(Herbst) mates both on the surface near female-defended burrows and ..... This effect of burrow diameter could lead to size assortative mating in. U. tetrugonon ...
Journal
ELSEVIER
Seiji *Department
of Experimental Marine Biology and Ecology, 196 (1996)
Mate acceptance
131-143
and guarding by male fiddler crabs Uca tetragonon (Herbst)
Goshimaa.*,
of Marine
JOURNAL OF EXPERIMENTAL MARINE BIOLOGY AND ECOLOGY
Biological
Tsunenori
Kogab,
Minoru
Muraib
Science, Faculty C$ Fisheries,
Hokkaido
University,
Hakodate
041,
Japan hDepartment
ofBiology, Faculty of Science, Kyushu University, Fukuoka 812, Japan
Received 17 February 1994; revised I May 1995; accepted
13 June 1995
Abstract
The fiddler crab, Uca tetmgonon (Herbst) mates both on the surface near female-defended burrows and underground in male-defended burrows. In underground mating, which requires relatively high male investment, males attract both non-ovigerous and ovigerous females into their burrows by claw waving. Males aggressively expel some females soon after they enter their burrows and others after pair formation, but before females spawn. Finally males guard some females in their burrows until they spawn, which presumably ensures paternity. Males do not select mates of a particular body size, but they do differentially accept females with late-stage eggs, those about to release larvae and spawn another clutch. Except at the beginning of the reproductive season, few ripe non-ovigerous females are available because females spawn successively and only in moderate synchrony. By differentially accepting late stage ovigerous females, males may increase their fertilization rates because they minimize the time they spend guarding each of their mates to ensure their paternity. A male-biased operational sex ratio and a high last male advantage in sperm competition are two conditions that may have favored male choice based on female guarding time in this species. Keywords:
Egg stage; Fiddler crab; Mate assessment;
Mate-guarding;
Operational
sex ratio
1. Introduction Paternity may depend upon the ecological and social circumstances in which breeding occurs. Mate-guarding is one male behavior that may enhance paternity (see Ridley, 1983, for review). Mate guarding, however, costs time and energy and, therefore, *Corresponding
author.
0022-0981/96/$15.00 0 1996 Elsevier Science B.V. All rights reserved SSDf 0022.098 1(95)00127-l
132
S. Gnshimu
et ul. I J. Exp. Mar.
Bid.
Ed.
196 (1996)
131-143
reduces the time during which males can search for more females. The relative advantages of how long males of a species guard prespawning females will depend on the environmental conditions in which they breed and the strategies adopted by other males (see Yamamura, 1986; Yamamura and Tsuji, 1989, for theoretical discussion). Fiddler crabs of the genus Ucu mate on the surface near female-defended burrows or underground in male-defended burrows. In underground mating male reproductive investment is higher than in surface mating (Goshima and Murai, 1988), because a male must court vigorously to attract a female into his burrow, stay with her for a few days (Yamaguchi, 197 I Christy, 1987), and guard her until she lays eggs (Goshima and Murai, 1988). In contrast, surface mating takes only a few minutes and males do not invest in either pre- or post-copulatory guarding (Yamaguchi, 1971; Goshima and Murai, 1988). In some species, males attract only non-ovigerous females into their burrows (Yamaguchi, 197 1; Christy, 1982; Zucker, 1983; Murai et al., 1987), and matings with ovigerous females are seen only on the surface (Nakasone et al., 1983 Salmon, 1984 Salmon, 1987). Unlike other fiddler crabs, however, male Ucu tetrugonon (Herbst) attract both ovigerous and non-ovigerous females into their burrows. When a male mates with an ovigerous female, he has to wait, first for the release of the larvae fertilized by an earlier mate, then for the new brood. Mating with non-ovigerous ripe females should be more advantageous because guarding time to spawning should be less. The first aim of this study was to discover why U. jetrugonon males guard ovigerous females. Whereas males court and attract both types of females, they show some rejection behavior suggesting that they assess females by certain criteria. To determine whether males may assess females on the basis of their time until spawning, we observed the behavior of male CJ. tetrugonon toward non-ovigerous females and females with eggs in various developmental stages. Finally we examined if males select females as mates on the basis of their body size. Although pair formation may be the result of mate assessment by both sexes, we focus mainly on male behavior in this paper.
2. Materials and methods 2.1. Study sites und muting
beha\?or
The study was carried out in a pebble intertidal zone in Ao Tang Khem, Phuket, Thailand in November 1987, and from October to December 1990. The two types of mating behavior of Uca terrugonon may be summarized as follows. In surface mating a male courts a nearby resident female for several minutes, then copulates with her for an average of 96249 SD s (range = 48-238 s, II = 12), after which he moves away. In underground mating, a male attracts a wandering female into his burrow by claw waving. Either sex may enter the burrow first. Males frequently use their large chelae aggressively to expel females that have entered their burrows. If the female is not expelled, the male plugs the burrow with surface mud, and the pair stays underground for 5.2+ 1.5 SD days (range = 4-8 days, n = 12, Koga, Goshima and Murai, unpubl. data) during which time they presumably copulate (Koga et al., 1993). In some instances, males opened their burrows and were active on the surface, but did not
S. Goshima et al. I J. Exp. Mar.
Biol. Ecol. 196 (1996) 13-143
133
court while their mates remained underground. During this time, some females would attempt to come out of the burrow whereupon the male would rush into the burrow and plug the entrance. Crabs that have mated underground separate after an ebb tide when the female has newly oviposited eggs. Either sex may leave the burrow and seek a new one. If the male leaves, the female closes the burrow and incubates her clutch underground for at least a few days, after which she comes out and is active on the surface.
2.2. Egg number and incubation period The number of eggs attached to the female’s abdominal appendages (pleopods) was estimated by weighing the egg mass and a subsample of counted eggs and dividing the total egg weight by the estimated weight per egg. Egg developmental stages were determined microscopically and were classed into one of four categories (Henmi, 1989) for each brood: Stage I eggs are newly-deposited and completely filled with yolk. Stage I-l eggs have up to 16 cells and stage I-2 eggs have more than 16 cells. Stage II eggs have a yolk-free portion with the developing embryo but no eye pigment is present. This stage was subdivided into five sub-stages, each corresponding to a 10% decrease in yolk volume. Stage III eggs have their yolk reduced to less than 50% of the egg volume and the eye pigment is present. Stage IV eggs have a pigmented embryo with a well-developed abdomen and limbs and yolk divided into small globes. To determine the duration of each egg stage, 55 ovigerous females in various stages were kept individually in beakers containing sea water at approximately the same temperature as the sea during the study (average temperature = 28.3”C, range = 26.929.O”C). Egg developmental stage was monitored daily. Seven females were monitored from just after oviposition (no cell division) until hatching. On the assumption that the rate of yolk absorption is constant within the stage, we could then estimate the approximate last spawning date for any ovigerous female.
2.3. Field observations The carapace width (CW) of observed crabs was measured and females were classified as ovigerous (OV) or non-ovigerous (NO). To estimate the relative frequency of both types of mating, areas on the sediment of approximately 4 m* each were arbitrarily defined within the habitat. We observed each area for l-2 h at different times and days for a total of 22.15 h in November 1987. We caught 47 females from October to December 1990. They included wandering females which were potential mates (n = 20), females that were expelled after they entered male burrows both before (n = 10) and after (n = 7) males plugged their burrows, and females that stayed with males in closed burrows and were not expelled (n = 10). We measured (CW) all females and checked the egg developmental stages for all OV females. The CWs of the four groups of females were compared with the Kruskal-Wallis test. The reproductive state (NO or OV) of these females and egg
134
S. Goshimcr ct al. I J. Exp. Mar.
Bid.
Ed.
196 (1996)
3.3/L143
developmental stages of OV females were compared with the Fisher’s exact test (Sokal and Rohlf, 198 1). In 1990 we classified as NO or OV 63 females that mated underground and 100 females that mated on the surface. We caught and measured (CW) both males and females in 55 of the 63 underground matings. On October 21, 1990, about the middle of reproductive season, we collected a sample of crabs by digging, measured the CWs, and classified their egg developmental stages. This sample allowed us to estimate the percentage of OV females and predict their spawning date. 2.4. Field experiments To clarify the relationship between female CW, egg-carrying condition and pairing success in underground mating, we released 28 OV females with eggs in various stages and 13 NO females near waving males. After the release, the females were observed entering and leaving some male burrows. Two outcomes were recorded: non-pairing in which females briefly entered and then were aggressively expelled from burrows by males using their large claw (rejected, category I), and pairing in which females entered and stayed in male burrows and the males closed the burrows from within. We enclosed each burrow in which pairing occurred with a 20-cm diameter, 12-cm high plastic fence. We checked these enclosures daily to determine which sex was out of the burrow. If the female was out of the burrow, we judged her to have been expelled by the male. If the male was on the surface we judged that he had left the burrow. Two observations support these judgements. First, when placed in the same male burrow, females that had been on the surface again soon emerged from the burrow. Second, when the male was on the surface the female (OV) nearly always plugged the burrow from within. This behavior is common for many fiddler crabs. The pairing was further classified into two categories; the female was expelled after pair formation (out of the burrow, category 2), and the female was not expelled (in the burrow, category 3). Seven to 9 days after we released the females that were not expelled (category 3) we dug out the burrows, caught the crabs, and checked the females for eggs. We then classified these pairs into three types: (a) pair in the act of mating; (b) female without new eggs; (c) newly spawned female. Seven crabs escaped the enclosures; these pairings were excluded from the analysis. The CWs of the three groups of females (category 1 to 3) were compared with the Kruskal-Wallis test and the reproductive state (NO or OV) and egg developmental stages of OV females were compared with the Fisher’s exact test (Sokal and Rohlf, 1981). Because of small samples, we pooled the data of egg developmental stages I and II as “early stage” and stages III and IV as “late stage”.
3. Results 3.1. Frequency
of both types
of mating
We saw 23 pairings during the 22.15 h observation period, 18 (78%) on the surface and 5 (22%) in male burrows; crabs paired signiticantly more on the surface (X2-test,
S. Goshima
et al. I .I. Exp. Mar.
Bid.
15
Ed.
20
196 (1996) 131-143
25
13s
30
Carapace width (mm) Fig. 1. Relationship between female carapace width (mm) and the number of eggs in her clutch. The line is described by 4’ = - 12321 + 1027x (r = 0.706, n = 24, P < 0.05).
,y* = 7.348, P < 0.01). Both NO and OV females mated on the surface and underground.
Of the 63 underground matings observed in 1990, 12 ( 19%) involved NO and 5 1 (8 1%) OV females. A total of 100 surface matings were observed with 45 NO (45%) and 55 OV females (55%). 3.2. Clutch size and duration of egg developmental
stages
Fig. 1 shows the relationship between female CW and clutch size. There is a significant correlation (n = 24, r = 0.706, P < 0.05), but body size accounts for only up to 50% of the variance in egg number. Table 1 shows the durations of the different egg developmental stages. From stage I to IV, the total incubation period was about 28 days, similar to 28.322.3 SD days (range = 25-31 days, n = 7) for the crabs monitored successfully from the newly oviposited egg stage until hatching. 3.3. Pairing by body size In 55 pairs which mated underground, there is a significant positive between the CW of the two sexes (Y = 0.583, P < 0.05) (Fig. 2). However, 34% of the variance in female size is explained by male size. Table 1 Incubation
days (mean+SD)
correlation only up to
for each egg stage
Egg stage
Days
N
I II III
8.322.3 9.922.2 5.622.1 3.9kl.8 28.322.3
18 14 5 II 7
IV I-IV N = number of crabs
136
S. Coshima
E t 5 .T! s
et cd. I J. Exp. Mar.
Bid.
Ed.
I96 (1996)
281
13/-143
/
2624 22-
$ 3 5 Q) Tii E
14-
I?
12
/ 12
I
1
I
14
16
18
1 20
22
Male carapace
I
I
i
24
26
28
width (mm)
Fig. 2. Relationship between male and female sizes in underground pairings. The line is described y = 9.12 + 0.58.~ (r = 0.583, ,I = 55, P < 0.05). The dashed line indicates the same sizes for both sexes.
3.4. Pairing in relation to female
reproductive
by
condition
3.4.1. Field observations Table 2 shows the reproductive condition and the egg stages of the females that were wandering on the surface or attracted into male burrows. Wandering females that were potential mates included NO and OV females (A, B in Table 2). The females with early stage eggs (stage 1 and II) were rejected more frequently when they entered male burrows than were NO females (Fisher’s exact test, P = 0.049) and OV females with late stage eggs (stage III and IV, P = 0.014) (B, C in Table 2). The early stage OV females were also expelled more frequently after pair formation than were OV females with late stage eggs (P = 0.048). After pairing, NO and early stage OV females were expelled at the same rate (P = 0.155) (C, D in Table 2). No late stage OV females were rejected before pairing (B) or expelled after pairing (D). There was no significant
Table 2 Reproductive Egg stage
condition
of different Behavioral
types of females involved
in underground
mating
category
A
B
C
D
2 4 4
5 I
II III IV
I 4 I2 I 2
2 I 4
N
20
IO
IO
7
CW (mm)
19.622.3
18.7Z2.3
19.623.3
18.7k2.1
NO
1
2 2
A, wandering females; B, females rejected upon entering the male burrow: C, pairing females; D, females expelled after pairing. NO, non-ovigerous females; CW, carapace width (meanfSD); N, number of crabs
S. Goshimu
et al. I .I. Exp. Mar.
Biol. Ed.
196 (1996)
137
13-143
difference in the CW among the females belonging to different (Table 2, Kruskal-Wallis test, H = 2.084, P > 0.50).
behavioral
categories
3.4.2. Experiment There was no significant difference in the CW of experimental females with eggs in different developmental stages (Kruskal-Wallis test, H = 4.65, P > 0.30). There was also no difference in the CW among rejected, expelled, and non-expelled females (Kruskal-Wallis test, H = 1S46, P > 0.30) (Table 3). Table 3 shows the result of the release experiment. There was no significant difference in the relative frequency of pairing ((expelled + non-expelled females; both had been pairing)ln) between NO (7/ 12) and OV females ( 13/22) (Fisher’s exact test, P = 1.000). No difference was seen in the rate of expulsion ((rejected + expelled females)ln) between NO (1 l/ 12) and OV females with early egg stages (stage I and II, 14/ 16) (P = 0.611), but a significant difference existed between OV females with early (14/ 16) and late stage eggs (III and IV l/6) (P = 0.009). There was also a significant difference in the expulsion rate between NO (1 1/ 12) and OV females with late stage eggs (l/6) (P = 0.008). Hence, OV females with late stage eggs were seldom expelled by resident males. Among those that entered the male burrows, three had eggs in stage IV, two of them spawned 1 day later, the third spawned 3 days later, indicating that females spawn new clutches very soon after hatching. There was no difference in spawning rate between NO (O/12) and the early egg stage OV females (O/16) during 7-9 days experiment. Spawning rate differed significantly between NO (O/12) and the late stage OV females (3/6) (P = 0.049), as well as between the early (O/ 16) and the late stage OV females (3/6) (P = 0.026). Table 3 Pairing frequency
in accordance
Egg stage
Female
Female
Female non-expelled’
rejected”
expelledh
Pairing
NO
5
I
2 6 1
II III IV
with the reproductive
6 3 3
condition
of females in underground
Female only
N Spawned
1
2
3
I I
3
2
3
N
14
12
8
CW (mm)
20.2k2.7
21.4k3.1
19.2k6.1
mating
12 5 II 2 4
34
NO, non-ovigerous females; N, number of crabs; CW, carapace width (mean+SD). * Female rejected: the released females that briefly entered and then were aggressively expelled from the burrows by males. h Females expelled: the females that were out of the burrows in enclosures after they stayed in the burrow together for at least a day. LFemale non-expelled: the females that were not expelled by the males during the experiment.
138
S. Goshima et ul. / J. Exp. Mu-. Rid. Ed.
196 (1996) 131-143
(N = 86)
rl
r-l
R-l 26 Carapace Fig. 3. Carapace width frequency distribution (21~ October 1990). Percentage of ovigerous
3.5.
Percentage
qf ovigerous
width
(mm)
of the females sampled randomly at mid reproductive females above 14 mm in CW (shaded area) is XI %.
females
and spawning
season
date
Fig. 3 shows the CW frequency distribution of females sampled by digging during the middle of the reproductive season. OV females were 2 14 mm in CW and made up 81% of the sample. The egg stages and estimated spawning dates for these females are shown in Fig. 4. OV females had eggs in various stages (Fig. 4A). The predicted spawning frequency distribution was not uniform, nor was spawning confined to certain day(s). Rather. spawning showed moderate synchrony (Fig. 4B).
4. Discussion 4.1. Mate guarding
to ensure
paternity
Female crabs typically store in their spermatheca several spermatophores from more than one mate (Salmon, 1984; Diesel, 1988). Therefore, we can expect sperm competition among the rival males (Smith, 1984; Birkhead and Moller, 1992). In the spider crab, Inachus phalangium (Diesel, 1990) and the sand-bubbler crab, Scopimera globosa (Koga et al., 1993) only the sperms of the last mate fertilize the eggs. The oviducts of these crabs enter the spermathacae ventrally, near the opening to the vagina. This anatomical arrangement appears to be the usual pattern for ocypodid crabs (Diesel, 1991). Thus, we expect last mate sperm precedence in fiddler crabs (Koga et al., 1993). In underground mating, male fiddler crabs may ensure their paternity by guarding their mates against mating by other males until the female spawns. 4.2. Male fertilization
and mating
success
relative
to .female
body size
Larger females produce more eggs per clutch than do smaller ones in many crab species, and it is probably advantageous for males to mate with the larger females.
S. Goshima
et al. I J. Exp. Mar.
“-
Bid.
Ed.
196 (1996)
131-143
139
A
4
g
-
(N = 79)
20 1 ‘i5 6
10 -
z
NO I1 I2
II 1 II2 II 3 II4 Ii5 Ill IV
Egg stage
27 29 1 Sw
3
5
7
9
11 13 15 17 19 21 act
Date
Fig. 4. Egg stages of the females sampled at mid reproductive season, 21st October 1990 (A), and their spawning frequency per day calculated from estimated spawning dates (B). To estimate spawning date, we used the duration of each egg developmental stage shown in Table I. The duration used in stages I- I and II- I were 2.0 (+O.O SD, n = 3) and 4.0 days (+ 1.8SD, n = 12). respectively, while for other substages in stage II it was 1.5 day on the assumption that the rate of yoke absorption is constant within the stage. 0 and 0 indicate day of new and full moon, respectively. NO = non-ovigerous female.
However, variation in egg number within female size classes was larger in U. tetrugonon (Fig. 1, r2 = 0.50) than U. rapax (r2 = 0.72, Greenspan, 1980) and U. rosea (r2 = 0.84, Macintosh, 1979). This variation is probably caused by individual variation in nutrition together with successive spawnings in which the female may oviposit a new brood a few days after releasing zoea larvae (Table 3). OV females are less active than NO females and males, because OV females sometimes remain within a burrow while incubating. Individual differences in time spent underground (Koga, Goshima and Murai, unpubl. data) affect feeding time, leading to the variation in the energy and nutrients obtained for egg production. Male and female sizes were positively correlated in underground matings (Fig. 2). Size assortative matings are seen in many taxa, and there are many possible causes (Arak, 1983; Ridley, 1983; Crespi, 1989; Brown, 1990). Larger male U. tetrugonon tended to pair underground with females with a wider range of CW than did smaller males (Fig. 2). The diameter of the burrow entrance is correlated with crab size (Frith and Brunenmeister, 1980; Christy, 1982). It may be difficult for larger females to enter burrows of males 19 mm or smaller in CW, while burrows of the larger males are readily accessible to females of all sizes. A similar situation has been reported in U. pqilutor (Christy, 1983). This effect of burrow diameter could lead to size assortative mating in U. tetrugonon even though males may not select mates by size (Crespi, 1989). Indeed,
140
S. Goshimu
et cd. I J. Exp. Mar.
Bid.
Ed.
196 (1996)
I.~/-143
there were no significant differences between the sizes of females which males expelled and accepted for mate guarding (Table 2 and Table 3). Finally, large variation in egg number within female size classes would not strongly favor preferences for larger females. From the female point of view, she may show some selection for male or burrow size during pair formation (Christy, 1982; Goshima and Murai, 19X8), however, female mate choice in this species has not been studied. 4.3. Guarding
lute-stuge
ovigerous females
Males less often rejected or expelled after pairing females with eggs in the late compared to early stages. The larvae of late-stage OV crabs will soon be released. For egg stage III it takes a maximum of 9 days and for stage IV, 4 days until hatching (Table I). If the mating female is ripe, she oviposits new eggs soon after releasing the larvae (Table 3). Thus, males that accept females with late egg stages minimize the amount of time they spend in mate guarding, allowing more time for feeding and courting. Why do males infrequently guard NO females that may have ripe ovaries and spawn soon? Ripe NO females would require a short guarding time which would seem to be efficient for males. There may be two reasons: first, few ripe NO females may be available due to rapid sequential spawnings; second, there may be a male biased operational sex ratio (OSR) due to moderate reproductive synchrony. As described earlier, the females can lay eggs successively within a few days after pairing and thus, the period of the NO stage is usually a few days only. Further NO females that are about to ovulate may be unavailable to courting males on the surface because most are underground, having just released larvae and are guarded by their mates. The successive spawning we observed in U. tetrugonon contrasts with the case for some other fiddler crabs. In U. pugilutor (Christy, 1978) and U. lacteu (Murai et al., 1987) OV females remain in burrows while carrying eggs for about two weeks, and they are active on the surface only after releasing zoeas. They may take another several weeks to mature their ovaries for the next spawning. Hence, in these species only NO females are behaviorally receptive and are attracted and guarded by males. The moderate reproductive synchrony we observed in U. tetrugonon may lead to a male biased operational sex ratio (OSR). Many intertidal ocypodid crabs spawn synchronously and release larvae on days with large amplitude nocturnal tides (DeCoursey, 1983; Forward, 1987; Morgan and Christy, 1995); for example, I/. rosem spawn on only two or three days at every lunar spring tide period (Goshima and Kawai. unpubl. data). In such synchronously spawning species, the OSR may not be male biased; the ratio of ripe females to males is relatively large just before spawning and males have a relatively high probability of encountering ripe females, even if they are sequential breeders. This is not the case in U. tetrugonon in which a few spawners are always present in the habitat (Fig. 4). Hence, the OSR probably varies daily with variation in the abundance of receptive females but is always biased towards males. The OSR for underground mating may be further skewed by frequent surface mating and solitary breeding by females. In U. trtrugonon males have a relatively low daily probability of encountering ripe females compared to strongly synchronous species, and they might be expected to guard females for longer periods (Yamamura, 1986). In fact,
S. Goshima
et al. I J. Exp. Mar.
Biol. Ecol.
196 (1996)
131-143
141
the guarding time in U. tetragonon (5.2 days) is longer than that of U. pugilator (1.8 days, Christy, 1978) and U. lactea (2.0 days, Goshima and Murai, 1988) which are relatively synchronous spawners. The longer guarding time leads to inevitable pairing with females in the late OV stage in the sequential breeders like the present species but it still requires a shorter mate-guarding time than that for early stage OV females. Feeding on the surface by guarding male U. tetragonon, which is not common for other fiddler crabs, may compensate for the cost of the longer guarding time. These two reasons probably explain the low frequency of guarding NO females in U. tetrugonon. We suggest that the best tactic for males to enhance their fitness is to guard a mate for the shortest time that ensures paternity. In other fiddlers (Christy, 1982; Murai et al., 1987; Goshima and Murai, 1988) and in U. tetragonon in the early reproductive season when ripe NO females are available, this tactic may typically result in mating and guarding ripe NO females. As the season progresses, however, the rate at which male U. tetrugonon encounter ripe NO females decreases. Most males mate and guard late-stage OV females even though they guard these females longer than they would guard now rare NO females. The male mating tactic which we propose requires a mechanism whereby males assess how long it will take until a female that enters their burrow ovulates a new clutch. Our study indicates that such an assessment is made either when a female first enters a male’s burrow, or during the early stages of underground pairing (Table 2 and Table 3). The cues that males use to make this assessment are unknown. Male mate assessment leading to pairing and guarding OV females may enhance male fitness under conditions of a high sperm displacement rate and a male biased OSR, as predicted by the model analysis of Yamamura and Tsuji (1989). Female mate choice may also affect pair formation and should be clarified in further studies.
Acknowledgments We are grateful to K. Charoenpanich, U. Bhatia and S. Poovachiranon, Phuket Marine Biological Center, and S. Limsakil and C. Aryuthaka, Department of Fisheries, Thailand for making facilities available. Thanks are also given to the National Research Council of Thailand for the opportunity to research this study. We thank J.H. Christy for his critical reading and invaluable suggestions to improve the manuscript, and also thank S. Wada and N. Yamamura for comments that improved this paper. T. Kosuge and T. Komai kindly helped us with the field observations. This study was supported by a Grant-in-Aid for International Scientific Research from the Japanese Ministry of Education, Science and Culture (Nos. 62042019 and 01041069).
References Arak, A., 1983. Male-male competition and mate choice in anuran amphibians. In, Mate choice, edited by P. Bateson, Cambridge University Press, Cambridge, pp. 18 I-2 IO. Birkhead, T.R. and A.P. Mgller, 1992. Sperm competition in birds: Evolutionary causes md consequences. Academic Press, London, 282 pp.
142
S. Goshima et cd. I .I. Exp. Mar. Biol. Ecol. 196 (1996) 131-14.3
Brown, W.D., 1990. Size-assortative mating in the blister beetle Lyttu magister (Coleoptera: Meloidea) is due to male and female preference for larger mates. Anbn. Behuv., Vol. 40, pp. 901-909. Christy, J.H., 1978. Adaptive significance of reproductive cycles in the fiddler crab Uca pugilafor: a hypothesis. Science, Vol. 199, pp. 453-455. Christy, J.H., 1982. Burrow structure and use in the sand fiddler crab, Uca [xqilntor (Box). Anim. Behav.,Vol. 30, pp. 6877694. Christy, J.H., 1983. Female choice in the resource-defense mating system of the sand fiddler crab, Ilcn pugilator. Behav. Em/. Sociohiol., Vol. 12, pp. 169- 180. Christy, J.H., 1987. Female choice and the breeding behavior of the fiddler crab. (/(,a herhei. .J. Cnut. Rio/., Vol. I, pp. 6244635. Crespi, B.J., 1989. Causes of assortative mating in arthropods. Anim. Behubs., Vol. 38, pp. 980-1000. DeCoursey, P.J., 1983. Biological timing. In, The biology of Crustucw, edited by F.J. Vernberg and W.B. Vemberg, Academic Press, London, pp. 107-162. Diesel, R., 1988. Male-female association in the spider crab Inuchus phalangium: the influence of female reproductive stage and size. J. Crust. Biol., Vol. 8, pp. 63-69. Diesel, R., 1990. Sperm competition and reproductive success in a decapod Irzac~hu.~phulmgiurn (Majidae): A male ghost spider crab that seals off rival’s sperm. J. Zoo/. (Land.), Vol. 220, pp. 213-224. Diesel, R., 1991. Sperm competition and the evolution of mating behavior in Brachyura, with special reference to spider crabs (Decapoda, Majidae). In, Cncsrucrc~n sexual biology. edited by R.T. Bauer and J.W. Martin, Columbia University Press, New York. pp. 14S- 163. Forward, R.B., Jr.. 1987. Larval release rhythms of decapod crustaceans: an overview. Bull. Mar. Sci., Vol. 41, pp. 165-176. Frith, D.W. and S. Brunenmeister, 1980. Ecological and population studies of fiddler crabs (Ocypodidae, genus Uca) on a mangrove shore at Phuket Island, western Peninsular Thailand. Crustacrunu,Vol. 39. pp. 157- 184. Goshima, S. and M. Murai, 1988. Mating investment of male tiddler crab\, I/ccr Iwteu. Anim. Behav., Vol. 36, pp. 1249-1251. Greenspan. B.N., 1980. Male size and reproductive success in the communal courtship system of the fiddler crab Uca rqxx. Anim. Brhm., Vol. 28, pp. 387-392. Henmi, Y., 1989. Reproductive ecology of three ocypodid crabs. II. Incubation sites and egg mortality. &o/. Res., Vol. 4, pp. 261-269. Koga. Y., Y. Henmi and M. Murai, 1993. Sperm competition and the assurance of underground copulation in the sand-bubbler crab Su~pirnrru ,y/oho.m (Brachyura: Ocypodidae). J. Crust. Biol., Vol. 13, pp. I34- 137. Macintosh, D.J., 1979. The ecology and energetics of mangrove fiddler crabs (Ucu spp.) on the west coast of the Malay Peninsula. Doctoral Thesis, University of Malaya, Kuala Lumpur, 371 pp. Morgan, S.G. and J.H. Christy, 1995. Adaptive significance of the timing of larval release by crabs. Am. Nu., Vol. 145, pp. 457-479. Murai, M., S. Goshima and Y. Henmi, 1987. Analysis of the mating system of the tiddler crab, Uccl /ac,rrci. Anirn. Behas., Vol. 35, pp. 1334- 1342. Nakasone, Y., H. Akamine and K. Asato, 1983. Ecology of the tiddler crab Uca i’4cdns ~‘OC~IIS(Linneaeus) (Decapoda: Ocypodidae). II. Relation between the mating system and the drove. Galu,rxru, Vol. 2, pp. 119-133. Ridley, M., 1983. The e.rplanutiorz of orguoic diversity: The cnmparcrtive method crnd adaptutionr ,for mcmncq. Clarendon Press. Oxford, 272 pp. Salmon. M., 1984. The courtship, aggression and mating system of a “primitive” fiddler crab (Uctr ~‘occl,i.\: Ocypodidae). Tmns. Zoo/. SK. Lmd., Vol. 37, pp. I -SO. Salmon, M., 1987. On the reproductive behavior of the tiddler crab Uc,rr thtryeri, with comparisons to U. pu,qi/utor and U. L’O~OI~S: evidence for behavioral convergence. J. Crust. Biol., Vol. 7. pp. 25-44 Smith, R.L. (editor), 1984. S/rerm competitiotl und the rvolution of mimul mtrting swtum. Academic Press. New York, 687 pp. Sokal, R.T. and F.J. Rohlf, 1981. Biomrt~. W.H. Freeman, San Francisco, second edition, 859 pp. Yamaguchi, T. 1971. Courtship behavior of a fiddler crab. Ucu luctecz. Kurnnmoro J. Sci. Biol.. Vol. IO, pp. 13-37.
S. Goshima
Yamamura, Theor.
et al. I .I. Exp. Mar. Biol. Ecol.
N., 1986. An evolutionarily
stable strategy
(ES)
196 (1996)
131-143
model of postcopulatot-y
143
guarding
in insects.
Popul. Biol., Vol. 29, pp. 48-55.
Yamamura, N. and N. Tsuji, 1989. Postcopulatory strategy in a finite mating period. Theor. Popul. Biol., Vol. 35, pp. 36-50. Zucker, N., 1983. Courtship variation in the neo-tropical fiddler crab Uca deichmnni: another example of female incitation to male competition? Mar. Behav. Physiol., Vol. 10, pp. 57-79.
