THE COMPARATIVE BEHAVIOR OF THREE SPECIES OF BLENNIOID FISHES A Thesis Presented to the Graduate Faculty California State College, Hayward In Partial Fulfillment of the Requirements for the Degree ~~ster of Arts in Biology by David Gregory Lindquist July 1971 ACKNOWLEDGEMENTS I would first like to thank Dr. Boyd w. Walker, my former advisor at UCLA, for suggesting this group of fishes to study. I am indebted to Dr. John s. Stephens, Jr., of Occidental College, for his sugges­ tion of the comparative problem and for his encourage­ ment and enthusiasm during the study. Dr. George s. Losey, Jr., of the University of Hawaii, generously relinquished some of his unpublished methods for my use. Finally I wish to thank the members of my com­ mittee: Drs. James w. Nybakken, Jack T. Tomlinson and Edgar Yarberry for critically reading the manuscript. TABLE OF CONTENTS LIST OF 1rABI...:fSS • • • • • • • • • iii LIST OF FIGtJRES • • • • iv ABSfJ.'RACT • • • • • • • • • v Il!l'l'RODUCTION • • • • • • • • • • • • l I. GEl:lERAL PROBLEI;i • • • • • • • • • • l ~ II. REVIEI'J OF LITERATTJRE • • • • • " P-~ .. Systerratics • • • • 2 B. Behavior • • • • 7 .0-JETHGDS .t::-1:-JV MA'L'ER:W.LS • • • • • • • • • 9 a r. FISHES STUDIED • • • • • • ~ A~ Gulf California Fishes • • • • • • • 9 1. Ernblemaria hvpacanthus • • • • 9 2. Acanthernblernaria crockeri • • • 9 B. Honterey Bay Fishes • • • • 10 1. Neoclinus uninotatus • • • 10 c. Se]~Ual Dj,morphism • • • • 11 II. OBSER\.11\TION l·'ETHOD Al:ID EQUIPl·lliNT 13 p_ .. Laboratory Observations • • 13 B. Field Observations • • 16 i TABLE OF CONTEblTS (Continued) III. COLLECTION AND TRANSPORT~TION OF SPECIMENS 17 GENERAL ETHOGRAM •• . 0 • • 0 0 21 COMFORT f-10VEMENTS • • • • • • • • ~ • 0 • 33 ELIMIN!\.TORY BEHAVIOR •• . 35 ENVIRONMENTAL ALTERATION'S . .. • • • • • 0 0 • 36 TERRITORIAL BEHAVIOR ••• . 38 I. SPECIES DESCRIPTIONS • • • . 39 AGONISTIC BEHAVIOR . • Ill • • 45 I. SPECIES DESCRIPTIONS . Q • .. • 46 II. TEMPORAL ASSOCIATION OF AGONISTIC MOVEMENTS 56 III. MOTIVATIONAL ANALYSIS • • 0 • • 0 . .. 65 REPRODUCTIVE BEHAVIOR • • • • 77 COLORATION 0 • .. • • . 89 I. COLOR PATTERN DESCRIPTIONS 0 • • • 0 0 0 • • 89 II. COLOR PATTERN CHANGES 0 0 0 0 . 99 DISCUSSION AND CONCLUSIONS • • 8 8 D 8 8 8 8 8 8 0 8 109 LITERATURE CITED • • • • • • • • • • .. • o • o • • • 114 ii LIST OF TABLES Table !?age Relationships between the species studied 14 2. The differences between the species' repertories in the general ethogram 28 3. Occurrence of agonistic motor patterns and their mean durations 55 4. Frequency with Which the nine different movements precede or follow each other in the agonistic behavior of Eroblemaria hypacenthus 59 5. Spearman's rank correlation coefficient for the agonistic movements of Emblemaria hypacanthus 60 6. Frequency with which the nine different movements precede or follow each other in the agonistic behavior of Acantheroble­ m<lria crockeri 61 7. Spearman'o rank correlation ~oefficient for the agonistic movements of Acanthem­ bleP.taria crocker:i. 62 8. SWlil!lary of stn.t~.stically correlated move­ ment pairs for the chaenopsids 64 9. Fin Flicker subunit duration 80 iii LIST OF FIGURES Figure Page 1. Circumorbital bones used to separate the three species studied 5 2. Study area for Neoclinus uninotatus 18 3. Agonistic behavior of Emblemaria hypacanthus 31 4. Flee~ttack diagrams for Ernblemaria hypacanthus, Acanthemblemaria crockeri and Neoclinus uninotatus 71 5. 1~ Emblemaria hypacanthus displaying the courtship inviting action~ Fin Flicker 81 6. Stimulus-response nature of the male invitation action 83 7. Early larva of Neoclinue uninotatus 87 B. Color patterns for Emblemaria hypacanth~ 90 9. Color patterns for ~canthemblemaria crocker! 93 10. Color patterns for Neoclinus uninotatus 96 11. Color pattern frequency in Emblemaria hypacanthus and Acanthemblemaria crockeri 104 • Color pattern frequency in Neoclinus uninotatus 107 iv ABSTRACT Laboratory and field comparisons of the behavior, morphology, distribution, habitat, and ecology of the blennioid fishes Effiblemaria hypacanthus, Acanthemblemaria crockeri and Neoclinus uninotatus indicate the former two species of the family Chaenopsidae to be more specialized than N. uninotatus of the family Clinidae. The comfort, eliminatory, and environmental alteration behavior of the three species is similar. The agonistic behavior of each species is distinct. The chaenopsids have elaborate threat codes and combat behavior. The agonistic behavior of the three species is apparently a function of the flee and attack motivation interaction. The action pattern sequence of pure attack or pure flee motivation differs from the sequence of mixed flee and attack motivations. The species differ in rela­ tive threshold levels of the action patterns, degree of ritualization, and in the relative importance of sign stimuli. v Analysis of the preceding and following motor pat­ terns of the combat behavior of the two chaenopsids by the Spearman rank correlation coefficient reveals correlated action pairs that fail to fulfill the fixed action pattern. One explanation for this suggests that selection for a decrease in the monotony of sequences of agonistic action patterns is advan·tageous for occasions when novelty is necessary for arousal and attention; e.g. fishes that en­ gage in complex fights require that each animal continu­ ally and accurately assess the state of its opponent. The general reproductive patterns for the three species does not differ. However, differences were ob­ served in the manner of the male invi·ting actions. Intraspecific color pattern variations are apparently related to the behavior of the individual. Three basic color patterns for each species are demonstrated. The function of the color patterns appears to fall into three categori~s~ accentuation of postural displays, signaling the motivational state of the individual, and effecting reproductive isolation. vi I~""TRODUCTION I. GENERAL PROBLEM The purpose of this study is to contribute to the understanding of the behavior of three related blennioid fishes. The comparative approach to animal behavior can assist in the assessment of relationships among related species. Systematics too often evaluates the phylogenetic relationships among species on the basis of morphological characteristics •. Comparison of the behavior within a closely related group of animals often reveals subtle differences in action patterns which may be analyzed in terms of their origins and adaptive value. Knowledge of these origins is in­ valuable for assessing the degree of inter- and intra­ specific convergence and divergence of action patterns and the value of these actions as taxonomic characters. These comparisons have proven valuable in assessing phylogenetic relationships (Hinde, 1970). Furthermore, it is the author's intent to make the first step in contributing behavioral information that can 2. later be used to elucidate the behavioral evolution of the fish families Clinidae and Chaenopsidae. Thus the system­ atics of the two groups might well be clarified with the additional use of behavioral information. II. REVlEW OF LITERATURE A. Systematics The two fish families involved in the study are the Clinidae and the Chaenopaidae. Neoclinus uninotatus, Hubbs 1953 (Clinidae) , Emblemaria hypacanthus (Jenkins and Ever­ mann 1888) (Chaenopaidae) and Acanthemblemaria crocker! Beebe a:l".d Tee-Van 1938 (Chaenopsidae) are the species under consideration. Clark Hubbs (1952 and 1953) in his revision of the family Clinidae placed the subfamily Chaenopsinae (which contained all three of the species above) under the family Blenniidae rather than under the Clinidae. He based this placement on the extension of the lachrymal bone to a point below the middle of the eye and the absence of scales except in Neoclinua. Springer (1955) disagreed with HUbbs' 1953 classification. Springer be­ lieved the lachrymal extension to be a clinid character 3. rather than a blenniid character. Thus, according ·to Springer, the Chaenopsinae should go with the Clinidae. Bohlke (1957) noted that the position of the Chaenopsinae may be about the same in either Hubbs' or Springer's classification since it is probable that the Blenniidae are derived from a clinid-like ancestor. But if the two families were to be maintained Bohlke preferred to follow Springer. In addition, Bohlke supposed the position of Neoclinus to be primitive to the other chaenopsin genera. Stephens (1963) erected the family Chaenopsidae, com­ posed of 10 genera and 39 species. This assemblage is based on two bones in the circumorbital ring, a lack of scales, and the absence of a lateral line. Since Neoclinus has four circumorbitals (Fig. 1), is scaled, and has a well developed lateral line, Stephens prefers to include ~­ clinus within the Clinidae. In this scheme the chaenopsids are derived from an ancestry similar to Neoclinus. Springer (1964) questioned Stephens' classification because of the scarcity of investigations as to the nature and number of the circumorbitals in clinids and blenniids. The major point Springer made is "whether the nuniber of circumorbi­ tals is a phyletic character, as Stephens believes, or a pointe'-! out Stephans (3.970), lateral line c~racteristics. Almost no variability degree o£ lateral line reduction in chaenopsids was described in pore counts. side ia a present within too adapt.iv,;;; and 5. Figure l Circumorbital bones used to separate the three blenny species studies. A. NeoclimJs uninotatue, four circwnorbitals. B. Acanthemblema:ria