Changes in offspring vulnerability account for the increase in convict cichlid defensive behaviour with brood age: evidence for the nest crypsis hypothesis

Anim. Behav., 1995, 49, 1177–1184

Changes in offspring vulnerability account for the increase in convict cichlid defensive behaviour with brood age: evidence for the nest crypsis hypothesis ROBERT J. LAVERY* Department of Biology, Queen’s University, Kingston, Ontario K7L 3N6, Canada (Received 7 October 1993; initial acceptance 3 February 1994; final acceptance 15 April 1994; MS. number: 6820)

Abstract. Several hypotheses have been proposed to explain the increase in parental defensive behaviour with brood age: parent re-nesting potential, age-investment, predator revisitation and nest crypsis hypotheses. However, these hypotheses are difficult to separate experimentally because many variables covary with brood age and all hypotheses predict an increase in effort with brood age. In this study, an attempt was made to determine which of these variables convict cichlid, Cichlasoma nigrofasciatum, parents assess in determining their defence levels at each developmental stage: egg (embryo), wriggler (free-embryo) and fry (larvae). The type of offspring parents had at each stage was manipulated. The experimental group was repeatedly given eggs until the offspring in the control group reached the fry stage, at which time the experimental eggs were allowed to develop. At each stage, parental behaviour was observed to determine whether the groups differed in past investment. Subsequently, a model predator was presented to determine whether the groups differed in defensive behaviour. The results suggest that the increase in parental defence in this species is due to the increased vulnerability of their offspring at the fry stage (nest crypsis hypothesis). Numerous studies have shown that parents adjust care to brood age; most have found that older offspring are defended more vigorously by parents than younger offspring (for a review see Sargent & Gross 1986; Montgomerie & Weatherhead 1988). However, it is unclear how parents assess the change in brood value (Clutton-Brock 1991). Several hypotheses have been proposed to explain the results: predator revisitation (Knight & Temple 1986a, b, c), parent re-nesting potential (Barash 1975), nest crypsis (Harvey & Greenwood 1978) and age-investment (Andersson et al. 1980) hypotheses. Lavery & Colgan (1991) have shown that the number of model presentations has little effect on the defensive behaviour of parental convict cichlids; furthermore, the revisitation hypothesis has been contested on theoretical grounds (Coleman 1987; Montgomerie & Weatherhead 1988). The other three hypotheses have not been adequately tested, partly because they are difficult to separate experimentally in the field. If older offspring have a higher probability of survival (expected benefits), parents should increase *Present address: Department of Integrative Biology, University of California, Berkeley, CA 94720, U.S.A. 0003–3472/95/051177+08 $08.00/0

?

defence levels as the brood ages. But covarying with brood age is a parent’s re-nesting potential, which is a function of the time remaining in a breeding season and the amount invested in the brood (Barash 1975). Because older offspring have received more past investment and consequently require less future investment, they are relatively more valuable to the parent than younger offspring. In species that initially conceal their offspring, offspring often become more conspicuous to predators as they age (Armstrong 1956; Harvey & Greenwood 1978). This is true of biparental cichlids, where the embryos (eggs and wrigglers) are concealed in a cave. Convict cichlids prefer spawning in dark cavities with one entrance (Lavery 1991). After the young become freeswimming during the larval period, parents and young exit the nest and search for food (Keenleyside 1991). At this time, they are more vulnerable to predators (Barlow 1974, 1976). The observed increase in parental defence with brood age (Lavery & Colgan 1991) may be a response to the increased vulnerability of their offspring to predation (nest crypsis hypothesis); in that study, the increase in parental defence occurred between the wriggler and fry stages. If offspring are 1995 The Association for the Study of Animal Behaviour

1177

Animal Behaviour, 49, 5

1178 Group Control

Egg

Wriggler

Fry

Experimental

Egg

Egg

Egg

Wriggler

Fry

Observation times

T1

T2

T3

T4

T5

Figure 1. Diagram illustrating the experimental design and observation times. Data were collected from control pairs while the young were eggs, wrigglers and fry. Experimental group parents were observed at five different times. An observation time (e.g. T1) indicates that the parents were videotaped for 10 min and then presented with a model predator for 2 min.

concealed in a nest, like eggs and wrigglers, the benefits of defence will be lower than when they leave the nest and are susceptible to predation (Montgomerie & Weatherhead 1988). As the offspring become independent and can escape predation, the benefits of added parental defence drop and the potential costs rise. Ridgway (1988) has shown that the defensive behaviour of smallmouth bass, Micropterus dolomieui, initially increases with brood age, then decreases as the young approach independence (see also Colgan & Gross 1977). In this study, I attempted to hold past investment constant while manipulating the age and developmental stage of the offpsring. Parents in the experimental group were repeatedly given eggs for 8 days (Fig. 1). Noakes & Barlow (1973) have shown that cichlid parents will care for foreign offspring for up to 9 months (see also Krischik & Weber 1974). Control pairs had their young switched but the young were allowed to develop normally. In effect, at each observation time, parents in the two groups had young for the same amount of time but the stage of development was altered. I recorded parental care behaviour at each stage to determine whether the groups differed in past investment. Then, I presented a model predator at each stage to determine whether parents in the two groups differed in defensive behaviour. If parents assess their own effort, I predicted the two groups would not differ in defence levels. However, if defence levels were based on offspring age, I predicted the control group would exhibit higher levels of defence than the experimental group at both the wriggler and fry stages. The nest crypsis hypothesis would be supported if the

control group exhibited higher defence levels than the experimental group only at the fry stage when the young become vulnerable to predation.

MATERIAL AND METHODS Study Animal and Aquaria The fish were five to six generations removed from Costa Rica; they had been outcrossed with fish obtained from local pet stores. The individuals were inexperienced breeders that had been kept in large holding tanks (246 and 511 litres). Fish were randomly assigned to each group. In this study, the mean (&) total length of males was 8·7&0·1 cm, while females averaged 7·3&0·1 cm (N=24). There were no differences in fish size between treatments for both males (t-test: t=1·12, df=22, P=0·275) and females (t=1·53, df=22, P=0·139). I added females to the tanks 4 days before males to facilitate pair formation. All pairs were kept in 59-litre breeding tanks (61#31#31 cm). Each tank contained gravel, a clay flower-pot for spawning, and an airstone. A room heater kept the water temperature at 24)C; lighting was provided by overhead fluorescent tubes on a 12:12 h light:dark cycle. Each fish was fed a pellet of food a day (Tropic Aquaria, stock no. A149). Experimental Groups I observed 12 pairs of cichlids in each of the two groups: experimental and control. The experimental group was given a new batch of eggs every 4 days until the young in the control group reached

Lavery: Parental defence and offspring vulnerability the larval period (fry). Control pairs were also given foreign offspring, but they were allowed to develop through the three different behavioural stages: egg, wriggler and fry (Fig. 1). Because the idea of extending the observations of experimental pairs beyond the egg stage occurred after four pairs were tested, only eight experimental pairs were observed at the wriggler and fry stages. Procedure Two days after spawning, I videotaped pairs in both groups for 10 min. Immediately after the first observation period, I introduced a model predator for 2 min (observation time: T1). The predator was a preserved Gobiomorus maculatus (total length=13 cm). It was suspended from the top of the breeding tank by a thin wire, which was attached to a piece of wood that rested on the aquarium trim. The model was presented 10 cm from the brood. One day after the first observation period, I replaced eggs of all groups with foreign eggs. However, the experimental group was given newly spawned eggs while the control group was given eggs at the same stage of development. A day later, eggs in the control group hatched while the experimental group continued to care for eggs. Two days later, I videotaped all pairs for 10 min and conducted a 2-min defence test (T2). The next day, I removed wrigglers from the control group and replaced them with foreign wrigglers. Experimental group pairs received newly spawned eggs. In approximately a day, young under the care of control pairs began to free-swim (larvae). Two days after young in the control group became free-swimming fry, I observed pairs in both groups and conducted a 2-min defence test (T3). I removed pairs in the control group after this test. The next day eggs in the experimental group hatched. Two days later, I observed experimental group parents with wrigglers for 10 min and conducted another 2-min defence test (T4). I also videotaped the experimental group and conducted a defence test 2 days after the young developed into freeswimming fry (larvae; T5). Behaviour Patterns From the 10-min observation periods on videotape, I scored the following parental behaviour

1179

patterns: time spent away from the brood (two parental body lengths), time spent fanning, and the frequency of digging, foraging and mouthing embryos. Mouthing is an embryo cleaning behaviour. The purpose of these observation periods was to determine whether the groups differed in past investment. From the 2-min defence test tapes, I scored the time spent away from the brood and predator (greater than two body lengths), the number of bites, frontal displays and headshakes. I also recorded the latency to exhibit an aggressive act. Because bites, frontal displays and headshakes exhibited the same trends, I combined these scores into one measure: total defensive behaviour patterns. The purpose of the defence test was to determine whether the groups differed in their commitment to the brood at each stage. Statistical Analyses and Specific Predictions I used ANOVAs with group (experimental versus control group) as an independent factor and stage as a repeated measure to compare the experimental and control groups. When significant differences were found, I used t-tests to determine where the differences occurred. The data from males and females were analysed separately because it was unclear whether the sexes behaved independently. If needed, the data were ln(x+1) transformed (Zar 1984). I compared the three test periods of the control group with the first three test periods of the experimental group (T1, T2 and T3; Fig. 1). If parents can assess their own past investment, the defensive behaviour of experimental and control groups should not differ during the first three tests. If the control group exhibits higher defensive behaviour at the wriggler and fry stage, the results would suggest that parents assess the age of their young. I then compared the latter two test periods (T4: wriggler and T5: fry) of the experimental group to the same offspring stages of the control group (T2: wriggler and T3: fry). The age-investment hypothesis would also be supported if the defensive behaviour of the prolonged egg-stage treatment during the wriggler and fry stages was similar to the behaviour of control pairs during the same periods. The nest crypsis hypothesis would be supported if defence levels only increased in both groups while parents had free-swimming fry.

Animal Behaviour, 49, 5

1180

Table I. Comparison of the behaviour of control and experimental females (X&) Control

Experimental

P1

P2

Dig

T1 T2 T3 T4 T5

0·75&0·66 0·67&0·50 1·82&1·23

0·50&0·29 0·58&0·34 0·25&0·25 0·38&0·26 5·13&3·30

0·36

0·47

Forage

T1 T2 T3 T4 T5

0·25&0·18 0·58&0·36 5·00&3·77

1·17&0·82 0·42&0·42 0·58&0·19 1·13&0·55 1·75&0·62

0·64

0·68

Time away (%)

T1 T2 T3 T4 T5

21·60&5·83 34·43&7·01 14·38&2·68

26·21&6·05 21·11&5·49 34·97&8·29 32·21&7·71 15·48&3·32

0·85

0·79

Fan (%)

T1 T2 T3 T4 T5

61·28&5·56 21·56&6·13

48·29&5·60 38·18&4·74 30·18&6·70 10·48&5·11

—

—

P1: P-value for experimental and control group comparison: T1, T2 and T3. P2: P-value for group comparison: control T2, T3 and experimental T4, T5.

RESULTS Observation Periods There were no statistically significant differences between the experimental and control groups in the frequency of digging (females: F1,21 =0·87, P=0·362; males: F1,21 =0·40, P=0·531) and foraging (females: F1,21 =0·22, P=0·640, Table I; males: F1,21 =1·35, P=0·259, Table II). However, experimental males spent significantly more time away from the brood than control males during the latter two egg stages (T2: t=2·38, df=22, P=0·026; T3: t=7·54, df=22, P<0·001; Table II); as is also indicated by a significant interaction between stage and group (F2,44 =9·38, P<0·001). After the young hatched (T4, T5), the time experimental males spent away from the brood did not statistically differ from the levels observed in control males at the wriggler (T2) and fry (T3) stages (F1,18 =1·22, P=0·285). There were no statistical differences between experimental and control group females in time spent away from the brood (Table I). However, experimental females continued to fan their brood until they became free-swimming (Table I). Given

that fanning is energetically costly, experimental females may have expended more effort than control females. Defence Test Males and females adjusted defence levels to the developmental stage of their offspring. A significant interaction between group and stage for both males (F2,44 =8·22, P=0·001) and females (F2,44 =8·15, P=0·001) revealed that defensive behaviour only increased in the control group at T3 while the young were free-swimming fry (males: t=2·24, df=22, P=0·036; females: t=3·24, df=22, P=0·004; Fig. 2). Defensive behaviour of experimental females only increased when their young became free-swimming (T5) and it did not differ statistically from the levels observed in control females at T2 (wriggler) and T3 (fry; F1,18 =1·01, P=0·327; Fig. 2b). However, experimental males exhibited higher levels of defensive behaviour at the wriggler (T4: t=2·82, df=18, P=0·011) and fry (T5: t=3·80, df=18, P=0·003) stages than control males at T2 and T3 (F1,18 =12·17, P=0·003), suggesting that males

Lavery: Parental defence and offspring vulnerability

1181

Table II. Comparison of the behaviour of control and experimental males (X&) Control

Experimental

P1

P2

Dig

T1 T2 T3 T4 T5

1·75&0·96 1·25&1·07 0·18&0·12

1·00&0·69 1·25&0·66 2·42&1·23 1·50&0·96 4·13&3·98

0·53

0·23

Forage

T1 T2 T3 T4 T5

0·25&0·18 0·42&0·26 3·55&2·87

0·33&0·33 0·50&0·42 0·25&0·13 1·13&0·67 0·38&0·26

0·26

0·70

Time away (%)

T1 T2 T3 T4 T5

89·51&6·78 65·82&9·69 33·32&5·10

90·72&4·71 96·78&1·88 94·38&2·32 74·29&5·77 17·25&4·34

0·001

0·29

Fan (%)

T1 T2 T3 T4 T5

0·63&0·63 4·71&2·19

2·24&1·22 0·06&0·06 0·33&0·33 1·15&0·70

—

—

P1: P-value for experimental and control group comparison: T1, T2 and T3. P2: P-value for group comparison: control T2, T3 and experimental T4, T5.

may be responding to more than the state of their offspring (possibly past investment). Males in the control group spent more time with their brood and attacking the model during the fry stage (T3) than did experimental males during the same period (group*stage: males: F2,44 =7·60, P=0·001; T3 comparison: t=4·36, df=22, P<0·001; Fig. 3a). Control females spent significantly more time away from the brood and predator at T1 (t=3·25, df=22, P=0·004) and less time away at T3 (t=1·61, df=22, P=0·121; F2,44 =5·20, P=0·009; Fig. 3b) than experimental females. The T1 results are due to experimental females hiding inside the flower-pot near their eggs. Control males and females also took less time to display at the model during the third defence test (T3) than did experimental pairs (group*stage: males: F2,44 =7·28, P=0·002; T3 comparison: t=2·72, df=22, P=0·013, Fig. 4a; females: F2,44 =6·21, P=0·004; T3 comparison: t=2·50, df=22, P=0·020, Fig. 4b). Experimental pairs with wrigglers (T4) and fry (T5) did not differ statistically from control pairs with young at the same stage of development (T1 and T2; males, time away: F1,18 =0·06, P=0·813, Fig. 3a; latency to display: F1,18 =2·98, P=0·102, Fig.

4a; females, time away: F1,18 =0·36, P=0·554; Fig. 3b; latency to display: F1,18 =0·46, P=0·507, Fig. 4b).

DISCUSSION The results suggest that the defensive behaviour of parental convict cichlids is largely dependent on the stage of development of their young. Despite changes in brood age, males and females in both groups only increased their defensive behaviour as the young developed into free-swimming fry. It is at this time the young exit the nest to feed and are vulnerable to predation (Keenleyside et al. 1990; Rangeley & Godin 1992). Because of high predation rates on fry, cichlid parents may be selected to increase their effort at this time (Rogers 1988). In Lake Jiloa, Nicaragua, McKaye (1984) observed that parental attacks directed at predators were highest during the fry stage for both Cichlasoma rostratum and Neetroplus nematopus (see also Neil 1984: Cichlasoma meeki). The benefits of increased defence at the fry stage, when the young are vulnerable to predation, are greater than when the young are

Animal Behaviour, 49, 5

1182

100

30

(a)

(a) 25

80

20

60 40 10

20

5 0

T1

T2

T3

T4

T5

30 (b)

% Time away

Defensive behaviour patterns

15

0

T1

T2

T3

T4

T5 (b)

50

25

40

20

30

15 10

20

5

10

0

T1

T2

T3 Time

T4

T5

0

T1

T2

T3 Time

T4

T5

Figure 2. Mean (&) defensive behaviour of (a) males and (b) females. .: Control group; /: experimental group. See Fig. 1 for a guide to the observation times.

Figure 3. Mean (&) percentage of time spent away from brood and predator by (a) males and (b) females. .: Control group; /: experimental group. See Fig. 1 for a guide to the observation times.

concealed in a cave (nest crypsis hypothesis; Harvey & Greenwood 1978; Montgomerie & Weatherhead 1988). As the young approach independence (juvenile period) and are capable of escaping predators, parents may decrease their effort with brood age; such a pattern has been observed in other species (see Colgan & Gross 1977; Huntingford 1977; Sargent & Gebler 1980; Neil 1984; Ridgway 1988). Unfortunately, I was unable to determine whether defence levels decrease as the young become juveniles. Cichlid parents care for their young for many weeks, and in my laboratory, they would quickly run out of tank space. Within 20 days of spawning, female defensive behaviour does not change within the fry stage, however, males eventually decrease their effort with offspring age (Lavery & Keenleyside 1990). Although these results indicate that the observed increases in defensive behaviour with

time are in response to the changes in offspring vulnerability, parental fish have been shown to base investment decisions on past investment (Lepomis macrochirus: Coleman et al. 1985; C. nigrofasciatum: Lavery & Keenleyside 1990). Also, experimental group males may have been responding to their own effort; they exhibited higher defence levels at the wriggler and fry stages than control males did while they had young at these stages. Investment in time may be important for males, as continued investment in the current brood limits future mating opportunities and reproductive success (Keenleyside et al. 1990; Keenleyside & Mackereth 1992). It is unclear whether parents do assess the age of the brood per se (Andersson et al. 1980; Thornhill 1989). Most studies to date that have addressed this question have at least one other confounding variable. For instance, Reid & Montgomerie (1985) have shown that

Lavery: Parental defence and offspring vulnerability

1183

ACKNOWLEDGMENTS

120 (a) 100

I thank Patrick Colgan, Raleigh Robertson and Joe Waas for discussion during the course of this study. I thank Barrie Frost, Laurene Ratcliffe, Raleigh Robertson, Mark Ridgway and Bruce Tufts for comments on an early version of the manuscript. This study was funded by an N.S.E.R.C. grant to Patrick W. Colgan.

80 60

Latency to display (s)

40 20 0

REFERENCES T1

T2

T3

T4

T5

120 (b) 100 80 60 40 20 0

T1

T2

T3 Time

T4

T5

Figure 4. Mean (&) latency to display at the predator by (a) males and (b) females. .: Control group; /: experimental group. See Fig. 1 for a guide to the observation times.

nest-defence intensity was positively correlated with brood age in Baird’s sandpipers, Calidris bairdii. Because the season is too short for more than one breeding attempt, the authors discounted the re-nesting potential hypothesis. However, as offspring age increases, past investment in the current brood increases. Given that current reproduction devalues future survival and reproduction (Sargent & Gross 1986), increases in both brood age and past investment lead to the same predictions. Given that offspring development in cichlids is dependent on temperature and nutrient reserves (Kamler 1992), I contend that basing investment decisions on stage of development and consequently offspring vulnerability may be the most reliable strategy to ensure high reproductive success in a given season. Thereby, parental defence is highest when the offspring need the most protection.

Andersson, M., Wiklund, C. G. & Rundgren, H. 1980. Parental defence of offspring: a model and an example. Anim. Behav., 28, 536–542. Armstrong, E. A. 1956. Distraction display and the human predator. Ibis, 98, 641–654. Barash, D. P. 1975. Evolutionary aspects of parental behavior: distraction behavior of the alpine accentor. Wilson Bull., 87, 367–373. Barlow, G. W. 1974. Contrasts in social behavior between Central American cichlid fishes and coralreef surgeon fishes. Am. Zool., 14, 9–34. Barlow, G. W. 1976. The Midas cichlid in Nicaragua. In: Investigations of the Ichthyofauna of Nicaraguan Lakes (Ed. by T. B. Thorson), pp. 333–358. Lincoln, Nebraska: University of Nebraska Press. Clutton-Brock, T. H. 1991. The Evolution of Parental Care. Princeton, New Jersey: Princeton University Press. Coleman, R. M. 1987. Does experience with predators affect parental investment? Auk, 104, 791–792. Coleman, R. M., Gross, M. R. & Sargent, R. C. 1985. Parental investment decision rules: a test in the bluegill sunfish. Behav. Ecol. Sociobiol., 18, 59–66. Colgan, P. W. & Gross, M. R. 1977. Dynamics of aggression in male pumpkinseed sunfish (Lepomis gibbosus) over the reproductive phase. Z. Tierpsychol., 43, 139–151. Harvey, P. H. & Greenwood, P. J. 1978. Anti-predator defence strategies: some evolutionary problems. In: Behavioural Ecology: An Evolutionary Approach (Ed. by J. R. Krebs & N. B. Davies), pp. 129–154. Oxford: Blackwell Scientific Publications. Huntingford, F. A. 1977. Inter- and intraspecific aggression in male sticklebacks. Copeia, 1977, 158–159. Kamler, E. 1992. Early Life History of Fish. London: Chapman & Hall. Keenleyside, M. H. A. 1991. Parental care. In: Cichlid Fishes: Behaviour, Ecology and Evolution (Ed. by M. H. A. Keenleyside), pp. 191–208. London: Chapman & Hall. Keenleyside, M. H. A., Baily, R. C. & Young, V. H. 1990. Variation in the mating system and associated parental behaviour of captive and free-living Cichlasoma nigrofasciatum (Pisces, Cichlidae). Behaviour, 112, 202–221. Keenleyside, M. H. A. & Mackereth, R. W. 1992. Effects of loss of male parent on brood survival in a biparental cichlid fish. Environ. Biol. Fish., 34, 207–212.

1184

Animal Behaviour, 49, 5

Knight, R. L. & Temple, S. A. 1986a. Why does intensity of avian nest defense increase during the nesting cycle? Auk, 103, 318–327. Knight, R. L. & Temple, S. A. 1986b. Methodological problems in studies of avian nest defence. Anim. Behav., 34, 561–566. Knight, R. L. & Temple, S. A. 1986c. Nest defence in the American goldfinch. Anim. Behav., 34, 887–897. Krischik, V. A. & Weber, P. G. 1974. Induced parental care in male convict cichlid fish. Devl Psychobiol., 8, 1–11. Lavery, R. J. 1991. Physical factors determining spawning site selection in a Central American hole nester, Cichlasoma nigrofasciatum. Environ. Biol. Fish., 31, 203–206. Lavery, R. J. & Colgan, P. W. 1991. Brood age and parental defence in the convict cichlid, Cichlasoma nigrofasciatum (Pisces: Cichlidae). Anim. Behav., 41, 945–951. Lavery, R. J. & Keenleyside, M. H. A. 1990. Parental investment of a biparental cichlid fish, Cichlasoma nigrofasciatum, in relation to brood size and past investment. Anim. Behav., 40, 1128–1137. McKaye, K. R. 1984. Behavioural aspects of cichlid reproductive strategies: patterns of territoriality and brood defence in Central American substratum spawners versus African mouth brooders. In: Fish Reproduction: Strategies and Tactics (Ed. by R. J. Wootton & G. W. Potts), pp. 245–273. New York: Academic Press. Montgomerie, R. D. & Weatherhead, P. J. 1988. Risks and rewards of nest defence by parent birds. Q. Rev. Biol., 63, 167–187. Neil, S. J. 1984. Field studies of the behavioral ecology and agonistic behavior of Cichlasoma meeki (Pisces: Cichlidae). Environ. Biol. Fish., 10, 59–68.

Noakes, D. L. G. & Barlow, G. W. 1973. Crossfostering and parent–offspring responses in Cichlasoma citrinellum (Pisces, Cichlidae). Z. Tierpsychol., 33, 147–152. Rangeley, R. W. & Godin, J. J. 1992. The effects of a trade-off between foraging and brood defence on parental behaviour in the convict cichlid fish, Cichlasoma nigrofasciatum. Behaviour, 120, 123–138. Reid, M. L. & Montgomerie, R. D. 1985. Seasonal patterns of nest defence by Baird’s sandpipers. Can. J. Zool., 63, 2207–2211. Ridgway, M. S. 1988. Developmental stage of offspring and brood defence in smallmouth bass (Micropterus dolomieui). Can. J. Zool., 66, 1722–1728. Rogers, W. 1988. Parental investment and division of labor in the Midas cichlid (Cichlasoma citrinellum). Ethology, 79, 126–142. Sargent, R. C. & Gebler, J. B. 1980. Effects of nest site concealment on hatching success, reproductive success, and paternal behavior of the threespine stickleback, Gasterosteus aculeatus. Behav. Ecol. Sociobiol., 7, 137–142. Sargent, R. C. & Gross, M. R. 1986. William’s principle: an explanation of parental care in teleost fishes. In: The Behaviour of Teleost Fishes (Ed. T. J. Pitcher), pp. 275–293. Baltimore, Maryland: Johns Hopkins University Press. Thornhill, R. 1989. Nest defense by red jungle fowl (Gallus gallus spadiceus) hens: the roles of renesting potential, parental experience and brood reproductive value. Ethology, 83, 31–42. Zar, J. H. 1984. Biostatistical Analysis. Englewood Cliffs, New Jersey: Prentice-Hall.