A BEHAVIORAL STUDY of the HAWAIIAN GOBY-SHRIMP RELATIONSHIP and the EFFECTS of Predanon on the SYSTEM
Total Page:16
File Type:pdf, Size:1020Kb
A BEHAVIORAL STUDY OF THE HAWAIIAN GOBY-SHRIMP RELATIONSHIP AND THE EFFECTS OF PREDAnON ON THE SYSTEM A THESIS SUBMITTED TO THE GRADUATE DIVISION OF THE UNIVERSITY OF HAWAII IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN ZOOLOGY AUGUST 2005 By Robert Paul Nelson Thesis Committee: Jim Parrish Julie Bailey-Brock Tim Tricas Acknowledgments: I'd like to thank first and foremost my advisor Dr. J.D. Parrish for all his help in organizing and funding the project and reviewing my manuscript. Thanks to my committee members, Dr. Tim Tricas and Dr. Julie Bailey-Brock. Dr. Andrew Thompson helped in reviewing the manuscript. Mahalo to Casey Kaneshiro, Tyler Bouland and Rami Huiguas for collecting most of the daily rhythm cycle data. I also wish to express thanks for help from Georgi Kinsela, Jan Dierking, Katja Wunderbar, and Jeff Whitehurst. 111 Abstract The belief that the relationship between certain gobies and snapping shrimp (Alpheidae) is mutualistic typically includes the assumption that predation is a selective force driving the co-evolution of the relationship. In this study, I first showed the importance of the Hawaiian shrimp goby (Psilogobius mainlandi) to the sheltering behavior of its associated alpheid shrimp. Shrimp spent 53.6 ± 21.8 percent oflight hours in the day outside burrows with gobies present, but only 6.9 ± 3.4 percent ofthe time outside without gobies present. I then examined effects of predation by experimentally excluding predators on gobies from several I.S-m square plots and observing the subsequent density and size of gobies. Over the 5 months of predator exclusion, no significant effect on goby density was detected (ANOYA; p = 0.345). The most conspicuous results of the exclusion were the changes in the size classes of gobies (ANOYA, P < 0.00 I). Plots with exclosures had a mean of 2.23 more large gobies (> 4 cm TL) than plots with no exclosures. In a final part ofthe study, I documented the daily cycle of activity of the snapping shrimp, Alpheus rapax, with an associated goby present. The results indicate that A. rapax increases foraging and burrow maintenance activities toward the end of daylight. Overall, this study was able to test the widely held assumption that predation can be a selective pressure on the goby-shrimp association. IV TABLE OF CONTENTS Acknowledgments........................................................................ iii Abstract iv List of Tables............................................................................... vi List of Figures vii Introduction 1 Methods 3 Results 9 Discussion 12 References 16 Appendix 31 v LIST OF TABLES Table Page 1. Dates ofdata collection, corresponding sample ID numbers, and timing relative to 25 exclosure events. 2. Duration ofbehaviors in seconds (mean ±SO) recorded for shrimp during five 26 1000-sec sampling periods in daylight 27 3. Means (± S.D.) from seven shrimp burrows with associated goby and seven burrows without: 'bout' length; percent oftotal time that shrimp were out ofthe burrow; percent ofshrimp emergences that involved burrow maintenance 4. Results from two-way ANOYA of numerical density ofgobies (all sizes combined) 28 for exclosure treatment and time. 5. Mean numerical densities ofgobies per plot, divided into three size classes, for 29 each ofseven data collection dates and three exclosure treatments (full, partial, none). (For lO#s, refer to Table I.) 6. Results from two-way ANOYA ofnumerical density of large gobies for exclosure 30 treatment and time. VI LIST OF FIGURES Figure Page 1. Placement oftreatment areas, plots, and assignment oftreatments in goby bay. 18 2. Layout ofplots within each treatment area. 19 3. Mean time (± S.D.) spent by gobies in each type ofbehavior at various times of 20 day. 4. Mean number (± S.D.) ofgobies (all sizes combined) recorded per plot for each of 21 three treatments from February to October 2004. Exc10sures were in place and functional just after ID#O and were taken down immediately after ID#5. 5. Mean number (± S.D.) oflarge gobies (> 4 cm) recorded per plot for each ofthree 22 treatments from February to October 2004. Exc10sures were in place and functional just after ID#O and were taken down immediately after 10#5. 6. Estimated total biomass ofgobies per plot for each ofthree treatments from 23 February to October 2004. Exc10sures were in place and functional just after ID#O and were taken down immediately after ID#5. 7. Changes in snapping shrimp (Alpheus spp.) behaviors throughout one day. 24 Vll INTRODUCTION Associations of shrimp with gobies are widespread across the tropics. Most of the work on these goby-shrimp relationships has been done in the Red Sea/Indian Ocean (Luther 1958; Magnus 1967; Karplus 1981; Karplus et al. 1981; Polunin and Lubbock 1977) and Japan (Harada 1969; Yanagisawa 1978, 1982, 1984). There have been a few other studies in the Atlantic (Karplus 1992) and the South Pacific (Cummins 1979; Thompson 2004, 2005), but only one study on the Hawaiian association (Moehring 1972; Preston 1978). Nearly every report of the goby-shrimp relationship has concluded that the specIes are mutualistic (reviewed by Karplus 1987). The alpheid shrimps dig burrows in the predominantly sandy habitats where they live, providing protection for the gobies. The gobies stay near the entrance ofthe burrow during the day, close enough to dart in for escape from danger. At night neither species emerges from the burrow. The gobies, which probably have much keener eyesight than the shrimp, apparently provide a kind of 'early warning system' by being able to see potential predators earlier (Karplus 1981). The goby relays this information back to the shrimp through (1) its rapid head-first entries, and (2) a series of tail flicks which the shrimp detects through its antenna that are 'continuously positioned on the fish's body' (Moehring 1972; Karplus 1981, 1987, 1992). Theories have been proposed as to how the relationship may have evolved (KarpIus 1992). Most notably, the idea has arisen that the relationship evolved because ofpredation pressure exerted on shrimps and gobies living in close proximity to each other. Until recently, only Yanagisawa (1982) had noted any effect of predation pressure on the shrimp gobies. This effect was cited as a side-note in a larger study of shrimp gobies. He noted that in areas where predation was naturally lower, shrimp and gobies ventured farther from the burrow, and seemed to be more abundant. This study, however, was not experimental, and the observations were not tested analytically. Seeing a need to better understand the effects of predation on this goby-shrimp relationship, Thompson (2005) conducted a detailed experiment in French Polynesia in which, among other things, he manipulated the density of predators with access to observed gobies. In areas with increased predators, he found no effect on total goby density, but a significant decrease in the proportion oflarge gobies with increasing predator density. Because goby-shrimp relationships are known in more than 70 species ofgoby living in coral reef habitats around the world, it is unclear whether the effect of predation is universal. In Hawaii, one shallow-water shrimp goby occurs, (Psilogobius mainlandi Baldwin 1972), which is endemic to the islands, and classified with only one congener (Watson and Lachner 1985). Psilogobius mainlandi, along with its associated snapping shrimp (Alpheus rapax and A. rapicida), have been examined previously only in the context of communication between them, and the details of all goby-shrimp interactions seem to have been generalized from these two species (Moehring 1972; Preston 1979). I used these two species to study daily shrimp activity cycles, the effects of goby presence on shrimp behavior, and the effects of predation on numerical density and size of gobies by experimentally decreasing predation pressure. This study consisted ofthree parts. (1) I studied the behavior and interaction ofthe Hawaiian shrimp goby (P. mainlandi) and snapping shrimp (A. rapax) and compared them to other well studied goby-shrimp relationships. (2) I experimentally removed gobies to look for effects on shrimp behavior. (3) Beginning with natural densities of gobies and shrimp, I experimentally decreased predation pressure on gobies to determine the effects on density and sizes of gobies. The predation results can be compared with those previously shown for increasing predator density (Thompson 2005). 2 METHODS The entire study was conducted between October 2002 and October 2004. The chosen site was a small (approximately 100-m by 100-m) bay (hereafter called "goby bay") on the western side of Coconut Island, Oahu, HI (Lat: 2F 26.2' N Long: 157 0 47.6' W). The bay is protected from wind and waves, and thus has silty-sandy sediment covering a coral rubble base. These conditions provide suitable habitat for shrimp gobies (Thompson 2004). Diurnal Rhythm The diurnal rhythm ofsnapping shrimp was determined from six full-day measurements over the course of six weeks in March and April 2004. Each full-day measurement was set up just before 0600 hours. A small underwater surveillance camera was set above the goby burrow and wired to a video monitor and recording device on shore. The activities of shrimp of a single burrow were then analyzed for 1000 sec at each of five set times over the course of one day (0600, 0900, 1200, 1500 and 1800). To facilitate comparison with results of Karplus (1976), the same times of day, lengths of observation, and measurements were taken. To record the daily activity rhythm, a data logging program called BEAST (Behavioral Events Analysis System, Windward Technology, Hawaii) was used. BEAST records the occurrence of a behavior when a key on a computer keyboard is pressed. Following the general terminology of Karplus (1976), shrimp behaviors were classified into one of four categories: J) Inside the burrow: This behavior occurred when the shrimp was inside its burrow and could not be seen by the observer.