.l ,4 ~z—‘\.« ,._ .. £11.. .~ Qua—“w“v4‘ i ... ,- “i v NURTURING EXPERMENTS WITH REGARD T0735; 5,1-3“ADULT IMPRINTING" AND RECOGNITIO IOF'YOUNG} MICHIGAN STATE UNIVERSITY ‘ : :7 41331:; ’7 3 HoIIIe L Collins f. : THESIS *J LIBRARY Michigan Stave University l L is... This is to certify that the thesis entitled NURTURING EXPERIMENTS WITH REGARD TO "ADULT IMPRINl'ING" AND RECOGNITION OF YOUNG IN THE CICHIJD SPECIES TILAPIA SPARRMANI AND AEQUIDENS LATIFRONS presented bg Hollie L. Collins has been accepted towards fulfillment of the requirements for Ph . D 0 degree in Zoology WGW James C. BraddOCk Major professor Date June 26, 1965 0-169 I mums EXPERD’ AND RECOGNIT] TILAPIA $3.1 This study is of cichlid fishes those of other si frtfl, and a few Tne tests were i nent of filial ' sane or another Other 8993, am Wrigglers. Eg respective de\ hex-even Intempec but wriggleI‘ another Spec erence. In hrooded pri the Species ABSTRACT NURTURING EXPERIMENTS WITH REGARD TO ”ADULT IMPRINTING" AND RECOGNITION OF YOUNG IN THE CICHLID SPECIES TILAPIA SPARRMANI AND AEQUIDENS LATIFRONS _____————' L c by Hollie La Collins This study is an attempt to discover how adult pairs 0f cichlid fishes distinguish young of their own species from those of other species. Tilagia sparrmagi, Aeguidens lati- IEEEEI and a few Hemichromis bimaculatus pairs were tested. The tests were fostering experiments which involved replace— ment of filial broods of eggs Or wrigglers by those of the same or another species. Eggs were exchanged only for other eggs, and wrigglers were only exchanged for other wrigglers° Eggs and wrigglers of various ages within their respective developmental stages were often exchanged, however. InterSpecific exchanges of eggs were never successful, but wrigglers were exchanged successfully. Wrigglers of another species were accepted with a definite order of pref— erence. In A. latifrons this was: younger than those being brooded prior to exchange; the same age; and older, Here the Species of the substituted young was 2, sparrmani. When T; M a similarI wt was eliCited° mani are large stage of devel‘ upon the faCto IntraspeCi cessful even u stage were gre and wriggler-I longed or sho: younger or 01 Visual stimul factors, poss synchronizing of the young‘ several pair: 0f their own nurturing du Prior br rEjection of 30th experie Other Specis Hollie L. Collins When 2, Sparrmani pairs were given A, latifrons wrigglers, a similar, but inverse, preferential order of acceptance was elicited. Since the developmental stages of I, sparr— mani are larger than those of g, latifrons at any given stage of development, these preferences appear to be based upon the factor of size° Intraspecific exchanges of eggs or wrigglers were suc— cessful even when age differentials within the developmental stage were great. Examples were noted where the egg fanning and wriggler-tending phases of parental behavior were pro— longed or shortened as a result of the acceptance of broods younger or older than those being tended prior to exchange. Visual stimulation by the young appeared to alter internal factors, possibly endocrine, within the brooders, thus synchronizing the behavior of the parents with the needs of the young. This situation is not a simple one since several pairs cared for as many as three successive broods of their own young simultaneously and performed appropriate nurturing duties to all of themo Prior brooding experience did not affect acceptance or rejection of young at subsequent matings in either speciesa Both experienced and inexperienced pairs reared young of the other species successfully in alternation with those of their own spec young of both imprinting . " Hollie L. Collins their own species or reared mixed broods consisting of young of both species. No evidence was noted for "adult imprinting." NURTURINC AND I m NURTURING EXPERIMENTS WITH REGARD TO ”ADULT IMPRINTING" AND RECOGNITION OF YOUNG IN THE CICHLID SPECIES TILAPIA SPARRMANI AND AEQUIDENS LATIFRONS by Hollie L3 Collins A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Zoology 1965 l The aut C. Braddock during the preparation Special Rollin H. B while servi extended to Coilege. Ch The aut‘ Henderson, , many favors tion fOr he, ACKNOWLEDGEMENTS The author wishes to express his thanks to Dr. James C. Braddock, of the Department of Zoology, for his guidance during the course of the experimental work and the preparation of this manuscript. Special thanks are extended to Dr. John A. King, Dr. Rollin H. Baker, and Dr. Eugene W. Roelofs for their help while serving as committee members. Thanks are also extended to the late Dr. Bernard Greenberg, of Roosevelt College, Chicago, Illinois for his helpful suggestions. The author would also like to thank Mrs. Bernadette Henderson, secretary of the Zoology Department, for her many favors. The author's wife, Barbara, merits appreciar tion for her encouragement during this research. ii INTRODUCTI METHODS AN RESULTS Genera Spawni The Re latifr The Re EEEEEE Parent Egg Me BQhEVj Involt Wrigg] Gener; Young an] Correy with c Adult DISCUS$10} SmhmRY EIBLIOGRA} TABLE OF CONTENTS Page INTRODUCTION . . . . . . . . . . . . . . . . . . . 1 METHODS AND MATERIALS . ‘.' . . . . . . . . . . . 8 RESULTS . . . . . . . . . . . . . . . . . . . . . 17 General Breeding Behavior of Substrate Spawning Cichlids . . . . . . . . . . . . . . 17 The Reproductive Sequence of Aeguidens latifrons . . . . . . . . . . . . . . . . . . 19 The Reproductive Sequence of Tilapia sparrmani . . . . . . . . . . . . . . . . . . 28 Parental Recognition of Eggs . . . . . . . . 37 Egg Measurements . . . . . . . . . . . . . . 43 Behavior of Adult Brooding Pairs in Tests Involving Exchanges of Young in the Wriggler Stage . . . . . . . . . . . . . . . 47 General Behavior of Aeguidens latifrons and Tilapia sparrmani Pairs Toward Introduced Young that Differed in Age from Their Own . . . . . . . o . . . . . . . . . . . . . 64 Correlation of Acceptance—Rejection Ratios with Comparative Wriggler Sizes . . . . . . . 65 Adult Imprinting . . . . . . . . . . . . . . 68 DISCUSSION . . . . . . . . . . . . . . . . . . . 73 SUMMARY . . . . . . . . . . . . . . . . . . . . . 82 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . 84 APPENDICES . . . . . . . . o . . . . . . . . . . 87 Table l. React Hemi< Brooc 2. Egg 1 Aegu: 3. Recog by B] When 4. Egg 1 Tile; 3- The l of 55 m 6. Resu: Broo< Their The I \‘Olv; @ Were Broo< FOste Pain Own I am 9' The ] Im'oj of ti Table LIST OF TABLES Reaction to Tilapia sparrmani and Hemichromis bimaculatus Eggs by Brooding Pairs of Aequidens latifrons Egg Exchanges Made Within the Species, Aeguidens latifrons . . . . . . . . . Recognition of Aeguidens latifrons Eggs by Brooding Pairs of Tilapia sparrmani When Exchanged for Their Own Eggs Egg Exchanges Made Within the Species, Tilapia sparrmani . . . . . . . . . The Mean Areas and Shape Indices of Eggs of Aeguidens latifrons, Hemichromis bimacu— latus, and Tilapia sparrmani . . . . . Results of Fostering Experiments Involving Brooding Pairs of Aequidens latifrons When Their Own Wrigglers Were Exchanged for Tilapia sparrmani Wrigglers . . . . . The Results of Fostering Experiments In— volving Brooding Pairs of Aeguidens latifrons When Tilapia sparrmani Wrigglers Were Exchanged for Part of Their Own Brood (i.e., Mixed Broods) Fostering Experiments Involving Brooding Pairs of Aeguidens latifrons When Their Own Wrigglers Were Exchanged for Tilapia sparrmani Wrigglers of Various Ages The Results of Fostering Experiments Involving Brooding Pairs of Aeguidens latifrons When the Foreign Wrigglers were of the Same Species as Their Own iv Page 38 4O 42 43 46 48 49 50 52 Table 10. Foste Pairs Aegui Ages ll. The F volvi Aeggi 12. The R volvi Chang Mixed ‘3- Posts Pairs Wrigg Volvi When Spec: Fests pairs wrlgg “£2 Ages 16. The P Cant FOste Age L Table 10. ll. 12. l3. 14. 15. 16. Fostering Experiments Involving Brooding Pairs of Aeguidens latifrons When Their Own Wrigglers Were Exchanged for Non—filial Aeguidens latifrons Wrigglers of Various Ages . . . . . . . . . . . . . . . . . . . The Results of Fostering Experiments, In— volving Brooding Pairs of Tilapia sparrmani When Their Own Wrigglers Were Exchanged for Aeguidens latifrons Wrigglers . . . . . . The Results of Fostering Experiments In— volving Brooding Pairs of Tilapia sparrmani When Aequidens latifrons Wrigglers Were Ex— changed for Part of Their Own Brood (i.e., Mixed Broods) . . . . . . . . . . . Fostering Experiments Involving Brooding Pairs of Tilapia sparrmani When Their Own Wrigglers Were Exchanged for Aeguidens latifrons Wrigglers of Various Ages . . . . The Results of Fostering Experiments In- volving Brooding Pairs of Tilapia sparrmani When the Foreign Wrigglers Were of the Same Species as Their Own . . . . . . . . . . . Fostering Experiments Involving Brooding Pairs of Tilapia sparrmani When Their Own Wrigglers Were Exchanged for Non—filial Tilapia sparrmani Wrigglers of Various Ages . . . . . . . . . . . . . . . . . . The Probability Values Computed for Signifi— cant Differences Between the Four Classes of Fostering Experiments at Each of the Three Age Levels . . . . . . . . . . . . . . . Page 53 54 55 56 58 59 66 17. The l Tiler Wrigg ments ‘ Wrigg ; Fema] 18. The E Pairs Made EEEEE 19. The E Pairs Made Table 17° 18. 19. -Page The Mean Total Length Measurements of Tilapia sparrmani and Aeguidens latifrons Wrigglers Based on One-Hundred Measure- ments Taken at Each of Three Ages on Twenty Wrigglers from Each of Five Different Females for Each Species . . . . . . . . . . 67 The Spawning Sequences of Aequidens latifrons Pairs in Which Fostering Experiments Were Made With Foreign Species Young (Tilapia sparrmani) . . . . . . . . . . . . . . . . 69 The Spawning Sequences of Tilapia sparrmani Pairs in Which Fostering Experiments Were Made With Foreign Species Young (Aequidens latifrons) . . . . . . . . . . . . . . . . . 71 vi as Frees F T A .l. 2 e e «L Mr. D We 04 n5 . l. . 1 _ F Di Figure 1. Figure 2. LIST OF FIGURES Page Fanning Time in Minutes Per Hour for Three Pairs of Aequidens latifrons. Average Male and Female Fanning Times and Their Totals for Each of Eight Subsequent Spawnings . . . . . 23 Fanning Time in Minutes Per Hour for Three Pairs of Tilapia sparrmani. Average Male and Female Fanning Time and Their Totals for Each of Six Subsequent Spawnings . . . . . . 29 Appendix .. l (U Appendix l° LIST OF APPENDICES Page The Distribution of Area Calculations (Width x Length x W/4) for Eggs of Tilapia sparrmani, Aequidens latifrons, and Hemichromis bimaculatus . . . . . . . . 87 The Distribution of Shape Index (W/L) Calculations for Eggs of Tilapia sparm ani, Aeguidens latifrons, and Hemichromis bimaculatus . . . . . . . . . 88 viii The Ci( the fresh v and Central United Stat several car unusual abi caused thei their adapt successful where they tein (Uchid Diversi ability to of many beh ESpects of . ship, terri' and imprint The cicl according t< though inter Categories ' INTRODUCTION The Cichlidae is a large family of fishes native to the fresh waters of Africa, the Near East, India, South and Central America, and the southwestern part of the United States. Most species are extremely hardy, and several can acclimate to brackish and marine water. Their unusual ability to adapt to many habitats has recently caused their adoption for use as food or game fish. Thus, their adaptability to culture techniques has led to many successful introductions in southeastern Asian countries where they have already become a valuable source of pro— tein (Uchidna, Richard N. and Joseph E. King, 1962). Diversity in the behavior of Cichlids and their suit— ability to laboratory conditions have made them the subjects of many behavioral studies. Most such studies have involved aspects of social behavior such as sex recognition, court- ship, territoriality, parental care, recognition of young, and imprinting. The cichlid family is often divided into two groups according to the manner in which spawns are brooded. Al— though intermediate examples are known, the two general categories "substrate spawners" and “mouth brooders" or bond. Bot the fry. become som tive tende African, Ne Members of to sPawning In some Spe SFarming ty female Emit intermitte of 999$ an brooding p incubation free Swimm ”oral incubators" are useful. The first group is comprised of those species that lay their eggs on a substrate. The eggs remain there throughout the incubation period. The parents remain together in at least some degree of pair bond. Both the male and female fan the eggs and care for the fry. The pair usually remains together until the young become somewhat independent and begin losing their aggrega— tive tendencies. The substrate spawning habit is charac— teristic of almost all the American and a few African and Indian species. The two species used in this study, Agggig dggg latifrons and Tilapia sparrmani, belong to this group. The second group (mouth brooders) includes most of the African, Near East, and at least one South American species. Members of a pair generally form a loose sexual bond prior to spawning after which only one parent cares for the young. In some species this is the male; in others the female. Spawning typically occurs on a substrate. Typically the female emits a few eggs and this act is accompanied by intermittent ejections of milt by the male. The mixture of eggs and milt is then taken into the mouth of the brooding parent where the embryos remain throughout the incubation period. Even after the young are released as free swimmers, in some species the parent‘s mouth remains a place of are treated reported on 993% Ble and Aronson Parenta many famili of a protec over most 0 parental be The Cichlid actively de months. Th gation of c. reCOgnition the reCOgni were Perfor inate betwe, and, if so, made t0 det edge gained COmbinat ion a place of refuge when danger ensues, whereas in others they are treated as food. Detailed laboratory studies have been reported on the black—chinned mouth brooder Tilapia macro- cephala Bleeker by Aronson (1945, 1948, 1949, 1951) and Shaw and Aronson (1954). Parental care of spawn and young has been observed in many families of fish. This varies from the mere construction of a protective nest to complex nurturing behavior extending over most of the developmental period of the young. Most parental behavior of fishes lies between these extremes. The Cichlidae approach the latter extreme, since they often actively defend and herd a brood of young for two or more months. This extended period of nurturing permits investi- gation of certain problems of socialization and species recognition. One problem studied by various workers is the recognition of young by adults. Such investigations were performed to discover if adult Cichlids could discrim— inate between young of different ages or of different species and, if so, the mechanism responsible. Efforts have been made to determine whether this is an innate pattern, knowl— edge gained by the experience of rearing a brood, or a combination of the two° i I Lorenz ducks and tions have (Collies, Alaskan fu 1945). Noble cichlids o biocellat learn to r recognize t that adopts thereafter had been a] matings . Lorenz, as cited by Tinbergen (1952), reported that ducks and pheasants will incubate the eggs of other species but they kill the chicks when they hatch. This phenomenon has been referred to as adult imprinting. Similar situa- tions have been reported in the following mammals; goats (Collias, 1956) (Hersher, Moore and Richmond, 1958); Alaskan fur seal (Bartholomew, 1959); and sheep (Scott, 1945)° I Noble and Curtis (1939) attempted to determine whether Cichlids of the species Hemichromis bimaculatus, Cichlasoma biocellatum, Cichlasoma bimaculatum and Cichlosoma cutteri learn to recognize their young. They reported that parents recognize their young by size and color pattern, Pairs that adopted a brood of another species at first spawning thereafter failed to raise their own. Control pairs, which had been allowed to rear their own young at the first mating, ate the young of foreign species substituted at subsequent matings. Baerends and Baerends von Roon (1950) made only a few tests similar to those of Noble and Curtis, They studies Hemichromis bimaculatus, Cichlasoma bimaculatum, and Cichla- soma meekii and affirmed that ”the experiments of Noble and Curtis cle recognitio (1952) and imprinting reports of Curtis. The la changes inv Aeggidens schools” (; do not imp: or subseque presence 01 tion of the 0f the pare imPlies thz less of Sp) their appe phase. At evidence 0 Litre. Ci Curtis clearly show that imprinting does play a role in the recognition of young in Hemichromis" (p. 166). Tinbergen (1952) and Thorpe (1956) reiterated the concept of "parental imprinting," apparently basing their conclusions upon the reports of Baerends and Baerends von Roon and Noble and Curtis. 1 The late Dr. Greenberg (1961), (1963a) made various ex— changes involving eggs and young of Hemichromis bimaculatus and Aeguidens portalegrensis and reported that ”experienced as well as inexperienced pairs raised foreign spawn successfully in alternation with their own or were induced to brood mixed schools" (p. 143). Thus he concluded that these two species do not imprint on the species characteristics of their first or subsequent young. Greenberg did, however, postulate the presence of an internal mechanism which determines the dura— tion of the phases of parental care and thus the reactions of the parents to particular stages of young. This proposal implies that eggs, wrigglers, or free swimming young, regard- less of species, would be acceptable in those instances where their appearance and behavior matched the parentsI brooding phase. At the same time, however, Myrberg (1961) reported evidence of ”parental imprinting“ using Hemichromis bimacu— latus, Cichlasoma bimaculatum, and Cichlasoma biocellatum. He found of Hemich young wou concluded own speci Myrberg a and Curti parents b of their (1962) f0 and gegui_ Greenl non in Ci a used ggggé found that CLHEEELIL It Or mixture iterated h respect to negative, be imprint their own inVestiga He found that Cichlasoma biocellatum reared alternate broods of Hemichromis bimaculatus, but having once reared their own young would never accept a foreign species again. Thus, he concluded that g, biocellatum imprint only on young of their own species and called this an “imprinting—like" phenomenon. (Myrberg also accepted the deduction based upon the Noble and Curtis data which implied that Hemichromis bimaculatus parents become imprinted upon the first brood reared, whether of their own or a foreign species. Collins and Braddock (1962) found no ”parental imprinting" in Tilapia sparrmani and Aequidens latifrons thus supporting Greenberg's findings. Greenberg (1963b) tested the ”imprinting—like" phenome— non in Cichlasoma biocellatum as described by Myrberg but used Aeguidens portaleqrensis as the donor species. He found that supposedly imprinted pairs of Cichlasoma big— cellatum maintained schools of Aequidens portaleqrensis or mixtures of these with their own young. He further re— iterated his original conclusion by stating, ”. . . with respect to parental imprinting our results are decidedly negative, neither Hemichromis nor Aequidens pairs seems to be imprinted on their first or subsequent young, whether their own or foreign" (p. 142). Thus all of Greenberg's investigations tend to cast doubt regarding the presence of imprinting studies. that excha of the pa The w attempt t their par the recog awareness were esse it became not in all (1963a) (1 young. Su formed as results ar imprinting in any of the cichlid species utilized in his studies. Furthermore, his results support the hypothesis that exchanges must involve young matching the brooding cycle of the parents in order to be successful. The work reported here was originally designed as an attempt to determine whether young Cichlids "imprint" on their parents or on foster parents of other species or if the recognition of parents by fry is based upon an innate awareness of species characteristics. Fostering experiments were essential to these studies, and, as these progressed, it became apparent that their results were generally, but not in all respects, consistent with those of Greenberg (1961) (1963a) (1963b) with respect to parental recognition of young. Subsequent fostering experiments were then per— formed as an attempt to resolve these differences and their results are reported here. w because th of timing breed unde young diff provide cl fish and h are as fol trasted to latter als the forehe Study and four foste All 56 the labora be known, member of USed for E meht Of pa Pairs Who's ,q uOHOrS On] METHODS AND MATERIALS Aequidens latifrons and Tilapia sparrmani were chosen because their brooding cycles display marked similarities of timing and.behavior. Also, they are easy to rear and breed under normal laboratory conditions. Finally, the young differ sufficiently in appearance and behavior to provide clues for species identification by the parent fish and human observers. These differences in appearance are as follows: dorso—ventral stripes in Aequidens as con— trasted to uniform olive green body color in Tilapia. The latter also have three or four horizontal black bars on the forehead. Two old pairs of a third species, Hemif chromis bimaculatus, were available for a short time in the study and their broods were used as the donor species in four fOStering experiments. All adult fishes used in the experiments were reared in the laboratory in order that their complete histories might be known. When a pair was terminated by the loss of a member of the pair the surviving member was never again used for experimental purposes. This avoided rearrange— ment of pairs and confusion of brooding experiences. Those pairs whose spawning histories were not known were used as donors only. Thirteen aquaria witk and pair tar partments bl titioning 5) ments were 1 four in the ments were nithe 40 9 100 gallon A number of were used i developmenx hospital 0: gallon met The su aquarium g especially plastic g5 green Ple) the indivi dichloride Thirteen 40 gallon and four 60 gallon rectangular aquaria with slate ends and floors were used as pairing and.pair tanks. These were divided transversely into com— partments by inserting dark green opaque Plexiglass par— titioning sheets. When necessary, as many as six compart— ments were set up in the 60 gallon aquaria and as many as four in the 40 gallon aquaria. For test pairs the compart— ments were reduced to four in the 60 gallon size and three in the 40 gallon size. Two aquaria, each of approximately 100 gallon capacity, were used as holding or rearing tanks. A number of aquaria of assorted dimensions and capacities were used for housing young fishes in various stages of development and some adult donor pairs. Others were used as hospital or recuperation tanks. Fifteen of these were 2.5 gallon metal framed aquaria. The substrate used in all tanks was washed whitish aquarium gravel approximately two inches deep. Live vegeta— tion (Vallisneria) was planted but was constantly uprooted, especially by the Tilapia. This was replaced by clumps of plastic Vallisneriawlike plants made of 0.020 inch thick green Plexiglass. These were cut from Plexiglass sheets, and the individual fronds were fastened together with methylene dichloride and then twisted under hot water. Each 0: equipped w the filter about one these filt supplied t Early was suppli heaters. possible, tioner was then range Each 5 glass to } 5'! Ol‘erhg: large aqua double de( Pended re: lighting x of these 0th3r the thirtEEn 10 Each of the larger aquaria (10 gallon and larger) was equipped with an outside Le Bern filter using glass wool as the filtering medium. The water intake funnel was situated about one inch above the substrate. Air was supplied to these filters by a compressor unit. An air stone was also supplied to each aquarium and/or aquarium compartment. Early in the course of the observations each aquarium was supplied with one or more thermostatically controlled heaters. The temperature was held as near to 800 F. as possible. The heaters were discontinued when an air condi~ tioner was installed in the laboratory. The temperature then ranged between 80 and 820 F. Each aquarium was covered with double strength plate glass to keep evaporation at a minimum. Light was provided by overhead laboratory fluorescent units. Seven of the large aquaria which were housed as the lower tier of a double deck situation received additional light from sus— pended reflectors containing 25 watt showcase bulbs. A11 lighting was controlled by means of two time switches. One of these operated the overhead laboratory lights and the other the reflector lights. The lights remained on for thirteen hours (8:00 am - 9:00 pm) with some diffuse light entering Dark for spaw1 12 for dé were pla< holding t All c a day arc shrimp tw was Suppl as soon a was froze aVoided e: tated the Attem} thESe we“ range Of E was not 1 Q6 ll entering through the covered laboratory windows. Dark green Plexiglass plates (1/8 x 4 x 6 in.) were used for spawning sites in the pair and pairing tanks (see page 12 for definitions). Flower pots, both whole and halves, were placed on their sides in pairing tanks and the large holding tanks for places of refuge and spawning. All of the fish except the developing fry were fed once a day around midday. The young were fed newly hatched brine shrimp two to three times a day for the first month. This was supplemented with Wardley's Supremix Medium fish food as soon as the size of the young fishes permitted. The larger young and adults were fed a mixture of a ground raw shrimp, oatmeal, and Wardley's Supremix Medium. This food was frozen in patty form and fed in small chunks. This avoided excessive decay of uneaten foods and thus facili— tated the maintenance of clean water° Attempts were made to control the pH in the tanks, but these were quite unsuccessful and were abandoned. The range of pH in all the tanks was between 7.4 and 9.2. It was not determined if pH differences had any effect on spawning frequencies. Pairs were established by selecting five or six fish of the desired species from an overcrowded holding tank (over—err them in a patibilit either tr pair at a fishes in adjoining members 0 was found the same pair was energy to one four 1 Opaque Ple (I) an €XCelle of injury C‘C‘ttfact Q Separat lop t"e Pair e p: [b I} ~ Gen \‘ sometimes (- “Wetter w 12 (over—crowding inhibits fighting and spawning), and placing them in a pairing tank. As soon as courtship or close com— patibility was observed between two individuals, they were either transferred to a pair compartment (holding only one pair at a time) or were merely isolated from the remaining fishes in the pairing tank by herding the other fish into adjoining compartments. Hostility often broke out between members of a pair with resultant injury to one member. It (was found that the sight of another individual or pair of the same species somewhat alleviated this problem since the pair was then able to discharge some of their aggressive energy toward them. This situation was provided by cutting one_four by four inch opening in the bottom center of each opaque Plexiglass tank divider and inserting a glass pane. The use of these windowed compartments also served as an excellent means of separating members of a pair in cases of injury or incompatibility while maintaining their visual contact. Such maintenance of visual contact throughout the separation period tended to make the re—establishment of the pair easier. Reintroduced members of a broken pair that had been visually isolated during separation fought at first, sometimes to an extent that made it impossible to leave them together without injury. In contrast, reintroduced pair members ‘ courting The V fosterim with the aquarium herded tl secured 1 If eggs V removed soon as t were 311C Containir howeve r I own E'Ou n0 members that had been in visual contact while separated began courting and in most cases spawned within a few hours. The windowed compartments proved useful when performing fostering experiments. All exchanges of broods were made with the brooding test pair held in an empty adjoining aquarium compartment. The pair in question was gently herded through the opening in the aquarium divider and secured behind it by insertion of an opaque plastic plate. If eggs were being exchanged, the filial plate of eggs was removed and replaced by the plate of non—filial eggs. As soon as the pair resumed relatively normal activities they were allowed to return to their original compartment now containing the non-filial eggs. The majority of exchanges, however, involved larvae which subsequently shall be called wrigglers. The pairs were treated as in the egg exchanges, but the wrigglers were transferred by means of a siphoning process. This involved the use of a length of 3/16" inside diameter clear flexible plastic tubing as a siphon hose, and two clean battery jars. The wrigglers of the donor pair were removed and trans— ferred to the foster pair subsequent to the removal of their own young. The siphon was started and the receiving jar elevated t‘ 91ers. T‘ sacs and 0‘ the foster They were filial wri< adjust the As soon as the pair we procedure 5 holding COT Immedie haVior of 1 unless the Isrood was mtervals, irregularl§ :est day. to the test 14 elevated to allow a very slow gentle flow or water and wrig— glers. The wrigglers were quickly checked for ruptured yolk sacs and other noticeable injuries before being introduced into the foster aquarium and any damaged individuals were Removed. They were then siphoned into the same pit nest from which the filial wrigglers had been removed. Much care was taken to adjust the siphon flow to a rate that would not cause injury. As soon as the non—filial wrigglers were behaving normally the pair was released from the holding compartment. The entire procedure starting with the placement of the parents in the holding compartment usually took about fifteen minutes. Immediately after the exchange of eggs or young the be— havior of the foster parents was observed for twenty minutes unless the introduced brood was eaten at once. If a foster brood was accepted, brief observations were made at half hour intervals, for the next two hours. These were followed by irregularly spaced checks throughout the remainder of the test day. Periodic checks continued over the days subsequent to the test day until developmental stage changes occurred. If the young were present through the one day free—swimming stage they were checked thereafter only at the general morn— ing laboratory check and at feeding times. Foster broods were never removed until their members were well into the free-swimminr tion of yount developmenta The sole foster paren (eaten) or a young were c hatching. ; to those 6111' that :he “01' mat“)! the s Exchanged . tion on the 15 free—swimming stage° This was done to determine if rejec— tion of young was more apt to occur at any particular developmental stageo The sole criterion of treatment of young by actual or foster parents was simply whether the young were rejected (eaten) or accepted and cared for in the normal manner° The young were considered to be rejected as eggs if eaten before hatchingo All interspecific exchanges of eggs were limited to those differing in age by only a few hourso This insured that the non-filial eggs were scheduled to hatch at approxi— mately the same time as the eggs for which they had been exchanged. This prevented the possibility of discrimina— tion on the basis of an innate awareness of the incubation perioda Determination of rejection and acceptance of broods at the wriggler stage was based upon a survival time of 24 hours, If a brood was not eaten within 24 hours post ex- change, it was assumed that it had not been treated as foreign and was accepted, If the wrigglers were eaten within 24 hours, the brood was considered to be rejectedo Where a mixed brood was involved, rejection was recorded only if it was certain that the non—filial wrigglers had not been destroyed by the pair's own young, Mixed broods were of two typeso One of these res removal, glers and wrigglers with near second ty for two 5 Only wrig brood of Only in t the possi This was Cont: trol exis 16 these resulted from inadvertently overlooking young during removal, thus producing a brood consisting of filial wrig- glers and a larger number of newly introduced non—filial wrigglers. In a few instances mixed broods of this type, with nearly equal numbers, were assembled on purpose. The second type of mixed brood occurred where a pair was caring for two successive broods of its own young at test timeo Only wrigglers were exchanged and this resulted in a mixed brood of filial free swimmers and non—filial wrigglerso Only in the latter instances was it necessary to consider the possibility of predation by young on the test broods This was never observed. Controls were established in two ways, A natural con— trol existed when young of both the filial and non—filial broods were present since here the parents were provided with a basis for discriminatory behavioro In those instances where a mixed brood was not present a control test was per— formed as follows, When a non—filial brood was rejected, the pair was offered its own brood again by introducing them in the same manner as the test broodo In all instances where this control was employed the filial young were accepteda Special methods employed in the various experiments can best be described under the appropriate headings in the texto The t course of by activi character involves ing of cc from the entered :‘L‘Q ~-.L join male and Spawn l nq RESULTS General Breeding Behavior of Substrate Spawning Cichlidso The beginning of pair formation, as observed in the course of these studies and others, is usually initiated by activity on the part of the maleo His behavior is characterized by aggression toward other fish and this involves variable periods of chasing. A general darken— ing of coloration occurs and this makes him very conspicu— ous. The chasing bouts become increasingly shorter in distance and duration as an area of substrate is identi— fied as a territory, This territory usually centers around an object of spawning preference, which will usually become the Spot of eventual egg depositionc Interspersed with these aggressive displays is removal of mouthfuls of gravel from the future egg laying siteo After a gravid female has entered the territory, and attacks by the male have ceased, she joins in its defense, Aggressive displays by both the male and the female toward intruding fish alternate with digging at the spawning site. The tempo of digging increases as spawning approaches and is accompanied by activities more characteristic of courtship, eggs, “nipping” the spawning substrate, ”quivering,” and ”skimming” (Baerends l7 and Baer varies b ing foil immediat periods which wi they are nestso wriggler another; into the l8 and Baerends von Roon l950)° This prespawning behavior varies both with regard to species and individuals, Spawn— ing follows, and post spawning fanning of the eggs is then immediately begun by both parents“ Interspersed with these periods of fanning are periods of digging pits in the gravel which will later contain wrigglerso As the wrigglers emerge they are removed by both parents to one of the several pit nests. For the remainder of the pre-free swimming stage the wrigglers are periodically transferred from one pit nest to another° Whenever they stray they are retrieved and spat back into the nesto As the young approach the free swimming stage the parents” retrieving activities become ever more hectico At one day post free swimming the young swim well enough to maintain a compact school and this appears to result from their own activities rather than those of the parents. The school moves about in a mass and the parents keep it away from any other large fish that may be present in the aquariumo Three to five days of free swimming elapse before any notice— able response to parental signals is observed in the youngo After this they begin to follow the parents” Up to the first five days of free swimming the young spend most of their time hovering near the bottom but subsequently begin moving to higher levels, usually in the vicinity of the parents themselve considera young slc 25 days 5 parental val and 1 other ad1 :1 few pa reSUltin 3‘ inter little The ClEarlv 19 l 7 i ? themselves, After about ten days they can often be seen considerably above the adults and even at the surfaceo The young slowly become independent of the parents and at 20 to 25 days show very little response to parental signalso The parental interest also shows a waning throughout this inter- val and this allows the young to be subject to predation by other adult fish in the tanko The Reproductive Sequence of Aequidens latifrons The breeding and nurturing behavior of this species conforms closely to the general pattern of substrate spawn— ing cichlidso Their behavior differs from the mode primar- ily with regard to its marked placidness. Members of pairs were very compatible and often showed little aggression toward each other during months together in the pair tankso A few pairs, however, were terminated by death of a mate resulting from sudden outbreaks of fighting during court— shipo In most of these cases death was apparently caused by an internal injury as the result of buttingo Usually very little in the way of external wounds was produced other than frayed finso The approach of spawning in established pairs was clearly indicated by certain characteristic changes in the female‘ in the Spawni The C01“ pair 0f d< striping ( appeared i brightenil the sides thronghou fadlng an the pairi accompan: only rare quiverin vertical u vigoro members the subs as spawr 20 female: increased coloration, increased time spent clearing the spawning site, and obvious protrusion of the ovipositor. The color change was characterized by brightening of the single pair of dorsolateral ocelli and the darkening of the lateral striping consisting of blue—black bandsa The interstripes appeared paleo Further color changes involved a noticeable brightening of the iridescent blue streaks of the cheeks and the sides of the rostrumo This intense coloration remained throughout the first two weeks of brooding and then began fading until a condition of maximum pallor was reached° In the pairing tanks with other fish present the male first set up his territory and initiated courshipa Here, in the absence of intruders, it was the female who initiated courtship. She cleared a spawning site and then attempted to attrack the attention of the male by gentle nudging at his sides and head accompanied by a quivering motiono Mouth fighting occurred only rarely in response to this activityo Usually such quivering on the part of the female was followed by a near vertical head down position and nipping at the substrate with a vigorOus twist at the finish of each nipo In time both members of the pair could be seen quivering and nipping at the substrate simultaneously in response to one anothero As spawning approached the sequence became one of quivering, F nipping t] essential site on t] strate wi' egg depos, onset but during ac took placs teen with pelvic fix that the I OViPOsito: was no: t; egg—law ‘ H i -Lgs c [h n- altematm depositing eggs Simij prespawnir the incub; embrYOS we eggs Per 5 21 nipping the substrate, and skimming. The latter, which is an essential element of spawning, involves passage over the nest site on the part of the female. She actually touches the sub— strate with her belly. This occur prior to, as well as during, egg deposition. The skimming act involved quivering at the onset but became more steady as its frequency increased until during actual spawning only a very slow, deliberate skimming took place. The eggs are deposited in strings of five to fif— teen with the ovipositor just touching the substrate and the pelvic fins dragging passively along. Breder (1934) reported that the pelvic fins aid in squeezing the eggs out of the ovipositor during egg laying in Aequidens latifrons. This was not the case here. Spawning consisted of one to three egg—laying passes by the female with intermittent spray— ings of milt over the eggs by the male. The pair rarely alternated exactly during spawning. Subsequent to an egg depositing run the female would often direct nips at the eggs similar to those directed toward the substrate during prespawning courtship. This nipping continued throughout the incubation period. During this time dead and dying embryos were removed and probably eaten. An average of 372 eggs per spawn was counted for one spawning of each of ten differen (total 1 with the previous sisted 0 eggs for Spaw the pair shared b1 a time. were mort Over the ShallOw i be plaCec female as Eterage I fanning t was no 8] Spawnings 22 different females varying in length from 5.4 cm. to 9.5 cm. (total length). The size of the spawn obviously correlated with the size of the female and the lapse of time since the previous spawn, if any. The largest count in this study con- sisted of 609 eggs. Breder (1934) reported spawns up to 486 eggs for this species. Spawning lasted about one hour immediately after which the pair began fanning the eggs. The fanning duties were shared between the parents with only one members active at a time. (Pairs rearing eggs in the presence of other fish were more cautious during the period when one member took over the fanning duties from the other than were those rearing spawn in a pair tank.) As the nonfanning member approached the egg site the fanning member would quickly dart away and out toward the periphery of the territory. The off—duty parent spent its time foraging and digging shallow pits in the gravel in which the hatched young could be placed. An analysis of the fanning bouts revealed the female as the most active in this phase of care with an average fanning time of 26.8 minutes per hour. The average fanning time for the males was 19.1 minutes per hour. There was no significant change in fanning time with subsequent Spawnings (Figure l). Fanning time per hour for the males Figure l. Fanning Time Per Hour In Minutes Ave ra ge 3nd femal 23 Figure l. ,Fanning Time in Minutes Per Hour for Three Pairs of Aequidens latifrons. Average Male and Female Fanning Times and Their Totals for Each of Eight Subsequent Spawnings. 50 H 45 F 9+O‘}?=45.9 (1) Q4 40 _ (1) 5% 35 — 94-) mg 30 - .3}; 9 2:26.23 E c 25 ' £3“ - A _ o S 20 - /// \‘\ ’,-——' “‘~~ 5 x :1901 v, mo gm 15 g - 5 O _ J I I I I I I I Sequence of Spawns and females of three pairs over a sequence of eight spawn- ings is shown. These data represent the averages of two one-hour timings for each pair at a given spawning experi» ence. All timings were made in the afternoon of the day after spawning with a one hour interval between the two timings. The effect of fanning duration upon hatching found that the was not determined. Van Iersel (1953) duration of individual fanning bouts rather than the overall duration fanning was very hour (Fi Hatc imately long (Ta the pare previous Pit youn Onset of irbitrar is very SWimming Swimm;3£ forward free Swi Of the 3 Thrc guarding Ally that 24 duration of fanning influenced hatching. The totalled fanning times of all three pairs at each spawning time was very consistent with an average of 45.9 minutes per hour (Figure l). Hatching occurred after an incubation period of approx— imately 51 hours. The newly hatched wrigglers were 2.4 mm long (Table 17). They were removed from the egg site by the parents and transferred directly to one of the pits previously prepared. All were placed in the same pit. This pit young or wriggler stage lasted from 92 to 98 hours. The onset of the free swimming condition was identified in an arbitrary way since the transition from the wriggler stage is very gradual. Thus, the young were said to be free swimming as soon as they were able to maintain a normal swimming position ab0ve the substrate and propel themselves forward a short distance. In most instances these early free swimming movements were very brief due to the weight of the yolk sac. Throughout the wriggler phase the parents took turns guarding the nest and maintaining the young in a compaCt mass. Periodic mouthing of the young in the Pit occurred. Any that strayed were immediately retrieved by one or the other of the parents. At irregular intervals they were transfe ible pa wriggle. particn each la] the pare laboratc time in separate €Xperime picked u natural see whet is certa time and Parents. The a 409 mma mine the 25 transferred from one pit to anothero There was no discern— ible pattern of transfer or use of pits except that all the wrigglers were housed in only one pit at a timec Both parents participated in these activitieso Breder (1934) stated that each larva was carried separately to another pit and that the parents often carried young to separate pits, In our laboratory pairs carried more than one stray wriggler at a time in certain instances where the young were deliberately separated or when exchanges of young were made in fostering experimentsa Either of the parents en route to the pit picked up and carried any wrigglers encounteredo During natural transfers to a different pit it was impossible to see whether one or more young was being moved at a time. It is certain that only a few wrigglers were transferred at a time and that these were placed in a common pit by both parents. The average length of the newly free swimming young was 4‘9 mmo (Table 17). During the early hours of free swim— ming the parents kept the fry in a compact school near one The fry remained near the bottom of the tank of the pits° and in a school for almost five dayso Thereafter they began responding more to parental movements and frequented higher levels of water, mostly in the vicinity of the parents. free—swim response i other timi days. It parent re result fr as report m this spec The sharp reached i larges: L' Our fry, CY Baerer Visual a< this may KUhme (l< to visual: Ma Th. and ram As t‘ were Str 26 } parents° During the interval from five to twelve days post free—swimming the fry exhibited a greater degree of fright response to movements outside their tank than during any other time. This reached a peak at about eight or nine days. It was at this same time that a maximum young to parent relationship seemed to exista This interplay may result from the late maturation of visual acuity in the fry as reported by Baerends, Bennema and Vogelzang (1960) for Aguidens portalegrensiso They found the visual acuity of this species to vary with the body length and thus with age. The sharpness of visual perception in their test situation reached its peak in fishes nine centimeters long, The largest used were 1105 cmo, while the smallest were 390 cm. Our fry, although much smaller than the smallest observed by Baerends et al (1960) may have undergone significant visual advancement at the age of eight or nine days, and this may account for their fright behavior at this age. Kfihme (1962) also reported an age dependent fright reaction to visual stimuli in the cichlid fish Cichlasoma biocella— Egm, The fright response reached a maximum at seven days and returned to normal at twelve dayso As the fry approached 20 days post free swimming, they were strong swimmers and largely independentg Each went its own way This sit \ the aqua \ curred i l The among pa Spawning moval of or two ; noted mc Several than fit a ten mc Several spawned 0Ccurrec in Othel spawns V Spawn w} CeSSiVe Single i hibited 27 own way in the tank, and thus schooling behavior diminished. This situation may have resulted from the limited size of the aquarium and some degree of aggregation might have oc- curred if more space had been available° The interval between spawnings was extremely variable among pairs but was relatively consistent within a pair. Spawning usually occurred within two weeks after the re- moval of young, but instances were recorded where a month or two passed before another spawning took place. This was noted more frequently where the older pairs were involved. Several pairs had spawning histories consisting of more than fifteen spawns. One pair spawned eighteen times over a ten month period thus averaging one every seventeen days. Several pairs ate their free swimming young and then spawned again within a matter of hours. This behavior occurred quite consistently in a few pairs and occasionally in others. One interval as small as six days between spawns was observed. It was not unusual for a pair to spawn while still rearing a brood. As many as three suc— cessive broods of young were observed being reared by a single pair and appropriate nurturing behavior was ex— hibited toward each of the three. Th The I many res; Cichlids. species a aggressix aggressr patibili bility w few broo spawning subseque gressior first 3F tending Spawning duties, warded ( fanning each su Vicious approac Wise d‘l 28 The Reproductive Sequence of Tilapia sparrmani The reproductive behavior of this species varies in many respects from that of the typical substrate spawning Cichlids. The primary behavioral difference between this species and Aequidens latifrons is its greater intraspecific aggressiveness and greater general activity. The factor of aggressiveness resulted in a very short period of pair com— patibility as compared to that of A. latifrons. Compati— bility was greatest during courtship generally and the first few brooding periods but became much less in subsequent spawnings. The decrease in duration of compatibility in subsequent spawnings was largely the result of increased ag— gression by the female while tending a spawn. During their first spawning the members of a pair would share the brood— tending duties with very little incident. During later spawnings the female attempted to monopolize the brooding duties, especially during the egg fanning period. She warded off the male when he attempted to take his turn at fanning (Figure 2). Female aggressiveness increased at each successive spawning with her attacks becoming more vicious and more frequent. The only time the male could approach the nest was when the female was feeding or other— wise distracted from guarding it. As soon as the male was WOUDCeZ CH H50: ,HUQ 05H? GCHCCQL UFGHO>< terminate: The majori siderably E n . 2 d e e C r .l U t g o .l n F 29 noticed near the eggs he was driven off. Many pairs were terminated when the male began responding with attacks. The majority of the pairs contained males which were con- siderably larger than the females and thus fights usually Figure 2. Fanning Time in Minutes Per Hour for Three Pairs ‘ of Tilapia sparrmani. Average Male and Female Fanning Time and Their Totals for Each of Six Subsequent Spawnings. 55 —‘ H 50 _ 9 + d x = 46.8 s o 31'. 45 - 23 g. 40 — s )2 = 31.3 g m 35 r -H o E *5 own 30 - as a 25 c s (U H \ In. 20 — x‘ A g \xvx/l \\\ a 15 _ \\\ l»———— d )2 = 15 5 m V” > 10 ‘ 4 5 _ O i l 41 I l I I Sequence of Spawns ended in progressi pairs end terminate apparent] have ends separate< was save< had been nearly e followed Such Out aggressi when the resPOnse 3O ended in fatal injury to the female. As a result of this progressive increase in female aggressiveness most Tilapia pairs ended after a few spawnings. Twenty four pairs were terminated when the male killed the female, and this was apparently the result of this situation. Other pairs might have ended in the same manner if their members had not been separated in time. Many occasions arose where an individual was saved by separation of a pair before extensive injuries had been incurred. One male was killed by a female which nearly equalled the size of the male. This pair was only in its second brooding experience. A second critical period of aggressiveness sometimes followed periods of compatibility in the absence of a brood. Such outbreaks were Characterized by an increase in male aggressiveness but diminished in intensity and frequency when the female responded with courtship. Aggressive responses on the part of the female however, led to con— tinued attacks by the male. Five pairs were terminated in this manner. Four of the individuals that died as a result of precourtship aggression were females. The single in— stance in which a male was killed by the female during pre— courtship occurred after only two brooding experiences. .___'__—-_—_— Crii could be ness by Thus, ml damage The with re with la In cont by SEpe seriOug 31 Critical periods of increased aggression of this nature could be predicted early on the basis of lack of responsive- ness by the female to the courtship behavior of the male. Thus, members of pairs could be separated before extensive damage occurred. The aggressive nature of Tilapia sparrmani, particularly with regard to the despotic nature of the female broodiness with later spawns, led to rather short spawning histories. In contrast Aequidens latifrons pairs wEre rarely terminated by death resulting from fighting. The greatest number of spawnings by any one pair of Tilapia was ten. In the cases of the longer spawning histories the pairs were maintained by separation during highly aggressive outbreaks. The most serious fighting between pair members usually took place dur— ing the second day after spawning. If the pair was main— tained until after hatching without the male counter—attacking, the aggressive displays were suppressed somewhat. The fe— males directed increasingly less aggressiveness toward the males as they became increasingly more occupied with re— moving the hatching young to a pit nest and retrieving them as they spilled over the edges. The males were then often able to contribute to this activity with little difficulty. Wit tioned, i conform | patterl specif: but we] spawnir quiver: ing the were of exagge] in the This 5; gravel, the gra COHSide the glg of gra\ the gre the Obs ad30in: direct6 partit; 32 With the exception of the differences previously men— tioned, courtship and spawning behavior of this species conforms quite closely to that described for the general pattern of substrate spawning Cichlids and that described specifically for A, latifrons. Exceptions have been noted but were mainly of a quantitative nature. Thus, pre— spawning behavior of this species was nearly devoid of the quivering element of the typical courtship pattern. Nipp- ing the proposed egg site, skimming, and gravel digging were of the ordinary type except that the latter was mudh exaggerated and seemed to replace the quivering response in the prespawning sequence as exhibited by A, latifrons. This digging consisted of taking up large mouthfuls of gravel, rushing toward the aquarium glass, and expelling the gravel about four to five inches above the bottom with considerable force. The pebbles were actually blown against the glass, and this resulted in the formation of a ridge of gravel on the bottom of the aquarium. The majority of the gravel was placed against the aquarium glass nearest the observation post. When a pair occupied a compartment adjoining that containing another pair, the gravel was directed toward the 4” by 4" glass window in the separating partition. This behavior should be useful in future investi furnisl Dig and th: When a took p the ba Sp erence A.M. differ Sequer which A reas Spawn ten f5 lengt1 was p; Size ( feren Cichl that 33 investigations involving species recognition where it could furnish a measure of the stimulus value of various species. Digging and gravel moving were exhibited by both sexes, and this produced a large bare area around the spawning site. When a spawning slate was used, a large amount of digging took place around the slate, and eggs were often placed on the bare tank floor rather than on the slate itself. Spawning occurred at many times of the day, but a pref— erence was shown for the morning daylight hours before 11:00 A.M. The actual process of spawning was not appreciably different from that described for Aequidens pairs. The sequence, however, lasted only about 35 to 45 minutes, which in many instances was only half as long as in Aequidens. A reasonable explanation is that the number of eggs per spawn was consistently less (§-= 231 eggs per spawn for ten females with a size range of 6.7 cm. to 11.7 cm- total length). The spawn size ranged from 158 to 329 eggs which was proportional to the size of the females. The spawn size of this Tilapia species was also considerably dif— ferent from that of a related Tilapia substrate spawning cichlid as reported by El Zarka (1956). He indicated that spawns of Tilapia zilli number from 1,000 to 1,500 eggs. mani Md mentior counts are re] consist eggs of nature. AS began i for mos no long he dug flnd ft 34 eggs. I, zilli attains a much larger size than T, spagg— mani which probably accounts for this, although he did not mention the actual sizes of the females from which his counts were taken. The individual eggs of T. sparrmani are relatively large and olive green in color which is consistent with what El Zarka reported for T. zilli. The eggs of A, latifrons are small and of an almost transparent nature. As soon as spawning was completed the female immediately began fanning the eggs whereas the male continued digging for most of the first day of incubation. This digging was no longer limited to the vicinity of the spawn. Instead, he dug random scattered pits in the gravel. These would find future use as pit nests for hatched young. As the second day of incubation approached, the male made a greater effort to fan the eggs. As mentioned earlier, and as shown in Figure 2, this was the time when the female became more aggressive with successive spawns. The duration of incuba- tion was not different from that of A. latifrons. The average of ten such incubations at 800 F. was 53 hours. El Zarka (1956) reported a 74 hours incubation period for three spawns of T, zilli at temperatures of 75, 80 and 820 P. which is considerably longer than observed for T, sparrmani. as they ‘ transfe | on the l directe ically fers we Then, i new p11 Parent: 91ers ; Activi 35 The newly hatched wrigglers averaged 3.6 mm. long (Table 17). They were removed to a single pit nest by the female as they hatched. The male rarely took part in this initial transfer of wrigglers. At this time the females' attacks on the males diminished as their attention was increasingly directed toward the active wrigglers. Both parents period— ically moved all the wrigglers to another pit nest. Trans— fers were made at irregular intervals and had no discern— ible pattern. The only situation observed as a possible triggering mechanism for a transfer of wrigglers occurred when a few aimlessly moved into an adjoining pit nest. Then, the pair would often move the entire brood to this new pit, carrying several wrigglers at a time. Thus, parents apparently were unable to recognize numbers of wrig- glers or to distinguish one nest from another. Some other triggering phenomenon must exist since irregularly spaced transferrals also occurred over long distances. El Zarka (1956) reported that the larvae of I, gilli_moved from pit to pit without the aid of the parent. Although the wrig— glers have head glands by which they attach to the substrate, several members of a brood may always be observed wandering. Activity increased as the free swimming stage neared. This l .. allowed with th the fem used ir bites c ing it young 5 patible wriggh F. Fr length school if the retrie then s the sa mt their 36 allowed the males to enter into more frequent encounters with the wrigglers even though interference on the part of the females still occurred. Most of the males' time was used in digging and moving gravel. Often they took large bites of gravel and mumbled it in the mouth before releas- ing it as if sifting for food. The termination of the pit young stage marked the beginning of one of the most com— patible periods for brooding pairs. In ten broods this wriggler stage lasted approximately 76 to 82 hours at 800 F. First day free swimming fry averaged 6.5 mm. total length (Table 17). They were maintained in a compact school in the vicinity of one or more of the pit nests if these were close together. Wandering individuals were retrieved by either member of the brooding pair and were then spat into the school. Subsequent events were much the same as noted for A, latifrons except that the If spggr- magi fry were larger and much stronger swimmers. Thus, their schedule was about two days ahead of that for A. latifrons fry. Spawning intervals varied considerably but in unmolested "good” pairs 25 to 35 days usually elapsed between succes— sive spawnings. Spawning sometimes took place in the presence of a re: tagonim other he been ea was not teen da observe creased months. (I) ((1 ((1 8 W6 a “erg (I 37 of a resident brood but in such instances considerable an— tagonism was usually shown toward the older Young. On the other hand, where broods were removed from a pair, or had been eaten early in a particular spawning experience, it was not uncommon for spawns to occur consistently ten to four— teen days apart, and an interval as 10w as eight days was observed. Spawnings occurred throughout the year but in— creased in frequency furing the spring and early summer months. This was possibly due to the increasing diffuse light entering the covered laboratory windows at that time. Parental Recognition of Eggs Aequidens latifrons: Ten brooding pairs of Aequidens latifrons were tested with eggs of Tilapia sparrmani. Two of these were tested twice giving a total of twelve tests. A distinct discrimination against Tilapia eggs was observed (Table 1). In nine tests the eggs were eaten within a few hours. In the other three, they were reared for 20, 23.5 and 28 hours before being eaten. Each of the last three figures are estimates within two hours of exactness. When possible, exchanges of eggs were made within twelve hours after spawning. Green— berg (l963a) found that foster eggs presented to a pair early in the Foster wriggl wriggl ; ' eggs a of the One te the pa Table ~—___ Previc broodi exper: \ C) C) 43 cl CD P4 9) r) k4 :1 JL (3 ha CD CD 0. X- )F *Nona Contr S:JUS 38 in their fanning phase were accepted and reared normally. Foster eggs presented to a pair in the late egg-care or wriggler tending‘phase usually were eaten or treated as wrigglers. In these investigations pairs never treated eggs as wrigglers, i.e., transferred them to a pit. Six of the twelve tests involved the first spawning experience. One test involved the thirteenth brooding experience of the pair. Table 1. Reaction to Tilapia sparrmani and Hemichromis bimaculatus Eggs by Brooding Pairs of Aequidens latifrons. Previous Age of eggs in days at the brooding time of exchange Remarks experience Age Filial Non-filial difference 0 1 1 0 Fanned 23.5 hrs. 0 1/2 1/2 0 Eaten immediately 3 1/4 1/2 +l/4 Fanned 28 hrs. 4 l l O Eaten immediately 0 1/2 1/2 0 Few fanned 20 hrs. 6 1/4 1 +3/4 Fanned 3 hrs. 0 1/2 1/2 0 Eaten immediately 0 1/4 1/4 0 Eaten immediately 12 S 1/4 +l/4 Eaten immediately 0 1/4 1 +3/4 Fanned 3 hrs. 4 1 1—1/2 +l/2 Eaten immediately 2 2 1—1/2 —l/2 Eaten immediately *1 1/4 1/4(J) 0 Fanned 29 hrs. *2 1/2 l/2(J) 0 Reared 14 days *3 l 1-1/4(J) +l/4 Reared 10 days *Non—filial eggs of jewel fish (Hemichromis bimaculatus) Controls: All cases of eggs returned were accepted S=Just spawned Cont turning eaten. not useé been use Thre pairs a1 the bro swimmin same pa were re they we Sex Three Older Velved Cuttir eggs t ‘Lwh i Q] on the 39 Controls employed in nine cases simply consisted of re— turning the pairs' own eggs after the non—filial eggs were eaten. All were reared without incident. Controls were not used in the other three tests, since the filial eggs had been used in other experiments. Three tests were made involving Aequidens latifrons pairs and Hemichromis bimaculatus eggs (Table 1). Two of the broods were accepted and reared well into the free swimming stage. Both of these tests were made with the same pair. In the third test situation the foregin eggs were reared for at least 29 hours but were eaten before they were due to hatch. Seven tests involving five adult pairs of Aequidens latifrons were made with non—filial eggs of the same species. In all of these the substitute eggs were accepted (Table 2). Three of these tests involved exchanges of non—filial eggs older by at least one day than the filial eggs. One in— volved an age difference of two days. This resulted in cutting short the fanning phase. The pair brooding the eggs two days older than their own fanned the wrigglers (which all hatched in five hours) while they were still on the slate. The wrigglers were not carried to a pit nest unti involved least one similar 5 She spawr fanned t} abandonix glass wlI Ten < from the minutes ( successfl Table 20 \ PfelegS brooding experiem \ ,_.. r.)<fitA)O'\r—‘ru y—J 4O nest until some time the following morning. Two of the tests involved exchanges of eggs younger than the filial eggs by at least one day and those exhibited an extra day of fanning. A similar situation was observed in a non-experimental female. She spawned while separated from her male pair member and fanned the unfertilized dead eggs for 3—1/2 days before abandoning them. She was able to see the male through a glass window throughout the separation period. Ten controls were employed by separating the parents from the filial eggs and reintroducing them after fifteen minutes of separation. All ten controls were reared successfully. Table 2. Egg Exchanges Made Within the Species, Aequidens latifrons. Previous Age of eggs in days at the Days retained brooding time of exchange post exchange experience 7 Age Filialeon—filial difference 1 l~l/4 1/4 -1 14 2 1/4 1-1/2 +1—1/4 l6 1 S 1 +1 5 6 1/4 1/4 0 36 13 l S -l 10 5 S 2 +2 10 2 1—1/2 1 -l/4 —l/4 18 S=Just spawned Ten eggs 0f total 0f ‘ ination The ages matched. were eat their in for abou four cas eggs wer mm pair eggs pri tests Wd own eggs all but returned 41 Tilapia sparrmani: Ten brooding pairs of Tilapia sparrmani were tested with eggs of Aequidens latifrons. One was tested twice giving a total of eleven tests. In all of these an obvious discrim— ination occurred against the heterospecific eggs (Table 3). The ages of the filial eggs and non—filial eggs were closely matched. In six of the test situations the non-filial eggs were eaten in a very voracious manner immediately after their introduction. In a seventh case they were mouthed for about ten minutes and then swallowed. In the remaining four cases variable intervals of time elapsed before the eggs were eaten. The maximum time involved lay between four and six hours. Four of these tests involved inexperienced pairs, while seven involved pairs that had mated previously. One pair had had six brooding experiences with their own eggs prior to testing. The control employed for all eleven tests was simply the return and acceptance of the pair's own eggs after the substituted eggs had been eaten. In all but one instance this reSulted in the acceptance of the returned eggs. The other control was not a success because the male parent was killed by his mate and the nest was destroyed before the eggs reached the wriggler stage. Table 3. _— Previous brooding experier N.L>|——IJ>4>-OONOOJ> one to were y( the fi Stitut 42 ‘ Table 3. Recognition of Aequidens latifrons Eggs by Brooding " Pairs of Tilapia sparrmani When Exchanged for Their in age of the eggs at the time of exchange. This varied from one to 1-1/4 days. In three instances the non—filial eggs were younger, while in the other two they were older than Own Eggs. 1 Previous Age of eggs in days at the brooding time of exchange Remarks experience Age Filial Non—filial difference 3 4 1/2‘ _ 1 +1/2 Eaten within 2 hrs, 0 1/4 1 +3/4 Eaten immediately 0 1/4 1 +3/4 Eaten immediately 6 2 2 O Eaten immediately 0 l 1 0 Eaten immediately 4 l/2 1 +1/2 Eaten within 4 hrs. 4 1/2 1 +1/2 Eaten within 6 hrs. 1 1/2 1 +1/2 Eaten immediately 4 2 1 —1 Eaten within 1 hr. 2 1 1—1/2 +l/2 Eaten immediately Ten pairs were tested against non—filial eggs of their own species and all of these tests were successful (Table 4). In five of these experiments there was a distinct difference the filial eggs for which they were exchanged. When the sub— I stituted eggs were younger, the fanning period of the brood— ing cycle was extended beyond its normal length. In the two 1 instances where the non—filial eggs were older than the filial eggs the fanning phase was shorter than normal, since wrigglers arrived earlier than would otherwise have been the case. . Fiv involve eggs fo not dis ‘ introdt Table 4 Previor broodir experie mwwc‘JI—‘I—Ji—J/ O ,_. (II 43 Five controls were employed for the above tests. These involved the temporary separation of the pair from their own eggs for fifteen minutes in order to determine whether or not disturbance alone would affect the responses to the re— introduced eggs. All five were successful. Table 4. Egg Exchanges Made Within the Species Tilapia sparrmani. Previous Age of eggs in days at the Days retained brooding time of exchange , post exchange experience Age Filial Non—filial 'difference 1 S 1 +1 17 1 1/4 1—1/4 +1 15 l 1 1/2 -l—l/4 19 O l S —l 6 3 1—1/4 S —1-l/4 12 1 1/2 1/2 0 8 2 3/4 1 + 1/4 27 0 1—1/2 1—1/4 0 14 5 S S O 21 1 1 1 O 7 S=Just spawned Egg Measurements The results observed in the egg exchanges suggested that a comparative analysis of size and shape of the eggs of the species involved be made. Twenty eggs were measured from each < the tr The 11 spawn: ‘ ocula: readi l the In ments of tin restr compu~ A shar width taken 44 each of five spawns of five different females for each of the two species, Tilapia sparrmanni and Aequidens latifrons. The length and width of the eggs were measured on the day of spawning by means of a binocular microscope containing an ocular micrometer so calibrated that the measurements were readily converted to the metric scale. All were taken to the nearest one—hundredth of a millimeter. These measure— ments were then utilized in computing the area and shape of the individual eggs as observed from the top in the normal resting position. Thus the sectional area of each egg was computed as the area of an oval (length x width x n/4). A shape index was also computed in terms of the ratio of width to length (shape=w/l). The same measurements were taken on 20 eggs from each of the two fl. bimaculatus pairs. The size of Tilapia sparrmani eggs was found to be much greater than those measured for Aequidens latifrons and Hemichromis bimaculatus (Table 5). Thus, a comparison of mean lengths and widths finds the eggs of the latter two species very similar, while those of Ty sparrmani deviate from them significantly. The mean length of the T. sparr- mapi_eggs was 1.76 mm. as compared to 1.24 mm. and 1.27 mm. for A. latifrons and fl, bimaculatus respectively. In con— trast the widths of the eggs of the three species exhibited 45 less difference. The mean width of the Tilapia eggs was 1.21 mm. as compared to 1.06 mm. for A. latifrons and 1.13 mm. for A. bimaculatus. The comparative egg sizes are illustrated by the area computations in Table 5.* The egg area measurements of A. latifrons and A, bimaculatus overlap considerably (means 1.04 mm? and 1.12 mm? respectively) and are completely separate from the T. sparrmani curve which has a mean area of 1.69 mm? The shape differences are shown in Appendix 2 as they departed from roundness or unity calculated as the relation— ship of length to width. The shape index curves of Aeggi— dggs and Hemichromis eggs overlap considerably. That of the latter species most closely approaches the round with a mean shape ondex of 0.892. The mean shape index of Aeggi— dens and Tilapia observations in Appendix 2 show overlap, and this can probably be attributed to deformed eggs re— sulting from the females attempting to wedge them between those already deposited. In such instances of overlap it appears that a few Aequidens eggs were elongated during the spawning process. * See appendices l and 2 for graphs showing dispersal of measurements. s I DWIHSUMEflQ mHEOHSUHEUE HO WOO 202,... Cliefimw W3 1% O W Ho Frau. CHCQEHHUQMM OH, 044.1: UGO ii;i|. .awmflw. TH WCUC a DWQAN :17 fTCHAN 315?...“ 11:. . \ .u_1.__ orb mN¢0.0 + 00.0 HmH.0 + EE0©.H 46 ,mHm0.0 + 00.0 haaoo + NEENH.H mm¢0.0 + 00.0 wmmoeo + NEEv0.H Xopnfl ommfim mon< shoema\snoas u woven mangoes anew spans x chosen u enema 0000.0 + .EEHN.H 0000.0 + m0®d.0 + .Efimaefi 0mm0.0 + 0©¢0.0 + .EEOO.H NH®0.0 + been: .flcmEHHMMm Mammafle psm .msumesomfiflfl mHEonnoflEom .mQOHMHumH msopflomofl mo mmmm mo mooflch ommfim pew mmoum coo: mSB .Efiomoa fismfissmmm mHmmHHB .EEBN.H mopmasomfifla mHEOHQUHEom .EEem.a maoemanma msopfismom summon .m OHQmB Aeguidc a} 47 Behavior of Adult Brooding Pairs in Tests Involving Exchanges of Young in the Wriggler Stage Aeguidons latifrons: Twenty-eight tests were made in which Aequidens Aggie frons wrigglers were replaced by non—filial Tilapia sparr— magi wrigglers. In ten of these tests the non—filial wrig— glers were accepted, while in eighteen they were rejected (Table 6). Thirteen additional tests were made in which A. latifrons pairs reared mixed broods consisting of their own and introduced If sparrmani wrigglers. In eight of these tests the non—filial young were accepted while in five they were rejected (Table 7). Thus a total of 41 tests were made involving A. latifrons brooding pairs and T. sparrmani wrigglers. Tests were made on pairs with and without previous brooding experience with their own young, and the results appeared to be indepent of this factor (Tables 6 and 7). These tests were of several types. In certain instances wrigglers were replaced by others of their own stage of de— velopment. Thus, the young ones were replaced by others equally young, and older stages were replaced by their equivalents in development. In addition, wrigglers were replaced by those of older or younger developmental stages Table Previ brood l exper ~)—4wf\)wr—'(X)tu |_”. 3. 5.1.1.4 C) l\.) ran I») C) {\J J; lu‘ l—-‘ ()J 48 Table 6. The Results of Fostering Experiments Involving Brooding Pairs of Aequidens latifrons When Their Own Wrigglers WEre Exchanged for Tilapia Sparge mani Wrigglers. Previous Age of eggs in days at Results. Brooding the time of exchange ' experience ‘ Non— Age Filial Filial difference Accepted Rejected 3 H 1 +1 ' X 8 H 1 +1 X 1 H 3 +3 X 3 H 3-1/2 +3—1/2 X 2 3-1/2 1-1/2 —2 x 3 3 1/2 -2—1/2 X 1 3+1/2 1 -2—1/2 X 11 2—1/2 H —2+1/2 X l 3 H -3 X 1 2-1/2 1/2 -2 X 4 4 H -4 X 2 4 1/2 —3—l/2 X 0 3—1/2 1/2 +3 X 1 6 1 —5 X 1 1 1 O X 4 2+1/2 2 -1/2 X 2 4 3-1/2 —l/2 X 0 3-1/2 .3 —1/2 x 3 H H O X 3 1 1 O X 1 H H O X 6 1 1 O X 7 H H O X 10 1 1 0 X 1 4 3-1/2 —1/2 X 0 H H O X 0 H H O X 2 3-1/2 1/2 ~3 X H=Just hatched 7 .L I E O .1 ‘1 t . C .E E e .l d r 2 3 0 3 co 1. .1 A. 3 3 r: i. 0 S l L e S .J n: 1 V 0 e u R t w n E. 1 b e O 0.. T o G C E :1. v . L C. u N «a “I r X I n Q1, Q\ 3» T Md fix T P .C at ., . . 1}. i 1 x a e r i ‘1 : 1 1|" 1|. ‘ l A 49 Table 7° The Results of Fostering Experiments Involving Brooding Pairs of.Aequidens latifrons When Tilapia sparrmani Wrigglers Were Exchanged for Part of Their Own Brood (i.e°, Mixed Broods)° Previous Age of eggs in days at Results brooding the time of exchange experience Non- Age ' Filial filial difference Accepted Rejected 2 H 3—1/2 +3-1/2 X 3 H 1—1/2 +1—1/2 x 0 H 3-1/2 +3~l/2 x 3 1/4 3-1/2 +3—l/4 x 8 3 H —3 x 3 3-1/2 1/2 —3 x 1 4 1 —3 x 4 4 3-1/2 —1/2 x 3 H H o x 3 1/2 1/2 0 x 6 H H o x 4 1 1 0 x 0 H H o x H:Just hatched than their owno Table 8 summarizes the acceptances and re— jections on the basis of age differences of the wrigglerso The test situations with at least a one day difference in age between filial and non—filial wrigglers are summarized in terms of age differential° Any exchange where the age dif— ference was less than one day is designated as a test involv- ing young of the same ageo A total of nine acceptances and four rejections was observed in exchanges involving :0 a W t O w .1 r O b e u e d .4 H C an x r e n n n e e r a C. 0 HA 1 C 3 .1 i O \t C S S .1 a 1 r W t w A n. .. 0 Y S O 3 \r A f I‘ll 1 50 sparrmani wrigglers younger than the filial wrigglers° Seven acceptances and thirteen rejections were recorded in exchanges involving young of the same ageo Two accept- ances and thirteen rejections were recorded in exchanges involving young of the same age. Two acceptances and six rejections occurred when young older than the filial brood were introduced° Eighteen of the 23 rejections were con— trolled by reintroduction of the pair's own wrigglers in Table 89 Fostering Experiments Involving Brooding Pairs of Aeguidens latifrons When Their Own Wrigglers Were Exchanged for Tilapia sparrmani Wrigglers of Various Ages, The Results are Tabulated as Filial Broods Accepted (Reared) and Rejected (Eaten) Under the Appropriate Age Categories. Prob— abilities Computed by the Fisher Exact Method, Age of non—filial young relative to the filial brood Accepted Rejected Older 2 6 Younger 9 4 Same 7 13 Older x younger p = 00056 Older x same p = 0.052 Younger x same p = 00047 Total controls 18 Return of own wrigglers after rejecting the foreign brood. Accepted 18 Rejected O the ma rearec made cessf tests ‘71‘\ ”Acre ance O: Q ‘Y\~\. 51 the manner of the original test, and all were accepted and rearedo The five other tests where rejection took place and which were not controlled in this way involved mixed broods where the parents continued to rear their own young after eating the non-filial youngo Twenty—three intraspecific fostering experiments were made on Aeguidens latifrons, seventeen of which were suc- cessful and six unsuccessful (Table 9)o Fifteen of the tests involved exchanges for young that differed in age from the filial brood» Table 10 summarizes the results with respect to the age differences of the wrigglers involvedo Five of nine tests showed acceptance of young older than the filial broodo Four of seven tests involved acceptance of foster young younger than the filial broodo Two of the former and one of the latter age categories involved mixed broods of nonmfilial young and filial youngo In all eight tests where the substituted young were of the same age accept— ance occurred and thus these might be considered as a type of controlo None of the six tests where rejection occurred involved mixed broodso Table Previc broodi exper: l mgmnp-wmk—‘fiNl-J 52 Table 9. The Results of Fostering Experiments Involving Brooding Pairs of Aeguidens latifrons When the Foreign Wrigglers Were of the Same Species as Their Own. Previous Age of eggs in days at the Results brooding time of Exchange experience Non— Age Filial Filial difference Accepted Rejected 1 l 5 +4 X 2 H 4 +4 X 4 l/2 2 +l—l/2 X 1 1/2 3 +2 —1/2 X 6 H 4 +4 X 3 2 4 +2 X 1 H 2—1/2 +2-1/2 X 0 l 4 +3 X 0 1/2 3 +2-l/2 X 1 3—1/2 H —3-l/2 X 4 3—1/2 H -3-l/2 X 0 4 H -4 X 2 4 H -4 X 6 2 H —2 X 4 4 2 —2 X 1 2 O O X 5 l—l/2 l — l/2 X 3 3 3~l/2 + l/2 X 0 1 1/2 - 1/2 X 7 H H O X 2 2 2 0 X 1 H H O X 0 2—1/2 2 - 1/2 X H=Just hatched ‘ Table Age c young the i OldeJ Young Same Olde: Olde: Young 53 Table 10. Fostering Experiments Involving Brooding Pairs of Aeguidens latifrons When Their Own Wrigglers Were Exchanged for Non—filial Aeguidens latifrons Wrigglers of Various Ages. The Results are Tabulated as Non—filial Broods Accepted (Reared) and Rejected (Eaten) Under the Appropriate Age Categories. Probabilities Computed by the Fisher Exact Method. Age of non-filial young relative to the filial brood Accepted Rejected Older 5 4 Younger 4 2 Same 8 0 Older x younger p = 0.377 Older x same p = 0.052 Younger x same p - 0.164 Tilapia sparrmani: Twenty—five tests were made with Tilapia sparrmani pairs in which there was a complete replacement of their own young by Aeguidens latifrons wrigglers. In nine of these acceptance occurred, while in sixteen the non—filial wrig- glers were rejected (Table 11). Eight additional tests were made in which 1. sparrmani pairs reared mixed broods consisting of their own and introduced A, latifggn§_wrig- glers. In three of these the heterospecific young were accepted, while in five they were rejected (Table 12). Thus a total of 33 tests were made. Ju thei Tat Pre bro exp I‘llll‘ j 54 Table 11. The Results of Fostering Experiments, Involving Brooding Pairs of Tilapia sparrmani When Their Own Wrigglers Were Exchanged for Aequidens lati— frons Wrigglers° Previous Age of eggs in days at the Results brooding time of Exchange experience l—‘OI—‘l—‘OONNNNOWOI—‘l—‘OvaF-‘LJJOOWl—‘O Non- Age Filial Filial difference Accepted Rejected 1/2 3—1/2 +3 X ‘ H 3—1/2 +3—1/2 X 1/2 4 +3—l/2 X 1/2 3-1/2 +3 X 1/2 4 +3—1/2 X H 3 +3 X 3—1/2 1/2 —3 X 3—1/2 1/2 —3 X 3 1/2 —3 X 3 H -3 X 3—1/2 1/2 —3 X 3-1/2 1/2 -3 X 1 1—1/2 + 1/2 X 1/2 1 + 1/2 X l l 0 X 1 1/2 - 1/2 x 1/2 1/2 0 X 3—1/2 4 + 1/2 X 3 3—1/2 + l/2 X H H O ' X H H 0 X 3—1/2 4 + 1/2 X 1/2 H — l/2 X H H O X H H 0 X H=Just hatched Pairs with and without previous brooding experience with their own young were tested an d the results appeared to be Tabl Prev broc expe 55 Table 12. The Results of Fostering Experiments Involving Brooding Pairs of Tilapia sparrmani When Aegui— dens latifrons Wrigglers Were Exchanged for Part of Their Own Brood (i.e., Mixed Broods). Previous Age of eggs in days at the Results brooding time of Exchange experience ' Non— Age Filial filial difference Accepted Rejected 6 1/4 4 +3-3/4 x 1 H 4 +4 x 2 4 2 —2 x 2 3 1/2 —2—1/2 7 1 1 o x 0 H 1/2 + 1/2 ‘x 2 3—1/2 4 + 1/2 x 2 1 1 o x H=Just hatched independent of this factor (Tables 11 and 12). These tests were of several types. In certain instances wrigglers were replaced by others of their own stage of development. Thus young ones were replaced by others equally young, and older stages were replaced by their equivalents in development. In addition, wrigglers were replaced by those of older or younger developmental stages than their own. Table 13 summarizes the acceptances and rejections on the basis of age differences of the wrigglers. The test situations with at least a one day difference in age Age you the Old You Sam Old Old You Tot rej bet ter 56 Table 13. Fostering Experiments Involving Brooding Pairs of Tilapia sparrmani When Their Own Wrigglers Were Exchanged for Aeguidens latifrons Wrigglers of Various Ages. The Results are Tabulated as Foreign Broods Accepted (Reared) and Rejected (Eaten) Under the Appropriate Age Categories. Probabilities Cbmputed by the Fisher Exact Method. Age of non-filial young relative to the filial brood Accepted Rejected Older 8 0 Younger 0 8 Same 4 13 Older x younger p = 0.000 Older x same p = 0.000 Younger x same p = 0.188 Total controls 16 Return of own wrigglers after rejecting the foreign brood Accepted 16 Rejected 0 between filial and non—filial wrigglers are summarized in terms of age differential. Any interchange involving young of less than a one day difference is designated as one in— volving young of the same age as the filial brood. No acceptance and eight rejections were observed in exchanges for Aeguidens wrigglers younger than the filial wrigglers. Four acceptances and thirteen rejections were recorded in- volving young of the same age. Eight acceptances and no rejections occurred when young older than the filial brood were inl control. the man: were ac curred mixed b own you Fif 57 were introduced. Sixteen of the 21 rejections overall were controlled by reintroduction of the pair's own wrigglers in the manner that the original test was performed, and all were accepted. The five other tests where rejection oc— curred and which were not controlled in this way involved mixed broods where the parents continued to rear their own young after eating the non—filial young. Fifteen intraspecific fostering experiments were made with I, sparrmani, all but one of which were successful (Table 14). Table 15 summarizes the results with respect to the age differences of the wrigglers. Six of these tests involved exchanges for young older than the filial brood, while three involved wrigglers younger than the filial brood. Two tests in each age difference category involved mixed broods of filial and non—filial homospecific young. All six of the tests involving young of the same age were successful. The lone unsuccessful test cons1sted Of a mixed brood in which the exchange involved young that were older than the resident brood. Ten controls were employed to determine whether dis— turbance alone had any effect on the acceptance of broods. These were carried out in the same manner as described for Ae uidens and in all instances the restored filial wrigglers __g______ Tabl Pre\ broc expe H=J1 were aCn 58 Table 14. The Results of Fostering Experiments Involving Brooding Pairs of Tilapia sparrmani When the Foreign Wrigglers Were of the Same Species as Their Own. Previous Age of eggs in days at the Results brooding time of exchange experience Non— Age Filial filial difference Accepted Rejected 0 3 4 +2 X 1 H 3—1/2 +3-1/2 X 2 H 3 +3 X 0 l 3 +2 X 0 l/2 3 +2-l/2 X 4 1/4 3-1/2 +2-l/4 X 0 3—1/2 l/2 -3 X 3 3-1/2 1 —2—l/2 X 3 . 3 l —2 X l 3 2—1/2 — l/2 X 2 H H 0 X 3 l 1 0 X 2 1/2 1 + 1/2 X 0 l/2 H — 1/2 X l 2 2 X H=Just hatched were reared without incident. Three ”natural experiments" occurred accidentally when foreign young entered a pair compartment through small gaps in the plexiglass partitions. The parents had no young of their own at the time the heterospecific young moved into their compartments. Other such infiltrations of young across aquarium dividers may have gone unnoticed where Tabl Age your & Olde YouI Same Olde old You] 59 Table 15. Fostering Experiments Involving Brooding Pairs of Tilapia sparrmani When Their Own Wrigglers Were Exchanged for Non-filial Tilapia sparrmani Wrigglers of Various Ages. The Results are Tabulated as Non-filial Broods Accepted (Reared) and Rejected (Eaten) Under the Appropriate Age Categories. Probabilities Computed by the Fisher Exact Method. Age of non—filial young relative to the filial brood Accepted ‘ Rejected Older 5 l Younger 3 0 Same 6 0 Older x younger p = 0.666 Older x same p = 0.500 Younger x same p = 1.000 adjoining compartments contained pairs of the same species. The only way such successful migrations could have been reCOgnized was when the two groups of young were different in age and size. No mixed schools of this sort were ever observed. If other interspecific migrations occurred, the intruders must have been promptly eaten by the resident pairs as mixed broods resulting from this type of situation would surely have been detected. The chance of the latter occurring was unlikely since only in a few instances did adjoining pair compartments house pairs of different species, and then this was only temporary until space permitted their separation. Ca on a 1 12th. hatch comp: 60 Case one: On February 10, 1962, Tilapia pair 9 spawned on a plastic spawning plate. The eggs hatched on February 12th. At the morning check on February 15th all of the wrigglers were gone and the female had been killed in fight— ing. On February 18th two day free swimming (six days post hatching) Aequidens latifrons had begun moving into the compartment with the male Tilapia and were being herded by him. If his own young had survived, they would have been three days free swimming or only one day older than the newly acquired Aequidens young. The male spent a consider— able amount of time mumbling the young in his mouth and carrying them about for long periods with no apparent in— jury to them. These young were reared until March 3rd when they were removed. Case two: At 10:00 A.M. on March 9, 1962, Tilapia pair 15's foster brood of 29 day old A, latifrons young, which they had been brooding for 28 days, was removed. At 2:30 P.M. on March 12 the pair had accepted two day old free swimming 5. latifrons that had crossed into the Tilapia pair compart— ment from the adjoining compartment when gravel had been removed at the bottom of the tank divider. The young made their initial entrance sometime between 10:00 A.M. and 61 2:30 P.M. when they were first observed. In summary, the Tilapia pair accepted two day old Aequidens free swimming young three days after a brood of 29 day old Aequidens young were removed. Case-three: On the morning of April 26, 1964, the members of Tilapia pair 101 were separated due to fighting. The pair had newly hatched wriggling young of their own at the time. The male was removed and placed in an empty pair compartment between two occupied compartments. On one side resided a pair of Tilapia visible through a 4 x 4 inch glass pane while the opposite adjoining compartment housed a brood of newly hatched Hemichromis bimaculatus wrigglers without parents. There was a 4 x 4 inch glass pane in that side as well, but the wrigglers were in a deep depression in the gravel in the far corner of the tank and presumably could not be seen by the male in the middle compartment. At 11:30 A.M. on April 30th Hemichromis young, that became free swimming that same morning, were observed passing beneath a corner of the plastic tank divider. The male had been without young for four days nearly to the hour, and the non—filial young were only twelve to fifteen hours younger than his own would have been. The behavior of the L S u n LL v1 1. C . a a 0 e e e l e n a a .d h l r .1 1 m C Y H m r w w l w. b e T n a a t p T; C t C 62 male was similar to that described earlier for the male of case one while brooding young by itself except that the young were carried about in the mouth almost constantly. He would pick up all of the 50 to 60 young in his compart- ment and maintain them in his mouth while attempting to retrieve those remaining on the other side of the divider window. While carrying the brood aggressive movements were made toward the pair of Tilapia in the other adjoin— ing compartment. At feeding time when a pellet of food was dropped in the tank he would nudge the food, then move back and deposit the mouthful of young along the bottom edge of the tank divider, and return to eat the food. The Hemichromis young were carried in his mouth over a nine day period but became progressively harder to catch and the male seemed progressively less intense in his attempt at retrieving them. The young were removed on the 20th day of May. The young Hemichromis trapped on the other side of the plastic divider were definitely attracted to the male Tilapia. They spent all of their time at the window. When the male moved from the back of the aquarium to the front, the young followed his movements. Only occasionally did one of the young stray from the window but all of them w l l a T tri ing the wer 63 always remained on that side of the compartment. When the Tilapia male-moved toward the young in an attempt to re- trieve them, they showed no fear and responded to his call- ing movements. Theiirbehavior was consistent even when their brood mates were not in view on the other side or were housed in the male's mouth. -‘ .——..__....___.. lit swj not of res wrig type ment 64 General Behavior of Aequidens latifrons and Tilapia sparrmani Pairs Toward Introduced Young That Differed In Age From Their Own The test pairs of both species involved in brood ex— changes for wrigglers younger than their own appeared little disturbed when the foster young did not become free swimming at the time their own would have. There was no noticeable increase in frequency of mouthing or transfer of the wrigglers. There was behavioral stress in the in— stances of mixed broods where the younger foster wrigglers replaced about one half of the filial brood. Here, when the resident young began free swimming they were initially constantly retrieved and spat into the wriggler pit, but after they were sufficiently mobile to move around the aquarium they were only occasionally returned to the pit nest. Considerably more time was spent with the foster wrigglers than with the filial free swimming young. This type of behavior was observed in mixed interspecific experi— ments as well as in mixed intraspecific experiments. A more extreme ambivalent behavior was exhibited by pairs brooding a mixed school of filial wrigglers and older foster wrigglers. They attempted to return the straying newly free swimming young for a much longer time than would other rest retri was d typic exper l-l the p merel U ‘11. ( r- f 65 (otherwise have been the case and showed considerable un- rest during the first day of brooding. After their retrieving efforts subsided, most of the parents' time was devoted to the filial wrigglers. This behavior was typical of both intraspecific and interspecific fostering experiments of this type. In complete exchanges for older non-filial wrigglers the pairs showed no noticeable behavioral unrest. They merely returned the more active straying wrigglers to the pit nest. Correlation of Acceptance—Rejection Ratios with Comparative Wriggler Sizes Table 16 summarizes the probability values computed for significance tests made among the four classes of fostering experiments shown in Tables 8, 10, 13, and 15. Probability computations were made at each of the three age differentials by the Fisher Exact Method for small cell frequencies when using fourfold contingency tables. The acceptance~rejection ratios at the three age differentials made measurements of the wrigglers neces— sary. The total length of 300 wrigglers was determined for each of the two species, Tilapia sparrmani and Table 0);. s THOR 11 O _:Ha4u|:o: O V. DOOHQ seesaw one On 9r~3nfh mew a S Heme/N Ae uid three Cu NV. 1% ments v; the tch a S 66 ’ Table 16. The Probability Values Computed for Significant Differences Between the Four Classes of Fostering Experiments at Each of the Three Age Levels. Com— puted from the Fisher Exact Method. Tilapia Aequidens Tilapia x Tilapia Tilapia Aequidens Tilapia Aequidens g Older 0.252 0.429 m E Younger 0.417 0.006 0 H '3 Same 1.000 0.002 0‘ O s n :3 .-Q 0 >3: 3 : Tilapia; x Aequidens Aequidens X Aequidens g E Aequidens Tilapia Aequidens Tilapia I I: (D éfi Older 0.004 0.181 1 c o i % u Younger 0.002 0.395 ‘ g Same 0.216 0.002 < Aequidens latifrons. One hundred were measured at each of three ages. In doing this 20 wrigglers at each of these ages were measured from each of five females. The measure— ments were obtained in the manner of the egg measurements reported earlier. Thus 100 measurements were taken at each of the following ages: newly hatched, 2—1/2 days after hatching, and beginning free swimming stage. The latter was approximately 3—1/2 days after hatching for Tilapia sparr total E Table i 551 Newl hat: ‘ 2—1/ i post Free swim F 201‘ dev The rat £01 Of 67 sparrmani and four days for Aequidens latifrons. The mean total length at hatching (Table 17) was 2.41 mm and 3.59 mm Table 17. The Mean Total Length Measurements of Tilapia sparrmani and Aequidens latifrons Wrigglers Based on One—Hundred Measurements Taken at Each of Three Ages on Twenty Wrigglers from Each of Five Different Females for Each Species. Aequidens Tilapia Length Age latifrons sparrmani difference Newly hatched 2.4lmm. __0.0131 3.59mm. __0.0071 1.18mm. 2—1/2 days post hatch 3.33mm. _ 0.0082 4.63mm. _ 0.0340 1.39mm. Free swimming 4.94mm. __0.0283 6.46mm. _ 0.0146 1.52mm. for A. latifrons and T, sparrmani respectively. At 2-1/2 days after hatching the mean total lengths were 3.33 mm and 4.63 mm, while at free swimming these measurements were 4.94 mm and 6.46 mm respectively. The measurements of the newly free—swimming stages correlate with equivalence of developmental stage rather than a strict time sequence. These data can be correlated with the acceptance-rejection ratios of the interspecific fostering experiments. It was found that A, latifrons pairs accepted more foster broods of T. sparrmani that Were younger than their own than they 1 i | l wher Pei: own 68 did broods the same age or older. The opposite trend was shown by T, sparrmani pairs. They accepted A, latifrons wrigglers older than their own more often than young of the same age or younger. The wriggler measurements (Table 17) reveal the sizes of the two species to be most similar when exchanged at the preferred age differences. Adult Imprinting Since Noble and Curtis (1939) first suggested that in— experienced cichlid pairs are imprinted on the first spawn raised through the free swimming stage there has been a con— siderable number of contrasting reports. Some insight into this hypothesis of adult imprinting may be gained by analysis of spawning histories of certain of the pairs used in these investigations. Interspecific substitutions of wrigglers or eggs were given to pairs of Aequidens latifrons and Tilapia sparrmani. Egg exchanges were never successful between these two species, but wrigglers were exchanged with vary- ing acceptance depending largely upon correspondence in size. Table 18 summarizes the spawning histories of seventeen pairs of A, latifrons. Nine of these reared young of their own species during their first successful spawning experience 69 Table 18. The Spawning Sequences of Aggpig§p§_;ap;f£pp§ Pairs in Which Fostering Experiments Were Made With Foreign Species Young (Tilapia sparrmani). Sequence of Spawnings Pair 1 2 3 4 5 6 7 A OS OS FS FU/OS B FU* OS MU/OS FU/OS OS MOS C FU*/OS FU/OS OS FS MU/OS . D os os 05 FU*/OS os 05 FU/OS E OS FS OS MS FU/OS F FU* FS OS OS FS FU/OS MOS G FU*/OS FU/OS OS MS H FU*/OS OS FS FU/OS 05 os 08 I FU*/OS_ FS OS MU/OS FU/OS OS MS J MOS FU/OS OS OS MS K FU/OS OS FU/OS L MS FS OS FS FU*/OS OS M FU/OS OS FU/OS OS N MU/OS OS OS MS 0 OU FU*(Hb) OS P os 03 FS*(Hb) OU FS*(Hb) 08 Q OS/FU 0U 0U MU/OS FS FU OS OU = Egg Exchanges — Foreign species reared successfully — Foreign species unsuccessfully — Own species reared successfully Own species unsuccessful ’U 5.2) I—‘ yoruozzwwmx-omotumuowrv _ latifrons .5 Were Made L sparrmaniL M% 05 FU/OS U) FU/OS mm 0305 sosMS OS 03 O n 0 Z 3 H N M H m D W H U 0 w m a [—1. H MS MU MOS MOU FS(Hb) 8 9 10 ll 12 13 14 FU/OS OS 0U ' FU/OS OU FU*/OS FU/OS 0U 0U OS MS OS OU FS 2 Mixed species successful = Mixed species unsuccessful = Mixed brood own species and successful = Mixed broow own species and unsuccessful = Foreign brood successful FU*/OS OS (Hemichromis bimaculatus) and 5 fully and t of it and t can y pairs those Your 70 and subsequently reared young of the other species success— fully. One pair reared the young of the other species first and then those of its own species. One pair reared a mixture of its own young and those of the foster species initially and then reared broods consisting exclusively either of its own young or of young of the other species. The other six pairs reared their own young initially but never accepted those of the other species during subsequent spawnings. The spawning sequences of pairs E, F, H, I, L, and P are of particular interest since each of them alternately raised broods of their own young and young of the other species three times or more. Case H was exceptional in this respect in that broods were reared as follows: own, other species, own, mixed, other species and finally own. The other species involved in the exchanges in the spawning sequence for pair P was Hemichromis bimaculatus. Table 19 summarizes the spawning histories of eighteen pairs of T, sparrmani. Eight pairs reared young of their own species during their first successful spawning experience and subsequently raised young of the other species. Three pairs reared young of the other species first and then those of their own. The other seven pairs reared either their own young or young of the other species at the first successful . -___ .. Mega! Table "0 LL) '1... 1 L0 (X) \J O“ m 4L (.KJ d\_) I"J ,1 7 . r—- r-‘ I4 * (J! We 9) Ex.) r4 C) l-" : on ‘7. r‘) m '11 '71 MC MC 71 Table 19. The Spawning Sequences of Tilapia sparrmani Pairs in Which Fostering Experiments Were Made With Foreign Species YOung (Aegpidens latifrons). Sequence of Spawnings Pair 1 2 3 4 5 6 7 8 1 FS '08 FU/OS OS FU*/OS 2 FU*/OS os FU/OS DU 3 FU*/OS FU/OS OU OS FU/OS OS MS MU/OS 4 FS 08 FU/OS OS OS MOU/OS FU*/OS 5 OS FS FU/OS 6 MU/OS OS FU/OS OS 7 FU/OS OS OU FS 8 FU*/OS FU/OS OS FS FU*/OS DU 9 08 OS MU/OS FS FU*/OS 10 OS FU/OS MU/OS OS OU 11 OS FU/OS FU/OS OS 12 FS 0U 13 F5 FS OS OS OR 14 OS MS MS MOS 15 FU/OS FU/OS OS OS FU/OS OU 16 OS OS MU/OS MOS FU*/OS 18 OS 0U OS OS OU FS(Hb) OS 0U * = Egg Exchanges FS = Foreign species reared successfully FU = Foreign species unsuccessfully OS = Own species reared successfully OU = Own species unsuccessful MS = Mixed species successful MU = Mixed species unsuccessful MOS = Mixed brood own species and successful MOU = Mixed brood own species and unsuccessful FS(Hb) = Foreign brood successful (Hemichromis bimaculatus) . .— —_. __ __.___-__._—.__—.,__— _.__ EZZZ::—______"""""""""——r—------———————r—————————_____1_______llll Spaw when ampl freq thus that for wher D) 5. TWO 72 spawning but failed to raise young of the other species when offered at later spawnings. Most of the latter ex— amples lack real significance, since pair incompatibility frequently resulted in the early death of one pair—mate and thus effectively limited the possible number of matings. If further exchanges could have been made, it is probable that acceptances would have been observed. The necessity for a multiple spawning sequence is illustrated by pair 3 where young of the pairs“ own species were reared initially and a mixed brood was not accepted until the 7th spawning. Two Aequidens broods were rejected in the interim. 6 E .T. E 1 L1. T u S 8 i d t a. t E O. 10 1 g r S I C e a C LL 9 1 U h S __L m“ n T g E a. a u an 13. C a a _ : S r 1. r P d 1 U :1 e C D V S 11‘. C O 11 . . 1 U C a p: 3 DISCUSSION The results of this study led to at least a partial understanding of methods by which brooding pairs of Cichlids (Tilapia sparrmani and Aeridens latifrons) discriminate:the eggs and wrigglers of other species from their own under certain conditions. The criterion for preferences was based upon acceptance—rejection ratios resulting from various types of fostering experiments. Total rejection of heterospecific eggs occurred when such substitutions were made for either of the two species. The wriggler stage was never attained in any of these ex— changes, even when the eggs were of the same age. Rejection occurred whether the foreign eggs were presented early or late in the egg fanning phase of the pairs" brooding cycle. Greenberg (1963a) reported: ”...when eggs were presented early in the pairs" fanning period they were accepted and usually raised through the free swimming stage, regardless of the previous experience of the pair” (p. 141). On the other hand, he found that when eggs were given to pairs al— most ready to transport wrigglers to a pit, they were usually eaten or treated as wrigglers. Myrberg (1961) reported almost complete acceptance of a large number of egg exchanges 73 amox C , r r 1 a t O e I r C C E 11 y .1 . r1 :1 a a b C. a a S O _ n wn h E 0.. C T. O 0 1 0. O 5 b r , d t 1 e S t C a; S 74 among the cichlid species, Hemichromis bimaculatum, and Cichlasoma biocellatum. The data reported here revealed that A. latifrons and T. sparrmani recognize heterospecific eggs by the differences of size, shape, or color of the eggs or by some combination of these factors. The olive green eggs of I, sparrmani are larger and more elliptical than the rounder nearly trans— parent eggs of A, latifrons. Lowe (1955) reported that the oblong, olive green egg is characteristic of substrate spawning Tilapia species. The typical egg of the oral incu- bating Tilapia is bright yellow, pear shaped and larger than those of the substrate spawners. The eggs of the Cichlids used in fostering experiments performed by other investiga- tors and considered here are of the small, round, transparent type. T. Sparrmani and A, latifrons exhibited a marked prefer- ence for eggs of their own species. This was decisively shown when they reared their own eggs returned to them after they had eaten a test brood of the other species. The ac— ceptance of two Hemichromis bimaculatus spawns by a pair of A. latifrons lends support to the hypothesis that this species can recognize heterospecific eggs only if they differ 75 markedly from their own. These two cases of acceptance, as well as those obtained by Greenberg and Myrberg, indicate that olfactory recognition of eggs is probably minimal. Myr— berg (1964) reported that pairs of Hemichromis fasciatus rejected Cichlasoma biocellatum eggs in four of five tests. He mentioned that H, fasciatus has the larger eggs and that rejection of the smaller 9. biocellatum eggs may have occurred because of this. Differences in spawning configuration were also suggested as a possibility. No such differences were noted in the three species studied here. Greenberg"s results (1963a) led him to conclude that in order to obtain successful egg exchanges the heterospecific eggs must correspond to the brooding phase of the parents. Thus, he suggested the presence of an internal mechanism by which the duration of the stages of parental care are regulated. The intraSpecies fostering experiments performed in this study showed that the fanning phases of T, sparrmani and A, latifrons pairs are extremely labile. Thus, cases occurred where eggs were fanned for more than a day longer or shorter than normal. Eggs were never treated as wrig~ glers. Thus, the behavior of pairs toward eggs was guided more by the sight of the eggs than by an internally limited fanning phase. There was a positive reinforcement of the thz i i bet T01 wh i by 76 egg fanning phase when eggs were present for a time longer than normal. The exact means by which the recognition of eggs by the two species of parents used in this investiga- tion might best be determined is by exchanging the filial eggs for odorless gelatinous beads or similar models. These could be varied in shape, size and color. These three fac- tors are not greatly different for the species used by Green— berg and Myrberg, and this probably accounts for the ease with which their exchanges were made. Greenberg (personal communication) stated that exchanges were most successful between the species he used when made during the egg stage. This is well taken since eggs exhibit no behavior upon which recognition could be based. This fact, accompanied by the minimal differences in appearance would cause a more frequent acceptance of eggs as compared to exchanges of broods made at any other stage of development. The results of interspecific fostering experiments involving broods of the wriggler stage were highly predict- able when exchanges were made at various levels of age dif- ference. A preference for certain heterospecific broods was exhibited which was distinctly related to size. A. latifrons brooding pairs reared proportionally more broods .\ of the He: 0t} 77 of T, sparrmani that were younger than the filial brood than heterospecific broods either the same age or older. Thus, there was a linear preference as follows: older than their own; the same age as their own; and younger than their own. This is more easily understood when the size of the wrigglers at the age differentials are considered. T, sparrmani is the larger of the two species at any given age, and this probably explains the preference by Aequidens parents for T, sparrmani wrigglers younger than their own. Here the young are most closely matched in size. On the other hand, the number of acceptances of older T, sparr- mani wrigglers by A, latifrons pairs was least frequent and it is at this age differential that the size differ- ences are greatest. Intraspecific fostering experiments with A, latifrons yielded only a few rejections when exchanges were made for wrigglers differing in age. Thus, it seems that a proper balance must exist with regard to sign stimuli such as size, behavior, and the coloration of the young. Size equivalence is the most important of these factors in interspecific exchanges. Size is less important, however, in intraspecific exchanges where the species coloration, and possibly shape, reinforce acceptance. of pr hibit while Them size thes are quen 78 T. sparrmani pairs exhibited an opposite linearity of preference for A, latifrons wrigglers. Thus, they ex— hibited a marked preference for those older than their own while totally rejecting wrigglers younger than their own. These findings can also be correlated with the comparative size of the two species at the stages of development where these preferences were noted. The sizes of the wrigglers are most closely matched where the acceptances were fre— quent and most different where rejection was greatest. Intraspecific fostering experiments involving T, sparr— mani showed no preference based on age differences. This resembled the situation with regard to A, latifrons. Pairs of both species accepted foster broods of their own species that differed in age (thus size) on the basis of species characteristics independent of these two factors. The limits of the brooding phases in the species studied here were not as stringent as was indicated for those studied by Greenberg (1963a). There may well be an endocrine con- trol of parental phases, but visual stimulation by a specific developmental stage of young can cause the adults to adapt to that stage. This may result from a hormonal triggering stimulated by sign stimuli provided by the young. The fact that your len WE {- 1W 79 that pairs raised two or three successive broods of their own young simultaneously and performed the proper nurturing responses to all indicates that an interpretation on the basis of exclusive hormonal control is not adequate. The three cases where adults accepted free swimming young in the absence of a brood of their own indicate that broodiness can be initiated, or at least reactivated, if the pair had been brooding young in the not too distant past (p. 49). The length of time of retention of broodiness after removal of . a brood of young could easily be tested by offering young i after different time lapses. "Reflex" secretion of hormones in response to external stimuli has been well documented in other vertebrates. The sight of a nesting female pigeon is a sufficient stimulus for a male to develop a lactating crop and to sit on the eggs when the female leaves them (Patel, 1936). In the ring dove, visual stimulation provided by a courting male, ac— companied by the presence of a nest-bowl and nesting materials was necessary for the onset of incubating behavior (Lehrman, 1958). It is possible that the brooding alterations found in the Cichlids studied here involve similar neuro-endocrine relationships. The endocrines of fishes are not fully known I y. i t. m :L 80 but a triggering of hormones similar to those postulated for birds and mammals is plausible, e.g., elaboration of prolactin and the subsequent effect of progesterone on broodiness. In those instances where more than one suc- cessive brood was reared simultaneously a general endocrine controled broodiness is suggested but with direct neural responsiveness to different stages of young. The brooding responses of the pairs were highly ver— satile. Thus, they readily accepted young whose stage of development, and consequently behavior, differed markedly from that of those for which they were exchanged. Size was the primarly limiting factor in exchanges between species. In contrast, species markings or silhouettes reinforced acceptance and tended to mask differences in size when exchanges of wrigglers were made within a species. Several pairs of both species were tested for evidence of possible imprintability. Exchanges were made between A. latifrons and T, sparrmani where pairs of both species freely accepted each other”s young alternately with their own. Since there is little similarity between the markings of the young of these two species it appears that a basis for imprinting could exist. The young, however, were interchangeable. The spawning histories of some of the Ru. and 81 pairs seem to indicate the presence of imprinting (Tables 18 and 19). This was not because of imprinting on the previous young reared but because the subsequent broods rejected were eggs which were always rejected. Still others involved young that differed sufficiently in size to cause rejection. Our findings support those of Greenberg (1961, 1963a, 1963b) in that no evidence of parental imprinting was found in T, spaggr man; and A, latifrons. Subsequent rejection of young of a species other than that reared at the first mating is not adequate evidence of adult imprinting. Sequences have been reported here where this occurred but were followed later by acceptances. Recog— nition of young by parents involves such factors as size, markings, shape, and characteristics of movement. In order to be accepted, young apparently must provide those stimuli which can be minimally substituted. The relative values of the various stimuli may vary in different situations. Thus size acquired high importance in interspecific exchanges while markings, shape and movement had greater value in in— traspecific exchanges. SUMMARY Mated pairs of Tilapia sparrmani and Aequidens Tatif ££9n§_were able to distinguish between their own eggs and those of the other species. This occurred without regard to the age of the eggs or to correspondence in the stages of development between the broods exchanged. Each species accepted the wrigglers of the other pro— vided they matched in size the resident brood for which they were exchanged. At any given developmental stage T, sparrmani is larger than 5, latifrons. Thus, members of this spe— ‘ cies demonstrated a preference for A, latifrons wrig— glers older than those of their own broods at the time of exchange. The Opposite was true in the case of A, latifrons when given T, sparrmani wrigglers. When broods were accepted which were older than those for which they were exchanged, the foster parents ac— celerated their brooding sequence and vice versa. This occurred with regard to intra- and interspecific wrig— gler exchanges. Although attempts to exchange eggs between species never succeeded, intraspecies exchanges were successful for 82 I'll) 83 both species. These succeeded even when the eggs were in different developmental stages and thus the fanning phase of the adults was prolonged or shortened. In a number of instances adults changed from one brood— ing phase to another as the result of contact with young in a developmental stage other than that of those previously brooded. This suggests a possible endocrine response to sign stimuli offered by the young. Since a number of examples were noted where young of two or more developmental stages were brooded properly and simultaneously, factors other than a simple endocrine cycle must exist. Analysis of the subsequent matings of pairs experienced with young of their own or the other species provided no evidence of adult imprinting. Ar< Ba Ba: Ba BIBLIOGRAPHY Aronson, L. R. (1945), Influence of the stimuli provided by the male cichlid fish Tilapia macrocephali on spawning frequency of the female, Physiol. Zool. 18:403. Aronson, L. R. (1948), Problems in the behavior and physi— ology of a species of African mouth breeding fish, Trans. N. Y. Acad. Sci., 2(2):33-42. Aronson, L. R. (1949), An analysis of reproductive behavior in the mouth breeding cichlid fish Tilapia macrocephali (Bleeker), Zoologica, 34:133. Aronson, L. R. (1951), Factors influencing the spawning frequency in the female cichlid fish, Tilapia macro— cephali, Amer. Mus. Novit. 1484:1—26. Baerends, G. P. and J. M. Baerends van Roon (1950), An intro- duction to the study of the ethology of cichlid fishes, Behavior, Suppl. 1, 243p. Baerends, G. P., B. E. Bennema, and A. A. Vogelzang (1960) Uber die anderung der Sehscharfe mit dem Wachstum bei Aequidens portaleqrensis (Hensel) Pisces, Cichlidae), Zool. Jahr. 88(1):67—78. Bartholomew, G. A. (1959), Mother-young relations and the maturation of pup behavior in the Alaskan fur seal, Anim. Behav. 7:163—172. Breder, C. M. Jr. (1934), An experimental study of the re— productive habits and life history of the cichlid fish, Aequidens latifrons (Steindachner), Zoologica, 18:1-42. Collias, N. E. (1956), The analysis of socialization in sheep and goats, EcolOgY: 37:228—239. Collins, H. L., and J. C. Braddock (1962), Notes on fostering experiments with the cichlid fishes, Tilapia sparrmani and Aequidens portaleqrensis, Amer. Zool. 2:400. 84 Gree Gree Gre< Gres Her Kiih Leh MV] My j 85 El-Zarka, S. (1956), Breeding behavior of the Egyptian cichlid fish, Tilapia zilli, Copeia, 2:112-113. Greenberg, B. (1961), Parental behavior and imprinting in cichlid fishes, Amer. Zool. 1:450. Greenberg, B. (1963a), Parental behavior and imprinting in cichlid fishes, Behavior, 21:127—144. Greenberg, B. (1963b), Parental behavior and recognition of young in Cichlasoma biocellatum, Anim. Behav. 7(4):578- 582. Greenberg, B. (1964), Professor of biology, Roosevelt C01- 1ege, Chicago, Illinois. Personal communication. Hersher, L., A. U. Moore, and J. B. Richmond (1958), Effect of post partum separation of mother and kid on maternal care in the domestic goat, Science, 128:1342—1343. Kfihme, Von Wolfdietrich (1962), Das Schwarmverhalten eltern— gefuhrter juncichliden (Pisces), Zeit. F. Tierpsychol. 19(5);513—538. Lehrman, D. S. (1958), Induction of broodiness by partici- pation in courtship and nest—building in the Ring Dove (Streptopelia risoria), J. Comp. Physiol. Psychol., 51: 32—36. Lowe, R. H. (1955), The fecundity of Tilapia species, E. Afr. Agric. J. 11(1):45-52. Lowe, R. H. (1959), Breeding behavior patterns and ecological differences between Tilapia species and their significance for evolution within the genus Tilapia, Proc° Zool. Soc. Lond. 132(1):l-30. Myrberg, A. A. Jr. (1961), An analysis of the preferential care of eggs and young by adult brooding cichlid fishes, Ph.D. thesis, Library of U.C.L.A. Myrberg, Arthur A. Jr. (1964), An analysis of the preferential care of eggs and young by adult cichlid fishes, Zeit. F. Tierpsych. 21(1):53—98. Nob Pat Sha Th< T11 Uc Va 86 Noble, G. K and B. Curtis (1939), The social behavior of the jewel fish, Hemichromis bimaculatus, Amer. Mus. Nat. Hist. Bull. 76:1—46. Patel, M. D. (1936), The physiology of the formation of "pi— geon's milk," Physiol. Zool., 9:129—152. Shaw, E. S. and L. R. Aronson (1954), Oral incubation in Tilapia macrocephali, Amer. Mus. Nat. Hist. Bull. 103: 375-416. Thorpe, W. H. (1956), Learning and Instinct in Animals, 1 ed., 493p. Harvard Univ. Press, Cambridge. Tinbergen, N. (1952), The Study of Instinct, 228p. Univer— sity Press, Oxford. Uchida, R. N. and Joseph E. King (1962), Tank culture of Tilapia, Fishery Bull. 62:199. Van Iersel, J. J. A. (1953), An analysis of the parental behavior of the male three-spined stickleback (Gasteros— teus aculeatus L,) Behavior, Suppl. 3, 159p. sou A¢\t x Lumsoq x su Urru+flurrr<flc§wrfi (rilrxtu‘rxku 111»; 0H XHUQ00Q< 87 H1} mequHHHflE ohmsqm.sfl mwu< 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 2 7 2 7 2 7 2 7 2 7 2 7 2 7 2 7 2 7 2 7 2 7 2 7 O 9 9 8 8 7 7 r0 r0 5 5 4 4 3 3 2 2 l l O 0. 9 9 8 I I O G. I O O O I O O I O O O O 9 I D I I O O 2 l l l l l l _l. l 1 1|. 1 l l l l l .|_ l l 1“" O O O _ . _ . _ _ _ _ _ _ q _ _ . _ _ _ _ _ _ _ / magmaoomfifln mHEOHQUHEom / Hsmfihnwmm mammafla mQOHMHpMH msmcflsmw< .wsumasomfifla wHEonfioHEwm can .mQOHMHumH msopflsmw< .flsmEHQMMm memwaea mo mmmm Mom Av\t x Qumsmq x flupflzv wGOHpmHoono mosd mo conusflfluumflm oQB .H xflpsomm< NH fla 0H ma om NN um s663 go IsqmnN .WCCL.H+.r+fi_. UC.(..&.rI((<. . HHUWVMQ @efifi .mSHMHSUNEHQ mHEOH£UHE®E UGO AQ\3v XQUCH @QQQM MO COHDSQ MO mmmm HON mCOHHMHSUHmU .N XeUCOQQQ I| i HI} Ammmspssom EOHM Q0H¥MH>QQV wasH on3m 0 9 6 0 _ ‘O.990 - 0.970 — 0.950 — 0.930 - 0.910 - 0.890 - 0.870 - 0.850 - 0.830 - 0.810 - 0.790 - 0.770 - 0.750 - 0.730 - 0.710 ‘ 0.670 - 0.650 — 0.630 0-610 0.590 / mSumHSUmEflQ meEOHfioHEmm 88 wcouwfluma mswpflsoo< Hsmfinhmmm memwafle OH NH fla 0H ma ON .NN Wm mm, mm .msumasomEHQ mflfioufloflfiwm was .mcoswflpma mcopflsmo< .flcmfisummw memmasa mo mmmm How macsnmasuamo AQ\BV xwpsH modem mo cesusnflupmflo one .N xflpcwmmd s663 go xeqmnN LIBRQRIES milliiflifllfl‘flifliflifllfl‘ m a) M] 5 89