequency of these 122

wanderings could not be determined during this study, and it is not clear whether or not these larger areas should be considered "home ranges". Male

~.

brevis appear to defend the entire border of

their territories against intruding conspecific males. Exceptions occur when the resident male is tending eggs and is not in a position to observe an intruder.

Females,

likewise, defend territories against intruding conspecific females.

The size of female territories is significantly

larger than those of males with nests (P < .05, GT2 test) but is not significantly larger than the territories of males without nests.

Finally, although males without nests had

territories which averaged more than twice as large as those of males with nests, the difference was not statistically significant (P > .05). It has been hypothesized that defense of limited resources is the primary purpose for maintaining a territory but identifying the critical resource or resources is usually difficult.

Among fishes, these resources have been

hypothesized to include food, nest sites, refugia, and mates or access to mates. Living coral is a potentially limiting food resource for ~.

brevis.

It is also a defensible resource as demonstrated

by Tricas (1986, 1989) and Hourigan (1987).

Do~.

brevis

defend territories to maintain exclusive feeding areas? Several lines of evidence suggest that this is not an 123

adequate explanation.

First, coral does not appear to be a

limited resource at Hanauma Bay or at Kahe Point.

Both study

sites include large areas of living Porites lobata and Pocillopora meandrina which are not utilized as food by brevis.

~.

For example, the large coral heads at (L.0,15.0) and

(I.0,12.0) never had any resident

~.

brevis during this

study. Secondly, the amount of coral within each territory appears to be far in excess of what is required by each individual.

The smallest territories are those defended by

males with nests and are about 6.5 m2•

These territories are

usually large Porites lobata coral heads and can be considered to have at least 6.5 m2 of living coral tissue, and probably much more if the three dimensional aspects of the coral are considered.

Males, on average, feed at a rate

of about 14 coral-bites per hour, and each bite averages 2.04

± 0.42 cm2.

This amounts to 28.56 cm2 of coral tissue

removed each hour, or 314.16 cm2 per day, given an 11-hour feeding period (Table 2.3).

Hourigan (1986) reported that

coral scars about 1 cm2 regenerate in about 2 weeks, however, the

~.

brevis feeding scars tracked by me were still visible

after 20 days and remained barely visible after 50 days (Figure 2.10).

Using a 50-day figure, each

male~.

brevis

can remove 1.57 m2 of coral tissue during this period before the oldest scars have completely healed.

124

It is thus apparent

that even the smallest territories have far more living coral than required as food by The fact that

~.

~.

brevis.

brevis never chased food competitors

such as Chaetodon mUlticinctus is further evidence that brevis does not defend food resources.

Female~.

~.

brevis on

rare occasions chased the blue-eye damselfish, Plectroglyphidodon johnstonianus which is a food competitor, but these chases were short in duration and never involved any contact or overt aggressive acts.

Tolerance of the "dear

enemy" (Thresher, 1978) is another possible explanation. DeMartini (1976) undertook an extensive investigation of territoriality in the painted greenling, Oxylebius pictus and also concluded that defense of food resources was not the primary explanation for territorial defense in this species. He concluded that male territoriality in Q. pictus functioned as defense of refugia as well as monopolization of certain spawning sites which are preferred by females in areas of greater water movement on reefs.

Female aggression was

observed to be invariably associated with proximity to shelter holes and was interpreted as territorial defense of the shelter hole. will be true for

It is possible that a similar explanation ~.

brevis, but a more precise evaluation

will require further field work and experimentation of the kind carried out by DeMartini.

125

Nest sites Among demersal spawning fishes, the location of nest sites are sometimes very specific and predictable.

The

proposed explanations for such specificity vary with the species under study.

Anemonefishes, Amphiprion spp., prepare

nest sites near the base of their anemone hosts (Allen, 1980), and the anemone presumably plays some passive, protective role in the incubation of the eggs.

McDonald

(1981) reported that nesting males of the blue-eye damselfish, Plectroglyphidodon johnstonianus, are located almost entirely along the reef slope and he believed this facilitated dispersal of larvae into open water.

DeMartini

(1982) demonstrated that painted greenling, Oxylebius pictus, select nest sites which maximize exposure to the surge and thus to oxygen availability.

DeMartini determined that

oxygenation of the eggs was undoubtedly a factor involved in the deposition of painted greenling eggs. All of the above explanations are important for all eggs and larvae:

they must be protected from predators and other

threats, they must receive adequate oxygen for respiration, and the larvae must be able to find their way into open water after hatching.

The degree to which each of these factors is

important in nest site selection will undoubtedly vary among species and localities. I found that easy to locate.

~.

brevis nest sites were predictable and

It is immediately apparent from Table 2.6 126

that nests occur most often on dead coral substrate but more importantly these sites were located around the perimeter of massive colonies of living Porites lobata (see Figure 2.13). Nesting males also defended smaller territories than nonnesting males but continued to feed at the same rate (Chapter IV).

As a consequence, feeding scars were more concentrated

in their territories, and thus, a large number of closely spaced feeding scars often indicated the presence of a nearby nest.

It is also apparent from Table 2.6 that nests almost

invariably opened onto small channels in the reef.

Only

three nests occurred in areas surrounded by dense coral cover. Finally, I observed that the compass orientation of the nests was not random (Figure 2.16).

The majority of nests

was situated either perpendicular to, or facing the predominant ocean swell direction; only 2 of 47 nests were observed on the "leeward" side of coral heads.

This seems to

support DeMartini's observation that nests are oriented to maximize exposure to ocean swells.

However, the reason may

not be because of an increase in the availability of oxygen for the embryos. Exallias brevis do not establish their nests in lowrelief areas of the reef but prefer large coral heads or boulders.

This is not because of inadequate food resources

in the low-relief areas because these areas are utilized by females.

The massive coral heads and rocks do offer high 127

vantage points for early detection of ripe females but the actual nest sites on these coral heads rarely seem to face in the direction of nearby female territories.

The nest sites

are often under ledges or facing away from the direction of most approaching females.

Finally, virtually all

~'

brevis

nests are exposed and it seems unlikely that any nest would suffer oxygen deprivation, although leeward nest sites might result in increased siltation. A possible explanation for the observed orientation and location of

~'

brevis nests is similar to McDonald's

explanation for the location of blue-eye damsel fish (£. johnstonianus) nests.

Locations bordering open surge

channels may improve the chances of larvae being swept away from the reef and any potential predators.

Thus the location

of nest sites in areas which appear to maximize their exposure to surge may indeed increase their exposure to oxygen but may also aid in the dispersal of larvae. Hawaiian vs. non-HawaiiaTI Exallias brevis There appears to be an ecological difference in nest site location between Hawaiian

~.

more tropical Pacific localities.

brevis and those of the I have observed

~.

brevis

nest sites at Palau, Kwajalein Atoll, Fiji, and Papua New Guinea.

At each of those locations nests were established on

the branches of large Pocillopora sp. (cf. eydouxi) corals. Eggs were deposited directly on sections of the coral

128

branches which were white and apparently recently killed and cleaned by the male

E. brevis.

Pocillopora sp. (cf. eydouxi) colonies usually occur below the reef crest on the seaward side of coral reefs throughout the south and western tropical Pacific. almost always encountered both male and female

I have

E. brevis

living among the branches of this coral although sometimes they frequent Acropora spp. observed scuth Pacific

~.

Significantly, I have never

brevis living or nesting in massive

Porites colonies such as Porites lutea, except in Hawaii. All of the smallest juvenile

E. brevis I have observed

in Hawaii were residing in Pocillopora rneandrina just as their south Pacific relatives reside in Pocillopora eydouxi. But adult Hawaii

~.

brevis is too large to continue residing

in this coral, and Pocillopora eydouxi is not common on Hawaiian reefs.

It seems that

~.

brevis in Hawaii has had no

"choice" but to adapt to living, feeding, and spawning on Porites corals which are the dominant corals in Hawaii but appe~r

brevis.

to be only incidental food items for south Pacific (Note:

observed one male

~.

after the preparation of this chapter, I

E. brevis with a nest in the branches of a

small Pocillopora sp. (cf. eydouxi) coral at a depth of about 15 m off Waikiki on May 10, 1990. observed many

Between 1977 and 1983, I

E. brevis nesting on Porites lobata and £.

compressa, which were the dominant corals on this reef, but during Hurricane Iwa in November, 1982, virtually all corals 129

in this area were destroyed.

Pocillopora damicornis is now

the dominant coral on this reef with a few rare interspersed.

E. eydouxi

This is my first observation of an g. brevis

male nesting in pocillopora in Hawaii.)

130

CHAPTER III PATTERNS OF REPRODUCTIVE SUCCESS IN A POPULATION OF THE CORAL BLENNY, Exallias brevis Introduction Charles Darwin (1859, 1871) described sexual selection as a struggle among individuals of one sex for mating opportunities with the other sex. to sexual selection: choice of mates.

He recognized two aspects

male competition for females and female

Male competition has been referred to as

intrasexual selection; epigamic selection includes selection which one sex exerts on the other (Huxley, 1938) A consequence of sexual selection should be variation in reproductive success among individuals.

Bateman (1948) made

important generalizations about reproductive strategies in sexually reproducing organisms as a result of laboratory studies on Drosophila.

He demonstrated that variation in

reproductive success was greater among males than females and that males could increase their reproductive success by increasing the number of copulations.

The same was not true

for females which could produce only limited numbers of eggs. Trivers (1972, p. 139) restated "Bateman's Principle" in terms of parental investment which he defined as "any investment by the parent in an individual offspring that increases the offspring's chance of surviving, at the cost of the parent's ability to invest in other offspring ..•.

131

The

sex whose typical parental investment is greater than that of the opposite sex will become a limiting resource for that sex.

Individuals of the sex investing less will compete

among themselves to breed with members of the sex investing more, since an individual of the former can increase its reproductive success by investing successively in the offspring of several members of the limiting sex." Field studies documenting intersexual and intrasexual variation in reproductive success are rare.

To record such

data, studies are required in which individuals are followed through time.

This method is referred to as a horizontal or

longitudinal study (Lande and Arnold, 1983; Arnold and Wade, 1984; Clutton-Brock, 1988), although Endler (1986) finds this terminology ambiguous and prefers "cohort analysis".

I will

use the term "longitudinal study" in this paper because it has the widest useage in the current literature. Longitudinal studies are difficult to implement because of the logistical problems of identifying and tracking individuals.

Clutton-Brock (1988) assembled 25 longitudinal

studies of insects, amphibians, birds, and mammals.

Reptiles

have also been studied (Tinkle, 1967; Trivers, 1976). Patterns of reproduction for fishes have usually been described on the basis of repeated sampling of a population or populations at different points in time.

This method is

referred to as a "vertical" or "cross-sectional" study (Lande and Arnold, 1983; Arnold and Wade, 1984; Clutton-Brock, 1988) 132

although Endler (1986) prefers the terminology "age-class" or "life-history stage" study.

Usually this method cannot be

used to examine differences in individual reproductive success. Longitudinal studies on fishes are difficult because of the lack of effective tagging methods, especially for small fishes; the abundance of individuals in many populations; the difficulty in observing spawning events and thus measuring reproductive success; and the limited amount of time an observer can work underwater.

Furthermore, females of many

fish species are indistinguishable in the field from nonnesting males or immature individuals.

As a consequence,

data are usually only reported for spawning males with scant attention given to reproductive patterns among individual females - a situation which Wasser (1983) emphasizes is the case for most studies of vertebrates.

Long-term,

comprehensive studies of coral reef fish describing annual or seasonal fecundity, spawning intervals, mating patterns, and reproductive success among individuals of both sexes are rare (e. g. Oichi, 1991). The principal areas of investigation in this study of reproductive success in 1.

~.

brevis are listed below:

Do the same individuals spawn continuously or do spawning periods vary among individuals?

Cross-

sectional studies of coral reef fishes have provided data on the mating systems of pomacentrids, labrids, 133

chaetodontids, and other families (reviewed by Thresher, 1984).

For some tropical species spawning occurs, or is

suspected to occur, year-round (Tricas, 1986; Lobel, 1988). 2.

Do females spawn with only one male or with multiple males?

If they spawn with more than one male is the

pattern random?

Among some polygynous species showing

parental care, males care for multiple clutches obtained from several females (Sale, 1971). 3.

What is the interval between successive spawning events for individual females?

Ovaries of reef fishes often

have several size classes of ova (Swerdloff, 1970; Bouain and Siau, 1983; Tricas, 1986), but there is only speculation about how often and at what frequency individual females are spawning. 4.

Can seasonal, annual, or lifetime fecundity be estimated for a fish with year-round spawning?

Estimates of

fecundity for coral reef fishes are usually limited to batch fecundity for females or estimates of egg number from clutch areas for nesting males (e.g., Helfrich, 1958; Ralston, 1975; Moyer, 1975; Bell, 1976; Tricas, 1986; Barlow, 1987; Robertson, et al., 1988).

Estimates

of seasonal and lifetime fecundity are rare for coral reef fishes (Honda and Imai, 1973; Bouain and Siau, 1983; Ochi, 1985a). 5.

What is the extent of variation in individual 134

reproductive success within and between the sexes and how can this be described?

Methods Tagging procedure At Hanauma Bay, an attempt was made to capture and tag all E. brevis within the 2,375 m2 study site.

A few

individuals encountered outside these boundaries were also tagged.

Each captured individual was anesthetized in a 500-

ml clear plastic jar of seawater mixed with 5 cc of a stock solution of quinaldine (1:10 quinaldine: isopropanol) .

The

anesthetized fish were measured to determine standard length, and the dorsal or pectoral fin spines (or rays) were clipped for identification (Figure 3.1).

A colored glass bead was

also attached to the fish as a second tag.

The bead was

strung on silk suture thread which was loosely tied around the base of the first dorsal spine (Figure 3.2). The tagging and measuring procedure was performed underwater and took no more than five minutes, by which time the fish began to recover from the anesthetic.

The fish

were allowed to completely recover inside a transparent, flow-through plastic container and then each container with fish inside was placed next to the spot where the fish was originally captured.

The container lid was gently removed,

and the fish allowed to exit.

Observations continued for

another five minutes to ensure that the fish was behaving 135

Figure 3.1.

Fin clip on female 'F' 16 months after tagging. The fifth dorsal ray was clipped and is visibly shorter than the other rays. Photographed in February 1982.

136

137

Figure 3.2.

Male 'H' tagged with a white glass bead. Photograph by Dr. John E. Randall, November 1980.

138

139

normally; no complications were ever observed.

At the end of

about one year, as many tagged individuals as possible were recollected, measured, and released. Daily observations Whenever possible, daily observations were made using scuba, from October 22, 1980, through May 9, 1982.

All

observations were made between 0900 - 1200 hours, although each dive lasted no more than one hour.

During each dive, a

path was followed around the study area in an attempt to relocate each tagged individual and record its location using the map coordinates (Figure 1.3).

Individual observations

usually lasted about five minutes.

Once each week, a

thorough search was made throughout the study area for additional, untagged fish or tagged fish which had moved. This search also extended beyond the study area boundaries to the north, east, and to the west. I was unable to make observations during most of the month of May, 1981.

A trained assistant was able to identify

individual males, and record spawning events and clutches during this period.

He was unable to recognize most females

and spawning observations of females during this month are therefore limited. Spawning patterns and male fecundity All known males were observed during each dive.

If

spawning was in progress, an attempt was made to identify the female.

Since females often hid in nearby holes when 140

disturbed, a small flashlight and dentist's mirror were used to aid in observing fin clips and bead color.

If the female

could not be identified, a question mark was recorded on the daily spawning observations (Appendix B). For males with nests, a drawing was made during each dive showing the number of clutches present and the color of the eggs.

The area of each clutch was not measured

because observations in 1978-1979 suggested that nest disturbance might result in nest abandonment. Although clutch areas were not measured during the 19801982 study period, these data were collected each month from May, 1978, through April, 1979.

Clutch areas were obtained

by measuring the approximate length and breadth of each clutch with a 30-cm ruler, and the area calculated as an ellipse (Robertson, et al., 1988). Samples from five clutches were obtained at Kahe Point and brought to the laboratory to obtain an estimate of clutch density (eggs/mrn 2 ) , and clutch size (eggs/clutch). The area of each sample was determined by tracing its outline onto a piece of paper, and the area of the tracing was determined using a polar planimeter.

All of the eggs on each

sample were counted and, to avoid counting an egg twice, each was punctured as it was counted

(=

Method A).

Five

additional samples were counted by employing a 1 cm2 hole cut in a piece of thin cardboard.

141

This was set near the center

of a complete clutch of eggs and each egg within the opening was counted and then punctured

(=

Method B) .

Female batch fecundity Batch fecundity refers to the number of eggs ovulated on a given day and can be determined by counting eggs in the ovaries (see Discussion).

Since~.

brevis in Hanauma Bay

could not be removed from the Bay, female (and male) brevis were collected at Kahe Point, Oahu.

~.

Specimens

collected at this location were preserved in 10% formalin for at least one week prior to examination.

The standard

length of each fish was measured using dial calipers, and its total weight measured to the nearest 0.01 gm using an analytical balance. The left and right ovaries were excised from each fish, blotted dry, and weighed separately on an analytical balance to the nearest 0.001 gm.

A small subsample of each right

ovary was removed by sectioning it near its mid-point using a scalpel.

The weight of the sample was indirectly determined

by reweighing the sectioned ovary and recording the difference in weights. Eggs from the ovary sections were counted by placing each sample in a shallow dish of 10% formalin and gently teasing apart the eggs from associated tissue. made under 35x magnification.

Counts were

An ocular micrometer was used

to measure egg diameters, although for most samples only the largest size-class of ova was tallied. 142

Results An average of 4.5 observation dives per week was made at Hanauma Bay between October 22, 1980 and October 31, 1981 (Table 3.1).

The number of dives each month varied with the

fewest made during the month of May, 1981.

An average of 1.9

dives per week was made during the period from January 1, 1982 through April 24, 1982.

An additional 73 dives were

made between May 22, 1978 through April 27, 1979 to collect clutch area data. At Hanauma Bay, 63

~.

brevis were captured, measured,

tagged, and released, and of these, 55 were mature at the time of capture, or were observed spawning shortly thereafter (Table 3.2).

Part A: Patterns of Male Reproduction Male spawning periods Spawning activity for all males in the popUlation from November, 1980, through April, 1982, is shown in Figure 3.3. A maximum peak in spawning activity occurred in April, 1981 (92%), and a smaller peak in October, 1981 (56.1%) separated

by reduced spawning in July, 1981 (34.8%).

Between April 8-

12, 1981, all of the males in the population spawned.

Minimal spawning occurred in November, 1980 (23.5%), and again in November, 1981 (20.9%).

The only period when no

males spawned occurred between November 19 - 23, 1981. Between December, 1981 and January, 1982, the percentage of 143

Table 3.1. -- Observation dives made each month at Hanauma Bay beginning October 22, 1980, through April 24, 1982. All dives were made between 0900 and 1200 hours, and the duration of each dive was about one hour. =============================================================

Year 1980 1981

1982

Month No. dives October 6 November 22 December 16 January 24 February 15 March 19 23 April May 13 June 22 19 July 21 August 21 September October .-li Subtotal 240 November 7 December 6 January 7 February 8 March 13 __ 9 April Total 290

144

Table 3.2 Standard lengths (SL) in millimeters and capture dates for Exallias brevis tagged at Hanauma Bay =============================================================

A

B C

0

E F G

H I J K

L M N

0 P Q

R S T U V

Tagged 10/27/80 10/22/80 11/13/80 11/07/80 10/22/80 10/29/80 10/31/80 10/27/80 10/29/80 10/29/80 10/30/80 10/31/80 11/05/80 11/07/80 11/12/80 11/12/80 12/02/80 11/14/80 11/17/80 11/18/80 12/03/80 12/09/80

MALES Recapture SL 110 119 10/14/81 III 2/22/82 114 118 113 115 107 110 105 112 114 12/09/81 117 10/29/81 64 1/30/82 III 104 116 78 108 114 12/16/81 114 68

SL 120 115

125 120 105

118

W 12/17/80 108 X Y

Z

2 3 4

1/22/81 3/03/81 2/03/81 10/20/81 4/14/81 10/15/81

116 107 117 116 97 108

5 10/19/81 108 6 10/14/81 100 7 10/15/81 95 8 1/25/82 93

3/15/82

12/09/81

120

104

145

FEMALES SL Recapture A 121 122 B C 65 0 116 E 117 F III 5/09/82 G 103 12/16/81 H 91 I 114 J 65 K 122 L 114 M 118 N 104 0 104 P 104 124 Q R 116 S 114 4/06/82 T 129 117 U V 87 3/03/81 V 1/25/82 W 12/03/80 121 72 X 12/03/80 4/23/81 Y 12/17/80 116 1/25/82 Z 1/22/81 111 2 not captured 3 3/05/81 88 4 3/10/81 81 4/24/82 4 5/29/84 5 4/14/81 90 Tagged 10/22/80 10/27/80 10/27/80 10/28/80 10/28/80 10/30/80 10/31/80 11/03/80 11/03/80 11/03/80 11/05/80 11/05/80 11/07/80 11/07/80 11/12/80 11/12/80 11/13/80 11/17/80 11/17/80 11/18/80 11/19/80 11/24/80

SL

120 115

123 95 113 91 124

113 125

Figure 3.3.

Percentage of all Exallias brevis males at the Hanauma Bay study site observed spawning each month from November 1980, through April 1982. The dots represent monthly means; vertical lines represent the range in daily percentages during each month; the numbers above each dot represent the total number of males observed spawning during each month.

146

16

100 90 CJ)

Q)

eu E

15

70

--J

~

16

/t/

0>

..... .;..

\1

80

.s 60

"I

" 11 10

c

1~I~j~I

11

9

;:

~ 50

(f)

~

40 30

20

NO

1980

J F 1981

M

AM

J

J

AS

Months

0

NO

J F 1982

M

A

spawning males increased rapidly and remained high but variable through the end of the study in April, 1982. The spawning periods for individual males are illustrated in Figure 3.4.

Males which obtained the greatest

number of clutches are shown at the top of Figure 3.4, while less successful males are listed below them.

None of the

males spawned continuously, however male 'e' spawned at least once during every month of the study period, and male 'B' spawned every month except for October, 1980, and December, 1981.

Males such as 'F',

'A', 'J' and 'I' spawned

continuously for 4 to 8 months and then disappeared.

Other

males such as 'N' and '6' rarely spawned during the first few months but showed increased spawning in later months of the study. Male spawning events A total of 687 spawnings was recorded for individual males from November, 1980, through October, 1981, with a maximum of 85 observations for male 'C', and no observations for seven males (Table 3.3).

The mean number of spawning

observations per male was 22.16 (s.d.

=

23.57, n

=

31).

The number of clutches obtained by males from November, 1980, through October, 1981, is shown in Table 3.4.

The mean

number of clutches for all males during this period was 55.74 (s.d.

= 56.46, range =

° - 219,

n

= 31).

The mean number of

clutches for only those males which were successful in spawning was 66.46 (s.d.

= 55.43, range = 3 - 219, n = 26). 148

Figure 3.4.

spawning periods for individual male Exallias brevis at Hanauma Bay from November 1980, through April 1982. Each bar represents a period of uninterrupted spawning.

149

-

B C

2 F T X U

M J

....

a

L Cf)W ~ 4 CO I ~ 5

-

II-

S H y E

-

K N

-

•~ •l-

-





~

~



-

~



-l-

-

i 7

o N 1980

D

J

1981

F

..

M

- -----•

.



.• --

••

...• - - -

--•

Z

o

••

Q

G



~

A

U1



A

.- -

M J

J

.

A

S

~

-

0

N

..

X

U

L

~

W 4

5

- •

DJ

1982 Months

T

M

I-



B

C 2

F

M

.

A

N 6 7

Table 3.3 Observed spawnings for adult male Exallias brevis at Hanauma Bay November 1980, through October 1981 ====================================================== = ==~=

Month

Male

B C

N 2 0

D

6 0

6 0

2 6

J 9 12

F

7 1 1 2 7 0 7 9 0

6 3 4 1 11 4 9 5 0

A 7 8 5 10 7 2 4 10 4 9 8 0

5 8

M 7 6

2 F

T

0 0 1 1

0 5 4 4 1 0

7 1 0 0 13 2 13 12 0

6

5

6

5

7

3

8 4 0 0

1 1 0 0 0

12 2 7 5 11

3 5 4 0

0

X U

A M J L

W 4 I 5 Q G

S H K Y E Z 0 N 6 D p

a

Totals

M 8 4 3 7

J 6 5 9

J 8 2 8

A 11 12 15

S 12 10 9

0 7 12 9

a

a

4 0

6 6

0 1 3

3 0 4

3 1 0

1 0 0

3

5

2

0

0

5

5 1

2 1 10

a a 13

3 0 3

3 5 0

0 4 1

a

a

0

5

8

0

0

0

a

2 6

2

5

2

0 1 0

1 2 0

6 0 0

0 0

0 0 0

0 0 0

a

a

0

0

87

35

50

37

84 85 58 49 19 19 20 46 29 43 49 11 27 32 13 24 12 13 13 16 7 7 7 4

5

7

0

v

0 0

0 0

1 0

0 0 0

0 0 0

0

a

3 7 35

35 113

70

79

151

a 0

0

a

0 0

a a

0 0

0

a

0

0

51

48

0 47

a 687

Table 3.4 Clutches per month for adult male Exallias brevis at Hanauma Bay from November 1980, through October 1981 ===========================================================

Month

Male B C 2 F T X

N 8 1

D

J

F

7 11

14 19

16 10

M 15 12

14 0

16 0

0 0 1 0

0 12 7 12 0 0

20 0 1 0 25 6 21 22 0

11 9 4 13 14 0 19 15 0

14 13 16 11 21 9 13 7 3

25 18 12 22 13 11 17 16 14 19 15 10

3

12

16

11

15

4

12 11

3 1 0 0

22 2 0 0

6 13 13 11

0

0

0

0 0

2 0

0 0

0 0 0

U

A

M J L W

A

M 25 18 18 6 0 18 5 3 17 6 13

4 I

5 Q G

S

H Y K E Z 0

N 6 D p V

Totals J

J

A

S

0

26 20 17

28 7 16

24 25 26

21 27 21

10 29 25

0 23 19

2 2 5

27 0 13

25 12 6

8 6 4

15

7

0

0

10

2 4 20

0 0 24

6 0 16

6 18 0

0 25 4

8

1

0

20

21

0 3

8 8 9

14

3 1 12 5

2 0

0 3 0

1 5 0

12 0 0

0 0

0 4

0

0

0

0 0

0 0

0 0

0 0 0

0 0

0 0

0 0

0

0

3

14

3 7 64

83 208 153 183 220 129 158

152

0 92 137 156 145

219 197 135 103 97 93 93 91 85 85 79 73 64 61 50 43 27 24 23 21 19 14 13 12 4 3 0 0 0 0 0 1728

There are two reasons for describing the spawning observations and clutch totals in Tables 3.3 and 3.4 for only the period from November 1, 1980 through October 31, 1981. First, prior to November 1, 1980, not all males and females were tagged.

The subsequent 12 months span one spawning

cycle which is a reasonable period for analysis.

Second,

after November 1, 1981, fewer observations were made and females became increasingly difficult to identify. Therefore, by using only the period from November 1, 1980, through October 31, 1981, the between-months comparisons, as well as male-female comparisons, are more precise.

Inclusion

of data from November, 1981 through April, 1982, does increase the absolute number of observed spawnings, and clutches obtained by each male, but it does not significantly change the order of the most successful to least successful males.

Only male '8' is not listed in Table 3.4 as a result

of this procedure, since it was tagged after November, 1981. Male fecundity Fecundity is defined as the potential reproductive capacity of an organism or population, measured by the number of gametes or asexual propagules (Lincoln, et al., 1982). For female fishes, this usually refers to the number of eggs she produces.

For males, fecundity rarely, if ever, refers

to the number of sperm produced but to the number of eggs fertilized.

This latter definition of "male fecundity" will

be used herein. 153

The estimate of number of eggs/mm 2 in a clutch obtained by Method A was 1.50 eggs/mm 2 (n= 5, s.d.= 0.069).

The

estimate by Method B was 1.58 eggs/mm 2 (n= 5, s.d.= 0.108). These means were not significantly different (Student's ttest for comparison of means with equal variances, p.

o

70

C

(])

:J

cr

60

(])

~

u..

50 40 30 20 10

25 75 125 225 325 425 525 62? 725 825 925 1025 175 275 375 475 575 675 775 875 975

Clutch size (eggs x 10)

156

Table 3.5. -- Mean monthly clutch sizes (number of eggs/ clutch) for male Exallias brevis at Hanauma Bay from May 1979, through April 1980 ===========================================================

Month May Jun. Jul. Aug. Sep. Oct. Nov. Dec. Jan. Feb. Mar. Apr.

Mean 3413 2870 2687 3086 3125 3794 4858 4724 4500 4386 4445 5171 Total

=

n s.d. 57 1854.37 107 1887.18 81 1702.38 194 1815.95 207 1704.25 57 1995.55 33 2250.61 57 2225.26 36 2619.13 90 1874.23 52 2310.30 71 2025.22 1042 clutches

157

Table 3.6. -- Comparison of monthly mean clutch sizes for Exallias brevis males at Hanauma Bay May 1978, through April 1979. Tukey-Kramer test for unplanned mUltiple comparisons among pairs of means. Minimum differences at the p ' 6 o c 4 0,) 2

JAN.

:::J

CO,) L.

LL

28 26 24 22 20 18 16 14 12 10

FEB.

8

6 4 2 16 14 12 10

MAR.

8

6 4 2 9 10 11 12 13 14 15 16 1718

Days 169

Figure 3.9.

Observed spawning interval frequencies for female Exallias brevis at Hanauma Bay from April through October, 1981. The month of May has been omitted due to insufficient observations.

170

24 22 20 18 16 14 12 10 8 6

APR.

4

2

8 6 4 2

>.

u

C

(])

5-

(]) ~

JUNE

8 6 4 2

JULY

LL

10 8 6

AUG.

4 2

8 6 4 2

SEP.

8 6

OCT.

4

2 3 4 5

6 7 8 9 10 11 12 13 1415161718

Days 171

intervals resulting from missed observations of spawning events.

Spawning was obviously occurring much more often

than observed:

1728 clutches were recorded during the year

but only 690 spawnings were observed.

Thus 60.1% of all

spawning events were not observed, which implies that females were indeed spawning frequently over short intervals of 3-4 days rather than 6 days or longer. The frequencies of the 6-,7-,8- and 9-day intervals varied in proportion with the frequencies of 3- and 4-day intervals.

For example, in March when 4-day intervals were

most common, 8-day intervals were also common, while 7- and 9-day intervals were rare.

In September only 3-, 6- and 9-

day intervals were observed.

Therefore, to estimate the mean

spawning interval each month, all 3-, 4-, and 5-day intervals were summed, while 6-, 7-, 8-, and 9-day intervals were included as multiples of either 3 or 4.

While it is possible

for an 8 or 9 to represent a multiple of 5 plus either 3 or 4, this is unlikely given the rarity of observed 5-day intervals.

Intervals of 10+ days were not included in

estimating mean monthly spawning intervals. Given the above assumption, the mean monthly spawning intervals are listed in Table 3.8.

The minimum mean spawning

interval, 3.00 days, was recorded for September, while the maximum interval, 3.90 days, was recorded during March. Significant heterogeneity was detected among these values (Goodness of Fit test, G

= 188.4, P «.001) with the months 172

Table 3.8. -- Frequencies of 3-,4-, and 5-day spawning intervals for all 3-,4-, and 5-day observations plus 6 (=3+3), 7 (=3+4), 8 (=4+4) and 9 (=3+3+3) day intervals. ============================================================

Mon. Nov. Dec. Jan. Feb. Mar. Apr.

3 8 5 20 5 5 17

4 8 9 52 41 43 44

Jun. Jul. Aug. Sep. Oct.

15 14 34

9 6

3

23

0

o

~_1

Q 9

166 216

1.

5

Total

Mean interval 3.59

1 1

17

15

3.73

4

o

76 46

o

1

48 62

3.79 3.89 3.90

o 1 1

24 21 38 23

-il

3.74

3.38 3.38 3.13 3.00 3.05

391

Homogeneity among all months tested for Goodness of Fit: G-statistic = 188.40, P«.OOl

2.

Sets of homogeneous months (unplanned tests using Gstatistic, 5% significance level): Nov. Dec. Jan. Feb. Mar. Apr. Jun. Jul.

Aug. Sep. Oct.

===========================================================

173

of November through April being significantly different from the months of June through October.

A value for May, 1981,

could not be calculated due to limited observations. The spawning interval appeared to be the same for all individuals regardless of size.

For example, during April,

1981, the smallest individual (81 mm SL), female '4' spawned on a 4-day cycle, which was identical to one of the largest individuals (121 mm SL), female 'WI.

Similar

obsel~ations

were made during other months for other individuals. Female fecundity A total of 63

~.

brevis was collected at Kahe Point by

two collectors during 12 dives between July 14, 1984, and March 27, 1986. were females.

Of those collected, 31 were males and 32 Of the females, data on batch fecundity were

obtained from 23 individuals; the remainder were either immature or the ovaries were poorly preserved (Table 3.9). Four or five size classes of ova were found in the ovaries of most females.

A frequency distribution of these

size classes obtained from one female with a batch of fully hydrated eggs is shown in Figure 3.10.

A significant

relationship between batch fecundity and standard length was obtained using simple linear regression as illustrated in Figure 3.11 (r 2 =.67).

A better fit was obtained after

applying a log transformation to the data (r 2 = .74, see Table 3.10).

174

Table 3.9. -- Standard lengths (SL), gonad-free weights (GFW), gonad weights (GW), and batch fecundity (BF), for male and female Exallias brevis from Kahe Point, Oahu. ========================================================

SL 76.9 77.7 82.2 86.2 89.7 90.7 91.1 91.7 93.2 94.1 97.4 98.9 98.9 99.9 102.4 103.4 103.6 104.0 104.0 104.8 105.0 105.4 105.7 105.8 106.2 106.5 107.5 109.6 110.8 111.2 111.3 114.1 115.0

FEMALES GFW GW 19.66 0.028 20.01 0.068 23.69 0.289 0.352 29.30 30.61 1.401 33.92 0.357 32.39 0.581 31.48 1.056 37.32 0.075 33.06 0.194 38.86 0.445 41.52 0.568 46.70 0.923 42.22 0.865 45.22 1. 040 46.75 0.788 47.88 0.772 1. 373 47.87 0.570 47.08 50.50 0.507 52.03 0.240 49.89 0.599 48.58 0.684 45.79 0.950 48.53 1.134 54.09 0.498 1. 379 52.73 1. 700 51. 58 0.787 59.23 0.471 60.85 1.672 58.26 1. 092 68.43 1. 794

BF 1351. 5 2334.3 2618.0 2657.6 2586.6 2370.0 2386.2 2639.0 3894.8 3506.1 2307.5 4152.0 3200.0 3501.2 3743.4 3945.0 5609.4 3621. 5 3702.5 3894.8 3541. 5 5046.4 6315.0

175

MALES SL GFW 68.6 14.35 2.4 27.56 84.4 28.11 86.6 26.98 87.8 31. 05 90.0 31.85 92.4 37.18 93.8 37.14 95.1 32.88 95.4 33.01 97.3 37.69 97.7 39.70 98.0 44.43 98.1 34.42 98.6 43.62 98.6 39.58 100.0 41.62 100.8 46.40 100.5 36.50 101.0 44.47 102.7 42.03 103.4 46.21 103.6 48.02 103.6 44.06 105.2 47.39 106.8 51.21 107.0 50.83 107.3 54.90 109.5 57.97 109.7 55.59 114.9 59.23

GW 0.002 0.050 0.016 0.080 0.051 0.016 0.040 0.125 0.043 0.092 0.104 0.153 0.107 0.028 0.094 0.110 0.120 0.054 0.130 0.120 0.104 0.172 0.133 0.079 0.141 0.149 0.104 0.133 0.197 0.160 0.108

Figure 3.10.

size distribution of ova from one female Exallias brevis, standard length = 107.5 rom, collected at Kahe Point, Oahu. Ovarian sample weight = 0.0255 gms, total ovary weight = 1. 3794 gms.

176

30 25 ~

-..,J -..,J

>. U 20 c (])

5-

15

(]) ~

LL

10 5

~ 4 5

6 7 8 9

Diameter (1 unit=.0256mmJ

Figure 3.11

Batch fecundity (BF) of female Exallias brevis regressed against standard length (SL). Dashed line represents simple linear regression: BF = -7483.47 + 1087.31(SL in em). Solid line represents regression after log transformation of both batch fecundity and standard length (in mm): BF = (7.73 x 104)SL3.31

178

70

,

,',,

,



,,'-

, ",' ,

60

" " "

• ,......,

, ,, • ",I'

50

0 0

x

'--'

. ,-,. ,-.. " . ,,.

>. :t:: "0 C :J 0

(])

40

,

• ,I' ••



,,

30

"

,,'-

.c

• e. ",

,

0 ..-



~

CO

~

,'-

~

"

.





20 ------. B F = - 7 4 8 3 .47 + 10 8 7 .3 1 SL 10

80

-

90

BF = (7.73

100

110

Standard length

179

X

10-

120

4)

3 31

SL .

130

Table 3.10 Regression analyses of batch fecundity (BF) as a function of length (SL in cm) and weight (WT). ============================================================

Model: BF = a + b(SL) Regression equation Variable length BF = -7483.47 + 1087.31(SL) weight BF = - 500.98 + 90.57(WT) Model: BF = a(SL)b Variable length weight

Regression equation BF = 1.58(SL)3.31 BF = 43.65(WT)1.15

180

r2 .67 .71

r2

.74 .74

This relationship is described by the power function: BF = (7.7328 x 10- 4) x SL3.3118 Where BF is batch fecundity and SL is standard length in millimeters.

Discussion spawning season Johannes (1978) noted that a collective spring spawning peak occurs in fifteen of eighteen tropical coastal locations, and in five instances there is a second spawning peak in the fall.

Walsh (1984, 1987) reviewed reports on

spawning for 48 species of Hawaiian fishes and concluded that while some species have protracted spawning seasons, a distinct spawning peak occurs between April and July, with a secondary smaller peak occurring shortly after the late summer temperature maximum. Walsh speculates that water temperature and/or photoperiod are among the most important proximal factors controlling reproduction.

Lobel (1988) suggests that the

ultimate causal factor is the seasonal pattern of ocean currents which affect survival of larvae and recruitment. While the exact timing of peak reproduction varies among species of Hawaiian fishes, an often unstated observation is that the period of minimal spawning is the same for many species and occurs in the fall around November.

181

strasburg (1953) examined the gonads of 1010 zebra blennies, Istiblennius zebra, and determined that spawning in this species occurs throughout the year but peaks in the spring and summer.

The lowest percentage of mature eggs

(Strasburg's stage 3A) were observed in January, August, and November.

McDonald (1981) recorded the lowest number of

brooding males of the blue-eye damsel fish, Plectroglyphidodon ;ohnstonianus, occurring in January, July, and November; and little or no spawning in the Pacific gregory, stegastes fasciolatus, during the period from June through November. The Hawaiian sergeant, Abudefduf abdominalis, spawns throughout the year with a period of increased reproductive activity starting about mid-December and continuing through July (Helfrich, 1958).

The low ebb in spawning activity for

this species occurs in October and November which Helfrich attributed to the sharp drop in the minimum water temperature at this time. The spawning season for the milletseed butterflyfish, Chaetodon miliaris, was determined by Ralston (1975) from the relationship between gonad weight and body weight (gonadal index

= gonosomatic index of other authors).

He reported the

spawning season for this species to extend from November to perhaps June, with a maximum near late February or early March.

Tricas (1986) reported that spawning occurs

continuously from fall through spring in g. multicinctus, and Hourigan (1987) found that energy investment in the ovaries 182

of two additional Chaetodon species was highest in the spring. Among individual male

~.

brevis, there was considerable

variation in the periods of time during which they spawned. But when observed at the population-level, the pattern of annual spawning was similar to most other Hawaiian fishes,

i.e., the population showed a peak in spawning activity in the spring, a second peak in early fall, and minimal or no spawning in November (Figure 3.4). A significant positive correlation exists between the

size of males and the total number of days they maintained active nests with eggs (data square-root transformed, r .430, n

= 27, P

< .05).

=

The number of days was calculated

from the original daily observations including all months of the study (October, 1980, through April, 1982).

Daily gaps

were counted if spawning was observed before and after the gap.

Gaps up to five days are possible with no loss of data

since clutches present in the nest require up to six days to hatch.

This correlation was calculated using only males

which spawned and using the initial sizes recorded for each individual. For females, the correlation between size and the number of months each was observed spawning, was not significant (data square root transformed, r

= 0.212, n = 28, P

> 0.05).

This correlation is improved somewhat by deleting female 'T' which was the largest female in the population but only 183

spawned once before disappearing.

However, the correlation

remains statistically non-significant (data square root transformed, r

=

0.331, n

=

27, P > 0.05).

The overall population pattern of annual spawning thus appears to be the result of large males which spawn frequently or continuously throughout the year and smaller males which increase their reproductive activity during the months of January through April but show sporadic, or no spawning at other times of the year.

Among females, both

small and large individuals tend to spawn for extended periods during the year but the greatest number of females spawn between January and April.

The factors responsible for

initiating or terminating spawning periods in individuals are not apparent from this study. Mating patterns Thresher (1984) has pointed out that data on the spawning patterns of small site-attached reef fishes are sparse and available only for the pomacentrids.

He goes on

to note the long-standing assumption that both sexes are probably promiscuous but that evidence of female promiscuity has not yet been obtained.

He predicted that both sexes in

promiscuous species may actually show high degrees of mate fidelity.

The data collected on

~.

brevis thus provide some

of the first observations on the individual mating patterns of both sexes for any coral reef fish.

184

A total of 508 spawning events is recorded in Table 3.11 where both the male and female were positively identified.

This figure is higher than the 478 spawnings

recorded in Table 3.7.

The difference (30 or 5.9%)

represents the number of times a female was observed with two males on the same day

an occurrence which increases the

"apparent" number of spawning events. Whether or not the female actually spawned with two males on one day could not be determined during the brief daily observations.

A female in a male's nest does not

necessarily imply that spawning is occurring.

During hour-

long focal-individual observations (Chapter II), I occasionally observed spawning females leave the male they were spawning with, and swim to the nest of a second male, particularly when males' nests were in close proximity.

This

occurred while females made forays away from the first male's nest to feed and were courted by other males.

I never

observed any spawning occuring with the second male and the female always returned to the first male within a minute or less, and continued spawning with him. Exceptions did occur when the female was evicted from the first male's nest by a second, usually larger, female. The evicted female would either swim to a second male and spawn with him, or would remain in the vicinity of the first male's nest.

occasionally they would remain near the first

male even when vigorously courted by nearby males. 185

The 5.9%

Table 3.11 f~om

spawning events between individual male and female Exallias brevis at Hanauma Bay through October, 1981

November, 1980

=================================================================================--==================

Females A B

0 E F G

u

A 1 1 9

B

1

4 17 9 5 5 1 3 2 12

1

1

1

13 1 4

I

J K L

....OJ 0\

C

E

K

L

Males M 0

2 2 1 2

S

T

U

W

X

y

Z

2

4

5

6 1

13 6 1 4 7

3

1 1

4 1

1

1

2

1 1 3 5

4 1

4

2 1

2

1 4

1 3 1

4

13

1 4 1

3 1 2 4

1

2

9

9

1 2

1

1 2

7

2 3 2

1 6

4

3 2 26

Q

1

3

1 1

5 3 2 1

1

3

1

3 3

1 3

1

3

U

4

1

3

2 1 1

4 15 1

2

X Y Z

J

I

1 2

Q

V W

n

1

0 P

S T

G

1

M N

R

F

12 9

2 3 4 8 _5__ 1 1 2 Spawns: 44 59 65 Mates : 14 10 10

2

2 1

2

1

3

3

2

1

3

1

3 6 1

1 7 35 16 6 8 6

1 21 37 1 11 8

3 7 31 17 5 5 5

3 24 14 12 11 3 6 6 6 7

2 10 1 3

9 4

4 1

28 4

9

4

9 38 15 2 1'4 16 22 43 1 3

4 8 3 1 6 7 10

14

2

1 3

Spawns Mates 3 3 9 4 3 12 38 6 38 6 20 9 8 16 33 8 1 1

3 8 1

6 1 9 45 24 6 8 6

2 5 2

30 4 44 34 7 7 15

---lL

9

1 1 5 8 3 4 9

2 4 5 __ 5

508

147

figure thus probably represents the maximum percentage of clutches which were divided between two males. While females rarely spawn with two males on any given day, the data in Table 3.11 are evidence that females do indeed spawn with more than one male during successive spawns. 2.74, n n

= 24).

The mean number of mates per female was 5.07 (s.d. =

=

=

29), and for males the mean was 6.12 (s.d.

3.12,

The difference between these means is not

significant (square-root transformation of data, variances equal, Student's t-test, 0.2

Smile Life

When life gives you a hundred reasons to cry, show life that you have a thousand reasons to smile

Get in touch

© Copyright 2015 - 2024 PDFFOX.COM - All rights reserved.