Aimal Learig & Behavior 1982,10 (2),145-151 Breedig, pateral behavior, ad their iterruptioi Betta spledes PAUL M. BRONSTEIN Uiversity ofmichiga, Flit, Michiga Seve experimets examied the reproductive activities of Siamese fightig fish (Bettasp/edes). Spawig occurred after heterosexual pairs were together for about 24 h, ad males eared for eggs, est, ad fry thereafter. The visual cues provided by a itruder male, but ot a female, stimulated aggressio i the male breeder, ad these agoistic behaviors competed with breedig to cause a decremet i reproductive efficiecy. Males were foud to protect eggs ad fry by prevetig the growth of a fugus lethal to their offsprig. Several lies of evidece suggest that Siamese fightig fish (Betta spjedes) breed whe isolated from adult cospecifics. First, at temperatures coducive to breedig, males costruct ests i relatively isolated areas where detectio by other fish could be relatively ulikely (Brostei, 1981a, Experimet 4). Nest sites are spawig sites also. Secod, followig the costructio of a est, males become "attached" to their est sites after they observe a male cospecific (Brostei, 1981a, Experimet 3). Third, the frotal, gill-cover display of males, elicited by the approach of a cospecific to withi 12 em (Brostei, 1981b), causes the escape of the fish at the start of a male-male ecouter (Brostei, 1981a, 1981b). Fourth, the extesio of the male's operculum is a effective "keep-out" display (Brostei, 1981a, Experimet 2); fish that most persist i this gill-cover display are able to provoke escape amog males that display less teaciously. Thus, the fish with the more persistet display almost always gais cotrol of boudary areas whe two males see each other. Fifth, whe housed i groups, a pair of bettas will sometimes remove themselves to a relatively isolated part of their tak ad breed (Goldstei, 1975). Sixth, ad fially, the presece of a male i breedig coditio appears to excite fightig activity amog adult females; however, the evidece o this last poit is relatively weak (Brostei, 1980). The evidece justpresetedsuggests that i domesticated bettas physical isolatio amog cospecifics is associated with breedig. Furthermore, the males have a set of behaviors (display, escape, est-site te- This paper was supported by Grat MH33389-o1 from the Natioal Istitute of Metal Health ad by a grat from the Faculty Developmet Fud of the Uiversity of Michiga-Flit. I thak Aura Star for her aalysis of the fugi i Experimet 7 ad Collee Pace for her secretarial assistace. Requests for reprits should be set to: Paul M. Brostei, Departmet of Psychology, Uiversity of Michiga-Flit, Flit, Michiga 48503. acity) capable of producig such isolatio. The curret set of studies was desiged to test the coverse case. Namely, sice isolatio regularly accompaies breedig, I sought to determie whether the absece of social isolatio would be sufficiet to reduce reproductive efficiecy i bettas. EXPERIMENT 1 Subjects ad Appatus. Adult Betta spledes, purchased locally or bred i the laboratory, were used i all studies. The aimals were at least 8 moths old ad oe showed ay fi damage, makig it ulikely that the fish had bee fightig recetly. For at least 2 weeks prior to the start of each experimet, males were maitaied i at least 900 ml of water, visually isolated from other fish. Females were housed i a idetical maer, but ot i visual isolatio. Lights wet o at 0800 h ad off at 2200 h. Temperature was maitaied at 28 C by a space heater, ad the fish were fed either dry food (Wardley's) or recetly thawed brie shrimp (Artemia salia) twice daily. Test eviromets for all studies were glass aquaria, 40.4 cm log, 25.0 cm high, ad 20.3 em wide. Udyed gravel (2 em deep) covered the bottoms of the taks. Tap water (ph = 7.0), left stadig for at least 24 h ad i which o fish had swum, was used, ad there were 15-20 liters of water per tak. Six modificatios of each tak were made: (I) Thermostatically cotrolled heaters were added to elimiate mior fluctuatios i water temperature. (2) A clear glass plate covered the taks to esure relatively high levels of humidity immediately above the water surface, a coditio which my ow upublished observatios had suggested facilitates the maiteace of bubble ests. (3) Four strads of Elodeadesa were plated i a vertically orieted buch ear the ceter of each tak. (4) A triagular itruder compartmet, extedig the full height of each tak, was created by placiga pieceof glass (15.2 em wide, ad servig as the hypoteuse of the triagle) i a comer of each aquarium. The legs of the triagular cross sectio (the walls of the taks) were 8.4 ad 12.7 em log. (5) Taks were wrapped i brow paper o all but the oe short side where the itruder compartmet was located. (6) The aged tap water usedi the breedig taks was softeed by filtratio through a io-exchage resi (Aquarium Pharmaceuticals, Ic., Perkasie, Pesylvaia). Aquarists recommed this procedure whe breedig bettas (cf. Maurus, 1976). Procedure. A differet male was placed i each of 32 taks, ad, after he had had 72 h of adaptatio, a female was added to Copyright 1982 Psychoomic Society, Ic. 145 0090-4996/82/020145-07$00.95/0
146 BRONSTEIN each tak ad the idepedet groups were established. For cotrols ( = 12 pairs), the itruder compartmet remaied empty; these fish were isolated from other cospecifics durig spawig. For Group M (males; = 11), a adult male was placed i the itruder compartmet. For Group F (females) (=9), a adult female was placed i the itruder chamber. The itruders were added withi 2 mi after the female breeder was. Feedig cotiued o a twice-a-day schedule, ad the itruders were removed after 2 days' residecy. After the males were isolated, the taks were observed for 7 days for ests, spawig behavior, ad recetly hatched fry. Nests were defied as ay collectio of surface bubbles. Spawig was judged to have occurred either if eggs were clearly visible i the bubble est or if the male had bee observed to clasp the female. The existece of fry was established by visual ispectio. I additio, taks judged ot to cotai hatchligs were disassembled immediately followig the study. All water was the removed ad agai ispected, this time i volumes of about 1.5 liter ad agaist a white backgroud. Fry, which are black, were easy to detect i this eviromet; however, oe of my iitial judgmets was revised as a result of these additioal, post hoc observatios. addiscussio Fisher's exact probability test was used for data aalyses. All males had bubble ests withi 24 h of their itroductio ito the breedig taks. Nests were located agaist at least oe wall of these aquaria. With oe exceptio, i which the results were ambiguous, these origial est sites were used for breedig; ew locatios were ot chose. As show i Table I, betwee 73% ad 100% of each group spawed, ad spawig always bega withi 24 h of the females' beig preseted to the males. There were o differeces betwee groups o this measure. Relative to cotrols, the additio of a male itruder sigificatly reduced the probability of fry beig foud (p =.027). Female itruders did ot have this disruptive effect (p =.45), while Groups M ad F differed reliably (p =.024). With oe exceptio, a 4th-day hatch, all hatchigs occurred o the 3rd day after females were placed with the males. It is ot clear whether the male itruders disrupted reproductio by actig o the male ad!or the female breeders. Neither is it clear by what sesory pathways this effect might be achieved. The experimet was desiged to maximize the visual presece of itruders, but chemical commuicatio was also possible. Pheromoal cotrol over social behavior i bettas is kow to occur (e.g., Igersoll, Brostei, & Bovetre, 1976; Lee & Igersoll, 1979). Although Table 1 of Experimet 1 Total Number Number of Number of Pairs Pairs With Group of Pairs Spawig Hatchligs Cotrol 12 10 10 Female 9 9 8 Male 11 8 4 the mechaisms are ukow, breedig pairs of bettas, as hypothesized, curtail reproductio whe deied social isolatio from males. Barash (1976) has documeted similar strategies i other species. EXPERIMENT 2 This study was desiged to determie whether a male itruder would iterrupt reproductio through ochemical meas, with the itruders beig preseted i sealed bottles. Secod, the study also addressed the questio of whether breedig pairs require privacy throughout the etire breedig sequece i order for uiterrupted reproductio to occur. Third, more detailed behavioral records were maitaied tha i the first study. I this ad all subsequet studies, the clear glass partitio used to isolate the itruders of Experimet 1 was ot used. All other details of the subjects ad apparatus were ualtered. Male fish were etted ad placed i idividual aquaria for 72 h. Oe female was the added to each tak, ad the idepedet groups were established. For the cotrols ( = 23 pairs), square, clear-glass jars (8 em o a side ad cotaiig 800-900 ml of water) were placed approximately 15 em from each male's est withi 2 mi after the female's trasfer. Group M ( = 24 pairs) was exposed to the same type of cotaiers, except that each held a adult male cospecific. These jars were left udisturbed for 60 mi, after which they were removed for 5 sec ad the retured to their origial positio. Fially, Group I-H ( = 20 pairs) had a male added for oly the first 60 mi the breeders were together. These itruders ad their jars were replaced by empty vessels after 1 h with a heterosexual pair. Thus, the Groups M ad I-H experieced the same patter of physical disruptios-jars added ad replaced at the same times. However, the M group ecoutered far more social stimulatio tha did the I-H subjects. Twice-a-day feedigs cotiued throughout the breedig cycle i this ad all subsequet studies, ad taks were observed at least three times a day for the 3 days followig the start of cohabitatio. Aquaria were observed for 10 sec, ad the positios ad behaviors of the fish, alog with the presece/absece ad coditio of ests, eggs, ad fry, were recorded. The umber of fry i each tak was recorded 72 h after pairs were formed. Hatchligs were collected by rapidly slidig a beaker uder each bubble est, thereby draiig fry ad est ito that vessel. This procedure was usually adequate for gatherig all the youg i ay tak. Visual reispectio by three observers, alog with the occasioal emptyig of aquaria (as i Experimet 1) cofirmed the reliability ad completeess of this harvestig procedure. Multiple collectios of fry ad the removal of adult fish ad plats were occasioally required to complete the gatherig of hatchligs. Whe youger tha 48 h posthatch, most fry react to local vibratios by sikig slowly to the bottom of the tak, ad, after remaiig statioary for approximately 10 mi, swimmig back to the surface of the water. These characteristics, alog with their dark coloratio, were exploited i the collectio ad coutig of the hatchligs. Fry that were disturbed, but that evaded iitial collectio attempts, soo retured to the water surface, where they could be trapped. Followig harvest, the fry were poured ito white cotaiers, where they remaied motioless ad easily visible o the substrate for several miutes. Two traied observers, uiformed about the group from which ay clutch was take; were able to cout hatchligs to withi 5070 accuracy.
REPRODUCTION IN Betta spledes 147 Icidetal observatios suggested that a ecouter with a opposite-sex cospecific rapidly produced est attachmet amog the males ad a patter of vertical stripes amog the females. These effects appeared to occur withi miutes, or eve secods, of the aimals' iitial cotact. Durig the 1st day of cohabitatio, cotrol males stayed with their ests ad occasioally bit ad chased the females. The females escaped durig their 1st day with a male, but the teded to follow males to their ests ad spaw. Most egg layig occurred at the start of the aimals' 2d day together, ad the rate of spawig did ot vary as a fuctio of treatmet coditios (see Table 2). The presece of a male itruder stimulated agoistic behavior amog the male residets, ad betwee 63070 ad 79% of the male breeders were see emittig frotal displays (i.e., gill-cover erectios) at each observatio. While the itruder did ot affect the speed with which copulatio was iitiated, the additioal male did reduce the attetio paid to est sites by males durig every observatio [x2(2) ~ 9.19, ps <.02}. As oted i the cetral portio of Table 2, the 1-H ad cotrol groups were ot statistically differet, while the itruder did iterfere with est-site attachmet i the male coditio. Fially, the M group produced reliably fewer fry per clutch tha either of the other two groups [F(2,64) =9.76, p <.005}. Post hoc t tests showed that the cotrol ad 1-H groups were ot statistically differet while comparisos betwee the M group ad the other two coditios were highly sigificat (ps <.001). Also, more of the M pairs had o hatchligs at all tha did either of the other two [TOUPS [x2(l) ~ 8.18, ps <.01}. Half of the breeders i the M group produced five or fewer offsprig. Discussio Take together, these data replicate the fidigs of Experimet I-a male itruder ca reduce the breedig efficiecy of a pair of bettas without alterig their rate of spawig. Furthermore, three ew coclusios emerge: (1) The itruder effect is maifest whe clutch sizeas wellas the umber of clutches is determied. (2) The itruder is disruptive eve i the absece of chemical commuicatio betwee himself ad the breedig pair. (3) I order for a male itruder to iterrupt breedig, he must be preset for some period other tha just the first 60 mi of courtship; the 1-H group behaved idetically to the cotrols. Male itruders attract male breeders, ad the aggressio that emerges appears to affect the postspawig caretakig. There were 10 pairs i the M group whose ests either disappeared or became scattered after spawig. Such deterioratio of ests was ever observed i ay other group. There was a obvious loss of eggs from the ests of 12 pairs i the M coditio. Observatios also revealed some effect of the itruder o the quality of spawig. Males were occasioally see to copulate, but the to display to the itruder while the eggs just expressed from the female fell to the substrate, uatteded. EXPERIMENT 3 This study was udertake to determie whether, as suggested by Experimet 2, the presece of a male cospecific was sufficiet to produce a patter of high-cotrast vertical stripes amog females. The subjects were 24 female bettas. They were housed i idividual itruder jars (as used i Experimet 2) ad these cotaiers were placed adjacet to each other. Fish could see ad be see by females i adjacet jars. Each aimal's patter of coloratio was judged o four occasios, ad each assessmet was made idepedetly by two observers who agreed fully. First, a judgmet was made while each aimal was left udisturbed i its home vessel. Secod, the jars housig the females were placed, oe per tak, i the stadard aquaria used for breedig, ad female coloratio was assessed 60 sec after trasfer. (The taks lacked vegetatio ad had a illumiatio of 344 Ix at the water surface.) Third, a adult male, also i a itruder jar, was added to each aquarium, ad the females were observed after the pairs had bee together for 5 mi. Fourth, a fial judgmet was made after both jars had bee i the same tak for 30 mi. Three groups of eight females were used. I oe coditio, males were placed 23 em away from the subjects. I aother coditio, the adjacet jars cotaiig a male ad a female, respectively, were touchig. A itermediate coditio was created by puttig the male cotaiers 12.7 em from the female jars. Table 2 of Experimet 2 Fry Available at 72h Cumulative Percetage of Pairs With Eggs Percetage of Males at Their Nests After Pairig Hours Sice Pair Formatio Hours Sice Pair Formatio Fry/Clutch Group 3 19 22 26 45 50 68 3 19 22 26 45 50 68 Mea SEM p* Male 24 0 4 54 75 88 92 92 24 46 33 54 63 29 42 46 24 154 41 58 Cotrol 23 0 17 69 78 78 87 87 23 96 61 96 100 100 91 100 23 374 40 92 1 h 20 0 5 65 90 95 100 100 20 95 75 85 100 100 100 90 20 373 44 95 *Percetage ofpairs with fry.
148 BRONSTEIN I the absece of males, o female exhibited the high-cotrast, vertical lies. Durig the first two assessmets, some fish had logitudial stripes, some had pale vertical lies, ad others showed uiform coloratio, ofte with a pale vetral crest ear their geital papillae. After 5 mi of exposure to a earby male, a patter of thick dark-ad-light vertical bars appeared. All females that were closest to a male showed this chage, while oly two females i the 23-cm group had this stripig. The differece betwee these two extreme groups was reliable (Fisher's exact probability test, p <.01). At a distace of 12.7 cm betwee aimals, 75lt/o of the females showed thick vertical stripes, ad o other itergroup differeces were sigificat. The fourth set of judgmets produced the same results as the third, Females show vertical bars whe exposed to a proximate male. Chemical commuicatio betwee aimals is ot required to produce this patter. This high-cotrast patter, i cojuctio with a slackfied posture typical of spawig females, reduces the itesity of a male betta's attacks (Robertso & Sale, 1975). This form of disruptive coloratio is also thought to serve as effective camouflage i may species (e.g., Cott, 10). EXPERIMENT 4 This study was ru to determie whether stimulatio from cospecifics produces est-site attachmet amog residet males ad also to fid out whether a male itruder alters the umber of eggs released by females ad placed ito bubble ests by the males. Males were placed i 15-20-litertaks of water, ad 72 h later a female was added to each tak. Withi 2 mi of each heterosexual pairig, a itruder bottle was placed approximately 15 cm from each male's est. For the M group ( = 18 pairs), the jar cotaied a male cospecific; for the cotrol group ( = 15 pairs), it cotaied oly water. The itruder vesselswere ot moved oce they were iserted i the taks. The aquaria were observed 5 mi prior to male-female pairigs, 30 mi after the bottles were added, ad at 2-h itervals thereafter, util lights were shut off. Eggs were gathered by draiig ests ito a submerged beaker 30 h after the breeders were brought together. After collectio, the eggs, which are white ad opaque at this stage, were trapped i fie mesh, rised ito a Petri dish, ad placed o a black backgroud for coutig. This procedure permitted two idepedet observers to agree o clutch sizes to withi 5070. addiscussio The fidigs of Experimets 1 ad 2 were recofirmed. As see i the left sectio of Table 3, spawig rate did ot vary sigificatly as a fuctio of whether a itruder was preset. Also, betwee 61lt/o ad lt/o of male residets were displayig to itruders at each observatio. I cofirmatio both of Experimet 2 ad pilot work reported previously (Brostei, 1981a), the appearace of a female was followed by est-site attachmet amog residet males. Oly oe residet from each group was see at its est just prior to the formatio of heterosexual pairs, while 93lt/o of cotrol residets ad 22lt/o of the male group were at their ests 30 mi after pairig. The presetatio of a male itruder resulted i reliably fewer males appearig at their ests o all but oe observatio (xz(l) ~ 8.26, p <.01]. Males i both groups were equally likely to be see at their ests 26 h after pairig (xz(l) = 1.39, p >.05]. (See the cetral portio of Table 3.) Itruders attracted male breeders away from their ests; however, this chage i behavior did ot affect the mea umber of eggs per clutch. Cotrols had umerically more eggs tha did M pairs, but the differece did ot approach sigificace [t(31)= 1.61, p <.20]. These fidigs suggest that a male itruder reduces the productio of hatchligs either by alterig the process of fertilizatio or by disruptig caretakig activities so as to reduce the survivorship of the fertilized eggs. The productio ad collectio of eggs appears to be little affected by a extra male's remaiig visible ear a residet's est. Furthermore, the deterioratio of ests durig the postspawig stage i Experimet 2 shows that itruders do affect caretakig. Oly oe pair i the M group of Experimet 4 showed ay sigs of est deterioratio. EXPERIMENT 5 The specific questio posed by this study was whether the iferiorcaretakigproducedby the extra male somehow ivolved the presece of the female breeder. It is possible, for istace, that the female devours some fertilized eggs while the male residet leaves the est ad displays to the itruder. Cumulative Percetage of Pairs With Eggs Hours Sice Pair Formatio Table 3 of Experimet 4 Percetage of Males at Their Nests Hours Sice Pair Formatio Eggs/Clutch 30 h After Pairig Group.5 1.5 3.5 5.5 7.5 9.5 24 26 28.5 1.5 3.5 5.5 7.5 9.5 24 26 28 Mea SEM Male 18 0 0 0 0 0 0 0 56 72 18 22 22 39 45 39 28 40 36 11 18 542 91 Cotrol 15 0 0 0 0 0 0 13 60 73 15 93 100 93 100 93 93 100 60 100 15 759 99
REPRODUCTION INBetta spledes 149 A M group ad a cotrol group were created, ad each was bifurcated to create a total of four idepedet groups. First, male breeders were isolated i stadard aquaria. After 48 h, a female was placed i each residet's tak. Two miutes later, a jar cotaiig a male cospecific was added to each Group M tak ad a jar cotaiig oly water was added to each cotrolgroup tak. Pairs were observed periodically, fed twice daily, but otherwise left udisturbed for the ext 30 h. After 39 h of pairig, the females were removed from 17 of the M-group taks ad 17 of the cotrol taks. Whe a female was ot take from a particular tak, a et was dipped ito that aquarium ad withdraw empty a few secods later. The taks were observed periodically for a additioal 48 h, after which fry were collected ad couted. O ay observatio, betwee 68070 ad 8S% of the Group M residets were displayig to the itruder. Furthermore, the probability of detectig a male at its est was reliably higher i the cotrol group tha i the M group o every observatio [x2(1) ~ 22.83, P <.001]. However, the uexpected fidig, oted o the left side of Table 4, was that the M pairs were reliably less likely to spaw ad produce eggs tha were the cotrol breeders. As a cosequece of this reductio i spawig rate, it was ot possible to explore the possible iteractio betwee the female ad the itrudig male i reducig the quality of egg care ad, perhaps, the umber of fry produced. The umber of M pairs with eggs was so few that the appropriate comparisos could ot be made. However, the role of the female durig caretakig amog the cotrol pairs was assessed. As see i Table S, est attachmet amog males was ot affected by the presece of a female. The female also failed to have ay reliable effect o mea clutch size [t(34)= 1.6S, p <.20], although the tedecy was for more fry whe the females had bee removed. EXPERIMENT 6 Experimet S is the first report of a reductio i the speed or probability of egg layig followig the appearace of a cospecific male. This reductio may have bee due to the reduced amout of time males werealoe i their taks prior to the additio of a mate ad the itruder. The males of Experimet S were isolated for oly2 days; the fish i Experimets 1,2, ad 4 were aloe for 3 days. The curret study examied the possible ifluece of this variable-time of prematig residecy-o the productio ofyoug. Two groups of breedig pairs were formed. Both had male itruders. The procedures used to create a M group i Experimet 4 were agai employed. For a 2-Day group (= 17 pairs), the males were isolated for 48 h prior to beig joied by females. For a 3-Day group (= 18 pairs), the residet males were kept aloe i their breedig taks for 72 h prior to the additio of females. As oted i Table 6, both groups behaved as had the fish exposed to itruders i Experimets I, 2, ad 4. Egg layig occurred amog at least 83% of the pairs i each group after 27 h of cohabitatio, ad the average clutch sizes, which did ot vary betwee groups [t(33)=.66] also resembled the values foud i Experimets 2 ad 4. Betwee 47% ad % of each group was displayig o ay observatio, but there were o reliable itergroup differeces o either this measure or o est attachmet (see Table 6) [x2(l) ~ 3.63, p <.10]. The lack of spawig i Experimet S appears ot to have resulted from the brief (48-h) residecy of the makes i that study. The 2-Day ad 3-Day groups were ot differetiable i Experimet S. Group Male Cotrol 34 36 *x 2 (l ) ;;' 1l.37,p <.001. Table 4 Comparisos of Male ad Cotrol Groups i Experimet 5 Cumulative Percetage of Pairs With Eggs 3 o 20 6 17 23* 21 75 27* 53 92 39* 53 34 36 Percetage of Males at Their Nests 3 21 81 20 32 100 23* 35 100 27* 41 100 39* 26 92 Table 5 Effects of Female Removal i Experimet 5 Percetage of Males at Their Nests Fry Available at 70 h After Pairig Hours Sice Pair Formatio Fry/Clutch Cotrol Group 42 45 62 65 69 Mea SEM p* With Females Females Removed 19 17 79 84 88 99 82 89 95 88 19 17 321 465 65 58 79 "Percetage ofpairs with fry.
150 BRONSTEIN Table 6 of Experimet 6 Fry Available at Cumulative Percetage of Pairs With Eggs Percetage of Males at Their Nests 72 h After Pairig Hours Sice Pair Formatio Hours Sice Pair Formatio Fry/Clutch Group 3 20 23 27 44 47 51 68 71 3 20 23 27 44 47 51 68 71 Mea SEM p* 3 Day 18 0 17 50 83 83 83 83 83 83 18 44 11 6 28 6 23 33 28 28 18 164 48 50 2 Day 17 0 18 59 88 88 88 88 88 88 17 24 12 12 29 0 29 36 59 53 17 205 57 65 *Percetage ofpairs with fry. EXPERIMENT 7 The fial study, doe i two replicatios, explored the fuctio of the male carig for ests, eggs, ad fry. Pilot work from my lab has show that eggs ca develop i the absece of adult caretakers. Furthermore, eggs floatig at the water surface appeared more likely to develop tha those submerged 10 em below at the bottom of a glass jar. It is possible that fertilized eggs are more buoyat tha ufertilized ova. O the other had, submerged eggs might develop ormally if maitaied at the water surface, i cotactwith moist air. The behavior of male breeders could serve this physical fuctio sice males frequetly express bubbles to the water surface, mouth the eggs, ad place the eggs ito the bubble mass. This sequece of est-maiteace actios could icrease the buoyacy ad, idirectly, the survival of fertilized eggs. This hypothesis was tested by removig both male ad female breeders followig spawig. The umber of fry from taks with a male caretaker were later compared with the cout from taks where either male or female remaied preset after egg layig. Idividual males were placed i separate breedig taks, ad a female was added to each vessel after 72 h. Followig egg layig, 30 h after the start of cohabitatio, a experimetal group was created by ettig ad removig both adults from some taks. A cotrol group was also created by removig females oly. Hatchligs were the removed ad couted. I the iitial replicatio, this procedure was accomplished 30 h after the removal of the adults; the delay betwee the removal of adults ad egg coutig was 54 h i the secod study. Group Experimetal Cotrol Table 7 of Experimet 7 Replicatio 1 Hatchlig Cout Replicatio 2 As see i Table 7, the presece of a male caretaker improved reproductive efficiecy. I each repli Mea SEM Mea SEM 11 234 69 12 146 36 10 352 58 9 338 113 catio, a greater umber of fry were foud i cotrol taks (with males) tha i the experimetal aquaria (without caretakers). The differece betwee groups was ot reliable i the first replicatio [t(19) =1.31] but was sigificat for the secod replicatio [t(19)=2.28, p <.05]. Cotrary to the iitial hypothesis, males protected their geetic ivestmet ot just by keepig their offsprig afloat. Rather, the presece of a male preveted the appearace of a saprolegiid fugus amog the eggs ad fry at the water surface. Noe of the cotrol taks cotaied visible sigs of fugus; ad o fugus had ever bee detected i approximately 300 prior breedigs i which a male had bee with its eggs ad est after copulatio. I cotrast, fugus appeared i all but oe of the experimetal taks i Replicatio 1, while ie taks were visibly cotamiated i Replicatio 2. The fugus appeared as a cotto-like white mass; samples were subsequetly idetified microscopically. Fially, while my procedure for coutig fry was too crude to distiguish reliably betwee livig ad dead hatchligs, may fry couted as the reproductive products of experimetal pairigs were either dead or appeared to be succumbig to the fugal attack at the time of their iclusio i the hatchlig cout. The effect reported here, although statistically reliable, is a uderestimatio of the importace of pateral caretakig i permittig the survival of offsprig. It is possible that males do keep their eggs afloat ad i a physical positio where they might develop most rapidly. Males also have a fugicidal or fugistatic beefit for those eggs ad fry that are adequately buoyat to float at the water surface. GENERAL DISCUSSION are cosistet with the hypothesis that bettas are territorial, ad that reproductio is reduced whe males are exposed to the visual presece of other males. Whe added to the taks of males, females rapidly develop high-cotrast vertical bars ad flee from the males, ofte restig amog plats. Both behaviors protect females from males. Equally quickly, the males begi displayig to the females ad also become icreasigly "attached" to their ests. These chages i behavior ad coloratio oc-
REPRODUCTION IN Betta spledes 151 cur withi miutes of iitial sightigs. Males sped the ext day alteratig betwee est buildig ad displayig ear the females. The females geerally flee from the approaches of males, but, after about 24 h, the females begi to follow the males to their ests, where spawig, lastig several hours, occurs. The spawig sequece icludes the female's lyig early horizotally ear the water surface ad beig ecircled by the male. Release ad fertilizatio of the eggs the occurs. Oe of the immediate atecedets of the female's assumig this spawig posture is ofte the male's mouthig of its mate's geital papillae. This tactile stimulatio appears to be a importat, perhaps reflexive, cause for the female's assumig a receptive orietatio. Males fid both eggs ad geital papillae attractive targets for oral maipulatio, ad both appear to be approximately the same size, shape, ad color. These facts are miimal evidece suggestig a model/mimic relatioship betwee ova ad papillae. Wickler (1968) provides further evidece for the existece of ovomimicry amog teleosts. For at least 3 days followig egg layig, males typically maitai their ests ad place eggs ad fry ito those ests. They might also attack ad spaw with females durig this stage. These paretal activities ca guard the offsprig from lethal attacks from fugi(experimet 7); however, it is ot kow whether this protectio occurs by chemical or physical meas. Also, the pateral presece at the est site probably protects the fry from disturbace by other adult bettas. The sight of a estig male ca provoke escape ad estig i other males (Brostei, 1981a), as well as the escape of females. The curret results show a icrease i reproductive efficiecy whe males or females (Experimet 5) are elimiated from the area ear the est of a pateral male. It should be oted, however, that breedig seemed to be uaffected by the additio of a female (behid glass) i Experimet 1. It may be that the cotiuatio of tactile iteractios betwee males ad females is required durig the paretal phase of reproductio i order for a reductio i reproductive efficiecy to occur. I Experimet I, either were such tactile iteractio betwee malesad female itruders permitted or were the umber ofeggsrecorded. As oted i Experimets 1, 2, 5, ad 6, the cotiued presece of a male itruder always led to a deterioratio of reproductive efficiecy. Chemical commuicatio betwee breeders ad the itruder is ot required for this effect, ad the disruptio ca occur durig courtship ad copulatio (Experimet 5, above). The effect is most typically see durig the paretal phase (Experimets 1, 2, ad 6). Persistet visual stimulatio by itruders produces aggressive display amog male breeders (Brostei, Note 1), ad the agoistic sequece typically leads to escape by a itruder, especially at the start of a male/male ecouter (Brostei, 1981a, 1981b). Whe escape is preveted, aggressio becomes itese (Brostei, 1981a, Note 1) ad fightig behavior appears to compete with reproductive activities. The rate at which fightig escalates is largely depedet upo idividual differeces i visual display amog males (Brostei, 1981a; Robertso & Sale, 1975) ad could determie the timig ad extet to which breedig is disrupted by a itruder. Give the eormous variability i fightig amog males (e.g., Brostei, 1981a), it is ot surprisig to fid variatios i the extet to which fightig iterferes with reproductio. The early oset of this iterferece see i Experimet 5 is likely to be a cosequece of subject selectio. Through radom evets, the males of Experimet 5 could have bee aimals with a high potetial for fightig. REFERENCE NOTE 1. Brostei, P. M. Agoisticsequeces i male Betta spledes. Mauscript submitted for publicatio, 1981. REFERENCES BARASH, D. Sociobiology ad behavior. NewYork: Elsevier, 1976. BRONSTEIN, P. M. Betta spledes: A territorial ote. Bulleti of the PsychoomicSociety, 1980, 16, 484-48S. BRONSTEIN, P. M. Commitmets to aggressio ad est sites i malebetta spledes. Joural ofcomparative ad Physiological Psychology, 1981,95,436-449.(a) BRONSTEIN, P. M. Social reiforcemet i Betta spledes: A recosideratio. Joural of Comparative ad Physiological Psychology, 1981,95,3-9S0.(b) COTT, H. B. Adaptive coloratio i aimals. Lodo: Methue, 10. GOLDSTEIN, S. R. Observatios o the establishmetof a stable commuity of adult male ad female Siamese fightig fish (Betta spledesy. A imal Behaviour, 1975, 13, 179-185. GOODRICH, H. B., & TAYLOR, H. C. Breedigreactios i Betta spledes. Copeia, 1934, 165-166. INGERSOLL, D. W., BRONSTEIN, P. M., & BONVENTRE, J. Chemical modulatio of agoistic display i Betta spledes. Joural of Comparative ad Physiological Psychology, 1976, 90, 198-202. LEE, C.-T., & INGERSOLL, D. W. Social chemosigals i five Belotiidae (pisces) species. Joural ofcomparative ad Physiological Psychology, 1979,93, 1171 1181. MAURUS. All about bettas. Neptue City, N.J: T. F. H. Publicatios, 1976. RoBERTSON, C. M., & SALE, P. F. Sexual discrimiatio i the Siamesefightig fish (Betta spledes Rega). Behaviour, 1975, 54,1-26. WICKLER, W. Mimicry i plats ad aimals. New York: McGraw-HilI,I968. (Mauscriptreceived September28,1981; revisioacceptedfor publicatioapril S, 1982.)