Can Paradise Fish (Macropodus opercularis, Anabantidae) Recognize a Natural Predator? An Ethological Analysis

Ethology 94, 127-136 (1993) 0 1993 Paul Parey Scientific Publishers, Berlin and Hamburg ISSN 0179-1613

Department of Ethology, L. Eotvos University, God

Can Paradise Fish (Macropodus opercularis, Anabantidae) Recognize a Natural Predator? An Ethological Analysis ROBERT GERLAI GERLAI,R. 1993: Can paradise fish (Macropodus opercularis, Anabantidae) recognize a natural predator? An ethological analysis. Ethology 94, 127-136.

Abstract Predator recognition .md avoidance by paradise fish have been studied with allopatric species and model experiments. The effect of sympatric predators has not been investigated. Here I report that reactions of paradise fish towards a sympatric predator (Channa rnznopelres) are quantitatively different from those shown towards an allopatric predator o r different harmless species of fishes. I investigate the possible cues eliciting this differential response and show that visual as well as olfactory stimuli play roles. Olfactory stimuli from the sympatric predator alone elicit an elevated level of activity from paradise fish; the appearance of the sympatric predator (with o r without olfactory stimuli) results in an exploratory and display reaction. I speculate what visual stimuli may play roles in predator recognition in paradise fish and I suggest that previously asserted key stimuli such as the eyes of the encountered heterospecific fish may not differentiate the harmful species from innocuous. I conclude that the antipredatory behavior of paradise fish may be affected by both genetic factors and learning and that the relatibe importance of the former o r latter factor may vary depending on the situation. Mount Sinai Hospital, Research Institute, Division of Molecular Immunology Robert GERLAI, and Neurobiology, 600 University Ave., Toronto, Ontario, Canada M5G 1x5.

Introduction Paradise fish (Macropodus operculuvis, small, 10-cm long insectivore) live in Southeast Asia in densely vegetated, shallow swamps and rice fields. The species shares its habitats with several predatory and harmless fishes (NICHOLS 1943). The advantage to a prey such as paradise fish of recognizing predator species is twofold. First, recognition facilitates avoiding predators by the prey before an attack occurs. Second, recognition means that the prey d o not engage in time- and energy-consuming avoidance behaviors with harmless species. CSANYI (1985) showed that paradise fish are capable of learning to habituate to or to avoid other fish species. Naive paradise fish exhibited approach behavior U.S. Copyright Clearance Center Code Statement:

0179-1613/93/9402-0127$02.50/0

128

ROBERTGERLAI

when shown other fish such as goldfish (Curussius uurutus) or satiated pike (€sox Iucius). This initial approach behavior habituated after subsequent presentation of the stimulus fish. CSANYI demonstrated, however, that if a hungry pike or a goldfish plus a mild electric shock were presented, paradise fish showed a rapid avoidance learning. The stimulus fish species used by CSANYI were allopatric to paradise fish. Therefore his study could not address the question of whether paradise fish are capable of recognizing a natural predator without prior experience with it. Several prey species have been shown to be able to recognize predators without previous exposure to them. For example, OWINGS & Coss (1977) found that neither recognition nor mobbing behavior of ground squirrels depended upon prior experience with predatory snakes. Similarly, prairie dogs will respond appropriately to a predator upon their first exposure to it (OWINGS & OWINGS 1979). Prey species will be able to exhibit appropriate avoidance reactions at the first encounter with predators only if they are sympatric with them, an assumption confirmed for example for pied flycatchers (CURIO1975) and for subspecies of mice (HIRSCH & BOLLES1980). Genetic predisposition for avoiding a natural predator may be advantageous for the prey, since it could reduce the risk of being caught during learning. Therefore, I hypothesize that the responses of naive paradise fish to a sympatric predatory species, the red snake head fish (Channu rnicvopeltes), will be different from those shown towards either an allopatric predator o r toward species that are harmless to paradise fish. Several modalities can account for predator recognition and avoidance learning in fish (for a review see FEDER& LAUDER 1986). Visual stimuli, such as the body shape and color pattern of a predator or its motor patterns, may all contribute to its recognition by the prey. Olfactory cues may be just as important for the prey when visual detection of the predator is difficult, for example in turbid or in densely vegetated water. The role of olfaction in predator recognition and detection has been shown for example in minnows (Phoxinus phoxinus; MAGURRAN 1989) and paradise fish (MIKL~SI & CSANYI 1989). Among the available visual stimuli, eyes have been asserted to play a key role. Eyes elicit fear responses from fish (COSS1979) as well as from other species such as birds (JONES 1980), and mammals (COSS1978). Eyespots on a dummy presented together with a mild electric shock evoked a strong avoidance reaction from paradise fish, while dummies without eyes did not (CSANYI1986). Eyespot wearing dummies elicited strong exploration by paradise fish and a defensive display reaction (ALTBACKER & CSANYI 1990). The question, however, of what the exact function of the eyespots is in the recognition of predators still needs to be addressed since in the natural environment both the predators and harmless fish species carry this stimulus. The present paper will evaluate whether the reactions of paradise fish towards a sympatric predator (Channu micropeltes) are different from those shown towards an allopatric predator and allopatric and sympatric harmless species of fish. I shall investigate whether visual, acoustical or olfactory stimuli of the natural predator can evoke a differential response.

Recognition of a Natural Predator by Paradise Fish

129

Materials and Methods Animals and Housing In the experiments, 3--3.5 cm long 2-mo-old paradise fish were used that had never seen fish other than their own conspe,:ifics before the testing. They originated from an outbred population of paradise fish kept at the Department of Ethology, L. Eotvos University, Hungary. This population was founded in 1978 with mixed stock from 16 different sources from 12 countries and has been bred on a rotational basis to minimize inbreeding. The diverse origins of the founding individuals and the & CSANYI 1985) conserved genetic variation low inbreeding rate, 0,003-C.006 per generation (GERVAI and made this artificial population similar in behavior to wild-caught fish (CSANYIet al. 1985a). The experimental fish were raised in groups of 35 in 90-1 glass aquaria (78 x 38 x 30 cm length/height/ width) which contained bushes of Rotala and Hygrophila. The water was filtered, the temperature held constant at 28 "C and a I 2/12 h light/dark cycle was maintained. The fry were raised o n hatched nauplii of brine shrimp (Artemia salina) and later on tetramin flakes, dried tubifex and raw chicken liver.

Behavioral Recording The behavior of the test fish and the stimulus fish was video recorded and also directly observed from behind a curtain. Species-specific behavioral motor patterns of the test fish were immediately 1992). The video recordings recorded with an event recorder computer program (GERLAI& HOGAN were later used to analyze the behavior of the stimulus fish. The definitions of the behavioral motor et al. (1985 a, b). Here I provide a brief description patterns of paradise fish have been given by CSANY~ of only those that occurred frequently or whose definition is different from that published before: Escape (ESC), is rapid swimming perpendicular to the glass wall; Swim (SWI), is fast locomotion using the caudal fin; Move (MOV), is slow locomotion without using the caudal fin; Quiet (QUI), is a motionless state in which the fish is floating, the pectoral fin is beating normally; Leaping (LEP), is a quick jump with a forceful lateral slap of the caudal fin; Approach (APR), is a slow movement in which the fish is either moving toward or around an object and its head is oriented toward the object, and the fins are not erected. Typical ocular movements are also observable. The last motor pattern to be defined is Approach with display (ADI). Its definition is new and since it is one of the most important motor patterns in the present experiments, I give a more detailed description. During AD1 the caudal, dorsal and anal fins are erected, the displaying fish is moving very slowly forward or sometimes even backward on the periphery of a circle around an object. The fish's head is bent toward the object and typical ocular movements are observable. The body coloration is pale. Paradise fish also et al. 1985a display to conspecifics but this display reaction (DIP, described in paradise fish by CSANYI o r in Siamese fighting fish, BetLa splendens, by SIMPSON1968), is different from AD1 both in terms of its form and the context in which it appears (aggressive encounters). During AD1 paradise fish never show body and fin undulation, tail flashing o r beating, and the body coloration never darkens, which are characteristic features of thc aggressive display (DIP). Furthermore, D I P is often interrupted with quick movements and D I P ahernates with frontal display (the erection of the gill covers and the brachiostegal membrane), which never occurs during ADI.

Statistical Analysis Durations of the motor patterns relative to the length of the recording sessions were analyzed except for LEP, whose frequency was used. Groups were compared with ANOVA.Where significant F values were obtained, post hoc Student-Newman-Keuls multiple range tests (with p = 0.05) were carried out to reveal possible differences between each group. Sample sizes were kept the same for each group (balanced design, 11 = lo), which makes the parametric tests such as the ANOVAless sensitive to the violation of normality and variance homogeneity criteria (PAGANO1990). Nevertheless, logarithmic scale transformations were applied to homogenize variances where necessary.

Procedure The behavior of paradise b h was observed in two-compartment tanks (Fig. 1) of the same size as the rearing tank (length of the test fish compartment was 29 cm and length of the stimulus-fish

130

ROBERTGERLAI Fig. 1: Two-compartment test aquarium

compartment was 46 cm). The two compartments were separated by two removable guillotine doors placed adjacent to each other. One of them was transparent and the other was white. The larger stimulus compartment was either empty or contained one of the following stimulus fish (Fig. 2): sympatric predator (Chunnu minopeftes),allopatric predator (Hupfochrornzs compressiceps), sympatric harmless species (Trichoguster trichopterus), allopatric harmless species (Curussius uurutus). Only one individual of each of these species was used as a stimulus treatment throughout the tests. The test fish were randomly assigned to one of the stimulus treatment groups (see below), i.e. each test fish received only one stimulus treatment. The sample size of each stimulus treatment group was 10. One test fish of each group was measured daily in a random sequence. The test fish were put in the test compartment singly where they were allowed a 30-min habituation period before the recording began. The first recording (visual isolation test) was carried out for 5 min with the white guillotine door lowered. There were 7 stimulus groups. In four groups the stimulus fishes had been living in the stimulus compartment for one week: SP (sympatric predator), AP (allopatric predator), SH (sympatric harmless fish), A H (allopatric harmless fish). The other stimulus groups were EM (empty, the stimulus compartment never contained any fish), EM'"'

SP AP

SH Fig. 2: Stimulus fishes: (SP) Sympatric predator: Chunnu mzcropeftes, (AP) Allopatric predator: Hupfochromis compressiceps, (SH) Sympatric harmless species: Trichogaster trichopterus, (AH) Allopatric harmless species: Curusszus uurutus

AH

131

Reccgnition of a Natural Predator by Paradise Fish

(empty with ‘smell’, the stimiilus compartment contained no fish but the water of the tank was mixed with water in which the sympatric predator had lived), and SP”’ (sympatric predator with ‘no smell’, the sympatric predator was put in the stimulus compartment 1 min before the recording was started). Since the white guillotine door was lowered during the recording, n o visual cues were available to the test fish; olfactory cues, however, might have been present as water was allowed to penetrate from one compartment to the other before the test fish were put in. In the case of the SP“‘ group, no odor cues were present since the predator was put in just before the testing and the doors were kept tightly lowered (after each test the predator was removed and the tank refilled with I-d-old tap water). Neither the stimulus nor the test fish were fed on the day of testing. After the visual isolation test, the white guillotine door was removed and the test fish were allowed to view the stimulus compartment for 5 min, during which the recording was made. The transparent door was left lowered to prevent possible physical damage to the test fish. Five stimulus groups were the same as in visual isolation (SP, AP, SH, A H , EM). Test fish of the sixth stimulus group were allowed to view a satiated snake head fish (SP”, sympatric predator, satiated) that was fed and put in the stimulus compartment just before the testing began. The stimulus fish were different in several characteristics, however, the particular individuals used were chosen so that the diameter of their eyes would be 6 mm. The body length of the stimulus fishes measured from the base of the caudal fin was the following: Channa, 9.8 cm; Haplochromis, 7.8 cm; Trichogaster, 8.1 cm; Carassius, 7.5 cm. The nonpredatory fish could not cause any damage to paradise fish of the size (3---3.5 cm) used in the tests, however, the two predators could easily catch the paradise fish if the guillotine doors were removed. I recorded two behavior patterns of the stimulus fishes: the amount of time they spent motionless (no locomotion) and the number of attacks (quick, straight swimming towards the prey) performed. These measures differentiated the stimulus fishes most, therefore I discuss these parameters only.

Results The visually isolated test fish showed no Staccato, Creeping or Oblique position, motor patterns interpreted previously as indicators of fear (see GERLAI Lk CSANYI 1990). The statistical analyses revealed significant differences between groups in only one motor pattern during the visual isolation tests: the fish whose tank contained a red snake head fish that had been living in the compartment (SP group) as well as those from the EMsmgroup remained motionless (QUI) for less time than the fish of the other stimulus groups (F6,63= 5.51, p < O.OO1) (Table I). Table 1: Effects of various treatments o n the behavior of paradise fish (group i above, SE below) during visual isolation test. Symbols see text Behavior

:.

SP

AP

SH

AH

EM

EMSM

SP”’

3.la 1.74

0.6a 0.57

1.3a 1.30

O.Oa 0.0

0.5a 0.38

2.8a 1.04

0.8a 0.32

15.3a 4.68

22.9a 8.13

20.7a 7.68

14.9a 3.64

25.la 7.94

18.2a 5.09

21.la 2.25

58.6a 8.76

37.8a 6.36

35.3a 6.36

35.4a 8.03

29.3a 8.41

52.la 8.02

3l.Oa 5.13

6.9b 2.15

32.4a 5.66

30.3a 6.99

51.2a 9.14

33.0a 10.85

8.5b 2.06

41.8a 3.07

This behavior was logarithaiically transformed before statistical analyses. Same superscripts of means denote nonsignificantly different ranges (p > 0.05) by Student-Newman-Keuls multiple range comparison test. n = 13 in each group.

132

ROBERTGERLAI

The elevated activity level of the SP fish during visual isolation might have been due to the ability of the test fish to perceive either olfactory o r acoustic stimuli of the sympatric predator. Because the activity level of the SPnSfish (sympatric predator with no smell) was not elevated during the visual isolation and because the behavior of the EMsmfish was not different from that of the fish of the SP group, I infer that olfactory cues alone were sufficient to cause the behavioral change observed in the test fish. During the visual contact test the test fish could view the stimulus fish. The behavior of the stimulus fish appeared to be different from each other. The hungry predators (sympatric and allopatric) attacked the test fish often (mean attack frequency during the 10 recording sessions: 1.8, 1.6 respectively) whereas the satiated sympatric predator and the harmless fish never did. The harmless fish remained motionless for only a short time (mean relative duration: 3.2 %, Trichogaster; 1.4 % Carasszus) not like the predators (59.2 %, hungry Channa; 68.1 YO, hungry Haplochromis; 61.2 %, satiated Channa). Despite this clear difference between predators and non-predators and the hungry and satiated Channa, paradise fish did not differentiate the allopatric predator from the harmless fishes and reacted differentially only towards the Channa, the sympatric predator (see below). Variance analyses revealed significant differences among stimulus groups in several motor patterns (ESC: F5,54 = 9.11, p < 0.001; SWI: F5,54= 11.30, p < 0.001; QUI: F 5 , 5 4 = 4.17, p < 0.01; LEP: F 5 , 5 4 = 8.13, p < 0.001; APR: F5,s4 = 4.37, p < 0.01; ADI: F5,s4= 13.21, p < 0.001). The significant F values Table 2:

Effects of various treatments on the behavior of paradise fish (group 2 above, SE below) during visual contact test. Symbols see text

Behavior

SP

AP

SH

AH

EM

Sp'"

2.6a 0.60

3.0a 1.77

8.2ab 1.90

5.6a 2.03

14.3b 2.53

0.6a 0.28

6.8ab 1.96

14.8b 4.18

16.9b 3.29

16.lb 3.54

31.5~ 1.90

3.9a 1.27

16.7a 2.72

20.8a 2.88

14.4a 2.41

15.2a 4.40

23.9a 3.35

3.5a 1.01

18.lb 3.96

4.8a 1.55

13.7ab 5.41

14.8b 2.97

9.5ab 2.28

1.9a 0.35

0.2b 0.13

0.5b 0.44

O.Ob 0.0

O.Ob 0.0

0.3b 0.22

28.0a 1.08

19.8ab 4.24

25.la 3.83

14.0b 4.43

11.7b 1.47

25.6a 2.53

32.2a 2.40

16.3b 3.51

18.6b 2.69

18.4b 4.46

0.5~ 0.32

22.2a 2.88

29.8a 3.64

'f This behavior was logarithmically transformed before the statistical analyses. Same superscripts of means denote nonsignificantly different ranges (p > 0.05) by Student-Newman-Keuls multiple range comparison test. n = 10 in each group.

Recognition of a Natural Predator by Paradise Fish

133

justified carrying out post hoc multiple range comparison tests. These comparisons showed that ESC and SWI were most pronounced in the EM group and the least in the SP and SPsa groups (Table 2 ) . The reduction of ESC and SWI is associated with increased APR and AD1 (Table 2 ) . The highest AD1 values were observed when the stimulus compartment contained the sympatric predator. The increased AD1 was typical of both the SP and the S P a groups despite the fact that in the latter case the sympatric predator could not provide the test fish with olfactory cues. Therefore, I infer that visual cues alone were sufficient to elicit the differential response from the test fish. These visual cues may involve eye spots, body coloration and shape, and fine motor patterns of the predator or the combination of these. Some role of motor patterns as cues is supported by the finding that the attacks by the snake head fish elicited increased LEP in the SP group, however, the attacks by the allopatric predator ( A H group) or the nonattacking snake head fish (SP””) did not (Table 2). Furthermore, attack by the snake head fish cannot be responsible for the increased AD1 of the test fish since the satiated predator of the SPSagroup that never attacked also elicited increased ADI. This latter behavioral effect must have been due to other visual cues. No aggressive display reaction (DIP) towards the stimulus fish was observed in any groups.

Discussion Paradise fish have a tendency to explore novel stimuli. GERLAI & HOGAN (1992) showed that seeing a stimulus fish can serve as a reward for a paradise fish in an instrumental conditioning paradigm. Based on the results of his model experiments, CSANYI (1986) argued that paradise fish have a genetic predisposition to explore other fish species and, depending on the outcome of an encounter with them, paradise fish will learn to identify which of them are harmless and which are predatory. Based on quantitative genetic analysis, GERLAIet al. (1990) concluded that the exploratory behavior of paradise fish has undergone stabilizing selection in the evolutionary past, and that exploring novel stimuli is an adaptive response by paradise fish. These results suggest that paradise fish that explore novel stimuli, e.g. heterclspecific fish species, and learn from the encounters enjoy a selective advantage. The present findings, however, unequivocally demonstrate that naive paradise fish respond differentially towards a natural sympatric predator, the red snake head fish, at the very first encounter. The naive fish become agitated when exposed to the odor of the predator, which I interpret as an elevated exploratory activity. Furthermore, the visual stimuli of the sympatric predator elicit increased approach and approach with display reaction. An increased exploration of a predator by its natural prey that are able to respond distinctively to wards the predator without prior exposure to it seems paradoxical. Nevertheless, predator inspection by the natural prey has been observed in several species of fish (for examples see DUGATKIN 1988; MAGURRAN & HIGHAM 1988) and birds (e.g. CURIO 1978). PITCHER et al. (1986) showed experimentally that the probable function of predator inspection for the prey is to

134

ROBERTGERLAI

“gain information about the location and status of the intruder . . .” Learning that the natural predator in sight does not represent immediate danger because it is satiated, may save the prey unnecessary costs associated with escaping. This assumption may plausibly explain why an exploratory response to a natural predator could be observed in paradise fish. HELFMAN (1989) demonstrated a threat-sensitive predator avoidance showing that the strength of response of damselfish to its natural predator, the trumpetfish, depends on the level of threat, i.e., the relative size of the predator and its status. HELFMAN’S findings confirm that exploring a natural predator may be an adaptive response which enables the prey to respond according to the type and magnitude of the predatory threat. During the encounter with the predator paradise fish showed an increased amount of ADI. The exact function of this reaction is not clear. Nevertheless, its adaptive nature has been implied by the results of ALTBACKER & CSANYI (1990) who noted that displaying paradise fish could successfully avoid predation by such a visually hunting predator as a pike (Esox luczus). Similar display reaction has been shown in sticklebacks by HOOGLAND et al. (1956), who argued that the sharp spiky rays of the dorsal and anal fins may frighten away the predator. SEGHERS (1975) suggested that displaying prey fish look larger thus decreasing the probability of attack. Both of these functional explanations may be relevant for our findings, however, their validity in paradise fish will have to be ascertained empirically in the future. Previously it has been shown that eye-spots can evoke strong exploratory and antipredatory behavior from paradise fish (ALTBACKER & CSANYI 1990; CSANYI 1986). However, the specificity of eye-spots as cues in the recognition and avoidance of the predators seems questionable, as CSANYI (1986) pointed out, since several sympatric harmless species of fish also carry them. This study also confirmed that different heterospecific fishes could elicit distinct responses from paradise fish despite the fact that they had similar eyes. A plausible explanation for this contradiction is that eye spots may only draw the attention of the prey and direct the prey’s exploratory activity to the potentially dangerous stimulus carrier, however, they do not distinguish the harmful species from the innocuous. According to this reasoning the eyes of the heterospecific fish species elicit exploratory behavior from paradise fish and the exploration facilitates recognition of stimuli to which paradise fish are genetically sensitive. Future studies may ascertain what exactly these stimuli may be. MAGURRAN & SEGHERS (1990) demonstrated population differences in predator recognition and attack cone avoidance in guppies. Their results suggested that guppies are more vulnerable to allopatric predators than to sympatric ones presumably because of lack of genetic predisposition that could have made them more “cautious”. Though it has not been investigated whether paradise fish are more vulnerable to allopatric predators, the present and previous paradise fish studies together show that paradise fish are both able (1) to exhibit a distinct antipredatory behavior as a response to a natural predator without any prior exposure to it and (2) to avoid other heterospecific fish, e.g. allopatric predators, on the basis of learning. Such a duality in the ability of paradise fish to respond to heterospecific fish species may be adaptive since it not only facilitates an appro-

Reccmgnition of a Natural Predator by Paradise Fish

135

priate response at the very first encounter with a predator but also allows some flexibility for paradise fish to learn from repeated encounters. These findings suggest that the antipredatory behavior of paradise fish may be affected by both genetic factors and learning and that the relative importance of the former or latter factor may vary depending on the type of predator paradise fish encounters. Acknowledgments This work was supported by an O T K A Grant (#2309) of Hungary. I thank Professor J. A. for their useful comments on HOGAN,two anonymous reviewers and Professor H. J. BROCKMANN whose ideas initiated this study. the manuscript. I am particularly grateful to Professor V. CSANYI

Literature Cited ALTBACKER, V. & CSANYI, V. 1990: The role of eyespots in predator recognition and antipredatory behaviour of the paradise fish, Macropodus opercularis L. Ethology 85, 51-57. COSS,R. G . 1978: Perceptual determinations of gaze aversion by the lesser lemur, Microcebus murinus: The role of i:wo facing eyes. Behaviour 64, 248-270. _ - 1979: Delayed plastici~yof an instinct: Recognition and avoidance of two facing eyes by the jewel fish. Devel. Psychobiol. 12, 335-345. CSANYI,V. 1985: Etho1ogic.d analysis of predator avoidance by the paradise fish (Macropodus opercularzs L.): Recognition and learning of predators. Behaviour 92, 227-240. _ - 1986: Ethological analysis of predator avoidance by the paradise fish (Macropodus opercularis L.): 11. Key stimuli in avoidance learning. Anim. Learn. Rehav. 14, 101-109. - -, T ~ T H ,P., ALTBACKEF., V., D ~ K A A. , & GERVAI,J. 1985a: Behavioural elements of the paradise fish (Macropodus opercularis). I. Regularities of defensive behaviour. Acta Biol. Hung. 36, 93-1 14. - - - - - -, - - & - - 1985b: Behavioural elements of the paradise fish (Macropodus opercularis). 11. A functional analysis. Acta Biol. Hung. 36, 115-130. CURIO,E. 1975: The functional organization of antipredator behaviour in the pied flycatcher: A study of avian visual perception. Anim. Behav. 23, 1-45. _ _ 1978: The adaptive sigriificance of avian mobbing. I. Teleonomic hypotheses and predictions. Z. Tierpsychol. 48, 17ii--183. DUGATKIN, L. A. 1988: Do guppies play tit for tat during predator inspection visits? Behav. Ecol. Sociobiol. 23, 395-399. FEDER, M. E. & LAUDER,G. V. (eds.) 1986: Predator-prey Relationships. Perspectives and Approaches from the Study of Lower Vertebrates. Univ. Chicago Press, Chicago, London. G E R L A IR., , CRUSIO, W. E. ik CSANYI,V. 1990: Inheritance of species-specific behaviors in the paradise fish (Macropodus opercularis): A diallel study. Behav. Genet. 20, 4 8 7 4 9 8 . _ _ 81 CSANYI,V. 1990: Genotype-environment interaction and the correlation structure of behavioral elements in paradise fish (Macropodus opercularis). Physiol. Behav. 47, 343-356. - _ & HOGAN,J. A. 1992: Learning to find the opponent: An ethological analysis of the behavior of paradise fish (Macropodus opercularis, Anabantidae) in intra- and inter-specific encounters. J. Comp. Psychol. 106, 306-315. GERVAI,J. & CSANYI,V. 1985: Behavior-genetic analysis of the paradise fish (Macropodus opercularis). I. Characterization of the behavioral responses of inbred strains in novel environments: A factor analysis. Behav. Genet. 15, 503-519. HELFMAN, G. S. 1989: Threat-sensitive predator avoidance in damselfish-trumpetfish interactions. Behav. Ecol. Sociobiol. 24, 47-58. HIRSCH,S. M. & BOLLES,R. C . 1980: O n the ability of preys to recognize predators. Z. Tierpsychol. 54, 71-84. HOOGLAND, R., MORRIS,D . & TINBERGEN, N. 1956: The spines of sticklebacks (Gasterosteus and Pygosteus) as a means o i defence against predators (Perca and Esox). Behaviour 10,205-236.

136

ROBERTGERI-AI, Recognition of a Natural Predator by Paradise Fish

JONES,R. 1980: Reactions of male domestic chicks to two-dimensional eye-like shapes. Anim. Behav. 28, 212-218. MAGURRAN, A. E. 1989: Acquired recognition of predator odour in the European minnow (Phoxinus phoxznus). Ethology 82, 2 1 6 2 2 3 . _ _ & HICHAM, A. 1988: Information transfer across fish shoals under predator threat. Ethology 78, 153-158. _ _ & SEGHERS,B. N . 1990: Population differences in predator recognition and attack cone avoidance in the guppy Poeciliu reticulutu. Anim. Behav. 40, 443-452. MIKLOSI,A. & CSANYI,V. 1989: The influence of olfaction on exploratory behaviour in the paradise fish (Macropodus operculurzs). Acta Biol. Hung. 40, 195-202. NICHOLS, J. T. 1943: The Fresh-water Fishes of China. Am. Mus. Nat. Hisr., New York. OWINGS,D. H . & COSS, R. G. 1977: Snake mobbing by California ground squirrels: Adaptive variation and ontogeny. Behaviour 62, 50-69. _ _ & OWINGS,S. C . 1979: Snake directed behaviour by blacktailed prairie dogs (Cynomis ludoviciunus). Z. Tierpsychol. 49, 35-54. PAGANO,R. R. 1990: Understanding Statistics in the Behavioral Sciences (3rd ed.). West Publ. Comp., New York. PITCHER,T. J., GREEN,D. A. & MAGURRAN, A. E. 1986: Dicing with death: predator inspection behaviour in minnow shoals. J. Fish Biol. 28, 4 3 9 4 4 8 . SEGHERS,B. H. 1975: The role of gill-rakers in size-selective predation by lake whitefish, Coregonus clupeuformis. Verh. Int. Ver. Limnol. 19, 486-489. SIMPSON,M. J. A. 1968: The display of the Siamese fighting fish, Betta splendens. Anim. Behav. Monogr. 1, 1-73. Received: September 17, 1992 Accepted: February 24, 1993 (J. Brockmunn)