VISUAL ASPECTS 0F 66 CARE BEHAVIOR IN CICHLASOMA NIGRDFASCIATUM (GUNTHER) Thesis for the Degree of Ph. D.. MICHIGAN STATE UNIVERSITY PETER G. WEBER 1958 m.» '7 I; THESIS I I l/‘Iichigan em: I .5: University I. This is to certify that the thesis entitled VISULL ASPECTS OF EGG CARE BEIL‘IVIOR III CI CIILiE) OI‘ A NICE BORISCIL‘TUII presented by Peter G Weber has been accepted towards fulfillment of the requirements for PhD degree in Zoology Date _"3_I_I +_._ —__.m 0-169 . II_4I_I I_._.—e——. 4 4 EImmune av IIDAB & SDNS' BDIIK ‘BINDERY_ IND. PLEASE NOTE: Not original copy. Some pages are cropped. Filmed as received. UNIVERSITY MICROFILMS. ABSTRACT VISUAL ASPECTS OF EGG CARE BEHAVIOR IN CICHLASOMA NIGROFASCIATUM (GUNTHER) By Peter G. Weber The egg care behavior of Cichlasoma nigrofasciatum (Gunther) was studied with regard to the role of the spawn as stimuli. It was postulated that egg size is an impor- tant aspect of the spawn. Egg size was manipulated by the use of artificial "eggs" while maintaining all other visual aspects of the spawn as constant as possible. Two sets of experiments were performed. In the first, the artificial "eggs" were exchanged for each female's own, and several parameters of egg care behavior towards these were recorded. Egg care behavior was compared under two experimental and three control conditions. The experimental conditions consisted of presenting larger and smaller "eggs" than the fishes' own. The con- trol groups included a baseline recording of egg care towards their own eggs, a situation with no eggs, and one with artificial "eggs" similar in size to natural eggs. The results of this non-choice experiment showed that "eggs" larger or smaller than the natural were Peter G. Weber discriminated,since a considerable decrement in several parameters of fanning occurred when these results were compared either to the baseline or to the "egg" groups similar in size to the natural eggs. The total absence of eggs resulted in the greatest decrement in fanning. Under this circumstance fanning was not completely lost. A second experiment allowed each female to choose between her own eggs and artificial "eggs" similar to these in size. This was a test for discrimination and/or preference. It was found that females were unable to distinguish simulated from natural "eggs" early in the egg care cycle. At this stage the two types are visually similar. As development proceeded, the two egg types became distinct visually and discrimination was noted in favor of the natural eggs. This was indicated by fanning which was directed more often at these than at the artificial eggs. From the results of these two situations it appears that egg size is indeed one of the factors influencing egg care behavior in g. niggg— fasciatum. The presence of simulated "eggs" which did not hatch extended the egg care period appreciably. It was supposed that the inter-spawn interval would also be extended. The results, although they suggested this, were not significant. Peter G. Weber Several aspects of the structure of fanning were also recorded and described. These included the possible relationships between tempo and beats, tempo and bouts, intervals preceding and following bouts and the duration of a bout, and beats and bouts. The only correlation established was that between beats and bouts. This was inverse early in the fanning cycle and direct later on. These results are discussed in detail. The presence of simulated "eggs" when a female was ready to spawn appeared to cause her to deposit her eggs among them. This may be a mechanism restricting the females' choice of egg deposition sites. VISUAL ASPECTS OF EGG CARE BEHAVIOR IN CICHLASOMA NIGROFASCIATUM (GHNTHER) By \. 1"\ Peter G. Weber A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Zoology 1968 ACKNOWLEDGMENTS For meticulously reading and correcting the manu— script, his encouragement, and most of all, for his patience under fire, I am indebted to Dr. James C. Braddock. Special thanks go to Dr. H. M. Slatis. He pro- vided useful advice, especially relative to the statistical methods used, as_well as critically reading the manuscript and offering suggestions as to its improvement. Drs. J. A. King, H. Band, and W. D. Collings, members of my committe, read earlier drafts of the manu- script and offered helpful suggestions. Dr. W. Cooper provided helpful suggestions with regard to the statistics, as well as allowing me the use of the Olivetti-Underwood Programma 101. I thank him for this. It would be difficult to acknowledge all of the students-who have in one way or another contributed to or listened to my ideas during the execution of this project. I must single out Miss S. E. Anderson who listened with patience and made helpful suggestions in reading portions of the manuscript. ii Charles Bursey and Alan Bancroft contributed in the photography; Miss Rosalind Price made a number of the drawings. Finally, I thank Mrs. Bernadette Henderson for her many little favors. iii TABLE OF CONTENTS Page ACKNOWLEDGMENTS O O O O O O O O O O O O C O O C C i i LIST OF TABLES 0 O O O O O O O O O O O O O O O O 0 Vi LIST OF FIGURES o o o o o o o o o o o o o o 0' o o Viii LIST OF APPENDICES . . . . . .'. . . . . . . . . . x INTRODUCTION 0 O O 0 O O O O O O O O O O O O O O O 1- Synopsis of the Reproductive Sequence in CiCh lids O O O O O O O O O O O O O O O O 3 METHODS AND MATERIALS 0 O O O O O O O O I O O O C 12 StOCkS C C O O O I O I O O O O O O O O O O O O O 12 Experimental Arrangement . . . . . . . . . . . . l3 Recording and Experimental Procedure . . . . . . 18 Spawn Exchange: The Non-choice Situation . . 18 Spawn Exchange: The Choice Situation . . . . 19 Apparatus o o o o o o o o o o o o o o o o o o o 2 0 Experimental Design . . . . . . . . . . . . . . 24 Spawn Position Preference . . . . . . . . . . . 28 Statistical Analysis . . . . . . . . . . . . . . 30 RESULTS 0 O O O O O O O O O O O O O O O O O O O O 3 1 Parameters Relevant to Spawning . . . . . . . . 32 The Choice of Spawn Site . . . . . . . . . . . 33 Egg Measurements . . . . . . . . . . . . . . . 33 Spawn Hatch Time . . . . . . . . . . . . . . . 39 iv The Effects of Egg Size Upon Egg Care Behavior Duration of Egg Care . . . . . . . . . Egg Care Measurements: The Non-choice Situation . . . . . Beats 0 C C O O O O O C O O O O O O Bouts O O O O O O O O O O O O O O 0 Duration 0 O O O O I O O I O O O I 0 Tempo . . . . . . . . . . . Beats/Bout and Duration/Bout . . . . The Choice Situation . . . . . . . . . Nips . . . . . . . . . . . . . . . . Other Aspects of the Structure of Fanning Beats vs. Bouts . . . . . . . . Interval Preceding and Interva Followin a Bout vs. Duration of Bout' . . Duration and Beats vs. Tempo . . . . Time Between Spawns . . . . . . . . . . Spawn Site Preference . . . . . . . . . DISCUSSION . . . . . . . . . . . . . . . . SUMMARY . . o . . . . . . . . . . . . . . REFERENCES . . . . . . . . . . . . . . . . APPENDICES . . . . . . . . . . . . . . . . Review of Literature . . . . . . . . . . Page 41 41 45 45 51 54 54 59 59 65 68 68 68 71 71 77 87 100 102 106 140 10. 11. 12. 13. LIST OF TABLES Choice of Spawning Site . . . . . . . . . . The Result of a Hierarchial Analysis of Variance of Shape Index of Eggs . . . . . . The Result of a Hierarchial Analysis of Variance of Sectional Area of Eggs . . . . . Basic Statistics for the Time of the spam to HatCh o o o o o c o o o o o o 0 Basic Statistics for the Duration of Egg Care . . . . . . . . . . . . . . . . The Result of a One-way Analysis of Variance of Egg Care Duration-in the Five Groups. . . The Result of a Tukey-T Test Testing all Means of the Egg Care Duration in the Five Groups . . . . . . . . . . . . . . The Result of a One-way Analysis of Variance of Fanning Beats Across All Groups at Twenty-four Hours 0 o o o o o o o o o a o The Result of a Scheffe's Test for Multiple Comparisons of the Mean Beats at TwentY'four HOurS o o o ‘ o o o o o o o o o o The Results of a One-way Analysis of Variance of the Egg Care Duration in the Choice Situation . . . . . . . . .». . . Basic Statistics for the Time Between Spawnings Under Several Conditions . . . . . Basic Statistics for the Number of Days Fanning Above 200 beats/ten Minutes and the Number of Days Between Spawnings . . . . . . The Results of t-tests of the Mean Spawn Interval Under three Conditions . . . . . . vi Page 34 38 39 40 41 44 45 50 51 65 74 75 76 Table 14. 15. 16. Order and Position of Successive Spawnings Order and Position of Successive Spawnings When Simulated "Eggs" Similar to the Natural Were Present . . . . . . . . . . . Order and Position of Successive Spawnings When Simulated "Eggs" Larger than the Natural Were Present . . . . . . . . . . . vii Page 0 78 . 82 . 85 Figure l. 10. 11. 12a 13. 14. 15. LIST OF FIGURES Spawn Configuration Diagrams . . . . . . . General Observational Set-up . . . . . . . Representative Records of Parameters Recorded . . . . . . . . . . . The Common Positions Assumed under the Lean-t0 o o o o o o o o c o O 0 Position of Own Spawn in Relation to the Simulated . . . . . . . . Shape Index of Eggs . . . . . . . . . . . Sectional Area of Eggs . . . . . . . . . . Cumulative Number of Females Which Stopped Fanning . . . . . . . . . . Mean Number of Beats per Ten Minute Observation Period . . . . . . . . Mean Number of Bouts per Ten Minute Observation Period . . . . . . . . Mean Duration per Ten Minute Observation Period . . . . . . . . Mean Tempo per Ten Minute Observation Period . . . . . . . . Mean Number of Beats per Five Minute Observation Period in the Choice Situation Mean Duration per Five Minute Observation Period in the Choice Situation Mean Tempo per Five Minute Observation Period in the Choice Situation viii Page 14 21 26 29 36 37 42 48 52 55 57 61 62 63 Figure 16. 17. 18. 19. Cumulative Number of Females which Stopped Fanning in the Choice Situation Nipping and Fanning Curves for a Representative Female . . . . . . . Beats vs. Bouts for Four of the Five Conditions . . . . . . . . . . . Number of Days Fanning Above 200 Beats/ten minutes vs. the Inter-spawning Interval ix 0 Page 64 66 69 72 LIST OF APPENDICES Appendix la. 1b. 2a. 2b. The Distribution of Sectional Area Calculation for Eggs of Tila ia sparrmani and Hemichromis Eimaculatus TData from H. ColIIns) . . . . . . . . . . The Distribution of Sectional Area Calculations for Eggs of Cichlasoma nigrofasciatum and Simulated "Eggs" . . . The Distribution of Shape Index Calculations for Eggs of T. s arrmani and H. bimaculatus (Data from H. CoIIins). The Distribution of Shape Index Calculations for Eggs of g. nigrofasciatum and Simulated "EggS" o o o o o o o o' o o a Basic Statistics for the Analysis of Egg Size (Sectional Area) and Shape (Shape Index) . . . . . . . . . . . Figure for the Mean Number of Beats/Bout/ten Minutes . . . . . . . . .1. Figure for the Mean Fanning Duration/Bout/ten Minutes . . . . . . . . Basic Statistics for Fanning Beat Data for all Groups . . . . . . . . . . . Basic Statistics for Fanning Bout Data for all Groups . . . . . . . . . . . Basic Statistics for Fanning Duration Data for all Groups 0 o o o o 0' o o o o 0 Basic Statistics for Fanning Tempo Data for all Groups- . . . . . . . . . . . Page 107 108 110 111 112 114 116 118 120 122 124 Appendix 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. Basic Statistics for Fanning Beats/Bout for all Groups . . . . . . Basic Statistics for Fanning Duration/Bout for all Groups . . . . . Basic Statistics for Nips for all Groups . . . . . . . . . Interval Preceding a Bout vs. Duration/a Bout for Female 29 . . . . Interval Following a Bout vs. Duration/a Bout for Female 20 . . . . Length.ofInterval Following a Bout vs. Duration/a Bout for Female 26 . . . . Length of Interval Preceding a Bout vs. Duration/a Bout for Female 26 . . . . Length of Interval Following a Bout vs. Duration/a Bout for Female 29 . . . . Fanning Duration/Bout vs. Tempo . . . Fanning-Bouts vs. Tempo . . . . . . . Review of Literature . . . . . . . . . xi Page- 126 128 130 132 133 134 135 137 138 139 140 INTRODUCTION Parental care in fishes varies from none to rather elaborate nest building and nurturing of the young. All cichlids, as far as is known, undertake parental care to some degree (Breder and Rosen, 1966). This study concerns one aspect of the parental care in Cichlasoma nigrofasciatum (Gfinther). Reciprocal spawn exchanges between various cichlid species have emphasized the importance of visual, rather than tactile or olfactory stimuli in mediating these fishes' ability to discriminate their own from other species‘ eggs (Greenberg, 1961, 1963 a, b, Myrberg, 1961, 1964, 1966, and Collins, 1965). Collins (1965) was unable to achieve ac- ceptance of eggs in exchanges between Tilapia sparrmani and Aeguidens latifrons, and he attributed this to the obvious difference in shape, size, and color of the eggs of these species. Collins not only mentioned these differences but also supplied egg measurements of 2. sparrmani, A. latifrons, and Hemichromis bimaculatus as part of his evidence (see Figures 1 and 2 of Appendix). From this work, however, it is not possible to determine which precise features of the eggs the fish were discriminating. Kfihme (1963,1964 a,b) demonstrated that H. bimaculatus parents are able to make fine discriminations between water which had contained their own larvae and free-swimming young and that which had contained those of conspecifics, as well as of other species, e.g., E} fasciatus. However, these fish gave no response to water which had contained eggs,~and made no discrimination between this and the filtered aquarium water control. Neither olfaction, gustation, or tactile information plays a major role in the egg care behavior of g. nigrofasciatum (Mertz, 1967). Vision, on the other hand, is important (Mertz, 1967). All of this evidence suggests that visual stimuli are more important than chemi- cal stimuli in eliciting parental care of cichlid eggs. The precise visual aspects of the spawn Whicqlpidduce-paf rental response were not determined by these studies. In the substrate spawning cichlids such as Q. nigrofasciatum, the substrate on which eggs have been de- posited, and cared for may conVeniently be exchanged for an identical one with dummy eggs. One may then systemat- ically vary egg characteristics such as size, shape, and color. In addition, the density of the eggs in the spawn and the spawn configuration can be manipulated. This technique was adopted for the present study. Size and shape of the eggs varies from species to species. Color varies among species and within a given spawn as development proceeds. Thus, 9. nigrofasciatum eggs are light grey and transluscent upon deposition and measure 1.6 mm2 in sectional area (L x w x 20 (Figure 7). As they approach hatching, the eggs become a darker grey due to the development of pigments in the embryo. These are espe- cially concentrated in portions of the yolk sac. Overall shape of the spawn (spawn configuration) appears to vary among species and also varies considerably among individuals and for the same individual with successive spawns. g. nigrofasciatum deposits an irregularily shaped spawn whose shape cannot be predicted prior to spawning. In contrast, H. bimaculatus consistently deposits an oval shaped spawn (Figure»l). The purpose of this study was to determine the ef- fects of egg size upon egg care behavior in,g..nigrofasciatum. It was hypothesized that the size of the eggs determines the frequency, duration and rate of fanning, guarding and nip- ping the eggs (egg care behavior). It was further hypothe- sized that the size of.the eggs could be altered sufficiently to eliminate egg care behavior. Synopsis of the Re roductive Sequence in Cichlids with SpeciaI Reference to Cichlasoma nigrofasciatum (Gfinther) Cichlids are known for their characteristic reproduc- tive behavior, which has been described for a number of dif- ferent species (see Review of Literature). Of the papers listed, the most relevant to the work reported here are: Breder (1934), Peters (1941), Aronson (1949), Baerends and DNil 6N4 "ties CN-B#__ Figure 1. diagram. NB-ll IBfi! Spawn configuration Spawns taken from four 9. nigrofasciatum (CN—) and four H. bimaculatus (HB-) females. Iliustrates differ- ences in spawn configuration between the two species. Baerends von Roon (1950), Kfienzer (1962), Collins (1965), Greenberg et al. (1965), Myrberg (1965), and Mertz (1967). Cichlid fishes can be classifed as two groups on the basis of their brooding behavior. These are the substrate spawners and the oral incubators (Baerends and Baerends von Roon, 1950). In the latter, one of the parents picks up the eggs either prior to or after fertilization, and then incubates them in the mouth cavity (Wickler, 1962). In certain spe- cies the male does this; in others, the female. The substrate spawning group deposits the adhésive eggs on a surface after which they are cared for by one or both of the parents. The following is a description of the behavior associated with pair formation, spawning, and care of the eggs and young in g. nigrofasciatum. This account is similar to that of other workers, many of whom worked with other species, since the behavior associated with re— production is qualitatively similar among cichlids (Baerends and Baerends von Roon, 1950). The cichlid reproductive cycle may be broken down into several phases: pre-spawning or courtship phase, spawning, and post-spawning or parental phase. The latter involves egg care (incubation), wriggler tending, and care of free-swimming young. Synchronization of reproductive behavior in fishes, as has been demonstrated for most vertebrates, is hormonally as well as neurally controlled (Noble and Kfimpf, 1936, Aronson, 1948, Fiedler, 1962, Blfim and Fiedler, 1964, 1965, Wai and Hoar, 1963, and Smith and Hoar, 1967). Different hormones probably affect the predominant behavioral patterns during various phases of the reproduc- tive cycle. For instance, the fanning and decreased aggres- sion associated with the incubation period is strongly in- fluenced by prolactin (Fiedler, 1962, Blfim and Fiedler, 1964, 1965). Male and female 9. nigrofasciatum form pair bonds. In most species the male initiates this cycle by choosing an area and defending it against other individuals, and these include females and members of certain other species (Collins, 1965, Mertz, 1967). A ripe female may eventually enter the male's territory, persist despite his attacks and finally pair with him. Both members of the pair then jointly defend the territory and prepare for eventual spawning through a period of mutual displaying. The pre—spawning cycle lasted from one to seven days. In this study, the same pair bond continued for several successive spawning cycles. This was partly a con- sequence of the same individuals being confined together continuously. This extended pair bond made it possible to use the same pairs throughout the study. How long pair bonds are maintained for extended periods under natural con- ditions is not known. During initial encounters between prospective mates, agonistic behavior predominates. This consists of mouth fighting, chasing, butting, and biting at the flanks. This is gradually replaced by courtship displays with their char- acteristic cichlid behavior patterns such as brief frontal displays involving spreading of the gill covers, head shakes, and repeated rushes at the mate. Lateral displays, which are stiff undulating side-to-side weaving movements with the median fins extended, also occur. Such displays may end with vigorous tail beating. At times, one member of a pair may demonstrate appeasement behavior, which in- volves a lateral display with the head up and pectoral fins folded. This occurs when an individual is threatened by a more aggressive mate. Certain other behavioral patterns are associated with the pre-spawning period. These include picking up of objects such as gravel, chafing, skimming, jerking, quivering, nipping off, and digging (Greenberg et al., 1965). During the initial courtship phase the female appears in the more dominant role. As the pair approaches spawning condition, these activities increase in frequency. Differences between the sexes are largely of a quantitative nature (Aronson and Holz-Tucker, 1949, Greenberg et al., 1965). An intensification of color accompanies these pre- spawning behavioral changes. This has been noted by nu- merous workers such as Kramer (1960), Neil (1964), and Collins (1965). In-g. nigrofasciatum this chiefly involves a heightening of the contrast in the vertically striped pattern, and, in the female, a brightening of an orange plaque on the flanks and of the iridescence in the fins and opercular regions. This lasts until the end of the egg-incubation phase. A more or less secluded spot is usually selected as a laying site. 9. nigrofasciatum and H. bimaculatus, as observed in our laboratory, prefer an enclosed location such as the underside of a flowerpot, the area under a slate, or a hole in a brick. The behavioral patterns noted were nipping off, skimming, and displaying at the chosen place. At this time the female spent most of her time in cleaning off the prospective site by means of vigorous nip- ping at the substrate. This preference for a secluded spot made it possible to induce the pairs to spawn in a place selected by the observer. .From a few hours to a day prior to spawning, the members of each sex protrude their genital tubercles, and the frequency of skimming movements at the spawning site increases. Skimming is very similar to fanning, but in the former the frequency of fin beats is higher and the belly and ovipositor are dragged over the substratum. Once a spawning spot has been chosen, the area is cleaned by nipping and skimming. Displaying becomes largely restricted to the area. Immediately prior to and during egg deposition, aggression between the mates decreases. Egg deposition is accompanied by behavior very similar to skimming but in- volves a smoother gliding movement with the head up and the ovipositor pressed against the substrate. The female makes a run which consists of gliding over the substrate by means of rapid coordinated beats of the pectoral, dorsal, and caudal fins. During this run she deposits a row of five to twenty eggs. The male typically follows close behind, and exhibits similar behavior while emitting sperm over the eggs. Spawning continues for one-half to one and one- half hours. This depends upon the size of the spawn which largely reflects the size of the female (Noble and Curtis, 1939, Collins, 1965). There was no apparent pattern during these experiments as to time of the photoperiod when E. nigrofasciatum spawned. In contrast, H. bimaculatus demon- strated a propensity to spawn in the late afternoon (3-6 p.m.). The parental phase of the reproductive cycle begins immediately after egg laying. Although the female is by far the more dominant member, both partners take part in egg care, which primarily consists of fanning, nipping, and guarding the spawn. It is this stage of the reproduc- tive cycle with which this study was concerned. At the termination of spawning the aggressiveness of the female towards the male increases considerably. She will often display vigorously and attack her mate, espe- cially threatening him with repeated rushes of frontal dis- plays. 10 During the parental phase individuals of both sexes attacked and displayed frequently to others in the adjacent compartments as well as towards one another. During the incubation period the pair dig pits in the gravel substrate, and the frequency of this behavior increases toward the end of this period (Greenberg et al., 1965). Pit digging, performed by either parent, involves picking up gravel in the mouth and spitting it out away from the area intended to be the pit nest. This behavior continues until a de- pression is created in the gravel substrate. When the eggs hatch into wrigglers they are immediately picked up by the parents and placed in the previously dug pits or onto cleared areas of the bottom (if no gravel is present). The parents do not continue to fan the wrigglers, as in other cichlid species, but they do occasionally mouth and move them from one pit to another until they become free- swimming larvae. In our laboratory the transition from prolarval wrigglers to free swimmers occurred at approxi- mately four days post—hatching. The free-swimming young are then herded into a dense school and are guarded by the parents until a new spawn is imminent. As they become in- creasingly more independent, the free—swimming young exhibit a progressive decrease in responses to the paren- tal signals. In our laboratory the mean interval between spawn- ings was ten days. This is variable because duration of 11 the egg and wriggler stages are temperature dependent (Peters, 1941, Cridland, 1962, and Greenberg, 1963). In. contrast, the transition from free-swimming larva to ju- venile stage is partly temperature dependent and is partly the result of feeding rate. METHODS AND MATERIALS Stocks Cichlasoma nigrofasciatum (Gunther) is a native Central American substrate-spawning cichlid. It was chosen for this study, in part, because the species was known to be a prolific breeder in captivity, and to be sufficiently un- aggressive to permit pair compatibility through a sequence of spawns. In this respect it was a better choice than the more aggressive Hemichromis bimaculatus (Gill), which had been used in earlier pilot studies. Q. nigrofasciatum has also been used for studies of parental behavior (Myrberg, 1964, Mertz, 1966, 1967). The original stock of g. nigrofasciatum was obtained from dealers in Lansing and Ann Arbor. The genetic history of-these specimens was not known. This species occurs in two distinct color varieties. One is dark and vertically striped, and is known to dealers in this region as a "con- vict" or "zebra" cichlid. The other is a white phase sold commercially as a "golden" cichlid. The experimental fish were first generation descendants produced by a series of white male x dark female matings. All of these were dark in color. Presumably the offspring were not as inbred as the parental populations. This was desirable because they 12 13 might be expected to exhibit behavior more closely approxi- mating that of the wild population. None of the pairs used in the study had ever spawned. They were removed from the stock tanks at sexual maturity. This was judged by size, behavior and (in the case of the females) color. Thirteen individuals were used throughout the study; three of these died and were replaced by other individuals from the stock tanks.* ExperimentalArrangement Seven 120 liter (76 x 42 x 40 cm.) and two 180 liter (121 x 56 x 34 cm.) rectangular aquaria with slate sides and floors were used for observation tanks. These were divided transversely intocompartments by opaque green plexiglass partitions (Figure 2). There were three com- partments in the 120 1. and five in the 180 l. tanks. Each compartment held a test pair. A sliding glass door was located at the bottom of each partition. The dimensions of these doors were 10 x 10 cm. This allowed visual con- tact between pairs. In addition, seven 120 1., two 180 1., and six 60 1., metal-framed aquaria were used as stock and rearing tanks. The substrate consisted of white aquarium gravel approximately 2-6 cm. in depth. No vegetation was present. Each test compartment contained a red clay brick (21 x 9 x 6 cm.) with two rows of five holes (approximately *In the non choice situation five spawns from each pair were used. 14 Figure 2. General observational set-up, showing 120 l. tank divided into three compartments, each containing a lean-to supported by a brick. 15 2 x 2.5 cm.) and a plexiglass lean-to (Figure 2). The lean-to was composed of three pieces of green opaque plexi- glass 3 mm. thick. The vertical section and diagonal was 8 x 20 cm., while the horizontal was 8 x 8 cm. The angle between the vertical section and the diagonal was approxi- mately 20-25°. The lean-tos functioned as spawning places, and spawning occurred on any of the three sections. Each of the individual plexiglass sections will henceforth be called a "spawning slate." Eighty-three percent of the Spawnings occurred within the lean-tos. Any side of a lean-to could be conveniently replaced either with an iden- tical piece for control purposes or with a section carrying a simulated spawn. The bricks were useful in two ways. First, they prevented collapse of the lean-tos resulting from disturbances caused by display activities inside. Secondly, the holes in the brick served as places of shel- ter when one pair-mate became overly aggressive. The males were often especially aggressive after the failure or the removal of a spawn. In such instances the females did not always terminate this male aggressiveness by appropriate submissive behavior. In such instances the females could escape, since they were small enough to enter the holes, while the larger males could not. Each experimental aquarium was supplied with one Le Bern outside gravity flow filter containing glass wool, filter floss, and charcoal. The water intake was located 16 2-3 cm. above the tank floor. Air to drive the filters and air stones was supplied to the entire laboratory via an air compression unit. Each test compartment contained one air stone. The water temperature was generally maintained at 27° i l supplementing the heating unit by an air condition- ing unit servicing the entire laboratory. Illumination was provided by overhead 40 W fluorescent units supplying the laboratory. The light source for the stock tanks con— sisted of 15 and 30 W tubes located approximately 60 cm. above each tank. A twelve-hour photoperiod (on and off at eight) was maintained by time switches. Some diffuse light entered through the covered laboratory windows. Pilot studies were started on 12-10-65, and until 6-8-66 all of the water in the tanks was aged city tap water supplemented by distilled water to maintain the water level. Beginning at the latter date, water in all tanks was re- placed with a mixture of distilled water and Utility Seven- Seas marine salt mix (5 g/l).* With this change of medium a definite increase in spawning rate and success of brood rearing occurred. There was also a marked improvement in the general health of the fish and a decrease in the amount of cannibalism by the parents of their own eggs and young. *Thanks are due to Dr. G. W. Barlow for suggesting this medium. 17 No attempt was made to control pH, which ranged from 6-7 as determined periodically with short-range alkacid paper. Free-swimming young were fed brine shrimp (Artemia) nauplii once per day. Adult fish were fed a mixture de- veloped by Collins (1965). This consisted of ground shrimp, oatmeal, and Wardley's Supremix (mixed approximately 1 : l : 1) frozen into patty form. Dried commercial prep- arations (e.g., TetraMin) were given as supplemental food. Test pairs from which spawns were needed were fed'primarily ground shrimp or frozen brine shrimp. All fish were fed once per day between 11 a.m. and 1 p.m. Pairing of test fish was accomplished by removing- apparently healthy males and females from the stock tanks and placing them together in a closed compartment of an experimental tank. 9. nigrofasciatum is markedly dimorphic. The females are characterized by an orange coloration of the flanks, blue in the head region (e.g., operculum, mouth), and iridescent turquoise blue or green on the anal, dorsal, and pelvic fins. The males which are larger, lack this coloration and have a larger, more pointed dorsal fin. Pairs usually spawned within a week after being placed to- gether. Attempts were always made to assemble pairs where the male was slightly larger than the female by approxi- :nmtely 25% since this size difference was more likely to .insure breeding success (Barlow, 1962, 1965). 18 Recording and Experimental Procedure Spawn Exchange: The Non-Choice Situation Thirteen pairs were used in the course of the study. Pilot investigations indicated that the males usually did ruot contribute significantly to spawn care, but could be a factor in increasing variability in the recordings. Therefore, males were removed by netting after spawning and were reintroduced when the eggs hatched and observations were terminated. For each pairing it was determined by a coin flip whether the spawn was to be exchanged for a con- trol or experimental (simulated spawn) slate or used for baseline observations. That short (10-15 minute) observa- tion periods, taken at various times of the photoperiod, accurately represent the real changes in temporal organiza- tion of fanning during the course of the egg-care cycle in g. nigrofasciatum, has been demonstrated by Mertz (1967). On the basis of this, as well as preliminary studies, it was felt that recordings taken for ten minute observation periods three times per day would yield reasonably accurate data. These occurred at 9 a.m., 12 noon and 3 p.m.. Each started one-half to one hour after completion of spawning after which interval it was certain that spawning behavior had ceased. Where exchanges were made, the parents were first observed for a ten minute observation period to in- sure that they were performing within limits comparable to those of the baseline group.. The exchanges were then made 19 as follows. The slate containing the pair's own eggs was removed and replaced either with an identical slate with no eggs ("control") or with an identical slate containing a simulated spawn (experimental). One-half hour after each such exchange, the first subsequent observation was taken and was followed by others as outlined above. These recordings were terminated for each pair when all eggs hatched. In the gage of pairs whose eggs had been removed, the time of hatching was established by rearing the spawns in one liter cylindrical battery jars in which the parental water was agitated by means of an air stone. When these "artificially incubated" eggs hatched into prolarvae, re- cording of the parental behavior was terminated. In the cases of the control pairs parental behavior generally ceased before the artificially incubated eggs hatched. In spawn exchanges involving simulated eggs, which ob- viously did not hatch, fanning often continued beyond the time when the "artificially incubated" eggs hatched. In such instances recording was continued for fifteen days, which exceeded the average interval between spawns. Spawn Exchange: The Choice Situation One-half to one hour after the completion of spawn- ing the male was removed, and fifteen minutes to one-half hour later the spawn was also removed. The latter was 20 duplicated "artificially" which generally took from 20-45 minutes depending upon the size of the spawn and then both the simulated and the pair's own natural spawn were re- turned. The simulated "eggs" were placed where the actual eggs had been deposited, and the slate with the natural eggs was randomly placed in one of the two remaining posi- tions under the lean-to (Table 12). The same recording procedures, schedule, and length of observation periods, as well as the same individuals were used as in the pre- vious experiment. In this situation, however, fans toward the natural and simulated eggs were simultaneously recorded on two different channels of the event recorder. Apparatus The initial recordings were made with a Gerbrand six-channel multiple-event recorder (Figure 3). The re- corder pens were activated by means of a manual keyboard; the chart speed was 2 mm/sec. The majority of the record- ings were made with a Rustrack four-channel multiple-event recorder (model 92) at a chart speed of 2 mm/sec. For the non-spawn control group, measurements were begun the day after the wrigglers became free-swimming. Young were always removed from the parents. In the cases where no exchanges were made, this was done on the day they became free-swim- ming. 21 Representative records of fanning (top channel), departures (middle channel), and nips (bottom channel) for female No. 37 which had its own spawn removed in-exchange for a blank "control" slate. Recordings were made with a Gerbrand multiple channel event recorder at a chart speed of 2 mm ps. The numbers below a bout of fanning give the number of beats (top) and duration . (bottom). A. First 10 minute recording, taken 1/2 hour after termination of spawn deposition. This is a base- line recording of behavior toward own spawn (7-23— 67; 10:15 a.m.; 10' record: 957 beats/12 bouts/449 duration/2.131 tempo; 0 nips). B. Measurements taken at 24 hours after first record- ing above; this shows behavior recorded toward blank "control" slate (arrow indicates incipient "dis- placement fanning") (7-24-67; 9:45 a.m.; 10' record: 534 beats/18bbuts/239 duration/2.234 tempo; 22 nips). C. Measurements taken at 72 hours after initial re- cordings above; of behavior towards control slate. Most of the fanning bouts shown are "displacement fanning" and only the left end of the recording shows any "normal" fanning. Displacement nipping does not show up on the recording as does the dis-* placement fanning (7-25-67; 9:45 a.m.; 10' record: 619 beats/20 bouts/202 duration/3.064 tempo; 41 nips). - Figure 3. A portion of the fanning record for female 37. Legend: tempo: fanning beats/duration departures: the frequency with which the female leaves the lean-to; not used in the analysis of egg care behavior. I See pages 25, 27, and 28 for detailed explanation of the behavioral terms used. 22 - - WNW mmnnnmmmmmmmmrmnnrmmmmmt";r: 2‘ ”7777’ PT”?! 22 107 21 9 53 9 “I: m m medium 1221‘: teams 4 7 2 8 42 2 0.5 1 2 1 7.5 -*-- -‘~» .7 - I»-_ A7,, -11- a 7 ..-r v ‘7 .. -——; e—Aw- . -- v——- —- ~—~—¥rfl 34 13 22 29 11.5 4.5 0 9 N. O” -— .qu' 7 % A! III 1111-..- , ,, I 1 1 _ 1.1 23 It was noted in certain pilot studies that both E. bimaculatus and g. nigrofasciatum would tend rather un- realistic spawns of Indian beads for periods as long as 24 hours. Where ”artificial" spawns were constructed of glass beads, however, these proved to be ineffective as egg-substitutes for more than one day probably because of the reflectance and refractance of glass, as well as lack of appropriate texture and color. A technique was then developed for making simulated cichlid "eggs." This con- sisted of dipping a glass rod, which had been drawn into a point, into melted paraffin and allowing the paraffin to drop off the point. This formed a ball. Size could be controlled by the use of rods drawn into variously sized points, as well as by controlling the amount of par- affin that collected on the point before dropping off. Shape could be controlled by varying the shape of the needle point and by manipulating the rod as the paraffin dropped off. Color similarity to that of actual eggs was achieved by addition of powdered charcoal to the melted paraffin. The paraffin used was commercial "Parowax" (Standard Oil Co.). I As soon as a pair finished spawning, a quick sketch was made indicating the position of the eggs on the plastic slate and the general spawn configuration. This was used as a model in constructing the simulated spawn WhiCh was made to duplicate the real spawn as closely as possible. 24 Experimental Design The following is.a description of the independent variables. There were two experimental groups as follows: 1. Larger than natural eggs: this involved the ex- change of a female's own spawn for a simulated spawn simi- lar in all aspects except that the "eggs" were larger than the "Own" (i sectional area = 7.20 mmz). 2. Smaller than natural eggs: similar to the above group except the exchanged "eggs" were smaller (i sectional area = 0.733 mmz). The control groups were as follows: 1. Same size as natural eggs: this, again, was similar to the above except that the own were exchanged for simulated "eggs" as close-as possible to the female's own in size. 2. Baseline: this involved the female's own spawn. No exchanges were made. 3. Blank slate: in this group the female's own spawn was exchanged for an identical slate without real or simu— lated eggs. Dependent variables: various parental motor acts were quantified. FOr the most part, these were established by other workers (van Iersel, 1953, Morris, 1955, 1958, Sevester, 1961, Barlow, 1965, Mertz and Barlow, 1966, Mertz, 1967) and are merely reviewed here. In comparing- parental behavior under different spawn conditions as well 25 as "control" conditions, aspects of three behavior pat- terns were analyzed: 1. Fanning is a motor act associated with care of eggs and larvae in nearly all substrate spawning cichlids. This behavior is very similar to swimming but lacks for- ward movement. The caudad movement of the pectoral fins is counteracted by movement of the caudal and dorsal fins The beginning and end of fanning were at times difficult to determine. Fanning was occasionally confused with swim- ming or hovering to which it was similar.v The latter two, however, were not directed at the spawn. In-practice, such movements were labelled fanning only when oriented toward the spawn, and these judgements were sometimes sub- jective. Positioning of the parent with respect to the spawn was generally as indicated in Figure 4, although virtually any orientation of close proximity to the clutch was adopted. 2. A beat was defined as a complete undulatory cycle of the pectoral fins. These alternated in their movement, but in recording fanning only the movement of one fin was counted. Occasionally a fish would place itself in a po- sition such that movement of the pectoral fins was diffi- cult to observe; and, since the undulations of the caudal anddorsal fins are in rhythm with those of the pectorals, their movements were counted instead. Actually, the over- all picture of a parent fanning suggests a pumping movement 26 Figure 4. The more common positions assumed udEr the lean-to during fanning of the eggs. Upper drawing is of Hemichromis, lower is of C. nigrofasciatum. _ 27 with the motion of the entire body reflecting the tempo of fanning. 2. The digital fanning tended to occur in distinct units which were termed bouts. On the basis of preliminary studies, and the work of others (Barlow,1964, Mertz and Barlow, 1966, and Mertz, 1967) a bout was considered as a group of beats separated from the preceding and succeeding groups by intervals of at least two seconds. Six parameters of fanning were either measured or derived: Measured a. Bouts: the number of fanning bouts per ten minutes. b. Beats: the number of fanning beats observed per ten minute observation. c. Fanning duration: the total time (in seconds) spent fanning per ten minute observation period. This was obtained by summing individual times for all bouts and was measured to the nearest half-second. Derived a. Tempo: the average number of beats per second for all fanning per ten minute observation period. This measurement was obtained by dividing the number of beats by fanning duration. 28 b. Beats per bout: the average number of beats per bout for all fanning during ten minutes. This was obtained by dividing beats by the number of bouts. c. Bout duration: the average duration in seconds of all bouts within the ten minute observation period. This was obtained by dividing fanning duration by the number of bouts. In addition to the various aspects of fanning, the frequencies of two other behavior patterns were recorded: 1. Nips: throughout the duration of egg care, the parents nip at the eggs and eat moldyvlooking, and oc- casionally, healthy eggs. All the nips directed at,.or within l-2 cm. on either side of the spawn were counted. 2. Departures: were the number of times a fish left the lean-to area during the ten minute observation period. Records were also kept of the time and place of all spawns, the time of hatching, and the time the wrig- glers became free-swimming larvae. Spawn-Position Preference The locations where successive spawns were deposited were recorded. When simulated "eggs" were allowed to remain with the pair until the next spawn, the pair often chose to spawn among.or near these "eggs." Figure 5 shows the loca- tion of the natural eggs with respect to the simulated eggs on the spawning slate. These diagrams were made after 29 22 37 RE. 9.. . ffi"‘—T . . . . Figure 5. POSltlon of "own" Spawn (thin line) in rela- tion to location of the already present simu- lated spawn (heavy line). Spawns from several different individuals are represented. The number alongside the spawn diagram is the pair number. 39 8 30 egg-hatching when the spawning slates had been removed from the tanks. The scars left by the natural eggs were filled in with India ink and their outlines, as well as those of the simulated eggs were then traced. Statistical Analysis Wherever feasible a t-test or an analysis of variance was performed. The results of both the choice and non-choice situations fall into a time sequence that does not lend itself to an overall statistical analysis. However, the standard error was calculated for each point ' in time, and graphs (e.g., Figures 9 to 12) have been drawn showing the means and one standard error on each side of each mean. Where two series of observations were so di- vergent that their standard errors did not overlap, it is assumed that they represented statistically significant differences. RESULTS The main portion of the results is concerned with» the effects of presenting various sized "eggs" on egg care behavior. The relationship between egg care and the egg stimulus was investigated by two techniques: (1) a non-choice situation where the real eggs were removed and simulated "eggs" put in their place, and (2) a choice situation where real eggs and simulated "eggs" were pre- sented simultaneously. Most of the data presented here concerns the former condition, since this did not involve the confounding effect of real eggs visually changing with time; the purpose of the choice situation was to test for discrimination and/or preference. The Non-choice Situation.--The exchange of natural eggs for simulated "eggs" of the same size resulted in no appreciable change from the baseline in the fanning param- eters measured; exchange for "eggs" either larger or smaller in size than the natural resulted in a decrease in all of the fanning parameters; removal of the eggs altogether resulted in the most rapid decrement of fan- ning but not in its total extinction. The Choice Situation.--Simultaneous presentation of natural eggs and simulated "eggs" of the same size 31 32 resulted in no preference between the two during the very early periods of the egg care cycle. As development pro- ceeded, the natural eggs become visually distinct from the simulated ones. The female at this time discriminated in favor of the natural eggs by fanning more at these than at the simulated "eggs." The presentation of simulated "eggs" of various sizes did not appreciably alter the temporal organization- fanning. The trends in this organization in the exchange groups, as well as the baseline, confirmed those found by Mertz (1967): Changes in number of beats and duration corresponded closely; also, changes in beats per bout duration paralleled one another. In all groups these four measures showed high values during the early period> (first twelve hours), and declined to a lower level which was maintained until just before hatching. At hatching' there was again a decline, and finally, cessation. Tempo differed from all of the other fanning parameters in its relative stability during the entire fanning phase. It showed neither the early decline or the rapid decrement at hatching. Parameters Relevant to Spawning The three sections which follow are intended to present background data relevant to the egg exchange experi- ments '. 33 Choice of Spawn Site Table 1 shows where the females chose to spawn- during the course of the study and the preceeding pilot study. Eightzthree percent deposited their eggs under the lean-to and this indicates its effectiveness as a spawning site. Egngeasurements Simulated "eggs" used in the exchanges were made by hand as previously described. Their sizes and shapes were visually determined. It was necessary to make them rapidly, since egg exchanges were made within a few hours of spawn deposition, and irregular configurations had to be duplicated. Because of this time factor the artificial "eggs" could not be made to resemble a fish's own in size, shape, and homogeneity as closely as would have been desired. Subsequently, an analysis was made to describe accurately the size (area) and shape (shape index) of both the real and simulated "eggs." Length and width were measured to the nearest 0.5 mm. with an ocular micrometer (Collins, 1965). Six samples, each consisting of eggs taken from one pair, or artificial "eggs" from one simulated spawn, were used. Thirty eggs were measured in each sample as observed from the top. These were chosen by haphazardly moving the spawning slate to which they were attached. The egg to be measured was 34 oga o.o owo m.o m.H m.H h.m nsmhuv%o.mm m.m m.¢m a.mm a.mm unmonom . sz me o o a N N e e. m mm m mm mm om macacsmmm . . mo HonEsz Eonumm. noon mooflm nopa>ao coaumooq mpmHm oumam . . czmmm l l a a t t o o T T . ouucmmq may meamuao ouncomq on» opHmcH muwm csmem Ho DOHOSUII.H OHQMB 35 the one closest to the micrometer. From the length and width measurements, sectional area and shape index (Width/ Length) could be calculated. The sectional area of each' egg was computed as the area of an oval (Length x Width x n/4). The latter was an indication of deviation from roundness. Shape Index.--An hierarchial Analysis of Variance was performed to determine homogeneity among egg types. Table 2 gives a summary of this analysis. Figure 6 shows that the simulated eggs approximated roundness (§'= 0.901, 0.937, and 0.947) much more closely than did the fishes' own eggs (It: 0.721). As is also evident from Table 2, there was a highly significant difference among egg types pg£_§g (p << 0.005) and samples (p << 0.005). The various simulated "egg" types were homogeneous with respect to shape index (p > 0.05). The variability within samples also indicated significant heterogeneity (p < 0.005). Sectional Area (Length x Width x n/4).--Again, an hierarchial Analysis of Variance was performed and both samples (p << 0.005) and egg types (p << 0.005) were sig- nigicantly heterogeneous. Table 3 presents the results of this analysis. Significant differences among.both samples and egg types are quite evident from Figure 7, since larger artificial "eggs" have a much greater sec- tional area (§'= 7.206 mm.2) than the similar to natural "eggs." The analysis was applied to demonstrate 1.000 0.900 0.800 O O \l O 0 Shape Index (W/L) C» ox O c> 0.500 0.400 0.300 36 NATURAL > NATURAL < NATURAL = NATURAL 3?" Figure 6.—-The mean shape index of natural eggs, and of simulated "eggs" of the same size, smaller, and larger than natural eggs. Vertical bars indicate two standard errors. The grand mean for each egg category is giVen on the right side of the figure. :-:n\ died !1\ Facapirw.‘ J‘ 2) Mean Area (mm M 37 ‘I’I’Im _ <> RR¢¢¢ OQ IMAM A >' NATURAL Figure 7,--The mean sectional area (mmz) of natural eggs, and of simulated "eggs" of the same size, smaller, and larger than natural eggs. Vertical bars indicate two standard errors. The grand mean for each egg category is given on the right side of the figure. 38 Table 2.--The results of an hierarchial Analysis of Variance of shape index for natural eggs, and simulated "eggs" of the same size, smaller, and larger than natural. Shape Index Natural vs. Similar to Natural vs. Smaller vs. Larger Source SS. .MS. df. F p Egg Type 6.038 2.013 3/20 80.520 << 0.005 Samples 0.498 0.025 20/696 6.250 << 0.005 Similar to Natural vs; Larger vs. Smaller Source SS. MS. df. F p Egg Type ' 0.200 0.100 2/15 1.176 > 0.050 Samples 0.128 0.085 15/522 18.889 << 0.005 heterogeneity among-all of the eggs except those which were larger than the natural ones. Figure 7 shows that "eggs" smaller than the natural ones were significantly smaller (i = 0.733 mm.2 ); the Analysis of Variance between these eggs (natural, smaller than natural, and equal to natural) showed a high significance for heterogeneity among egg types (p << 0.005) and samples (p << 0.005). An Analysis of Variance was also done for natural vs. similar to natural "eggsya’and, although these were roughly the same in sectional area (natural §:= 1.555 mm.2, similar to natural II: 1.766 mm.2), both egg types (p < 0.025) and samples p << 0.005 showed significant hetero— geneity. 39 Table 3.--The results-of an hierarchial Analy is of Variance of sectional egg area (mm. ) for natural eggs, and simulated "eggs" of the same size, smaller, and larger than natural. Sectional Area Natural vs. Similar to Natural vs. Smaller vs. Larger Source 53, ms, df.’ F p Egg Type 4733.92 1578 0/20 39.450 << 0.005 Samples 79.79 40 20/696 52.840 << 0.005 Natural vs. Similar to Natural vs. Smaller Source SS. MS. df. F p Egg Type 107.121 53.561 2/15 123.129 << 0.005 Samples 6.530 0.435 - 15/22 4.860 << 0.005 A Natural vs. Similar to Natural Source SS. MS. df. F P Egg Type- 4.001 4.001 l/lO 8.951 << 0.025 Sample5* 4.471 0.447 10/348 4.382 << 0.005 S awn Hatch Time: Parentally Attended vs. Artificially RaISed Eggs Hatching appears to be a critical event bringing about change in parental behavior. Whenever a spawn ex- change was undertaken (i.e., own spawn exchanged for a simulated spawn or blank slate), the real spawn was placed within a beaker of water from the parents' tank which was 4O agitated with an air stone. The clutch was thus raised without the parents. Time to hatching was recorded for spawns with and without parental care (artificially raised). This was begun when most of the eggs had been deposited on the slate and was terminated when the majority had hatched into wrigglers. The time when fifty percent had hatched was compared between parentally reared and artificially raised eggs (Table 4). A two-tailed t-test indicated that there was no significant difference between the means of the two rearing conditions (t = 1.784, 64 df., p > 0.05). Values from the two conditions were then com- bined, and the cumulative percent hatch calculated. IA cumulative percent hatch curve is plotted along with the egg care parameters in Figures 9-12. The mean hatching time for all eggs was 69 hours. The variability was great in that the earliest time of hatching was 39, the latest, 93 hours. Table 4.--Basic statistics for the time to hatch for paren- tally cared vs. artificially cared spawns; x = time (hours) when most (one-half) of the spawn hatched. xf xzf 82x Sx SE i N Parentally Attended 1710.00 123786.00 84.717 3.530 1.879 71.250 24 Eggs ArtifiCially 2793.00 190719.00 121.573 2.895 1.701 66.500 42 Raised Eggs The Effects of Egg 41 Size Upon Egg Care Behavior Duration of Egg Care.--Egg care was considered terminated when the female no longer fanned the eggs. This behavior ceased at different times depending upon the spawn condition presented to the female (Figure 8). Females presented with a blank slate were the first to terminate egg care (i'= 52 days). group was second (2’5 69 days). The baseline Egg care.ceased earlier when females were presented with artificial "eggs" larger than the natural (2'5 82 days) than when given smaller than natural ones (2'; 124 days) in exchange for their 0WD 0 Finally, the group which persisted the longest, in terms of when the first and last individuals stopped fan- ning was the one given natural sized "eggs" (Y'2 163 days). Table 5 gives the basic statistical data. Table 5.—-Basic statistics for the duration of egg care' for the five groups; x = the time (hours) when a female ceased fanning. Blank Slate Baseline "Control" = Natural < Natural > Natural fo 901 676 2121 1615 1072 szf 63185 38612 353715 208025 103106 E 69.308 52.000 163.154 124.231 82.462 5x2 61.564 288.333 638.808 616.026 1225.603 sx 4.736 22.179 49.139 47.387 94.277 SE 2.176 4.710 7.010 6.884 9.710 N 13 13 13 13 13 42 msoum imam: Hosanna v u 4 msoum 3mm: amusumc u G @5090 :mmw: amusumc A n . moosm Losucoo: mumam x393 n ‘ msoum mcflnmmmn n O accommq .smmfi man no mpam nonpam so HouHm Gunpcmum ado mum moawa Handcufinom .Amusoc :Hv mnmo mam mo Facaumusv mom unaccmm.pmmmoum gowns goddamn mo Hogans o>HumasEdo mcaul.m madman 43 ' 220 2io ‘ ' f fi ' I“ I“ V v ' V MINI of [CG CAIUMM j V I *f t v v v ‘ T v. i i i 6 T a z n " 3m 99! 034401: mm 331qu :0 mm ammnuno 44 Figure 8 shows egg care duration within each group. Curves for both blank slate and baseline were comparatively steep, indicating that the majority of females stOpped fanning within a short time period. Curves for the simu- lated condition were less steep, showing that the individ- uals exposed to these "eggs“ terminated egg care over a longer period of time. A one-way Analysis of Variance of the egg care dura- tion means showed that these were significantly heterogeneous (p < 0.005, Table 6). Table 6.--The results of a one-way Analysis of Variance of the egg care duration means under the baseline and four exchange conditions. Source SS. MS. df. F p Main Effect 33964 8491 4 4.911 < 0.005 Deviations 105476 1729 61 Total 139440 65 This analysis was further extended by means of a series of Tukey T-tests to determine which groups dif- fered from one another (Scheffé, 1959). The results are presented in Table 7. Significant differences (at the 0.05 and 0.01 levels) in the duration of egg care were found between the baseline, similar to natural, and smaller than natural 45 Table 7.--The results of Tukey's T-test, testing all pos- sible pairs of means of the duration of egg care. Blank Slate. ' "Control" = Natural < Natural > Natural Baseline 1.060 N.S. 5.750** 3.020* 0.833 N.S. Blank Slate "Control" 6.815** 4.430** 1.870 N.S. = Natural 2.390 N.S. 4.950** > Natural 1.950 N.S. N.S. denotes No Significance *denotes p < 0.05 **denotes p<< 0.01 groups; between the blank slate"control," similar to natural, and smaller than natural groups and between the similar to natural and larger than-natural groups. Egg care Measurements: Non-choice Situation Fanning, by virtue of its predominance in egg care, is here considered to be the most important aspect of clutch—oriented behavior. The sections to follow present the effect different egg sizes had upon the various facets of fanning. The Number of Fanning Beats.--The mean numbers of beats per 10 minute observation period did not differ Sig- nificantly in magnitude, and showed essentially the same 46 temporal organization whether females were presented.with their own eggs or artificial "eggs" similar in size. Figure 9 shows the temporal organization in the baseline and natural "egg" groups. This involves a high magnitude of beats early, a leveling off at about six hours, and then a dorp at hatching. The two groups were similar up to the time of hatching, at which time differences occurred due to the rapid decrement in the baseline beats. Fanning was maintained, though at-a gradually decreasing level in the simulated "eggs," which did not hatch. In the baseline' group the rapid drop in beats at hatching agrees with what has already been found for this species-(Mertz, 1967). It differs from what has been reported for Gaserosteus (van Iersel, 1953 and Sevester, 1961), Cottus (Morris, 1955), or B3§i§_(Barlow, 1964). No such rapid decrement in beats‘ occurred in any of the simulated "egg" groups at hatching time. The exchange of the females' own eggs for either a 'blank slate or a slate with large "eggs" resulted in the most rapid fanning decrement. There were no significant differences between these two groups at any of the points measured. The temporal pattern of fanning was,.neverthe- less, maintained in both. It is possible that in the‘ period immediately prior to hatching fanning towards the natural eggs may be less than towards simulated "eggs" of apprOpriate size. After three hours the level of fanning 47 in the blank slate groups was significantly lower than in either of the others. It remained low and extinguished early (51 hours). The introduction of "eggs" smaller than the natural resulted in no significant differences from the baseline condition until approximately half way through the fanning- cycle. At this time fanning beats toward the smaller "eggs" decreased to a level similar to that of the blank slate and large "egg" groups. The females maintained levels of fanning toward the small "egg" groups which were similar to the baseline level until about two-thirds through the egg care period. At this time fanning beats gradually began to decline, their magnitude becoming intermediate- between the blank slate "control" and baseline groups. After the time of hatching the three simulated groups still maintained fanning, though at a progressively. decreasing level. The similar to natural group was higher than the small or large "egg" groups. The latter two did not differ in beat magnitude from one another, which re- , mained low until the end of the recording period. Superficially, these data appear to be of the type that would be analyzed by an Analysis of Variance or of Convariance. However, there are several reasons why such analyses are inapplicable to the data. (1) The observa- tions at each point are correlated with those at the other points (e.g., a fish that beats at a high level at 48 cope: unmoumm o>wuma2§o . . . . . . D moonm Louvsoo: macaw xsman IIIII‘ msoum gmmmz Hmunumc v 4 maoum =mmw= Hmnsumc n IIIIIO 96.3 :09»: .Hmusumc A III. ocaammmb IIIIIO ”baboon .mmsmnoxm csmmm o» Hoaum cowum>ummno mafiammmn ouscflfi sou on» mmHMOfiocfl scammu teammaum «mmsmnoxw sammm mo med» mmumoapca Bound .cmmfi map mo.w©wm Hmnuflm :0 House UkuGMpm mso .onm mmcwa amuGONHHom .mmsonm Ham How wmo_nmm mmfiflu moan» cmxmu acaumm coaum>ummbo muscw8.cmu Hom.mummn mcacsmm mo gonads some mnBI|.m musmwm 49 - till 1 "3.1V" “3083:! EAILV'IflWflO mm ml was 9mm; mam 50 one time will generally also beat at a high level at a later time). (2) The general slope of the lines is down- ward, but there is no reason to believe that this lepe should follow any given pattern, so that the change from~ point to point is not predictable or testable. (3) Two of the five groups (baseline and blank slate "control") stopped maternal care at an early stage relative to the other three groups, and this causes a severe change in the variance of these observations, while the tests assume that the variance components remain constant. (4) Many of the tests that might be performed (at points where the results appear to differ enough to "warrant testing") are' an chosen on an E pgsteriori basis, rather than on/a priori basis. For all of these reasons, it seems wise not to place emphasis on complex tests of significance. Table 8.--The results of a one-way Analysis of Variance of fanning beats of the baseline, blank slate "con— trol," similar to the natural, smaller, and‘ larger than the natural "egg" groups at twenty- four hours. Source. SS. MS. df. F p Main Effect 3954368 988592 4 Deviations 2540845 41653 61 23.734 << 0.005 Total 6495213 65 51 To show that complex tests of significance can be applied to parts of the data, a single example was chosen. This example was the beat data for the five groups at twenty-four hours (Table 9 and Figure 9). It was chosen because there were definite differences between the low fanning and high fanning groups and it was of interest to see if the differences were significantly different.~ A one-way Analysis of Variance showed that these five groups were significantly heterogeneous (p << 0.005, Table 8) with respect to fanning beats. This analysis was extended by Scheffe's test for selected comparison-(Scheffe, 1959). This showed significant differences between the groups chosen for comparison (large "egg" and blank slate vs. baseline, similar, and small "egg" groups) at the 0.005 level. Table 9.--Results of a Scheffe test for multiple comparison. Means which did not differ at the 0.005 level are* underlined. M Blank Slate Base- "Control" >Natural auma5§do ....... msong zaonusoo=.mumam xGMHn msoum =mmm= Hosanna v maonm =mmw= Hosanna msonm =mmm= Humans: A 00044:! wcHHmmMQ upcmmmq .wmcmsoxo Gavan on Hoflum coaum>ummbo mafia lemon ounces sou ecu mmumowcc«,coammu cmammfium awmcmnoxo szmmm mo mafia mmumoacca 3ouu¢ .con on» yo moan nonuwm co House pumpsmum mco mum mouse amuGONflnom .mmsoum Ham Mom and mom mosey owns» coxmu pofiumm sowum>nmmno muscflfi so» you mason mo Hogans coma one||.oa mucosa PM" EXCHANGE 1 53 HalVl-I 11433834! amnnnwna S s a .2 a 3 s g g g ... A l j I L. A444 A l A A A A - If 1: T . " m m 150 J! 1F “mm 0131 "d 51008 TIME (hours) 54 The overall temporal pattern of bouts for all groups differed from that of the beat pattern in that there was only a slow decrease of bout frequency over time, ex- cept in the "control" group at hatching. This was in con- trast to the pattern of initial decrease, leveling off, and drop at hatching evidenced in the beat data for all of the groups. The baseline "control" groups, however, stopped fanning and therefore all bouts ceased at hatching time. Clear-cut differences between the group means at the various observation times did not occur as they had in the beat data. Immediately prior to hatching, however,. the order of magnitude of curves for the different groups was similar to that of the beats. The baseline group had the highest bout frequency which was not significantly dif- ferent from the lower natural or the small "egg" groups. The blank slate and large “egg" groups had the lowest bout frequency, often differing in their means from the baseline (Figure 10). Duration.--The mean length of time females fanned per 10 minutes followed the shape of the fanning beat curves (Figure 11). The order in which the five groups occurred at the different observation periods also fol- lowed that for the beats (Figure 11). Tempo (Beats per second).--The overall pattern of beat rate is similar to that of the bouts (Figure 12), 55 nouns unmonmm 0>HumHsEso ....... msoum =Houucoo= mumam xcman maoum =mmm= Hosanna v @5090 =mmm= amusumc msonm =mmm= Humans: A II 00044:: mcwammmn “ocmqu .mmcmnoxwvczmmm ou Moanm coaum>ummbo mafia lemon ouscfifi sou on» mouseapcfi.c0flmwu teammflum «wmsmcoxo czmmm mo mafia mmum0fipcfi 3onn¢ .cmms can no spam nonuam no sound Unmocmum mco mum mwcfla HmucoNfiHom .mmsoum.aam How amp mom mafia» wmucu cmxmu poflnmm coaum>uombo muscfla sou Ham coaumuso mcflccmw some mnBII.HH onsmflm - FAWN EXCHANGE 56 HOLVH LNBSUBJ BAILV'IflWflO I“ "1 ISI ‘ mmmmm 8.¥.f1§1?-§1?.515.?9 1% 10m 1031 udi-“mouvuno TflEflloon) 57 souos usoonoaio>uuoasaso ....... D msonm ...nonusoo= ouoam xsoab 4 msonm ..mmo.. Honsuos v 4 msonm ..mmo.. Honsuos n 0 msonm =mmo= Honsuos A . osuaomon O “Usomoq .omsosoxo szomm ou nounm sonuo>nomno osnaomon oussnE:sou osu mouooubsu sonon,conmHum «omsosoxo.s3omm mo osuu mouoonvsn sonnd .sooE osu mo ovum nosuuo so nonno onovsoum oso ono mosna Housonnom .mmsonm Ham now mop nom mofiuu oonsu soxou connom sOnuo>nomuo ousswfi sou nom Aa\mv omEou sooE ossll.ma onsmum “31'" “3083‘ BAILV'IflWflG #1476115? t ‘ 11712711311011! 5‘ N ‘NIW N31. Jul Odflal NV!“ 59 in contrast to the beat and duration measurements. This constancy of tempo throughout the period was not affected by the absence of eggs or by "eggs" of different sizes. The rapid drop in performance (characteristic of the other fanning measurements) during the initial hours of egg care (0-6 hours) and at hatching did not occur here in any of the groups. Beathper Bout (B/B) and Duration per Bout (D/B).-- B/B and D/B were very similar to one another in-bOth tem- poral organization and the mean level of performance at any given point of obseration. This similarity is due to the fact that tempo (B/D) does not change measurably with. time over the fanning cycle (Figure 12). Both the temporal pattern and the position of the different groups in rela- tion to one another were similar to that of the beat and duration data. Thus, these two parameters added no infor- mation to the latter two aspects of fanning. The basic statistics and figures are given in Appendices 4-12. Choice Situation: Real vs. Sim— ulated "Eggs" Presented Simul- taneously , The preceding sections have been concerned with fanning behavior toward eggs which were either real (base- line) or simulations of the real eggs which had been ex- changed for the female's own eggs. In this situation there was no opportunity for a comparison between the 60 effects of real and simulated "eggs." The results obtained by simultaneously presenting natural and simulated "eggs" of the same size follow. The purpose of this procedure was to test for the effects of visual changes which occurred in the natural eggs over time. Q. nigrofasciatum females, when-simultaneously pre— sented with simulated "eggs" of approximately natural size and with natural eggs, made no apparent discrimination be- tween these until after approximately one day of incubation. Although the variability was quite great in all of the measurements at any point in time, the pattern which ap- peared was one in which the fanning beats toward the sim- ulated natural size "eggs" started with a high frequency and decreased over time. Fanning beats toward the females' own eggs started at a significantly lower frequency, increased by the beginning of the third day to a level significantly greater than the beat level toward the simulated "eggs," and then dropped at hatching (Figure 13). The females appeared to have been discriminating between the real and natural size simulated "eggs." The above pattern is clearest in the beat data but is also evident in the duration, and, to a lesser extent, in the tempo measurements (Figures 14 and 15). In the choice situation the females stopped caring for the simulated spawn earlier than the natural spawn (X length of time fanning was about 51 hours for the 61 .sz ounm oamfiom osu ono msooE msoao mnonfiss «sooE osu mo ovum nosuuo so nonno pnooswum oso ouosoc mosua HousoNnnom .mop non mofiuu oonsu pocnooon souuo>nomno oussflfi o>uu o ousomonmon usnomsoom ..:onsuos osu ou noansum s3omm couoassnm u mononno somo .szoam s30 n moamsonnu womoHov szomm wouoasfinm Hoouusoon so pso.s3omm s30 onozou mamsoosouassnm pocnooon muoon msussom mo nonEss soos osanu.ma onswum Amnsosv ofiue we as 3 we 2. 2 «a Z. 3 a o m o . 11 , , , massage om 4 m 2: m 03 6 m m com m . o 0 35.42" m . .m com oov Mean Fanning Beats per 5 min. 62 .sz onnm oHQEom osu ono msooE msoao mnoQEss «sooE osu mo ovum nonuno so nonno onocsoum oso ouosoo mosua HousoNunom .woo nom moanu oonsu poonooon souuo>nomno oussuE o>um o musomonmon usnom zoom .Aaonsuos osu ou noHuEnm ssomm oouoasfinm n moHonuo somo .ssomm s30 u moamsownu oomoaov ssomm wouoasfium Hoowusoon so oso s3omm s30 ono3ou mamsosouoasfiwm woonooon.sonuonso sooE osalu.¢a onsmnm 1869681 mane Nb mm mm mv mv Nv vN HN ma m m m o 0-- M. ., ...I.V m , soonesz O m OOH omH com Mean Duration per 5 minutes 63 .sz ouum oamfiom osu ono osoos msoao mnonass «sooE osu mo ovum nosuuo.so nonno cnocsoum oso ouosoo mosua HousoNnnom .moo nom mosuu oonsu coonooon souuo>nomno oussue o>nm o musomonmon usuom soom .Aaonsuos osu ou noHuEum s3omm oouoHSEnm n moaonuo somox.s3omm s30 u monsonnu pomoaov.s3omm oouoHsEum HoOflusowu so oso szomm s30 ono3ou MansoosouHSEum coonooon omfiou sooE osB||.ma onsmum Amnsosvhofifle E 8 3 3 m6 .8 3 nm 3 V a m ?\\wo. coo; ommé com; ooo.~ qamoeaz m mxu . .Vmo oom.~ m QQMDHGZ fl ooo.m 804 Mean Tempo per 5 minutes 64 .sooE osu mo ovum nosuno so nonno onoosoum oso ouoonosu mosua HousoNunom .onsmum osu mo mou.osu uo so>um ono msonm.sooo nom msooz .haosoosouasfium uousomonm “mononno,somov ssomm oouoHsEnm o oso Apomoaov s30 onozou Amnsos swv msnssom ofiuu mo sumsoa.osu mo sOHuossm o mo msussom commoum sons3 moaofiom mo noness o>nuoHsEso osanu.oa onsmum Amnsosv msnssom oEuB mo sumsoq ooa .om om on on om ow om om mmmo Honsuo nonsumz n gmmmm= nmumnssnm a Cumulative Number of Females Stopping Fanning 65 simulated and about 66 hours for the natural spawn). Egg care dropped off faster and earlier towards the simulated than towards the natural spawn (Figure 16). However, a one—way Analysis of Variance of the length of time spent caring for the spawn (Table 10) demonstrated that there was no heterogeneity between the natural and simulated "egg" groups (p > 0.05). Table lO.--The results of a one-way Analysis of Variance of the egg care duration means for natural eggs and simulated "eggs" presented simultaneously. Source SS. MS. df. F p Main Effect 946 946 1 3.625 > 0.05 Deviations 3916 261 15 Total 4862 16 Nip§.--Nipping at the clutch was absent or occurred at a low frequency except when the natural eggs were hatch— ing. Nipping frequency had a very low level at all times in the egg exchange groups. Figure 17 shows beat and nipping curves for a rep- resentative baseline female, and was generally characteris- tic of most individuals. Thus, there were no appreciable changes in nips until hatching. At this time nipping fre- quency increased several fold. Concomitant with this in- crease, fanning beats toward the eggs decreased abruptly. 66 . .ouosnono oooosm an oouoouosn on msflsouos Ho ofiflu ouosnxonamm .msOflunosoo osuaomon nooss Ram oaosomv Honou>aosa o>uuousomonmon o nom muoon mswssom oso msummus now mo>nso||.na onsmnm 67 NUMBER of BEATS- CON 09¢ .00 3.: Gou— 00¢— .00— N? on 6.52.58.— 3 «a O— N— 0 n" l 1 d I I 1 r... 4/44\ 4/44\. «=3 2.5.: ....../\/\ n NUMBER of HIPS 68 Other Aspects of the Structure of Fanning While the purpose of this study was to determine the effects of egg size upon egg care behavior, it was also possible to obtain data relevant to certain aspects of the structure of fanning. The sections to follow present these results. Fanning Beats vs. Bouts.--Fanning beats were plotted as a function of bouts for each of the five conditions. Figure 18 gives the results for four of these. The general pattern, which appeared consistently in all groups, was one of an inverse relationship early in the egg care cycle and a direct relationship during the later phase. The corre- lation between beats and bouts was inverse when fanning .frequency was low and direct when fanning frequency was high. The change in this relationship occurred at approx- imately 200-300 beats per 10 minutes. Interval Preceding and Interval Following vs. Dura— tion of One Bout.--The inter-bout interval does not appear to indicate a work and rest relationship with fanning beats. The intervals preceding-(IPB) and the interval following (IFB) a bout were plotted against duration of bouts on a scatter diagram for seven different females. Figures are given (Appendices 13-17) for three representative indi- viduals of these seven. None of these demonstrated corre- lations. This held true for either individuals of long or 69 Figure 18.--Fanning beats as-a function-of bouts for the baseline and three of the four exchange con- ditions. The outlined areas show the highest density of points at each of three times during the fanning cycle. These times are early (0—24 hours) middle (stippled; 25-72 hours), and late (beyond 73 hours). 70 = NATURAL BASELINE BLANK SLATE "CONTROL" FANNING BEATS v v v v v 71 short fanning duration. No correlation was noted with either IPB or IFB and duration. Also, there were no evi- dent correlations among early, intermediate, or late points during the egg care period. ,The only pattern which appeared was that longer durations, by and large, were associated with shorter IPB and IFB, and less often with long inter- vals. Thus the shorter durations were associated with longer intervals preceding and following bouts. Bout Duration and Fanning Beats vs. Tempo.--Bout duration was plotted against tempo on a scatter diagram (Appendix 18). This was done to determine whether a relationship existed between the length of bouts and the speed of fanning. There was no evident relationship be- tween the two. Also, there were no apparent relationships between fanning-beats and tempo when plotted against one 7 another (Appendix 19). Time Between Spawns The period of egg care, as already noted, was con— siderably extended by the continual presence of simulated "eggs" (Figure 9) . Females that extended their period- of egg care appear also to have extended their inter-spawn interval. The time of spawn deposition, hatching, and the time wrigglers became free-swimming was recorded for all individuals which spawned. A spawn was considered to be‘ free-swimming when the majority of the young were able to propel themselves above the substrate. 72 .oooam su umoH ono3 Amy Honsuos osu sosu nomnoa oso .Amv Honsuos osu sosu noaaosm .AHHV Honsuos osu ou noHuEum =mmmo=.oouoHsEum onoss sOHuosuHm osu now .2 osu muoxoonn su mnonfiss msfiasomEoooo osu .msomE ono moaonno puaom .ms3omm soo3uon mwow mo non&ss.osu mo souuossm o no oouuoam moussufi sou nom muoou com o>ono msnssow mmoo mo nonfiss osBII.mH onsmam 73 mmsnsaomm soo3uom mmoo mo nosfisz on on 0H 00 0130 0.0 0H Number of Days Fanning Above 200 Beats per 10 min. 74 Table ll.--Basic Statistics for time (days) between spawnings (x) when the natural spawn was left with the parents and when it was exchanged for either a blank slate or simulated "eggs." No Egg Exchange 2x 2x2 5x2 5% s; xl z Spawn Either Re- moved as Pro- larvae or First 740 14004 16.198 0.395 0.628 18.049 41 Day Larvae, or - Eaten as Eggs by Female Egg Exchange Blank Slate "Control" 277 5025 31.427 1.849 1.360 16.310 17 Simulated "Eggs" 406 7796 14.450 0.657 0.810 18.455 22 The time between spawns was recorded for three categories. The first was a blank slate "control" group, which has been presented under this same heading in pre- vious sections (see page 24 of Methods and Materials).‘ Here, the natural eggs were removed one to two hours after the completion of spawning. The second category was a simulated group which included all the cases in which artificial "eggs" had been exchanged for the female's own. The third was a baseline group where the spawn was left with the female until the first day of free-swimming at which time the larvae were removed. Table 11 summarizes these conditions. 75 Table 12.--Basic Statistics for the number of days-spent fanning above 200 beats per 10 minutes and the number of days between spawnings when the natural spawn was exchanged for one of simu— lated "eggs." Number of Days Spent Fanning‘ Above 200 Beats per 10 minutes xl 2 S S— S- x x Simulated "Eggs" Equal to Natural 3.835 0.274 0.523 6.786 14 Simulated "Eggs" Larger than - Natural 0.2617 0.044 0.210' 4.333 6 Simulted "Eggs" Smaller than Natural 1.143 0.163 0.404 3.143 7 Days Between Spawns 2 . 2 S Si. S; xl z Simulated "Eggs" Equal to Natural 30.423 2.173 1.474 21.500 14 Simulated "Eggs" Larger than~ Natural 55.367 9.228 3.038 19.833 6 Simulated "Eggs" Smaller than Natural 34.333 4.905 2.215 19.00- 7 When the means of the three groups (baseline, blank slate "control," and simulated) were-tested against one another by a series of paired t-tests, it was found that the differences among these means were as expected, though 76 not statistically different (Table 13). The presence of simulated "eggs" tended to increase the time between spawns, while the removal of eggs soon after spawning tended to decrease the time (Table 12). The effect of leaving the spawn with the female appears to have been comparable to leaving the simulated "eggs" until broodi- ness was lost. Table 13.--The results of t-tests comparing the mean inter- spawn interval under three conditions (see text). df t p Baseline vs. Simulated 61 0.123 > 0.9 Baseline vs. Blank Slate "Control" 56 0.423 > 0.6 Blank Slate Control vs. Simulated 37 1.423 0.2-0.1 Fanning decreased with time (Figure 9). At very low beat frequencies it is difficult to be certain of the broody condition of a particular individual. Thus a cut- off point at 200 fanning beats per 10 minutes was chosen. as a criterion of broodiness toward the eggs. Days fan- ning above 200 beats per 10 minutes was plotted against days-between spawnings (Figure 20). A cutoff point was established at 28 days, since no females fanned above 200 beats between 25 and 29 days. Thus no data after 28 days . 77 were included in the analysis. For unknown reasons some pairs ceased spawning for considerable periods of time. Later, some of these resumed spawning. This seemed to have occurred independently of the conditions noted for their previous spawns.r Thus, when more than 28 days elapsed it was difficult to determine whether the extended period of time was due to a particular treatment given during the previous spawning phase, or to factors still unknown. Figure 20 shows a trend which is not quite significant (correlation coefficient = 0.3895, p > 0.05). This trend, a direct relationship between interspawn interval and time- spent caring for eggs, is due to a few low and high points and appears to be unimportant within the center of the range. Spawn Site Preference Most spawns were deposited within the lean-to (Table 1). These could have been deposited as follows: on the horizontal section within the lean-to (position A), on the vertical wall of the lean-to (position B), or on' the under surface of the diagonal slate (position C) (Table 14). The outer surface of the diagonal and sites away from the lean-to were infrequently chosen, although they made up most of the surface area available for spawning. There apparently was no preferred site for deposition of eggs within the lean-to.- In addition, females showed little if 78 .Amo.o A a .mon .mmm.n u «xv conumuomaxm was scum unmnmuone mausoonmusmnm uos mn.sous3 .co>nomno ono3 m.ma cso msOnunmom noHuEum sn ob caso3 ma uosu mo3 sonuouoomxo osu .mnuom mm osu mo .nmm.o n ramm.oc + Niamm.ov + mlmsm.ov an nuns conunmos nannsnm 8 no sunannmnoua was .sno>nuommmmmn o.ocmw.m .4 muonunmom an mm~.o cam .mmm.o .msm.o mo monososvonm co>nomuo osu o>os usu .coooam MHEocson.ono oHoEom so>nm o >9 ms3omm uosu souumfismmo osu so couoasouoo mos msouunmom noHuEum su mssomm o>nmmooosm mo wososvonm couoomxo one .A.oanou osu mo nosnoo uwoa noBoH ocu sn Haoo osu sn.s30sm ono msouunmom omosev .ounsooa osu mo U.sonuumom sn ha cso .m souuumom su cm .s sonunmom.sn mosnszomm mm ono3 onone .omsns3omm o>ummooosm mo sonuumom oso nocno||.va oanoe N 79 balm mm H0352 0H Mfimh msuszomm sunsom mswskomm onusa msnssomm csooom msuszomm umnnm mmlv , xOHnm 4H mmlm ma: 10H mmlv NNI ml I IHH one Hum balm omlm malm mlv ml mmlm mmlm NIHH I mun mmlm wNIm onlv Hmlm mm: film HHIm mun mNIN 4N mm mm ma OH a m h m H nonfisz oHoEom mmsusSomm o>Hmmooosm mo souunmom oso nonno 80 any preference for using the same position for two suc- cessive spawnings. However, if simulated "eggs" were pres- ent,-they tended to lay their eggs among them. Of the 59 spawns deposited within the lean-to, 37 percent were on the horizontal section, 33 percent were on the vertical wall, and 28 perCent were on the undersurface. This is not significantly different fiFom random (x2 = 4.31, 2 df, p > 0.05)° In the analysis that follows, the ob- served frequency of spawning at each site is used to pre— dict the probability that a fish will spawn at that site. Wherever two successive spawns of a given female‘ occurred on the same section of the lean-to, they were con- sidered to constitute a "similar" pair. Successive spawns on different sections of the lean-to were considered as "dissimilar" pairs. There were 16.5 "similar" and-22.5 "dissimilar" pairs. (Occasionally eggs of a single clutch were deposited on two different slates. If one of these sites was concordant with that of a previous egg deposition, the spawning was classified as being 0.5 "similar" and 0.5 "dissimilar.") From the distribution of site choices 0.337 is the expected frequency of choosing the same site on two successive occasions (Table 14). There is no significant evidence that one spawn was followed by another in-the same position of the lean-to (x2 = 1.328, 1 df, p > 0.05). During the latter portion of this study, simulated "eggs" were left under the lean-to for a time longer than~ 81 the usual fifteen days. A female when spawning again- usually deposited her eggs on the slate with the simulated "eggs," often right among them (Figure 5). She did this even when the "eggs" were quite moldy and misshapen, which they frequently became after remaining in the tank for a. long time. Whether or not the presence of eggs (artificial in this case) tended to significantly increase the bias to spawn on a special lean-to segment was tested by com— paring the frequency of "similar" vs. "dissimilar” pairs at the various spawning sites for three conditions: (1) successive spawns in a series of two or more spawns with- out the presence of artificial "eggs"; (2) successive- spawns in a series of two or more spawns where simulated "eggs" of the same size as the natural were present; and (3) successive spawns where simulated "eggs" larger than natural were present. There appeared to be a strong tendency for females to spawn on the same slate that contained simulated ”eggs" of natural size. The distribution of sites at which the females spawned was tested on the assumption that this distribtuion should have the same frequencies that were observed for spawns deposited when simulated "eggs" were not‘present (Table 15). According to this assumption, 0.337 of the spawns should have been in a "similar" posi- tion to the eggs, but 11.5 of the spawns actually were "similar" and only 4.5 were "dissimilar." This is a highly 82 .lmo.o A a .me n .4sm.m n no mocmnmuono unmonunc ImHm os mo3 onosu .osHHomon osu umsHomo coumou onoB sOHuHcsoo mHsu ono3 mnHom =noHHEHmch= oso gnoHHEHm= sosz .Amoo.o v m .mc H .Hmm.m u xv NH oHnoB sH mszomm mm ocu mo soHuannuch osu so comon sOHuou loomxo Eonm oosonoMMHc usooHMHsmHm o mH SOHS3 mnHom =noHHEHmch= m.v cso =noHHEHm= m.HH ono3 onose .usomonm ono3 Honsuos osu mo oNHm oEom osu mo =mmmo= couoHsEHm onocB mszomm o>Hmmooosm mo soHuHmom oso nocnoul.mH oHnoB 84 significant difference from random (X2 = 9.891, 1 df, p < 0.005L and indicates that females place their eggs on a slate bearing "eggs" with a much greater frequency than expected by chance alone. It was found that the frequency of "similar" parts was not significantly higher than the "dissimilar" when simulated "eggs" were present than under the baseline conditions (x2 = 2.874, 1 df, p > 0.05). A similar test was given where "eggs" larger than the natural had been left under the lean-to (Table 16). The sample was small in that it consisted of only 6 "simi- lar" and 4 "dissimilar" pairs. Spawning did not occur significantly more frequently with "similar" than with‘ "dissimilar" pairs (p > 0.05). When compared to previous spawns, the presence of larger than natural "eggs" on the slate did not significantly bias the placement of a 2 0.417, ldf, p > 0.05). succeeding spawn on that slate (X Further, there was no significant difference when the fre— quency of occurrence was compared relative to the natural size "eggs" (X2 = 0.393, 1 df, p > 0.05). Insufficient data prevented a test for simulated "eggs" smaller than natural or for instances where the spawn was consistently placed on the outside of the lean-to. 85 .lmo.oaA a .uo n .mam.o u xv wocmnmuune unmouuncmnm o: mo3 OmHo onusu usomonm ono3 =mmmo= coNHm Honsuos sons mnHom =noHHEHmch= oso =noHHEHm= mo wososvonm osu uosHomo coumou 60:3 .lmo.o A a .uo n .an4.o u axe moamnmuono unmonunamnm os mo3 onocu .osHHomon osu umsHomo coumou ono3 sOHuHcsoo mHsu nocss mnHom =noHHEHmch= oso =noHHEHm= sosz .Amo.o A m .mc H .NmH.m n mxv NH oHnoB Eonm mszomm mm osu mo sOHu Iannuch osu so comon sOHuouoomxo Eonm usonoMMHc MHusoOHst ImHm uos ono AUHsz mnHom =noHHEHmch= v oso =noHHEHm=mwono3 onose .usomonm ono3 Honsuos osu sosu oNHm sH nomnoH =mmmo= couoHsEHm onosz mmsHssomm o>Hmmooosm mo sOHuHmom oso noonOII.mH oHnoe REBEE‘I ooooo ooooo mmmmm DISCUSSION This study was concerned with an analysis of the relationship between an environmental stimulus, eggs, and the care behavior associated with this stimulus. Of the measurements taken, egg care was best defined in terms of fanning, a multidimensional behavior which circulates water over the Spawn as a result of the beating of the parent's pectoral fins. The work of Kfihme (1964) with Hemichromis and Mertz (1967) with g. nigrofasciatum, along with-the egg exchange results of Greenberg (1961, 1963), Collins (1965), and Myrberg (1965) with several cichlid species established that vision,.not olfaction, is most likely the major stimu- lus involved in egg care behavior. This is a question to which Noble and Curtis (1939) had directed their attention some twenty-five years earlier (in Hemichromis). They reached the same conclusion. Accepting the hypothesis that vision is primarily involved, it was inferred that exact, or close, visual replication of a species‘ spawn should not appreciably alter that species‘ behavior when presented with this simulated spawn. Furthermore, it was believed that changes in various critical aspects of the 87 88 spawn should result in quantitative, and possibly qualita- tive, behavioral changes in the parent's responses. It was further hypothesized that egg size might be one of the more important variables in influencing egg care behavior. Therefore, this was the only variable pur- posefully manipulated. According to the data presented here, fanning fre- quency decreases with time regardless of the size of the eggs presented to the female. The overall temporal pattern is identical to that which Mertz (1967) found for this species. This pattern across egg groups implies some sort of internal, possibly-hormonal effect, and/or some sort of stimulus satiation over time. Complete removal of the spawn resulted in the most rapid decrease. However, a low level of fanning was maintained and lasted for as long as three days in some females. This was directed toward the slate which previously held the spawn. The magnitude of fanning-after the eggs were removed is thought to be the level of fanning under hormonal control. The difference between this and the level when the eggs were present is possibly due to an additive effect of visual stimulation- byvthe'eggs° In other words, the magnitude of fanning in the absence of eggs should be similar to the level obtained by injecting non-broody g. nigrofasciatum with a physio- logical dose of prolactin to induce fanning, a procedure reported by Blum and Fiedler (1965) and Fiedler (1967). 89 This interpretation-differs from that of Mertz (1967) who offered two explanations for the persistance of fanning in g. nigrofasciatum after removal of all visual cues associ- ated with the clutch. First, fanning "motivation" might" be sufficient to cause its expresSion even in the absence of an adequate external stimulus. Secondly, the location; of the spawn might serve as an effective stimulus for fan- ning. Mertz emphasized the importance of "motivation" as a partial explanation. The basis for this explanation was his finding that fanning tended to "overshoot" the normal level when the appropriate stimulus again became available. Both of the above factors may contribute to the persistence of fanning. However, they do not adequately explain its two to three day persistence in the absence of the appro- priate stimulus which was found here for some females. A slow hormonal change underlying the fanning decrement might be a more parsimonious explanation. The non-choice situation was designed to avoid the confounding effects of a temporal visual change in the natural eggs. This occurred in the choice situation where natural eggs were left in view at the same time with the simulated "eggs." The data seem to support the conclusion that egg size is one of the more important factors which- influence the egg care behavior of g. gigrofasciatum fe- males. This is emphasized even more by the fact that the artificial "eggs" differed significantly in shape from the 90 natural ones in that they approached roundness much more closely (Figure 6). As predicted, the sight of "eggs" similar in size to the natural did not appreciably alter the organization or magnitude of fanning. There was no decrease in fanning toward natural sized simulated "eggs” (compared with the real) desPite shape differences. It did, however, result in the absence-of the beat and dura- tion decrement at hatching and inva prolongation of fan- ning. "Eggs," either larger or smaller, were disciminated. Their exchange for the real eggs resulted in a decrease in all of the fanning measurements except tempo. This decrease was such that the counts in these groups approached similarity to those for the blank slate "control," where no eggs were present. Larger "eggs" caused a more marked decrease in fan- ning than did smaller "eggs." This most likely was due to the fact that the larger "eggs" were larger than the natu- ral by a factor of about five, whereas the smaller "eggs" were smaller only by a factor of one-half. Removal of the clutch and its replacement with an identical slate minus real or artificial spawn shortly after spawning resulted in the most rapid decrement in fanning. These results are compatible with those of Collins (1965) who found "Total rejection of heterospecific eggs . . ." when eggs of Tilapia sparrmani were exchanged for those of Aeguidens latifrOns during any stage of the fanning phase of either species. 91 The eggs of these two species, as described by Collins, are vastly different. Not only do they differ in size and shape (see Appendices l and 2), but in color as well. The data are also in accord with the reciprocal spawn exchange performed by-Myrberg (1964) and Greenberg (1961, 63) with, several cichlid species. These workers achieved acceptance only between species whose eggs are visually quite similar. Collins (op. cit.) further stated that the only instance of egg acceptance he obtained was in two cases where Aequidens were given Hemichromis eggs.. Again, of the three species tested these two have eggs approximately similar in shape and size. It should be mentioned that other workers (e.g., Myrberg, 1966) have cited spawn configurational differences as possibly influencing spawn recognition. Collins (1965) remarked that ". . . no such differences were noted in the three species studied. . . ." As already stated, Figure 1 shows differences in spawn configuration between Hemichromis bimaculatus and g. nigrofasciatum. Although it is unlikely that this contributes a major effect in the control of fan- ning in most species, it is still a possible minor factor in-spawn-recognition. Natural eggs (at 26°C) hatch after about 72 hours. This event, alone, or in combination with visual and/or chemical changes in the eggs prior to hatching, has an effect upon egg care behavior. At the time of hatching, 92 fanning decreases abruptly, while nipping increases just as abruptly (Figure 17). Neither of these changes wasc observed in the situations where artificial eggs were used. Instead there was a gradual decline and disappear- ance of fanning. No increase in nipping frequency was observed. These differences imply some sort of influence of a change in the natural eggs upon the'egg care behavior. The major change which occurred involved darkening due to the development of chromatophores in the embryo as hatching- approached. Concomitant with this change is a possible increase in embryonic movements. It could be hypothesized that some aspect of both or either of these changes reaches a threshold level and releases the high frequency of nip- ping at hatching. A simultaneous inhibition of fanning might also be brought about. Another possible factor is a chemical change postulated to be associated with the breakdown of the egg membranes (Myrberg cited by Mertz, 1967). Nipping itself may be‘a behavioral adaptation aiding the larvae to emerge from the chorionic membranes. The results also indicate that the presence of simulated "eggs".not only causes 9. gigrofasciatum consid- erably to extend its period of egg care but also appears~ to lengthen the time between spawns.» Thus, the continued presence of eggs which neither hatch, nor present a temporal visual change, appears to maintain the female in an extended egg care phase of parental behavior. Such lability, although 93 not as extreme, was noted in the fanning phases of T. gparrmani and A. latifrons by Collins (1965). One might postulate a mechanism by which the visual presence of eggs or larvae inhibits the spawning readiness of the female. Unfortunately the data are not sufficiently extensive to carry such an hypothesis very far. The data from the spawn site choice situation sug- gest two conclusions. First, they mitigate against the. suspicion that the simulated "eggs" were~not treated as eggs, but rather, as artifacts. If treated as artifacts they would most likely have been removed by the females. This would occur when a pair becomes ready to spawn. Secondly, these data indicate that prior presence of eggs influences a female to deposit her eggs in the same place. This might possibly suggest a mechanism which limits sub- strate spawning cichlids to depositing eggs at a particular spot rather than over a diffuse area. Further work is indi- cated regarding this point. For example, one question con- cerns-the minimum number of eggs which could bias the female's choice of depositing more eggs in a specific spot. When several dimensions of fanning and their pos- sible interactions are examined, no relationship was noted between either the interval preceding or the interval following a bout and the duration-of that bout. This is identifical with the findings of Mertz (1967). Thus, absence of stimulation by the eggs, when the females left 94 the lean-to, did not result in an increase or decrease in fanning upon re-exposure to the eggs. _"Motivation" for fanning does not build up during inter—bout intervals. It is possible, however, that some other dimension of fanning is affected by the duration of the inter-bout intervals. For example, either the strength of the fanning beat or the tempo might be correlated with the duration of either the interval preceding or succeeding a bout. Of these, the former was not measured, and the latter was not measured precisely enough. There were no apparent relationships- between duration of bouts and the speed of fanning, in that females with a propensity to fan for a long time did not fan any faster than females exhibiting low measures of fanning beats and short bout durations. In examining fan- ning beats as a function of bouts, however, an unexpected relationship was found. Whereas one might expect bouts and fans to be directly proportional, this was only par- tially the case. Bouts and fanning beats were inversely proportional at the high beat levels, but directly related at low levels. The change in this relationship occurred at about 200 beats per 10 minutes across all egg groups. As Figure 18 shows, the two types of relationships occurred at different stages of the egg care cycle. The inverse rela- tionship occurred during the early stages; the direct cor- relation during the later stages. In Eggig, Barlow (1964) found a general negative correlation between bout duration 95 and tempo. The exception to this was theeresults.for day one. On~day one he found that above 4 beats per second, duration and bouts were inversely related, while below this tempo the relationship was reversed. A possible reason for this observation is as follows. During-the first twenty-four ' hours after egg deposition the motivation for egg care is high. In fact, some females fan almost continuously during this time. To get a high number of bouts, there must be intervals. These represent time taken away from fanning. Later in the cycle the female spends less and less time‘ fanning, and the intervals are far longer than the bouts. Therefore each bout represents some beats added to the basic nothing: more bouts, more beats. On the Function of Fanning Baerends and Baerends van Roon (1950) pointed out that cichlid eggs succumb to fungal infection if removed from the parents. This appears to have been demonstrated in the case of mouth brooding cichlids (Shaw and Aronson, 1950). Baerends and Baerends van Roon, however, concluded that fanning prevents fungal spores from settling on the clutch. This may well be a partial function of fanning in cichlids. The high degree of success with regard to the development of the artificially reared eggs, however, appears masomewhat mitigate against this hypothesis at least in the case of g. nigrofasciatum. Also, as cited by 96 Mertz (1967), and confirmed by this study, some of the eggs developed fungal infections while-receiving parental tending. He-further found (op. cit.) that g. nigrofasciatum deposits eggs preferentially on~a vertical surface. He states ". . . it is difficult to envision sedimentation as a factor critical to the survival of the young when clutches are preferentially deposited on a vertical surface." In this study, the lean-to presented three surfaces (excluding the outside) for egg deposition (Table 12, page 75). One of these was vertical, another horizontal, and one was an. angle between the two others. The latter was so arranged that in order to spawn on it the female had to deposit her clutch nearly in an upside down position. Spawnings oc- curred with nearly equal frequency on all of these surfaces. This contradicts the findings of Mertz. It should be further mentioned that fungae are only secondary invaders at least on the eggs of mouth-brooding Tilapia raised extra- orally (Shaw and ArOnSon, 1950). The initial cause of egg decay is bacterial infection. It might be more parsimo- nious, in support of sedimentation, to hypothesize that fanning in g. nigrofasciatum functions to maintain a more or less continuous water current over the clutch and that this prevents the settling of an excessive bacterial popu- lation. In fishes whose parental repertoire includes fan- ning one of the major functions suggested is facilitation 97 of gaseous exchange between the embryo and its environment. The total metabolic activity of the brood increases with the approach of hatching. This results in an increase in fanning which presumably removes metabolites and supplies the clutch with an increased oxygen source (van Iersel, 1953, Morris, 1954, 1958, Sevester, 1961, Barlow, 1964, and Mertz and Barlow, 1966). The species tested included Gasterosteus aculeatus, Pungitus pungitus, Badis badis, Florinella japonica, and Cottus gobi. The shape of the baseline as well as the simulated "egg" fanning curves, suggests that fanning in Q. nigrofasciatum is not regulated in accordance with the metabolic requirements or output of the young. This supported the findings of Mertz (1967) where the temporal structure of fanning did not-correlate with the supposed temporal changes needed for gaseous exchange. In-fact, fanning dropped rapidly at hatching when it presumably is needed the most. Furthermore, accord- ing to Mertz (1967), when olfactory cues are removed from the natural eggs, fanning is not affected. This was con- firmed here by the behavior towards the simulated group of "eggs" similar to the natural.- It therefore appears un— likely that fanning in most cichlid species serves the function of "aerating" the clutch. It seems probable that fanning contributes to the survival of the young. It has yet to be demonstrated, 98 (however, even on a statistical basis, that fanners produce more surviving young than non-fanners. This could easily be tested with eggs raised under various artificially tended conditions. These could then be compared with a baseline of parentally raised eggs. This would provide little, if any, information regarding the immediate func- tions of fanning which might be subtle. For example, a slight selective advantage might be accorded to fanners with regard to the prevention of sedimentation of detri- mental bacteria or protozoa which are a portion of the microecology of this species‘ habitat. The egg care system of Q, giggofasciatum is extremely complex. Understanding complex systems such as the behavior of whole organisms is best handled by dissecting the components into simply analyzable parts. One can then resynthesize these into a meaningful whole. This philosophy was the one underlying the above study. In conclusion, it might be added that, at present, the findings are at best a partial picture. To gain a more complete understanding of the multivariate egg care inter- action with the spawn stimuli, one needs to look at more of the critical spawn parameters. One of these might be color. Within the genus Tilapia alone there is great variability in egg color. For example, 3. macrocephala eggs are yellow-orange, T. nilotica light yellow to yellow, T. galilaea olive green. The species-T. tholloni, T. zillii, 99 and T. guinensis, all substrate spawners, range from green to brown (Dambach, 1963). Finally, configurational dif- ferences as already mentioned plus egg shape, density, and texture are factors which could take part in shaping the egg care system in g. nigrofasciatum, especially with regard to fanning. Also, the results presented here indi- cate that egg care behaVior is influenced by internal factors expressed in the presence of appropriate stimuli of the spawn. These are visual. The causal factors under- lying the expression of these behaviors are not known as yet. It is strongly suggested that these may be, at least in part, hormonal in nature. S UMMARY 1. This study was designed to test the hypothesis- that egg size influences various parameters of~parental. (egg care) behavior in Cichlasoma nigrofasciatum either by increasing, decreasing,.or extinguishing them. 2. The parameters examined were fanning, nipping, and duration of egg care, 3. The effects of different egg sizes upon these parameters were recorded under two experimental and three control conditions. The former involved presenting a female with artificial spawns as similar to the natural as possible except-in egg.size in exchange for her own. spawn. Three groups of egg sizes were presented: simi- lar, larger, and smaller than natural. Here females had no opportunity for making a choice. 4. It was found that "eggs" larger or smaller than natural, when substituted for the natural spawn, Specif- ically caused an appreciable decrease in several parameters of fanning--duration and beats. Removal of the spawn resulted in-the most profound decrement in fanning. No difference was noted between artificial "eggs” similar to the natural and the natural in terms of the fanning param- ,eters measured. This was as expected. This supports the 100 101 hypothesis that egg size influences egg care behavior in g. nigrofasciatum. 5. When the natural eggs and simulated "eggs" of the same size were presented simultaneously (choice situation), the females were unable to discriminate between the two until after approximately one day of the fanning cycle. 6. Certain details of the structure of fanning were also determined. It was found that the relationship be- tween fanning beats and bouts was inverse early in the fanning cycle and became direct later on.' 7. The presence of artificial "eggs" (which do not hatch) was expected to prolong the interspawn interval. The results appeared to support this, but were not statis- ticially significant. 8. Spawns were usually deposited under the lean-to and were randomly placed with regard to the three avail- able possibilities. The presence of artificial "eggs" on a particular section of the lean-to increased the probability that the succeeding spawn would occur in the same place. 9. These results are related to factors influencing organization, and functions of egg care behavior in g. nigrofasciatum. REFERENCES Aronson, L. R. (1948). Problems in the behavior and physiology of a Species of African mouth-breeding fish., Trans. N.Y.~Acad. Sci., 2(2):33-42. . (1949).. An analysis of reproductive behavior in the mouth breeding Cichlid fish Tilapia macro- cephalus (Bleeker), Zoologica, 34:1 - . Aronson, L. E. and A. Marie Holz-Tucker (1949). Reproduc— tive behavior in the African mouth-breeding fish, Tila ia macrocephala (Bleeker), Anat. Rec., p. 551, ABst. #158: Baerends, G. P. and J. M. Baerends Von Roon (1950). An In- troduction to the Ethology of Cichlid fishes, Behavior Suppliment, 1:1-243. Barlow, G. W. (1964). Ethol. of the Asian Teleost Badis. badis V. Dynamics of fanning and other parentaI activities, with comments on the behavior of the larvae and prolarvae. g. f. Tierpsychol., 21:99-123. Blum, V. and K. Fiedler (1964). Der Einfluss von Prolactin asuf das Brutpflegverhalten von S h sodon aegui- fasciata axelrodi, L. P. Schultz C1c 1 ae, Teleostei), Naturw" 51:149—150 . (1965). Hormonal control of reproductive be- havior in some Cichlid fish. J. Gen. & Compar. Endocrinol., 5:186-196. Breder, C. M. (1934). An experimental study on the repro- ductive habits and life history of the cichlid fish, Aequidens latifrons (Steindacher) Zoologica, N.Y., : - . Breder, C. M. and D. E. Rosen (1966), Modes of Repr. in Fishes, Nat. Hist Press, Garden City, N.Y. Collins, H. L. (1965). Nuturing experiments with regard to "adult impriting" and recognition of young in the Cichlid species Tilapia sparrmani and Aequidens 1atifrons., Ph.D. Thesis, Mich. State Univ. Libr. 102 103 Dambach, M. (1963). Vergleichende Untersuchunger uber das- Schwarmverhalten von Tila ia--Jungfischen (Cichlidae, Teleostei), g. f. Tierpsychol. 20 (2):267-296. Fiedler, K. (1962). Die Wirkung von Prolactin auf das Verhalten des Lippfisches Crenilabrus ocellatus (Forskal), 2001. JE. PhysioI., 69:609-620. Greenberg, B. (1961). Parental behavior and Cichlid fishes, Am. Zoologist, 1:450. . (1963). Parental behavior and recognition of young in Cichlasoma biocellatum, Brit. J. Anim. Behav., 11 (4):578-582. . (1963). Parental behavior and imprinting in Cichlid fishes, Behav., 21(1-2):127-l44. , J. J. Zijlstra, and G. P. Baerends (1965).» A quantitative description of the behavioral changes during the reproductive cycle of the Cichlid fish' Aquidens portalegrensis Hensel Konikl. Nederl. Akademie Von Weterschappen, Amsterdam., Repr. from Proceedings, Series C, 68(3):136-149. Iersel, J. J. A. van (1953). An analysis of the parental behavior of the male three-spined stickleback (Gasterosteus aculeatus L.), Behavioral Suppliment, 3:1-159. Kramer, S. (1960). Color changes correlated with parental behavior in Cichlid fish., Anatomical Rec., 138 (3):362-363. Kfihme, W. (1963). Chemisch ausgeloste Brutpflege—-und Schwarmreaktionen bei Hemichromis bimaculatus (Pisces), E. f. Tierpsychol., 20 (65:688-704. . (1964). Eine chemisch ausgeloste Brutpflegreak- tion bei Cichliden (Pisces)., Naturwiss., 51(1):20. . (1964). Eine chemisch ausgelaste Schwarmreaktion bei junger Cichliden (Pisces)., Naturwiss., 51(5):120. Mertz, J. C. (1967). The organization and regulation of the parental behavior of Cichlasoma nigrofasciatum (Gfinther) Ph.D. Diss., Univ. Ill. and G. W. Barlow.(l966). On the reproductive' behavior of Jordanella floridae (Pisces: Cypro- dontidae) withispecial reference to a quantitative analysis of parental fanning, g. f. Tierpsych01., 23(5):537-554. 104 Morris, D. (1955). The reproduction of the river bullhead (Cottus gobi L.) with Special reference-to the fanning actiVity, Behavior, 7:1-32. . (1958). The reproductive behavior of the ten— spined stickleback (Pygosteus pungitus L.), Behav. suEElo’ 661-1540 Myrberg, A. A. (1961). An analysis of the preferential care of eggs and young by adult brooding Cichlid fishes, Ph.D. Thesis, Libr. of U. Cal. at Los Angeles. . (1964). An analysis of the preferential care of eggs and young by adult Cichlid fishes., g. f. Tierpsychol., 21(1):53-98. . (1966). Parental recognition of young in Cichlid fishes., Animal Behav., 14(4):565-57l. Neil, E. H. (1964). An.analysis-of color changes and social behavior of Tilapia mossambica, U. Calif. Publ. Zoolg., 75:1-58. Noble, G. K. and K. F. Kumpf (1936). The sexual behavior and secondary sexual characters of gonadectomized fish,Anat. Rec., 67, Suppl. 113. and B. Curtis (1939). The social behavior of the jewel fish, Hemichromis bimaculatus, Bull. Am. Mus. Nat. His., 76:46. Peters, H. M. (1941). Fortpflanzungsbiologische u. Tiersoziologische Studien an Fischen I. Hemichromis= bimaculatus Gill. a. f. Mogphol. g. OekoI: Tiere. 37:387-426. Scheffe: H. (1959). The Analysis of Variance, Wiley and Sons, N.Y. Sevenster, P. (1961). A causal analysis of a displacement activity (fanning in Gasterosteus aculeatus L.) Behav. Suppl., 9:1-170. Smith, R. J. F. and W. S. Hoar (1967). The effects of prolactin.and testosterone on the parental havior of the male stickleback, Gasterosteus aculeatus, Anim. Behav., 15:342-352. 105 Wai, E. H. and W. S. Hoar (1963). The secondary sexual characteristics and reproductive behavior of gonadectomized sticklebacks treated with methyl testosterone., Can. J. Zool., 41:611-628. Wickler, W. (1962). "Egg—dummies" as natural releasers in mouth-breeding Cichlids, Nature, 194:1092- 1093. APPENDICES .LUI 3.225 3.025 A .ssv oonc HosOHuoom 2 2 2 2 2 2 2 . 2 2 2 2 2 2 8 .6 4. 2 O. 8 .6. .4 2 Au R. .b 4 o o o o o o .o o a o o o 0 tits. 4 .ee. msuoHsooEHn .m HsoEnnomw .9 moomunuon em . . hummpH .m msHHHoo sonm soxouv usuoHs :ooEHn.mHEonsomfiom oso .msonMHuoH msocusmos .Hsosnnomo .9 no ammo non hex: x cameos x capes. moon» :oHsoHoo oons mo sonusnnnuouc osauu.oH xncsomms 0.225 oH ,mH 0N mm om mm Number of Eggs ' v 7 - I . l l v 9 1 -‘ K o _' f-n... .’ 4' 4 N - - .‘. 'l .,*o | . ' . v .o. ‘r‘. L .- "— s-«u .99 ' l' 1 . . ‘ '~.. ‘ ._ _“- o. _ z '7‘ 7-. r _ “'4- gnaw-r" - -.-~ - \. “- I- I I . ’V' . a-..° _ l' ‘4 ' 4 . ‘ 7‘ a . — . a u _ ~ { I - . , q . . n . ‘- ~. ‘ . C ’. a '. - - . t . - ' . — . .tl' . 3.225 1U8 .EEV oons HosOHuoom 3.025 2.825 2.625 HdMDBdZ U Hammoamzv EsuoHomomoans . .ouHo sH mmmo o Honsuos sosu noHHoEm cso .sosu nomnoH .oEon m.mmmo= couoHsEHm csow.ssuonooomoans am no mmmo o Honsuos now Av>ux sockv.ms0HuoH usOHoo oons mo soHuannuch ose||.nH xHosommm HdfiDBflz A 0.625 0.425 0.225 0H mH om mm om mm Number Of Eggs Appendix lb.-(continued). 35 30 25 .LUS 12.925- 13.075 12.125- 12.275 < 110525- ‘ 11.675 10.875 ‘ ‘1 10.725- { 9.925- 4 10.075.~ (V AA mm. ea ( 9.125- . 9.275 ‘ 8.325—2 8.475,_I < NATURAL A tiona 7 675 A A Sec A as \l N U1 I ‘ 6.875 ‘ 5.925- 6.075 AVA 5.275 ‘ 4.325- 4.475 3.525- 3.675 10 O I!) N H Number of Eggs 110 1.010 “i“ n‘ un- . 0.970 Asumsoq\sucH3V xocsH omosm 0 0 0 0 0 0 0 0 0 0 0 3 9 5 l 7 3 9 S 1 7 3 9 8 8 8 7 7 6 6 6 5 5 O O O O O C O O O O O 0 o o 0 0 0 o 0 0 0 .0 ‘ . .eumesow..n .8. .1.) ‘u 1' '_. ‘3 .‘ '4. - I VIE- L ' ’ , ‘I. Hsosnnomo .9 ‘ moonmnuun .4 ”HmGaH ..m .mennnoo sonm common mononpwosuow.m new .meonmnuon .4 eHsoEnnomw .9 mo mono non msoHuoHsoHoo ls\zv xooen oaoem no eonuoonnumno oeeun.o~ anoooaoa 0.490 0.450 OH om om ov om Number of Eggs F.-.-f . .x . . .43 EOHUSQfianmulflmV on.~.l\onHN vaginucmuanuzcx [gfi‘Jl = 9133.0 111 1.010 0.970 isomeos\euonzv xooen mango 0 0.930 0.890 0.850 0.81 0.770 0.730 0.690 0.650 0.610 0.570 c 0.530 EsuoHomomon Hs am . 4 l o 4 HflMD9HuoHsEso ....... msonm =Honusoo= ouon xsoHn msonm emmoe Honsuos v msonm emmoe Honsuos msonm =00o= Honsuos A u 0.04‘0 osHHomon "tsomoq .omsosoxo szomm ou noHnm,soHuo>nomno osHHomon oussHB sou osu mouooHGsH sonon oonmHum «omsosoxo s3omm mo oEHu mouooHcsH 30nn¢ .sooE osu mo ovum nosuHo so nonno cnotsoum oso ono mosHH HousoNHnom .mmsonm HHo nom wow non moEHu oonsu soxou oOHnom soHuo>nomno oussHE sou non uson nom muoon msHssom mo nonsss sooE os9ul.v xHosomm< A l 115 HQLVH manna 3A|LV1flflflO A 7.59: wig. see.» .3. .T. P P I P D EEG Efififie T a» é .3 U * U h I 6 sun uvaw é'é am“ I. I 8 'NIW N31 " JMN-“W " .v - . ., ~ 2 ..\-~ .. runthmwmm\\&- ‘ ~’ ‘ ‘ " ' ' “cg-.539}; -, .-~N"H- > .. 00w .0 . . .4. -/.'.'.'. , .- r Ix}, -;-...._-'~:';;;£'#z uggDXw 23(5 ,. 116 souos usoonom o>HuoHsEso ....... msonm =Honusoo= ouon xsoHn msonm emmoe Honsuos v msonm emmoe Honsuos n msonm emmoe Honsuos A 00044:: osHHomon “wsoaoq .omsosoxo ssomm ou nOHnm sOHuo>nomno osHHomon oussHE sou osu mouoOHosH sOHoon conmHum «omsosoxo ssomm mo oEHu mouooHosH 3onn4 .soofi osu mo ome nosuHo so nonno cnoosoum oso ono mosHH HousoNHnom .mmsonm HHo now wow nom ooEHu oonsu soxou GOHnom soHuo>nomno oussHE sou nom uson nom msHssom mo soHuonso some os9ll.m vasommd 117 3:232... s 8 2 s 8 1r h .m— E 3. on 5 E 5 =— ' ' P a 32.2.. a a & . new" lNiOUid SAILVW'IWOS 3 8 3 a A j A_L _r 1 gl 4 A - h A I 1 A 3 1 A a l C .:=. 79. ugSOXw 23‘: mm 7431 mi mos ml uouvuno uvaw 118 o o ooo.oo ooo.ooo ooooo man H ooo.ooo ooo.oooo oooooo mm H ooo.m~ ooo.em9 momn mm N amo.mm emn.eo Hans om 4 oom.mo 4mm.oan 4mom~ mmm o mme.mom mom.9ne ammeen 94m 4 94~.Ho oum.o- osmv oem mH Ho~.omH mem.mmm mmemnn 9mm m 9m9.- oon.~m omao men mn mnm.oHH 449.9om ooeem mom a 4mm.m4 mo9.4mn Hmmmm mew mH mam.eo ~mo.ooHH oeeem now an emo.~m o99.mmH mHoom was mH Hmm.em oo9.emn HHHmm mom mn -9mm.om amm.mom HAHNe ems mu Hom.mo man.aem monno «em mu mmm.no mmm.n- momme one mn moe.Ha mmm.omm moemon omm mn omo.mo mam.om~ osmmm oma mH «no.9on ema.smm mmmomn mam MH mmo.99 mo9.99~ mmnee mam mn om~.99 mmm.mam mamas moon mu. anm.es Nun.oem comma mmon mn mm xm Mm m 2 no no mm m z =nonueoo= oumnm noonm osuHomom moussHE 0H nom.muoon 0sHssom .mmsonm o>Hm osu mo sooo now so>90 ono mnonno cnocsoum cso .sOHuoH>o© unocsoum .oosoHno> .sooE .oNHm onEom .9o0 nom moEHu oonsu.soxou sonuo>nomno oussHE sou o no uHsmon osu musomonmon ouov mo uom soom .ouoon 0sHssom now moHumHuoum OHmom|I.0 xHUsomm¢ 119 ooo.ooo ooo.oooo oooooo me n ooo.ooo ooo.oooo oooooo eo n .ooo.ooo ooo.oooo oooooo omn n mmo.mon mmm.mono~ mooeo omm m moo.om mmm.omom enmon mmm m moo.mm mmn.oemo mmmom mmn e mnm.mm oom.ommo memmm son e o noo.mon.omm.mmm monnon mmn o ooo.m ooo.nm mon om m mom.nm omo.mom mene een m mmo.on oem.mmm mmon no o o mme.mo mee.mom memmoe «on m mom.on noe.onn one em m ooo.oo ooo.oooo ooooo om n omn.mo mmn.omn omonm eon nn omm.o ooo.mm mom mm m oom.om oom.mmem oeoo mm m oom.me mom.oon oeomm mon mn nem.m ooe.no mme no m ooo.m ooo.m on om m nem.mo omm.eon mmmmmn.oo~ mn moo.nn nom.emn memn mm on mmo.em mmm.eoo enon em m nem.mo omn.omm nmomo mom mn mmo.en mmm.enm mmmm mo nn nmm.m mmm.em mon me m mme.oo omm.oom mmmoe onm mn mmn.om mmm.moe moee mmn nn moo.on mne.mnn ooe om e omo.oo mmm.mmm eneem mmm mn mmo.mo nmm.emem mone onm mn mmn.om omm.onm meom oon e moo.om omm.oom mmmmm ene mn mme.mm moo.noom oommm 9mm mn omm.oe ooo.mmon omoo onn e mom.mm men.oom mmooo nme mn mon.me omm.nnem momnm mem mn eme.oe oom.oonm eooon mnn m moo.nm eeo.omn enmnm mme mn moe.mo mmo.momm mmmoo mmm mn nom.nm oeo.omm momm mm on meo.me mom.onm momoe mom mn omm.om mmm.omme mmoeo mom mn omm.om mnm.oooo moooo mom on mnn.oo mmm.mmm emmmo noo mn. onm.mo ome.mmoe memmm ome mn oeo.no mem.ommm oomoe mon nn ~ne.eo nom.me~ omnmo nmm mn mnm.om moo.mmme omoeo noo mn omn.mo mom.mmmm mommm omn nn mom.eo mmm.mmn .nommm mnm mn. oem.om nme.nmmo moonm moo mn nne.mo moo.mooe onmmm mem mn onm.mo ooo.onm moooe moo mn ooe.mo nmo.nmmm momom mmo mn mon.om emo.omme mmoeo mom mn moo.mo oem.en~ oonoe omm mn moe.mm eme.ommm emmoe mom mn nom.mo omm.mooe mommo emm mn mmm.mon omm.mmo mmoomm mmm mn oeo.mo moo.oomm oommm mom mn omo.em mme.omoo onemm ooo mn mmo.om omm.omm omnom ooo mn ooo.mo nmn.emmm onmme mom mn moo.om nmm.mmom mmmmm mmm mn mom.oo oom.mom omoom eoon mn nmm.me mmm.mmmn nemmm moon mn mmm.no mom.onmm neome oonn mn onm.~m omn.mom ommoo omen mn so no mm m z no no Wm M 2 so no Wm M z Honsuoz sosu noHHoEm Honsuoz sosu.no0non Honsuoz mo oNHm,oEom .UossHusOUI|.0 vasommm 120 000.000 o o ooo.o ooo.oo ooo.mn n ooo.o coo.oo ooo.ooo ooo.m n oom.m oom.mn oom.em oom.mn m ono.m nmo.e oom.on oom.e e moo.m mmm.o omm.mm mon.on o oom.o mmm.nn mmm.nmn mmm.nn e mmm.n mmm.n noe.mo mom.mn mn omm.e ome.m oom.mo ooe.en o mmm.n ooo.e emo.mm mno.mm mn oom.m eom.m mme.mo moo.en m mem.n ono.m mom.me oom.om mn omo.n eoo.o eoo.mm moo.en mn ome.n nmm.o onm.om eon.om mn ono.n moo.m mne.mn mmm.mn mn nmm.n oom.o emo.me ono.mm mn mmm.n omm.e omm.on emn.on mn mmm.n emm.o oom.mm mno.n~ mn mmm.n omo.o mmm.om nmm.on mn mmm.n mmm.e eoo.om mmo.om mn moo.m mnm.m eon.mo moe.mn mn oon.m mmm.nn mme.mmn ono.mm mn ome.n mmm.m ooo.mm mom.mn mn eom.m mmo.m meo.mm eon.mn mn so so Mo 9 2 so no Wm M z .nonusoo. muono nonno oennomom moussHE on nom\musom .soHuoH>o0 wnowsoum .oosoHno> .sooE .oNHm onEom .mmsonm o>Hm osu mo sooo now so>H0 ono mnonno cnocsoum wso .9o0 nom moEHu oonsu ooonooon soHuo>nomno oussHE sou o no uHsoon osu musomonmon ouoc mo uom .0onnom sOHuo>nomno oussHE sou nom muson now monumHuouo OHoomI|.9 xnwsomm¢ room 121 omo.m ooo.oo oooo.ooo ooo.o n omo.m ooo.oo oooo.ooo omo.m n omo.m ooo.oo oooo.ooo ooo.on n moo.o nnn.om ommm.oon mmm.~n m mmn.m nom.m ommo.mm moo.on m nmo.m mmn.m ooom.om omo.mn e moo.m mmm.mn ommo.mm mmm.nn e nnn.m mmm.o ooom.me emm.m m o omn.n emm.m enmo.nn oom.o m omo.m oom.o omo.mn oom.o m o mem.m mme.e mnom.me nnn.m m moo.n oem.m emm.mn emm.m o omo.m ooo.oo ooo.oo omo.mn n nom.n mmm.m mooo.on omo.m nn mmm.n emn.n omm.o nnn.m m oom.e oom.om oom.oe mmm.nn m mmm.o ooe.m mnmm.nn mon.on mn omo.m mmo.o ome.e emm.m m oom.e ooo.n omo.m mmm.nn m one.n mom.e onoo.em mmm.nn mn mmm.n mon.n emo.on mmm.on m moe.n nnn.m mmm.o moo.nn m mon.n mnm.e meem.mn emm.mn mn oem.n nom.n nno.on emm.m on mmm.o nnn.m mmm.o moo.m m mom.o emm.m nomo.on oom.mn mn mmm.o mmm.o mnm.on mmm.m nn moo.n mmm.m mnm.on emm.m e nnm.n mme.e noom.mm mom.on mn mne.n moo.m mmo.mm meo.mn nn mmm.n omo.n moo.» oom.on e omo.m mmm.n mmom.mm oem.mn mn mon.m mme.e mmm.mo omo.mn mn neo.m mon.e moo.on omo.m e ome.n omn.m emoo.om eon.mn mn mno.n mmm.n mmm.mm.ooo.on mn oom.n oem.m emm.mn omo.m o emm.m non.m oomm.m ono.mn mn noe.n emn.m omo.mm nmm.mn mn omn.n mom.n omo.mn oom.o on emm.m mme.m momm.nn mmm.nn mn mmm.n omo.n mmn.nm mom.mn mn emm.m mnm.e mmn.me mmm.nn on mmm.n emm.m meem.mn mmm.nn mn mmm.n mnm.n omm.em omo.mn mn emm.m mmn.o mno.mm mmm.nn nn mno.n ono.o ommo.mm emn.mn mn mmm.n mme.m nmm.mm emm.mn mn mnn.m moe.e nmo.me mom.mn nn mem.o noe.m nmoo.nn oom.om mn moo.n mmm.m emm.mm mmm.on mn on~.m nnm.e mmm.mo emm.mn mn mnn.n moe.e eeoo.on mmo.mn mn mmn.n nom.n mmm.on mmo.mn mn oom.n emm.m mom.om mom.mn mn mmm.n omo.m omom.mm oom.mn mn mnm.n ooo.m emm.me mmm.on mn one.n non.m omo.mm omm.on mn noo.n one.o oomm.mm moe.en mn oom.n mmm.n neo.om oem.on mn emo.n moo.m mmo.em mom.mn mn mom.n mom.o nnne.oe mme.en mn omo.m nom.e eoo.oo oom.om mn mnm.n emo.m oom.me oem.nm mn oom.~ emm.m ommo.mm mom.mn mn oom.n moo.m mme.ee mme.en mn oom.n mmm.n emm.mm mom.mn mn mme.m oom.m mmmn.mm mom.mn mn mo xo Wm M 2 mo .xo. mo m 2 no no mm m. z HMHSHMZ CNS“ HOH ngm Honsuoz sosu nomnon Honsuoz mo oNHm ofiom .oossHusoull.9 xHUsomm< 122 o o omo.m ooo.oo mom.eoo eee.o n omo.m ooo.oo mom.eoo omo.m n mon.o mme.e omo.m omo.m m oeo.o mmm.n mon.n moe.e e mmm.n mme.m enm.nn emo.o o mmm.~ omo.m omm.em men.m e mmo.m omn.m mon.mm nom.nn mn onm.m mon.o oom.om oom.m o mme.n men.o mme.om oem.on mn mmm.n nmo.e mon.on mmm.o m mmo.m emm.mn eom.omn mom.nm mn mme.e mem.m meo.m oon.m an. omo.m mmm.n one.no moo.mn mn emm.m mom.m omm.mn moo.m mn meo.n neo.o mon.ee ooe.on mn mmm.n omo.e omm.om omo.m mn omn.e omo.mn moo.omm mom.on mn omm.n oom.o oe~.oe mmo.mn mn eoe.m moo.o onn.mm emm.mn mn moo.m eem.mn mom.eon emm.on mn moe.e oeo.en nme.enm mmm.nm mn nno.o nmm.om... om~.moe mon.nm mn mem.o omo.mm omo.mmm eoo.nm mn no no Mo _m 2 no no Wm M z eHonusoue ouon xson osHHomom moussHE 0H nom\uson nom soHuonsQ .mmsonm o>Hm osu mo sooo now so>Hm ono nonno Unowsoum Uso .soHuoH>o0 unocsouo .oosoHno> .sooE .oNHm oHQEom .9oo nom moEHu oonsu soxou sOHuo>nomno oussHE sou o no uHsmon.osu musooonmon ouoo mo uoo soom .soHuonsU msHssom now moHumHuouo UHmomI|.m thsommd 123 coo.oo ooo.oo mom.eoo mmm.e n ooo.oo ooo.oo mom.eoo omo.m n ooo.oo ooo.oo mom.eoo emm.m n mmm.n nom.m moo.on 9oo.o m mon.o mno.m m eoo.m mno.o e mmm.m nme.m e eom.n onm.m omo.m ooo.on m o omo.m omo.m mee.em moo.o m moo.o ome.o mmm.e mmm.m m o ooo.n omo.m eem.om oom.m m mmm.e neo.o emm.m mmm.m o omo.m mom.eoo mom.eooo omo.m n mno.n omo.m oem.o~ mmm.m nn mmm.m mmm.e mem.m eem.m m nom.n mee.n omo.m .mme.m m eoe.n neo.o .enm.mm mmm.m mn emm.m omo.m omo.m oom.m m nme.m .me~.o mme.e .moe.e m mno.n nme.m emm.mn nme.m mn H00.o 00H.0 000.H 000.0 0 H0v.0 Hv0.o 009.0 000.0 0. 0H0.H 0v9.¢ 00v.00 009.0 0H moe.e mnm.n mmn.m, omo.m on eem.o mme.e mmn.o. mmm.m m one.n mmm.m mmm.mm meo.nn mn mnm.o mmm.e omo.m eem.o nn omo.m nmm.o omo.m mmm.e e mno.n nme.m enm.mm mem.nn mn nmm.o omo.m eom.o . mmm.m nn mmm.m oem.o mmm.n emm.m e omm.m moo.m ,eno.oo eoo.nn mn ome.n omm.m mmo.om mnm.o mn mmm.n omo.n mem.o ‘mnm.o e mmn.n mmm.e mmm.mn mmm.nn mn mnm.n mmm.n mmm.mn emm.m mn mem.n omo.n mmm.m nme.m o ene.n mmm.m mmm.mm nnmxmn mn mme.n omn.m onm.om mon.m mn nmm.o moe.e mmo.m. omm.o on mmm.n omo.o omm.oe moo.on mn mon.~ noo.e moo.oo mmo.on mn mmm.n mnm.n mon.mn ooo.o on mmm.n mmm.m mnm.nm mmn.on mn eme.m mon.o momnoo nmm.mn mn oeo.o mnm.o mom.m nme.o nn non.e moe.mn mmn.omm moe.nm mn nme.m mmn.mn moe.mon omn.mn mn mmm.e omo.m omm.o mem.o nn omm.n ooo.e non.nm mom.on mn omo.e omo.nm mmm.nom nme.mm mn mnm.n one.n nmm.mn nme.m mn mmm.n ooe.o mmm.mne moe.mn mn emm.n mmm.m moo.oe moowmn mn noe.n omn.m mmm.nn mmm.m mn oon.m onm.m eom.ono emm.mn mn mnm.o eno.om mmn.ooe mem.mm mn mmm.m men.nn nmm.een moe.mn mn emm.mn moe.mm mmm.mmnnmom.om mn oon.m moe.e emm.mm mon.on mn mmm.m mom.en mmm.mmn moe.on mn oeo.o mmm.nn neo.ome meo.om mn mmo.e mmm.nm-mom.mom mmm.mm mn moe.m emm.m omn.mm ome.mn mn mme.e omm.on oon.mom moe.om mn onm.o mmm.mm emm.mme mmm.em mn oom.on mmm.emm omm.oeom mmm.ne mn emn.o nom.mm mom.ome emm.mm mn mo Uno MM m 2 so no- Wm. M 2 no uno .Mo m z Honsuoz sosu noHHoEm Honsuoz sosu.nomnon Honsuoz no oNHm oEom .oossHusOU||.0 xmcsommo 124 o o omo.m omo.m omo.m mom.n n omo.m omo.m moe.e eee.m n mon.o mme.e moe.e oeo.m m onn.o mmm.e emm.m nmm.m e mon.o mmm.e mmn.o emn.m o mmm.e mme.e mmm.e mmm.m e mon.o mmm.e mon.o emn.m mn omn.o mme.e mmn.o mmm.m m mnn.o moe.e oon.o nom.m mn enn.o nom.o nme.m ooe.m m non.e mom.o mmn.o mem.m mn nme.m enm.o omo.m emn.m mn omo.m mmm.e eoo.o nmm.m mn moe.e enm.o omo.m oon.m mn moe.e eem.o moe.e mmm.m mn moe.e onm.o oeo.o oon.m mn mme.e onm.e eeo.o mmn.m mn meo.o omn.o nme.m emm.m mn omo.m mmn.o mme.e mmm.m mn omo.m mmm.e mme.e emm.m mn nme.m mon.o emm.m mon.m mn moe.e mnm.o oeo.o mon.m mn emm.m omn.o omo.m emm.m mn no no Wm M 2 so no mm m z .nonueoo. muono nonno mmmmmmmm moussHE 0H nom.omEo9 .mmsonm o>Hm osu mo sooo now so>H0 ono mnonno unocsoum cso .soHuoH>oU unocsoum .oosoHno> .sooE .oNHo onEom .0o0 nom,moEHu oonsu wounooon soHuo>nomno oussHE sou o 00 uHsoon osu musomonmon ouov mo uom soom .coHnom soHuo>nomno oussHE sou nom omEou now mOHumHuoum onoomul.m xnosommm 125 omo.m omo.m omo.m oem.n n omo.m omo.m omo.m oom.n n omo.m omo.m omo.m eeo.n n omn.o mme.e mme.e mmm.n m mmn.o mno.o omo.m emo.n m onn.o eno.o moe.e oom.n e eon.o emm.m omo.m mno.m e omn.o mno.o omo.m mmm.n o o omn.o oem.o mon.o mmm.n m mmn.o mno.o omo.m mmm.n m o omn.o moe.e eon.o oom.n m mnn.o eno.o moe.e oom.n o omo.m omo.m omo.m nnn.m n omo.m mmm.e ooo.o mmm.n nn mnn.o eno.o mme.e nnm.n m meo.o moe.e emm.m mnm.n m omo.m nom.o omo.m mmm.n mn omo.m moe.e moe.e nnm.n m oom.n ono.o omo.m emm.n m omo.m mmm.e mon.o oem.n mn onn.o mno.o mon.o nmm.n m mmn.o nme.m emm.m eoo.n m onn.o ene.o omn.o omo.m mn men.o omo.m emm.m mmm.n on nmn.o mno.o nme.m omo.m m moe.e mmm.e meo.o meo.m mn mon.o nno.o onn.o nom.n nn mmm.e neo.o oon.o nmm.n e mme.e enm.o mon.o nmm.n mn moe.e moe.e omo.m mmm.n nn nnn.m mno.o nme.m nnm.n e nme.m mmm.e omo.m men.m mn mme.e omo.m omo.m omo.m mn moe.e emm.m ono.o oom.n e omo.m emm.m mmm.e mon.m mn omo.m omo.m mme.e mno.m mn omo.m omo.m mme.e nnm.n o omo.m mmm.m omo.m oen.m mn mme.e moe.e omo.m oeo.m mn mme.e ono.o mmm.e mmm.n on mmm.e mmm.e omo.m,mom.~ mn mme.e moe.e nme.m omo.m mn non.e ono.o eon.o emm.n on moe.e omm.o omo.m mmm.m mn moe.e emm.m omo.m emm.m mn ooo.o emm.m neo.o mmm.n nn meo.o mon.o emm.m emm.m mn moe.e moe.e meo.o oon.m mn emm.m omo.m noo.o mnm.n nn omo.m emm.m moe.e mmm.m mn moe.e emm.m moe.e mmm.m mn mmm.m moe.e moe.e mno.m mn nme.m mmn.o mme.e mem.m mn neo.o moe.e mme.e mnm.n mn nnn.m mno.o non.e mmn.m mn oeo.o mmn.o omo.m mmm.m mn meo.o moe.e emm.m omn.m mn moe.e emm.m meo.o mmm.m mn meo.o mon.o omo.m nom.m mn omo.m moe.e oeo.o mon.m mn mme.e moe.e omo.m mmm.m mn moe.e nmn.o omo.m mmm.m mn omo.m moe.e neo.o mon.m mn eoo.o emm.m emm.m mme.~ mn moe.e mmm.e nme.m mmm.m mn mme.e neo.o mno.o mon.m mn moe.e emm.m meo.o omo.m mn mme.e emm.m meo.o emm.m mn mo xo Wm M 2 no no mm m 2 no no mo m z Honsuoz sosu noHHoEm Honsuoz.sosu nomnon HMHSHMZ mm QNHm gmm .UossHusOUIu.0 xHUsommm 126 o ooo.oo ooo.oo mom.eooo mmm.en n ooo.oo .ooo.oo mom.eooo mmm.m n ooo.e noe.on mmm.mm eeo.m m mmm.n mem.m moe.en mnm.m e mmm.m mmn.om omn.moe mom.en o mmn.m mmm.on omo.mom ooo.om e nmm.e mmm.mn omm.mem omm.mn mn mon.o omm.mn moo.mom emm.mm ‘o omm.m oeo.o meo.em omo.em mn emo.m oem.m mmm.em omo.on m emm.o nmm.em oeo.mmo mme.ee .mn mmm.m mmm.m mmm.mo oem.on an eoo.o ome.om mmo.o~e emm.me mn eee.m mnm.m nmo.mm omm.mn mn moo.o mme.em omo.mne eoo.mm mn oom.m mmm.nn mmm.mmn enm.nm mn mmm.e nem.on emm.mmm emm.mm mn mmm.e meo.on moo.mmm mnm.mm mn omo.m emm.mn mon.omm moo.mm mn mmn.o mmm.mm moe.mmo omm.mm mn omn.m mmm.mm ooo.omon eoo.oe mn omm.nn mnm.me mne.ommn nom.mo mn ooo.en neo.mo moo.nmm~ moo.mo mn so no Wm M z mo xo Mo m z .nonusoo= muono neonm monnooom moussHE OH nom.uson nom|ouoom .mmsonm o>Hm osu mo sooo now so>Hm ono ononno cnousoum Uso .sOHupooU onovsoum .oosoHno> .sooE .oNHm onEom .mow nom moEHu oonsu soxou sonu Io>nomno oussHE sou o no uHsmon osu musomonmon ouoo mo uom soom .Uonnom sOHuo>nomno oussHE sou nom uson nom ouoon msnssom now mOHumHuoum unmomlhoH wasomm< 127 ooo.oo ooo.oo oooo.oooo omo.m n ooo.oo ooo.oo oooo.oooo omo.m n ooo.oo ooo.oo oooo.oooo moe.mn n mmm.e omo.em omno.mm eme.mn m mmm.m omm.mn ommo.mm oom.on m mnm.m mmm.m emoe.me mno.nn e men.m emm.o mmme.mm moe.mn e ome.m eoo.m mmem.mo mmm.on m o omo.e nm~.mn nmmm.omn mmm.on m mmm.o oom.o nmo.n omo.m m o mmm.e omm.mn mono.oon ome.on m mme.n onn.m omm.mn mon.m o omo.e ooo.oo one.ooo mmm.o n mmm.m eoo.on oono.onn mmm.on nn mmo.n oon.n eon.o emn.m m ooe.m mmm.o mnm.mn mmm.o m emm.m ono.m eomm.om mem.mn mn mmm.o mmm.o nme.m moo.m m moe.e mon.o mmm.o oom.m m onn.m mmm.on ooeo.onn one.on mn mmm.o moo.o moe.e emm.o m mmm.n oom.n emm.m mmm.o m noo.m mmm.mn moon.omn neo.on mn mem.o mmm.o mom.m nme.m on mmm.o eeo.o mmo.n neo.o m omo.m mmo.nn mmon.non emn.mm mn mmm.n mmm.n omm.mn mmm.m nn omo.e mme.e mmo.n mmm.m e omn.e oon.on emon.mmm mmm.om mn ome.n omm.m mmm.em mmm.on nn moo.n oee.m mom.m ome.m e mnm.m omm.mn mmoo.m~e omm.mn mn mmo.n emm.o moe.mm omm.mn mn mnm.m oeo.e mon.on mmm.on e mmm.m nmm.en mmon.mnn omm.em mn mmm.m omo.m mom.mmn omm.mn mn enm.m nom.e_ mom.em eoe.nn o nmo.m mmm.m mmon.em nmm.om mn emm.m mnm.mn mom.eon mnm.mn mn meo.m mom.mn ooo.o~n eno.en on mnm.m mon.en ommo.mmn mnm.mm mn mmm.o emm.em mmo.moe nom.om mn mmm.m eoe.on omm.mon omm.mn on mmm.e omm.mn omom.mem mnm.mm mn eom.e omo.em mmm.onm noe.mm mn mom.n moe.m mon.om moe.mn nn omn.m omo.mm ommm.emon nom.oe mn oem.m omm.mo noo.moo mnn.om mn mmo.n omo.m neo.m~ omm.mn nn omo.m omm.mn mooe.mnn noo.om mn omo.m mne.mo mom.emo mnm.mo mn mmm.m omm.o mmm.om nmn.on mn emm.m meo.mn oonn.omn oeo.om mn oon.e nom.mn ooo.om~ mmm.mm mn nmm.m mem.mn nem.oon mmm.on mn mmo.o mmn.em ommm.mmo mmm.me mn omo.mn mmm.mon mmm.omom emo.mo mn ooo.o one.em mmm.mee meo.om mn mmm.nm omm.mm momo.om~o omm.mo mn meo.e neo.om mem.oom omm.me mn omo.e neo.mm mmm.mmm mmm.mm mn mom.mn omn.me mmmm.nmmm omm.mo mn mne.on meo.mon men.nnen moo.oo mn eon.o omm.mm oem.mme ooo.me mn emo.nn mmm.me momm.mmon ome.mo mn omm.mn oeo.oen omo.mmon nom.om mn mme.o mmm.me mme.oem nmm.eo mn omm.mn mon.oo mone.meom eon.no mn no no No m z no no Mo m 2 so no Mo m z Honsuoz sosu noHHoEm Honsuoz sosu nomnon Honsuoz mo oNHm oEom .UossmusOUII.on chsomm< 128 . 0 0 000.00 000.000 00000 000.00 H 0000.00 000.000 00000 .000.0 H 000.0 000.00 00H 000.00 0 00oH.0H e0m.e0 000 000.00 e e0H.00 000.00 009e 000.0HH 0 0000.00 090.00H 00000 ,009.00H e 9e0.9m 090.00H o0m0H 000.00H m 0000.H0 00H.0HH 00e0H 000.0eH 0 090.0H HeH.9e 0000 00m.Hom m 0e00.9e. 009.00H 0900H meH.00H 9 0He.0H 00e.00 0Hee e0H.09m m 000H.o0 . 000.00 0000 009.0HH 0H 0eo.oH 900.00 0000 H00.e00 m 0009.H0 o9e.09 00H0 000.0HH _mn 000.00 0He.H0 0000 000.Hmm m 00e0.00 000.HOH H000H 000.00H 0H HHO.90 000.00H 90HOH oe0.90m 0 0000.00 000.00H 00HOH 000.0e0 0H 000.00 09e.HOH 909eH 0H0.Hmm m 00e0.e0 H00.0oH 0H000 0Ho.e00 mH 0He.0m 000.90H 0000H 990.000 0 e000.mm e90.00H mmoeH 990.00e mH 09e.Hm moe.mHH e900H 000.0ee 0 no no mo m 2 no no mo m z =Honusoo= ouonm nson osHHooom moussHE 0H nom.soHuonso .omsonm o>Hm osu mo sooo now so>H0 ono ononno cnovsouo vso..somuoH>o0 anocsouo .oosoHno>..sooS,.oNHm onEom .hov nom ooEHu oonsu soxou sonuo>nomno oussHE sou o 00 uHsmon osu ousooonmon ouov mo uom zoom .onnom sOHu Io>nomno oussHE sou nom usonnnom.s0HuonsU msHssom now oOHuoHuoum OHoomiuHH chsommd 129 000.00 000.000 00000 000.00 0 000.00 000.000 00000 000.00 0 000.00 000.000 00000 000.09 0 000.09 000.000 00000 900.000 0 000.00 000.000 0090 000.000 0 000.00 900.0000 9000 000.00 0 009.00 000.000 0000 000.00 0 000.00 000.000 00000 000.00 0 0 000.00 000.00 0000 000.09 0 000.0 000.0 0 000.00 0 0 900.00 000.00 0009 000.09 0 000.0 000.00 900 000.00 0 000.00 000.0000 00000 000.00 0 900.00 000.09 0000 000.00 00 000.0 000.00 000 000.90 9 000.00 000.090 0090 000.00 0 000.00 090.00 0000 000.00 00 000.0 090.0 00 090.00 9 000.0 000.0 0 000.00 0 000.00 000.00 0000 009.000 00 900.0 000.000 900 000.00 0 000.00 000.000 000 000.00 0 000.00 000.00 0000 009.000 00 000.0 900.00 000 000.00 00 000.0 099.0 00 000.00 0 000.00 000.90 0009 000.000 00 000.9 000.00 000 909.00 00 000.0 000.00 00 000.00 0 000.00 009.000 00000 000.000 00 000.00 090.090 0900 000.09 00 900.00 900.090 000 000.00 0 090.00 000.000 00000 000.000 00 000.00 900.000 0009 000.000 00 000.00 900.000 0000 000.00 0 000.00 900.000 00000 000.000 00 090.00 000.000 0990 000.000 00 000.00 000.000 9000 000.00 0 000.00 000.00 0009 000.000 00 000.00 000.900 0000 000.000 00 999.00 900.000 0000 000.90 00 000.00 000.000 00000 000.000 00 000.00 000.000 0000 000.090 00 900.00 000.0000 00000 000.00 00 900.00 000.00 0000 000.900 00 000.00 000.909 00000 000.000 00 000.90 000.099 0000 000.00 00 000.00 090.000 00000 000.000 00 000.00 000.0000 00000 000.000 00 000.00 000.000 9009 000.00 00 900.00 000.09 0090 000.000 00 000.00 000.0000 09000 000.000 00 900.00 000.0000 00000 000.000 00 000.00 090.00 0009 009.900 00 000.00 000.000 0000 000.000 00 000.00 000.000 00000 000.000 00 000.00 000.00 0009 000.900 00 900.00 000.000 00000 000.000 00 000.00 000.0000 00000 000.000 00 900.00 000.000 00000 000.000 00 900.00 000.090 0009 000.000 00 000.00 000.0000 09000 000.000 00 000.00 000.000 00000 000.990 00 090.00 090.000 0000 000.000 00 000.00 000.000 00000 000.090 00 000.90 090.90 0000 000.900 00 000.00 000.000 0000 000.000 00 000.00 000.000 0009 000.000 00 000.00 000.000 09000 000.090 00 00 x0 M0 m 2 mm x0 M0 m z 00 x0 M0 m z Hmnnumz away 0000080 amusumz can» 000000 Hmusumz mm @000 0800 .Omscflucooll.00 xflvcmmmd 130 0 0000.00 000.00 0000.0000 000.00 0 000.0 000.00 000.000 000.0 H 0000.0 000.00 0000.00 000.00 0 009.0 000.0 000.00 000.0 0 0000.00 000.90 0000.0000 000.00 0 000.0 000.0 000.00 000.00 0 0000.9 000.00 0000.009 000.00 00 000.0 900.00 000.000 000.00 0 0900.0 090.00 0000.000 900.00 00 009.0 009.00 000.000 900.0 0 0990.0 000.9 0000.00 000.9 00 909.0 000.00 000.000 090.00 9 0000.0 000.0 0000.00 099.0 0 000.0 990.0 000.09 900.9 0 0000.0 000.0 0900.00 000.0 00 000.0 009.0 000.00 000.0 00 0000.0 000.9 0090.00 000.00 00 HHo.H 000.0 000.0H 900.0 NH 0000.0 900.0 0000.00 090.HH 0 000.0 000.9 900.00 900.0 00 0000.0 000.0 0000.00 090.0 00 000.0 000.0 000.00 000.9 H0 000H.H 000.0 0000.0H 000.0 0 mm x0 0000 m. 2 mm x0 0000 m z =Houucoo= mMMHm xCMHm GGHHmmmm 0005808 00 0mm mm0z _ .mmsonm w>0m may no 5000 How m>00 000000 numvcmum can .:00000>mv vnmccmum .moa00um> .cmm8 .0000 mHQEMm .000 0mm 0080» mmunu cmxmu c0000>0mmno 009208 mm» m 00 u0dwmn mnn mucmmmummn mumo 00 0mm £00m .GOHHmm :00» um>nmmno wuscw8 awu 0mm 020mm0c mo mocmsvmum may now mUHumflumum owmmmlu.00 xflvcmmmd 131 ooo.o ooo.o ooo.oo ooo.0 0 oom.o om~.o 000.0 000.0 0 000.0 000.0 ooo.oo 000.0 0 000.0 000.0 000.00 000.0 0 000.0 909.0 emm.m 000.0 N mmo.0 900.0 om9.0. 000.0 0 000.0 000.0 emm.m 000.0 N a 000.0 000.0 om~.o oom.0 0 000.0 oqo.o 000.0 000.0 0 000.0 000.0 000.0 oom.0 0 090.0 000.0 900.0 om9.0 0 0 000.0 mmm.m ooo.9 900.0 m mmm.o 000.0 900.0 900.0 0 000.0 ooo.o 000.00 000.0 0 000.0 000.0 009.0. 000.0 0 000.0 000.0 900.0 900.0 0 ooo.o ooo.o ooo.oo ooo.0 0 000.0 000.0 oom.m 900.0 9 000.0 000.0 oom.o 000.0 0 000.0 000.0 000.00 000.0 0 000.0 oom.o mmm.o 900.0 9 000.o 0m0.o 000.0 000.0 9 000.0 000.0 000.00 000.0 0 000.0 000.0 000.0 000.0 m m00.o omo.e ~00.o 000.0 9 000.0 000.0 000.00 ooo.0 m w99.o 000.0 ooo.0 oom.n o0 ooo.o 000.0 ooo.c ooo.0 0 000.0 ooo.o 000.00 000.0 0 9mm.o 000.0 000.0 000.0 0 000.0 000.0 000.0 009.0 m ooo.o 000.0 000.00 000.0 0 000.0 mmm.m 009.00 000.0 m «90.0 09m.0 m0m.m 090.0 9 000.0 000.0 000.0 000.0 0 090.0 ooo.0 000.0 000.0 9 009.0 000.0 000.0 000.0 m oo0.o 000.0 oom.o oo0.0 0 900.0 omo.m 900.9 000.0 9 900.0 000.0 mm0.0 oo0.m 00 000.0 000.0 090.0 900.0 9 000.0 000.0 000.0 mmm.m 0 000.0 00m.o emm.m ooo.m 0 000.0 000.0 900.0 900.0 0 909.0 000.0 m00.0~ m9m.m m mmm.0 000.0 000.00 o09.~ 00 mvm.o 000.0 909.0 000.0 0 900.0 mmm.m 000.0 000.0 o0 900.0 000.0 000.00 ooo.0 m mmm.o 900.0 000.0 000.0 0 009.0 000.0 000.00 000.0 00 090.0 0mm.o 000.0 oom.~ 00 000.0 000.0 090.0 900.0 9 000.0 000.0 000.90 000.0 0 009.0 900.0 900.0 000.0 00 000.0 000.0 000.00 000.0 9 000.0 m03.m oom.m 000.0 00 009.0 o00.o 090.0 000.0 9 000.0 000.0 m00.00 9mm.m 9 990.0 000.0 000.00 000.0 00 000.0 mmo.m 000.00 mmm.0 a 000.0 9mm.0 900.00 oom.0 00 moa.o 000.0 000.0 000.0 00 000.0 000.0 000.00 000.0 00 000.0 000.0 000.00 000.0 00 009.0 990.0 909.00 000.0 00 000.0 mmm.o 090.0 mmm.m 9 009.0 000.0 ooo.mm 900.0 m mom.0 000.0 000.00 000.0 00 mm xm Wm M 2 mm xm Wm M 2 mm xm mm m z HMHDUMZ CMfi Hmflfimgm HMHSHMZ 0.0.0.5“ Hmmumg HMHd—hwmz mm mNHW mamm .00nz0u:00||.00 x0oc0mm¢ 132‘ 0500 000000000 00>00000 00 00000050 . O C O. 0 100. m Q 0 90 M0 O C .0 d‘ 400‘ d0 d0 .0 .m >3 ~©0HHOQ COHfiMNrHGmQO 00.050.00.00— OH 4 0000 000 8000.00000 .000080050008 0000 000 .000 .0 0 >0 000000>00000 000000800000 000 .0 00 00 000000000 000 000000 00000>000 :00 005008 00 00000 000 8000 00000 00008 10050008 00000 009 .00 000850 000800 000 .0500 0:0 00 000000 000 00 00000050 0 00 000000m,AmAHV 0500 0 000000000 00>00000 000 00 000000 000 00 8000000 0000000II.00 x00c0mm¢ O 00 00 00 00 00 Duration of One Bout 133 00005008 00 0000 000 00x00 000080050008 u 0 000 .000080050008 005008 o0 000000800000 000 n 0 .000000 00000>00000 005008 o0 00000 n < .om 000850 000800 000 000 0500 000 00 000000.000 00 00000050 0 00 0000000 AmmH~_05o0 0 000300000 00>00000 000 00 000000 000 00 8000000 0000000|Iow0 00000000 0500 000000000 00>00000 00 00000050 om on om cm 00 . om cm 7 00 m o .. o. . 0/0 800 gm 0 00.x“. O o o o o o m 0 . C O 0.03. o 0 .. .-O 0 a .- 0;. 0 0 AV 40 0 0 . o o on w 3 4 0000000.. . . .. q C Q ‘6 .. .. «.0 0 0. _ a 0 0 or .- om . 0 0 o d 04% o o 0 .0 o o, a on 0 0 o 0 o 0'0 0 o O o0 . . O om 0 o 0 C 00 Duration of One Bout OHH gscm UCTBCHHCk Hm>HmHCH mo ccfiwmnsc om on om om .mmuscHE 0H ummH mnn cmxmu mucmEmusmmmE n o cam .mucme ImHSmmmE mpacwe 0H mumwvmeumucw Ham u o .UOflHmm soapm>nmmno mascafi 0H umnflm u a .mm HmnEsc mamsmm mco you anon mco m0 numcma mnn mo coauocsm m mm Umuuon AmmHv uson m mca3oaaom ow om Hm>umucfl mnu mo ngmcma may no Emummac Hmuumomnl.ma xflccmmmd OH ON 0 m O ‘3‘ Duration of One Bout O I!) om mo 135 .o m an .wowumm Goa» Im>nmmno muscHE oa umma may Eonw cwxmu .mucmfimusmmmfi mama may .Ucm .0 m Mg mcoflpm>nmmno mpmawmaumucfl Ham .< am an cmumoavca mum coflumm coflum>nmmno muscafi 0H umuflm may Eoum cmxmu mucmEmHSmmmE uanmm was .mm “mass: mHmEmm Mom .uson mco mo sumcma mnn mo COHuUCSM m mm ampuoam AmmHv #503 m mcawmowum Hm>umucw mnp mo numcma may mo Emummaw umupmomluvoa xwvammmm 136 om on om 1‘ '.j usom mcflcmowum Hm>umucH mo coHumHso om O 8 cm on OH m o o. 0.0 o o 00 C O .0 . o O O 0. on 0‘ o O 4‘ ‘q o .0 a O o C .0 o o o ‘q o o o 00 0 ‘4 o f 4 . .- «anamw. OH om om ow om cm on Duration of One Bout 137 om on om om ov .mmusafls oa umma on» coxmu muses IoHSmmmE n 0 can .mucmEmHSmmoE ouscwfi OH mumacmsumucfl Ham u o .oofiumm cowum>ummno onscflfi oH umnflm n < .mm umnEsc mango“ mco How noon oso mo nuvcma may mo convocsw m mm omuuoam AmmHv poop m mcfi3oaaom Hm>umucfl an» no numcma mnn mo EMHmMflU Hmuumomllona xwocmmmd psom ocwzoaaom Hm>umucH mo cofiumndo om om OH . m 43“.. O. Iv O O mTV..- AV 00 @ oo _oo. o. .o Axvo AV 0‘ o 4‘ ,¢@.AV 0 AG 0 AG ‘0 4‘ 4‘ .Q OH ON on ow om ow on DD Duration of One Bout 138 ooo.m .mucmEmHSmme ocflammmn Scum :mxmu mama .omEmu mo coauocdm m mm #903 mom coaumuso mcflccmm mo EmumMflv Hmuumomlloma xflocmmmd oom.~ o _ 0 av 00 o o o mv_o o o oo owwooo o o o o o @% o no.0 0 , $0 on. o O 8 o o, o . o o oo o g 00 O o o o o. o o o o o o Aa\mv omEoB eoe.n oom.H ooo.H 0H om om ov om ow on om Duration/Bout 139 :mxmu muma eoe.n omucmEmHSmmmE mcflammmn Eonm .omfimu mo cowaocSM m mm mumwn mcficcmm mo Emnmmwv Hmuumomlu.ma xflwcmmm< A0\mv omfime oom.N a ooo.m oom.a ooo.H o om o o o o 2: O O O O o o o o o . O o ow o o a com 96, . O, O O 000 o o 000 O 0.0 0% o o o o o .unvv 0010 a. Au 0 O O 000. 00.0 0 o 000 0 .v o o 00 O O O O o 60 Am 0 o o O o O a O o O O m 0 o w o 83 o o o O O o o coma o o O o oo 0 oova Fanning Beats LITERATURE REVIEW Noble and Curtis (1939) observed the responses of parent Jewel fish to their eggs and established that visual cues were used in egg recognition. They found that Hemi- chromis bimaculatus could not discriminate their own eggs from those of certain other species. Reciprocal spawn exchanges between several different cichlid species in which eggs were accepted and the young raised demonstrated that some cichlids are unable to distinguish between eggs of their own and other species' (Greenberg, 1961, 1964 a, b, and Myrberg, 1964). The eggs of these species were not visually distinct to these observers. When reciprocal ex- changes were performed between certain other species, how- ever, the eggs were immediately eaten. Eggs of these spe- cies differed in size, shape, and color (Collins, 1962, 1965). The egg exchange studies cited implied that chemical' stimulation was not of great importance in egg recognition. This was further suggested since Hemichromis g2, parents, when presented with a choice between tap water and water Which had flowed over their own eggs, were unable to dis- criminate. "Fine" discriminations, however, were made between tap water and water which had come in contact with 140 141 either prolarvae or free-swimming larvae (Kfihme, 1963, 1965). Furthermore, it is possible selectively to vary a female's exposure to olfactory and visual stimulation pro- vided by the eggs. When this was done, there was no dif- ference in either the magnitude or temporal structure of fanning in cases where C. nigrofasciatum was presented with visual but no olfactory access, as opposed to both visual and olfactory stimulation by their eggs (Mertz, 1967). The conclusion reached was that parental fanning is caused at least, in part, by visual stimulation by the spawn. Fanning is not completely eliminated in the absence of all stimulation from the clutch (Mertz, 1967). Thus, it is possible to obtain fanning to a fixed point from non broody cichlids by injecting them with.prolactin (Blfim and Fiedler, 1962). This suggests that there are also internal factors which contribute to the expression of egg care behavior. 3