|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001670364
007 cr mnu|||uuuuu
008 051118s2005 flu sbm s000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001232
Collins, Angela Barker.
An examination of the diet and movement patterns of the atlantic cownose ray rhinoptera bonasuswithin a southwest florida estuary
h [electronic resource] /
by Angela Barker Collins.
[Tampa, Fla.] :
b University of South Florida,
Thesis (M.S.)--University of South Florida, 2005.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 97 pages.
ABSTRACT: Cownose rays are benthic, suction feeders whose foraging activities have been implicated in severe damage to commercial shellfish industries and seagrass habitat. With jaws highly modified for durophagy, it has been assumed that they are crushing specialists, feeding primarily upon hard molluscan prey. In addition, R. bonasus are believed to be highly migratory, transient residents of coastal inshore waters. However, minimal quantitative data exist regarding R. bonasus feeding or movement patterns in the Gulf of Mexico. Stomach contents from 50 cownose rays caught within the Charlotte Harbor estuary between July 2003 and July 2004 were analyzed using the index of relative importance (IRI). Crustaceans, polychaetes, and bivalves were the dominant groups present, with bivalves representing the smallest proportion of the three dominant groups. High dietary overlap was observed between sexes, size groups and seasons.Shoalmates exhibited significantly more similar diets to each other than to members of other shoals. Although currently believed to be a hard prey specialist, these results suggest the cownose ray may behave as an opportunistic generalist, consuming any readily available prey. Between July 2003 and November 2004, 21 cownose rays were tagged and tracked within Charlotte Harbor using passive acoustic telemetry. Residence time ranged between 1-102 days. No significant relationship was detected between activity patterns and tidal stage or time of day. Minimum convex polygons (MCP) and kernel utilization distributions (KUD) were calculated to demonstrate the extent of an animals home range and core areas of use. Daily MCPs ranged between 0.01 and 25.8 km2, and total MCPs ranged between 0.81 and 71.78 km2. Total 95% KUDs ranged between 0.18 and 62.44 km2, while total 50% KUDs were significantly smaller, ranging from 0.09 to 9.68 km2.
Adviser: Philip J. Motta.
Index of relative importance.
t USF Electronic Theses and Dissertations.
An Examination of the Diet and Movement Patterns of the Atlantic Cownose Ray Rhinoptera bonasus within a Southwest Florida Estuary by Angela Barker Collins A thesis submitted in partial fulfillment of the requirements for the degree of Masters of Science (MS) in Biology Department of Biology College of Arts and Sciences University of South Florida Major Professor: Philip J. Motta, Ph.D. Robert E. Hueter, Ph.D. Michelle R. Heupel, Ph.D. Susan S. Bell, Ph.D. Date of Approval: July 20, 2005 Keywords: batoid, myliobatiformes, feeding ecology, Charlotte Harbor, Index of Relative Importance, acoustic telemetry, passive tracking, home range Copyright 2005, Angela Barker Collins
For my parents, Bob and Susie Barker, for their unconditional love and support, and for teaching me that dreams are simply future plans. For my husband, Josh Collins, for his gentle nature and quiet humor and for reminding me that its all small stuff. And for all of the fishes in the sea, for letting us poke and prod, tag and track, hook and harass so that we can make an attempt to intelligently protect them for future generations.
Acknowledgements Charlotte Harbor National Estuary Program (CHNEP), PADI Project Aware, and PADI Foundation provided support in the form of grants awarded to A. Barker. Mote Foundation and the Mote Marine Laboratory/University of South Florida Graduate Student Fellowship provided financial support thro ughout the duration of this project. Additional funding for this research was provided by the Mote Scientific Foundation, NOAA/NMFS through the National Shark Research Consortium, Mote Marine Laboratory and the University of South Florida. Enormous thanks to Colin Simpfendorfer for his savvy statistics help and advice, and for writing the computer programs used in the movement pattern analysis. Thanks to Beau Yeiser for his exhaustive efforts to help catch rays for acoustic monitoring, and to Dr. Tony Tucker for lending me his net and for catching cownose when we couldnt. Special gratitude is extended to the Florida Fish and Wildlife Research Institute, especially Gregg Poulakis, Phil Stevens, Bob McMichael, Tim MacDonald and the entire Charlotte Harbor field lab, for providing specimens for stomach content analysis. Thanks to Jack Morris for help in maintaining captive cownose rays at Mote Marine Lab for over a year. Heartfelt gratitude is extended to Jay Leverone, Anamari Boyes, and Debi Ingrao for their patience and assistance in teaching me the beauty of benthic invertebrate identification. Thanks to Michelle Amato, Jim Gelsleichter, John Tyminski, Tonya Wiley-Lescher and numerous interns for field assistance. Sincere gratitude is expressed to Adam Summers, for providing a portion of the acoustic transmitters used in this project. Thanks to the entire Motta lab, and of course, to my committee Philip Motta, Michelle Heupel, Robert Hueter and Susan Bell for without their guidance and support, this project would never have reached completion.
i Table of Contents List of Tables................................................................................................................. ....iii List of Figures................................................................................................................ ....iv Abstract....................................................................................................................... .......vi Chapter One General Introduction.....................................................................................1 Study species............................................................................................................1 Study site..................................................................................................................3 Chapter Two Feeding Ecology..........................................................................................5 Introduction..............................................................................................................5 Methods....................................................................................................................7 Results....................................................................................................................12 Discussion..............................................................................................................24 Chapter Three Movement Patterns and Residence..........................................................31 Introduction ...........................................................................................................31 Methods..................................................................................................................34 Study area...................................................................................................34 Acoustic array............................................................................................34 Study species..............................................................................................35 Data analysis..............................................................................................37 Results....................................................................................................................41 Effect of transmitter attachment.................................................................41 Sampling results.........................................................................................42 Residence time...........................................................................................42 Activity patterns.........................................................................................45 Home range analyses.................................................................................47 MCPs..............................................................................................47 KUDs.............................................................................................51 Seagrass habitat overlap.............................................................................55 Distance between consecutive centers of activity......................................56 Discussion..............................................................................................................58 Chapter Four General Conclusions..................................................................................67 References..................................................................................................................... .....69
ii Appendices .................................................................................................................... ....95 Appendix A: Growth data for three captive R. bonasus used in transmitter attachment and retention studies............................................................................96 Appendix B: Minimum convex polygon and kernel utilization distributions in kilometers2 for all tracked individuals over 3 day, 7 day, 30 day and total monitoring periods.................................................................................................97
iii List of Tables Table 2-1. Monthly numerical and sexual composition of rays caught between July 2003 and July 2004............................................................................................12 Table 2-2. Prey groups identified for all R. bonasus collected from Charlotte Harbor, Florida...........................................................................................................15 Table 2-3. Mean Shannon-Weiner diversity index (H ) and Evenness (E) values for all, immature and mature R. bonasus stomachs examined...................................18 Table 2-4. Seasonal dietary overlap values for all examined R. bonasus .............................22 Table 3-1. Results for Chi-square analysis of the proportion of detections for R. bonasus within Pine Island Sound by hour, time of day and tidal height..................46 Table 3-2. T -test probability values ( p ) comparing activity space sizes between immature and mature R. bonasus over 3 day, 7 day, 30 day and total monitoring periods.....................................................................................................49 Table 3-3. T -test probability values ( p ) comparing activity space sizes between male and female R. bonasus over 3 day, 7 day and total monitoring periods........................................................................................................................ 50 Table 3-4. T -test probability values ( p ) comparing day and night KUD areas for immature and mature individuals...............................................................................54 Table 3-5. T -test probability values ( p ) comparing daytime and nighttime 95% and 50% KUDs between immature and mature animals...........................................55
iv List of Figures Figure 1-1. Map of Charlotte Harbor and its location along the Gulf coast of Florida........................................................................................................................ 4 Figure 2-1. Cumulative prey curve displaying total prey group versus total number of R. bonasus stomachs analyzed.................................................................13 Figure 2-2. Numerical and volumetric percent composition for the dominant prey groups and species comprising them.........................................................................16 Figure 2-3. Volumetric percent composition of the single dominant prey species within individual stomachs........................................................................................17 Figure 2-4. Shannon-Weiner diversity index and Evenness values for all R. bonasus stomachs examined......................................................................................18 Figure 2-5. Stomach fullness and stomach volume based on size of individual R. bonasus for all stomachs analyzed.............................................................................19 Figure 2-6. Linear regression of average prey size vs. straight disc width of R. bonasus collected in Charlotte Harbor.......................................................................20 Figure 2-7. Index of Relative Importance for eight major prey groups consumed by mature and immature and male and female R. bonasus.......................................21 Figure 2-8. Average similarity values for individuals within shoals and between shoals......................................................................................................................... .23 Figure 3-1. Map of Pine Island Sound and its location along the Gulf coast of Florida........................................................................................................................ 35 Figure 3-2. Method of attachment and site for external transmitter.....................................41 Figure 3-3. Daily presence of R. bonasus monitored within Pine Island Sound between June and November of 2003 and 2004........................................................43 Figure 3-4. Frequency histograms of residency times for R. bonasus monitored in Pine Island Sound......................................................................................................44 Figure 3-5. Hourly detection patterns...................................................................................47
v Figure 3-6. Frequency histogram representing the size of daily MCPs for immature and mature R. bonasus ...............................................................................48 Figure 3-7. Mean MCP areas for immature and mature R. bonasus over all monitoring periods (3 day, 7 day, 30 day and total)..................................................49 Figure 3-8. MCP areas over 3 day, 7 day, 30 day and total monitoring periods compared to R. bonasus straight disc width...............................................................50 Figure 3-9. Mean 95% and 50% KUD areas for immature and mature R. bonasus over all monitoring periods........................................................................................52 Figure 3-10. 95% and 50% kernel utilization distribution areas over 3 day, 7 day, 30 day and total monitoring periods vs. R. bonasus straight disc width....................53 Figure 3-11. Total 95% and 50% KUD areas compared to R. bonasus straight disc width...................................................................................................................54 Figure 3-12. Examples of 50% KUD overlap of seagrass habitat........................................55 Figure 3-13. Mean distance moved over 30-minute intervals vs. R. bonasus straight disc width......................................................................................................56 Figure 3-14. Frequency histogram of mean distance moved over 30-minute intervals for immature and mature R. bonasus ..........................................................57
vi An Examination of the Diet and Movement Patterns of the Atlantic Cownose Ray Rhinoptera bonasus within a Southwest Florida Estuary Angela Barker Collins ABSTRACT Cownose rays are benthic, suction feeders whose foraging activities have been implicated in severe damage to commercial shellfish industries and seagrass habitat. With jaws highly modified for durophagy, it has been assumed that they are crushing specialists, feeding primarily upon hard molluscan prey. In addition, R. bonasus are believed to be highly migratory, transient residents of coastal inshore waters. However, minimal quantitative data exist regarding R. bonasus feeding or movement patterns in the Gulf of Mexico. Stomach contents from 50 cownose rays caught within the Charlotte Harbor estuary between July 2003 and July 2004 were analyzed using the index of relative importance (IRI). Crustaceans, polychaetes, and bivalves were the dominant groups present, with bivalves representing the smallest proportion of the three dominant groups. High dietary overlap was observed between sexes, size groups and seasons. Shoalmates exhibited significantly more similar diets to each other than to members of other shoals. Although currently believed to be a hard prey specialist, these results suggest the cownose ray may behave as an opportunistic generalist, consuming any readily available
vii prey. Between July 2003 and November 2004, 21 cownose rays were tagged and tracked within Charlotte Harbor using passive acoustic telemetry. Residence time ranged between 1-102 days. No significant relationship was detected between activity patterns and tidal stage or time of day. Minimum convex polygons (MCP) and kernel utilization distributions (KUD) were calculated to demonstrate the extent of an animals home range and core areas of use. Daily MCPs ranged between 0.01 and 25.8 km2, and total MCPs ranged between 0.81 and 71.78 km2. Total 95% KUDs ranged between 0.18 and 62.44 km2, while total 50% KUDs were significantly smaller, ranging from 0.09 to 9.68 km2. Both MCPs and KUDs exhibited a significant positive relationship with residence time and with disc width. As mobile, pelagic swimmers capable of traversing large distances, these data also demonstrated that cownose rays may remain within relatively small areas for extended periods. These diet and tracking results provide insight to R. bonasus use of a south Florida estuary and allow predictions regarding the impact of this species in similar environments.
1 Chapter 1: General Introduction Study species Atlantic cownose rays ( Rhinoptera bonasus Mitchill 1815) occur along coastlines of the western Atlantic (from New England to Brazil), Caribbean and Gulf of Mexico. Reaching disc widths greater than 100 cm, R. bonasus are commonly documented within Florida waters. Cownose rays exhibit schooling and shoaling behavior (Pitcher and Parrish, 1993) and have often been observed traveling along coastlines in organized groups that can number thousands of individuals (Clark, 1963; Schwartz, 1965; Blaylock, 1989; Rogers, 1990). Schools and shoals appear to form based on size, and typically contain representatives of both sexes. These massive aggregations have been suggested to form during large-scale migrations, triggered by seasonal changes in water temperature or sun orientation (Schwartz, 1965, 1990). This idea is supported in many areas by the inshore abundance of cownose rays during spring and summer (Snelson and Williams, 1981; Smith and Merriner, 1987; Blaylock, 1992) and absence during fall and winter (Grant, 1983; Hoese and Moore, 1977). Schwartz (1990) suggested that there are two separate migrating populations of R. bonasus within U.S. waters. He proposed that the Atlantic population migrates along the east coast from New England to Brazil, and that the population in the Gulf of Mexico travels clockwise along coastlines in a circular pattern from the Yucatan Peninsula to Florida. Alternatively, it has been suggested that cownose rays may simply move offshore during the colder months (Smith and Merriner,
2 1987; Rogers, 1990). While data supporting these hypotheses are lacking, R. bonasus is assumed to be a transient, seasonal resident of bays and estuaries. Although pelagic swimmers, R. bonasus behave as benthic suction feeders (Sasko, 2000) and have been observed creating feeding pits up to 1 meter wide and 45 cm deep (Orth, 1975). Considerable evidence exists that rays (including R. bonasus ) can significantly impact the benthic environment, either directly through consumption of organisms (Smith and Merriner, 1985; Blaylock, 1992; Peterson et al., 2001), or indirectly via invasive feeding behavior (Orth, 1975; Valentine et al., 1994, Silberhorn et al., 1996). Along the eastern coast of the United States, their feeding activities have been implicated in severe damage to the commercial shellfish industry (Smith and Merriner, 1985, Blaylock, 1992, Peterson, 2001) and seagrass habitat (Orth, 1975; Hovel and Lipcius, 2001). All previous evidence indicates that cownose rays are hard prey specialists (Smith and Merriner, 1985, Blaylock, 1992, Silberhorn, 1996, Peterson et al., 2001). With jaws highly modified for durophagy (Summers, 2000, 2003), it has been suggested that they are stenophagous in their choice of bivalve mollusks (Smith and Merriner, 1985). Due to the tendency of R. bonasus to travel in large groups, impacts from their feeding behavior can be concentrated in relatively small areas, magnifying potential effects. Most of the information available regarding cownose ray ecology stems from research done along the Atlantic coast of the United States, where damage to seagrass habitat and commercial shellfish beds (Orth, 1975; Merriner and Smith, 1979; Smith and Merriner, 1986; Blaylock, 1992; Kraeuter and Castagna, 1980; Peterson et al., 2001) prompted some investigation of R. bonasus diet and distribution. It is important to note
3 that although considerable data have been gathered for R. bonasus along the eastern coast of the U.S., there have been no quantitative data collected regarding feeding or movements of cownose rays within inshore waters of the Gulf of Mexico. The goals of this study were to identify the diet and movement patterns of R. bonasus within Charlotte Harbor, a southwest Florida estuary, to help better define the ecological role of this species along the western coast of Florida. Study site Charlotte Harbor is an estuary on the west coast of Florida (between 27 05 and 26 27 N latitude, 81 50 and 82 30 W longitude). The shore of Charlotte Harbor is variable with highly developed canal systems interspersed with undeveloped, protected areas (e.g. National Wildlife Refuge). The aquatic habitat within the harbor varies from shallow sand and mud flats to seagrass beds and deep channels (Figure 1). Depths within the estuary range from zero to ten meters with a tidal range of 0.7 to 1.8 meters. Seagrass beds are typically found in areas less than 2 m deep and consist of three main species: manatee grass, ( Syringodium filiforme ), turtle grass ( Thalassia testudinum ) and shoal grass ( Halodule wrightii ).
4 Figure 1-1. Map of Charlotte Harbor and its location along the Gulf coast of Florida. Sampling zones for stomach content analysis are indicated by dashed lines and labeled A F. Hatched areas indicate seagrass beds and the bathymetry contour indicates depths greater than 4 meters.
5 Chapter 2: Feeding Ecology Introduction As an abundant predator that can significantly modify the benthic environment, the cownose ray has the potential to be a major determinant of community structure (Choat, 1982). Cownose rays are benthic, suction feeders (Sasko, 2000), capable of creating feeding pits up to a meter wide and 45 centimeters deep (Orth, 1975). Rhinopterids have been shown repeatedly to feed upon bivalve mollusks (Smith, 1907; Bigelow and Schroeder, 1953; James, 1962; Blaylock, 1992; Smith and Merriner, 1985). Smith and Merriner (1985) concluded that the cownose ray is stenophagous and consistent in its consumption of bivalves. Since their jaws are highly modified for durophagy (Smith and Merriner, 1985; Summers, 2000), it has been suggested that R. bonasus eat hard molluscan prey to the exclusion of all else, (Summers et al., 2003). Most of the information available regarding cownose ray feeding ecology stems from research done along the Atlantic coast of the United States, where the feeding activities of cownose rays have been implicated in substantial damage to the commercial shellfish industry (Orth, 1975; Merriner and Smith, 1979; Smith and Merriner, 1986; Blaylock, 1992; Kraeuter and Castagna, 1980; Peterson et al., 2001). Smith and Merriner (1985) have performed the only quantitative diet study to date, examining the stomach contents from 68 R. bonasus caught within Chesapeake Bay. Bivalves were the dominant prey item, making up 99.8% of the total volume examined. They observed an ontogenetic
6 shift in prey type, with adults consuming deeper, burrowing bivalves while juveniles ate shallower or non-burrowing bivalves. Bigelow and Schroeder (1953) described R. bonasus prey to include oysters, clams and other large bivalves, as well as crustaceans and gastropods. Blaylock (1992) reported considerable damage to oyster beds in Chesapeake Bay and Tiller et al. (1952) attributed the loss of 600 bushels of clams in 2 nights to feeding cownose rays. The minimal R. bonasus diet data that exist for the Gulf of Mexico are qualitative, but also suggest that bivalves are the predominant prey type (Wang and Raney, 1971; Livingston, 1984; Valentine et al., 1994). Charlotte Harbor, one of the largest and least contaminated estuaries within Florida (Pierce et al., 1986, 2003), supports significant and diverse populations of invertebrate fauna (Estevez, 1981, 1986), including commercial shellfish. Cownose rays have been documented within Charlotte Harbor throughout the year (R. Hueter, Mote Marine Laboratory, Center for Shark Research, unpublished data). Presence of R. bonasus within the area suggests they are using this habitat and will feed within this region throughout the year. The goals of this study were to describe the diet and distribution of cownose rays within Charlotte Harbor, giving consideration to size and seasonal differences, to better understand the ecological impact of cownose rays within estuaries along Floridas Gulf coast.
7 Methods Stomachs were obtained from R. bonasus captured within Charlotte Harbor zones A-F (Figure 1-1) during daylight hours over a twelve-month period (July 2003-July 2004) during routine sampling by both Mote Marine Laboratorys Center for Shark Research (Mote CSR) and Florida Fish and Wildlife Research Institutes (FWRI) FisheriesIndependent Monitoring (FIM) program. Rays were collected in entanglement nets, purse seines, or beach seines. No evidence of regurgitation was observed in any collected individuals. Animals were euthanatized upon capture with an overdose of tricainemethane-sulfonate (MS222) and placed on ice until returning to the lab. Rays were measured to the nearest centimeter (straight disc width, SDW), weighed to the nearest 0.1 kilogram, and sexed. Stomachs were removed from 50 individuals (28 adults; >70 cm disc width and 22 juveniles; <70 cm disc width) upon returning to the lab and stored frozen until transferred to 10% buffered formalin for fixation. Maturity was based on sampling data by Mote CSR (Collins, unpublished data) and judged by reproductive condition (clasper calcification in males and average minimum size at pregnancy in females). Size at maturity differs from Smith and Merriner (1986), who estimated maturity for R. bonasus in Chesapeake Bay to be >90 cm (females) and >84 cm (males), but conservatively agree with size at maturity for R. bonasus in the northern Gulf of Mexico (65 cm; Neer and Thompson, 2005). After fixation all stomach contents were rinsed and placed in 70% ethanol or isopropyl alcohol prior to sorting. Prey items were identified to the lowest possible taxonomic level. To determine whether sufficient stomachs were collected, a cumulative prey curve was constructed by plotting total number of prey families observed against the number of stomachs examined
8 (Kyne and Bennett, 2002). The number of stomachs included was deemed sufficient to adequately describe the diet when the curve reached an asymptote. Numerical and volume composition (Nc and Vc) and frequency of occurrence (Fo) were calculated for each prey type and used to determine the index of relative importance (IRI) (Pinkas et al., 1971; Hyslop, 1980): IRI = (%Nc + %Vc) %Fo Prey that were not whole were enumerated by defining a unique part (i.e. head of crustacean, whorl of gastropod, head of polychaete) as one individual. After enumerating all prey items from an individual stomach, each prey type from that stomach was volumetrically quantified using displacement. Prey types were put into a cylinder containing a known volume of seawater, and the volume displaced was used to determine volumetric composition (Vc) of prey types within each stomach. To facilitate comparisons among prey type, percent IRI (%IRI) was calculated (Corts, 1997). After identification to the lowest possible taxonomic level, prey items were then classified into nine major taxonomic groups: bivalves, brachiopods, crustaceans, chordates, detritus, echinoderms, gastropods, nematodes, and polychaetes. The Shannon-Weiner diversity index (H ) (Krebs, 1999) was calculated for each stomach to measure the diversity of prey items consumed. This index gives a value to the uncertainty of correctly guessing the next prey type to be found within a stomach in other words, it quantifies the heterogeneity of prey items within a sample. H was calculated for each stomach and averaged over all stomachs in each size class: H = !s i ip p1 2) )(log ( where H = Shannon-Weiner Index of diversity,
9 s = number of major taxonomic prey groups (i.e. bivalves, crustaceans ), and pi = proportion of total sample (individual volume) belonging to the ith group. Evenness (E) describes the equitability of distribution of individuals among major prey groups (Krebs, 1999). For example, a stomach containing only two types of prey would have a low evenness value, while a stomach containing equal proportions of all possible prey groups would display complete evenness (E=1.0). Evenness allows us to compare actual diversity to the maximum possible diversity, and was calculated for each stomach and averaged over all stomachs in each size class as: E = H / Hmax where H = Shannon Weiner Index of diversity, Hmax = log2s, and s = total number of major taxonomic prey groups consumed by one size class of rays (Gray et al., 1997; Krebs, 1999). To determine how stomach volume changed with ray size, individual stomach content volumes were measured for all ray stomachs containing prey, and qualitative stomach fullness (SF) values of 1-10 (10 being a completely full stomach) were assigned for each stomach. Linear regressions were performed to detect any correlations between R. bonasus size and stomach volume or fullness. The size of individual prey items was estimated to the nearest millimeter. Total body length for crustaceans, nematodes, polychaetes, gastropods and chordates were recorded, while total width was measured for bivalves and brachiopods. When crushed, gastropod, bivalve and brachiopod sizes were conservatively estimated by measuring intact visceral masses. The broken condition of all observed echinoderms forced
10 minimum estimations of their numerical composition by counting intact mouths as one individual, and measuring echinoderm size was not possible. Detritus and seagrass were excluded from this analysis because they could not be enumerated. Prey items were measured and an average prey size value was calculated for each R. bonasus stomach and a linear regression was performed to determine whether prey size increased with ray disc width. Dietary overlap was assessed for male and female R. bonasus immature and mature R. bonasus, and between seasons, using the Simplified Morisita Index of overlap (Horn, 1966; Krebs, 1999): CH = !! !n i n i ik ij n i ik ijp p p p2 22 where CH = Simplified Morisita index of overlap (Horn, 1966) between groups j and k pij = proportion resource i of the total resources used by group j pik = proportion resource i of the total resources used by group k and n = total number of prey groups. A value of zero indicates no dietary overlap, while a value of one represents complete overlap. The system used by Morte et al. (2002) was followed, where overlap index values were arbitrarily classified as low (0.00 to 0.29), moderate (0.3 to 0.6) and high (>0.60). When multiple rays from a single shoal were collected, these individuals were tested to determine whether animals shoaling together had more similar stomach contents
11 than those not shoaling together. This was calculated using the Simplified Morisita Similarity Index (Morisita, 1959; Krebs, 1999 p. 391): CH = )] / ( ) / [(2 2 2 2 k ik j ij ik ijN P N P P P! where CH = Morisitas index of similarity between sample j and k, Pij, Pik = volumetric proportion of species (or lowest possible taxonomic group) i in sample j and k and Nj, Nk = Pij 2, Pik 2 = total proportion of individuals in sample j and k Volumetric proportions were used rather than numeric proportions because these better describe the nutritional contribution of a prey item to its predator (Hyslop, 1980). CH values were calculated for each individual compared to all of its shoalmates, as well as to all of the other rays that were not part of its shoal. The similarity values both within and between shoals were then averaged and examined for differences.
12 Results Fifty R. bonasus were collected from Charlotte Harbor between July 2003 and July 2004 (Table 2-1) Straight disc widths (SDW) ranged from 37 to 78 centimeters. Thirty-seven out of 50 stomachs (74.0%) contained identifiable material. Of these 37, 16 were from mature (7 females, 9 males) and 21 (8 females, 13 males) were from immature R. bonasus Thirty-five of the 37 analyzed (94.6%) were collected in zones A, B or C. No rays were caught in zone D. The remaining two R. bonasus were captured in lower Pine Island Sound (zone E) and the Caloosahatchee River (zone F). Results from the cumulative prey curve analysis indicated 26 stomachs were sufficient to represent the diet of the cownose ray within Charlotte Harbor (Figure 2-1). Table 2-1. Monthly numerical and sexual composition of rays caught between July 2003 and July 2004. Month N Male Female July 3 2 1 August 2 2 0 September3 0 3 October 7 6 1 November 7 3 4 December 1 0 1 January 5 5 0 February 6 2 4 March 6 3 3 April 0 0 0 May 4 2 2 June 6 3 3
13 Figure 2-1. Cumulative prey curve displaying the total prey group (species = when prey could not be identified to species, the next lowest possible taxonomic classification was used) vs. the total number of Rhinoptera bonasus stomachs analyzed. A total of 92,577 prey items were identified and fell into nine major groups: crustaceans, polychaetes, bivalves, echinoderms, gastropods, brachiopods, chordates, nematodes, and detritus (Table 2-2). Nematodes, chordates, gastropods, and brachiopods all had a %IRI of less than 1. Nematodes were the least common item, with a single organism being present in only two stomachs. Chordates also had a low frequency of occurrence, found in only five stomachs. However, when present, chordates did at times occur in high volumes. For example, one stomach contained 66% (by volume) Branchiostoma sp. (lancelets) and another contained 18% tunicates. Gastropods and brachiopods had a high frequency of occurrence (43.2% and 37.8%, respectively) but occurred in low numbers or volumes relative to the dominant prey groups, making their relative IRI values low (0.84% and 0.71% respectively). Echinoderms, represented
14 primarily by the sand dollar Mellita sp., occurred in 10 stomachs and had a %IRI of 2.3. Although echinoderms occurred in 27% of the examined stomachs, the %IRI was low due to minimum estimations of this groups numerical composition. Detritus occurred in 64.9% of the examined stomachs and was classified as any organic, non-animal matter (usually wood, leaves or seagrass). Detritus represented 5.8% of total volume consumed. Its IRI value was not considered because detritus could not be accurately enumerated. Crustaceans, polychaetes, and bivalves were the dominant prey groups present across all stomachs examined (%IRI = 55.3, 25.2, and 12.6 respectively; Table 2-2), together making up 98.5% (numerically), 75.3% (volumetrically) and 93.1% of the total IRI. Overall, crustaceans had the highest volumetric composition (28.8% Vc), followed by polychaetes (26.0% Vc) and bivalves (20.5% Vc). Polychaetes and bivalves had the highest frequency of occurrence (81.1 and 75.7 %Fo, respectively), and crustaceans represented 70.3 %Fo. Numerical composition of prey items was extremely high (92,577 items) due primarily to the consumption of small cumaceans (~2mm), an epibenthic crustacean that can occur in high density swarms. A total of 71,484 cumaceans was consumed by 17 rays comprising 77.2% of the total Nc. Infaunal polychaete worms, Pectinaria gouldii, were the next most numerically abundant prey items with 14,068 (15.2% by number) consumed by 24 rays. These two prey items contributed to the high numerical values and comprised 92.4% of the total number of prey items recorded. Table 2-2. Prey groups identified for all Rhinoptera bonasus collected from Charlotte Harbor, Florida. Frequency of occurrence (F), numerical composition (N), volumetric composition (V), Index of Relative Importance (IRI), and their respective percentages are shown for major taxa.
15Prey F Fo (%) N Nc (%) V (ml) Vc (%) IRI IRI (%) Bivalves 28 75.68 2141 2.31 116.82 20.54 1729.45 12.58 Mytiloida 20 54.05 1520 1.64 52.24 9.22 Mytilidae 20 54.05 1520 1.64 52.24 9.21 586.79 4.27 Veneroida 14 37.84 513 0.55 43.08 7.60 Crassatellidae 2 5.41 11 0.01 1.09 0.19 Mactridae 4 10.81 17 0.02 2.40 0.42 Solecurtidae 1 2.70 7 0.01 0.10 0.02 Tellinidae 12 32.43 440 0.47 35.57 6.27 Veneridae 4 10.81 16 0.02 1.15 0.20 Veneroid family 2 5.41 22 0.02 2.77 0.49 Arcoida 4 10.81 47 0.05 2.61 0.46 Glycymerididae 4 10.81 47 0.05 2.61 0.46 Pterioda 1 2.70 8 0.01 14.00 2.47 Ostreidae 1 2.70 8 0.01 14.00 2.47 Unidentified bivalves 9 24.32 53 0.06 4.89 0.86 Brachiopoda 14 37.84 666 0.72 10.06 1.78 94.42 0.69 Lingulidae 14 37.84 666 0.72 10.06 1.78 Chordata 5 13.51 280 0.30 17.63 3.11 46.13 0.34 Branchiostomidae 2 5.41 240 0.26 15.55 2.75 tunicates 1 2.70 11 0.01 1.90 0.34 teleost scales 2 5.41 29 0.03 0.17 0.03 Crustaceans 26 70.27 73561 79.46 162.95 28.76 7604.89 55.31 Isopoda 9 24.32 162 0.18 5.27 0.93 Anthuridae 8 21.62 161 0.17 5.17 0.91 Idoteidae 1 2.70 1 0.00 0.10 0.02 Amphipoda 14 37.84 1047 1.13 9.99 1.76 Ampeliscidae 2 5.41 47 0.05 0.47 0.08 Oedicerotidae 10 27.03 572 0.62 5.59 0.99 Unidentified amphipods 2 5.41 428 0.46 3.93 0.69 Mysidacea 7 18.92 656 0.71 12.06 2.13 Mysidae 7 18.92 656 0.71 12.06 2.13 Cumacea 17 45.95 71484 77.22 133.19 23.51 4628.01 33.66 Bodotriidae 17 45.95 71484 77.22 133.19 23.51 Decapoda 5 13.51 9 0.01 0.69 0.12 crabs 5 13.51 8 0.01 0.65 0.11 shrimps 1 2.70 1 0.00 0.05 0.01 Maxillopoda 6 16.22 129 0.14 0.77 0.14 barnacles 6 16.22 129 0.14 0.77 0.14 Unidentified crustaceans 8 21.62 74 0.08 0.96 0.17 Echinoderms 10 27.03 19 0.02 67.01 11.83 320.26 2.33 Holothuroidea 1 2.70 2 0.00 1.80 0.32 Echinoidea 7 18.92 13 0.01 62.86 11.10 Ophiuroidea 2 5.41 3 0.00 2.30 0.41 Unidentified echinoderms 1 2.70 1 0.00 0.05 0.01 Gastropods 16 43.24 424 0.46 12.18 2.15 112.78 0.82 Cephalaspidea 4 10.81 126 0.14 1.74 0.31 Acteocinidae 2 5.41 107 0.12 0.83 0.15 Haminoeidae 2 5.41 19 0.02 0.91 0.16 Neotaenioglossa 2 5.41 2 0.00 0.08 0.01 Assimineidae 1 2.70 1 0.00 0.03 0.01 Strombidae 1 2.70 1 0.00 0.05 0.01 Neogastropoda 2 5.41 2 0.00 0.10 0.02 Columbellidae 1 2.70 1 0.00 0.05 0.01 Nassaridae 1 2.70 1 0.00 0.05 0.01 Unidentifed gastropods 10 27.03 294 0.32 10.26 1.81 Nemata 2 5.41 2 0.00 0.06 0.01 0.07 0.00 Polychaeta 30 81.08 15484 16.73 147.31 26.00 3464.59 25.20 Capitellidae 1 2.70 50 0.05 1.00 0.18 Goniadidae 6 16.22 206 0.22 4.09 0.72 Nereididae 16 43.24 133 0.14 2.59 0.46 Pectinariidae 24 64.86 14068 15.20 125.74 22.20 2425.42 17.64 Phyllodocidae 1 2.70 1 0.00 0.10 0.02 Spionidae 3 8.11 817 0.88 7.05 1.24 Opheliidae 1 2.70 94 0.10 3.30 0.58 Unidentified polychaetes 3 8.11 116 0.13 3.45 0.61 Detritus and Grass 24 64.86 n/a 0.00 32.93 5.81
16 A small number of species represented an overwhelming proportion of the three dominant prey groups (Figure 2-2). The cumaceans Cyclaspis sp. and Oxyurostylis smithi (family Bodotriidae) had a combined %IRI of 33.7 and represented 97.2% (numerically) and 81.7% (volumetrically) of all crustaceans present. Pectinaria gouldii was the dominant polychaete with a %IRI of 17.6, making up 90.9% (numerically) and 85.4% (volumetrically) of all polychaetes present. Of the bivalves, Amygdalum papyrium (paper mussel) was the most abundant, with a %IRI of 4.27, representing 71.0% (numerically) and 44.7% (volumetrically) of all bivalves consumed. Figure 2-2. Numerical and volumetric percent composition (%Nc and %Vc, respectively) for the dominant prey groups and species comprising them. Both cumacean species identified within R. bonasus stomachs ( Cyclaspis sp. and Oxyurostylis smithii ) are in the family Bodotriidae. The number of major taxonomic groups (i.e. polychaetes, echinoderms, etc.) per stomach ranged between one and seven, and the number of families per stomach ranged between two and 17. Although all stomachs contained multiple items, there was a tendency for individual stomachs to be dominated by a particular species. The volumetric
17 percent composition of the dominant prey species (or lowest possible taxonomic group) was never less than 40% of the total stomach volume (Figure 2-3). The average percent volumetric composition (%Vc) of the dominant prey item per stomach was 69.0%. Diversity (H ) varied between stomachs, and values ranged between 0 and 1.9. Overall, H was generally greater than 0.75 (mean = 0.94) (Table 2-3). No distinct pattern emerged when H was compared to R. bonasus disc width (Figure 2-4), and there were no obvious differences in diversity between immature and mature rays (Table 2-3). Evenness (E) ranged between 0 and 0.66, but was generally low (overall average E= 0.33). As with H E values did not display a distinct relationship with R. bonasus disc width (Figure 2-4, Table 2-3). A generalized linear model (Statistica) was used to further test for a association between H or E and disc width. No significant relationship was found for either value when regressed against SDW (H vs. SDW: p = 0.44; E vs. SDW: p = 0.29). Figure 2-3. Volumetric percent composition of the single dominant prey species (species= when prey could not be identified to species, the next lowest possible taxonomic classification was used) within individual stomachs. Dotted line identifies 40% volumetric composition of a single prey type.
18Table 2-3. Mean Shannon-Weiner diversity index (H ) and Evenness (E) values for all, immature, and mature R. bonasus stomachs examined. H E Total 0.940.33 Males (n=22) 0.990.34 Females (n=15) 0.870.30 Mature 0.870.29 Males (n=9) 1.040.35 Females (n=7) 0.660.22 Immature 0.990.35 Males (n=13) 0.950.34 Females (n=8) 1.050.37 Figure 2-4. Shannon-Weiner diversity index (H closed circles) and Evenness (E, open circles) values for all R. bonasus stomachs examined. Dashed line separates immature from mature rays. Mean stomach volume (SV) of all sampled R. bonasus was 17.36 ml, and the average stomach fullness (SF) value was 4.8. Immature rays had lower SV values than mature rays (Figure 2-5; r2= 0.64, p = 0.008). Prey size ranged from < 2.0 mm to 35.0 mm, and showed a general increase with ray disc width (Figure 2-6, r2 = 0.88, p=0.0018).
19 The largest item consumed was an oyster Crassostrea virginica (35.0 mm), which was found in only one ray (mature female, 77.6 cm DW). Lancelets ( Branchiostoma sp.) up to 35.0 mm long were found in large quantities in two stomachs (both mature females, 73.0 and 76.3 cm SDW). Polychaetes reached lengths of 30.0 mm, while crustaceans and gastropods fell within a size range of 2.0 12.0 mm. Figure 2-5. Stomach fullness, SF, (a) and stomach volume, SV, (b) based on size of individual R. bonasus for all stomachs analyzed. Dashed line separates immature from mature rays. There was no significant relationship between stomach fullness and disc width (linear regression, p=0.4496). There was a significant relationship between stomach volume and disc width (linear regression, p=0.008**).
20 Figure 2-6. Linear regression of prey size vs. straight disc width of R. bonasus collected in Charlotte Harbor. There was a significant relationship between prey size and disc width (linear regression, p=0.0018**). Dotted lines represent 95% confidence intervals. Dashed line separates immature from mature samples. Dietary overlap between immature and mature R. bonasus was high (CH = 0.87); Figure 2-7). For both groups, the highest %IRI values were for the same three major prey groups (crustaceans, polychaetes and bivalves). However, there were slight differences between groups. Immature rays were more likely to have detritus and grass in their stomachs (Fo = 76.2% immature, 50.0% mature). Bivalves occurred in 90.5% (Fo) of immature stomachs, but only 56.3% of mature stomachs. Crustaceans dominated numerically over both size classes (N = 53,457, Nc = 77.8%, immature and N = 20,104, Nc = 84.3%, mature). Crustaceans also dominated volumetrically for mature rays (Vc = 38.1%), but polychaetes were present in the highest volume for immature rays (Vc = 32.8%). Only mature R. bonasus had chordates (lancelets, tunicates or teleost scales) in
21 their stomachs. Teleost scales were present within two stomachs in low volume, but no other fish remains (i.e. vertebrae or muscle) were identifiable. Figure 2-7. Index of Relative Importance (%IRI) for eight major prey groups consumed by mature and immature (a) and male and female (b) Rhinoptera bonasus collected in Charlotte Harbor for stomach content analysis. Morisitas simplified overlap index was high for both comparisons: immature vs. mature CH = 0.87; female vs. male CH = 0.84.
22 Diet comparison among male and female rays also showed high dietary overlap (CH = 0.84). The only notable differences between male and female diets were that males had a higher %IRI for polychaetes (male = 37.5, female = 13.0), and females had a higher %IRI for crustaceans (female = 69.8, male = 41.8) (Figure 2-7). Examination of seasonal differences in diet showed mixed results with all dietary overlap values falling in the medium or high categories (Table 2-4). No seasonal dietary overlap fell within the low (CH = 0.0 to 2.9) category. There was high dietary overlap for all rays between spring and summer (CH = 0.86), spring and fall (CH = 0.78), and summer and fall (CH = 0.72). Dietary overlap was classified as medium between winter and all other seasons with CH values ranging from 0.35 0.56. The lowest dietary overlap was between summer and winter (CH = 0.35) and was due mostly to seasonal differences in crustacean and polychaete consumption: summer polychaete %IRI = 42.1, summer crustacean %IRI = 0.15. Values for these groups were almost reversed during winter: winter polychaete %IRI = 8.3 and winter crustacean %IRI = 52.0. Table 2-4. Seasonal dietary overlap values for all examined R. bonasus. Values were calculated using the Simplified Morisita measure for niche overlap (CH). Seasonal overlap is considered high when CH > 0.60, medium when CH falls between 0.3-0.6, and low when CH <0.3. Lowest overlap values were calculated for the winter comparisons. Fall (n=14) Winter (n=10) Spring (n=9) Summer (n=4) 0.7228 0.3497 0.8634 Spring 0.7767 0.3888 Winter 0.5648
23 Multiple individuals from the same shoal were collected on 10 separate sampling occasions. On two of these occasions, similarity values could not be calculated because all stomachs were empty, and on a third they could not be calculated because one stomach was empty. When prey items were present, dominant prey items within individual stomachs were more similar between shoalmates than between non-shoalmates (Figure 2-8). For example, examined individuals (n=3) from one shoal had each consumed >95% sand dollar ( Mellita sp.), and all had a CH= 1.00 when compared to shoalmates. Average similarity values (CH) within a shoal ranged between 0.45 and 1.00, while average similarity values between individuals from separate shoals ranged from 0 to 0.24 (Figure 2-8). Figure 2-8. Average similarity values (Morisitas simplified similarity Index, CH) for individuals within shoals (closed circles) and between shoals (open circles). Error bars indicate standard deviation.
24 Discussion These data indicate that Rhinoptera bonasus do not behave as obligate hard prey specialists within Charlotte Harbor. Although morphologically suited for durophagy (Summers 2000, 2003), R. bonasus within Charlotte Harbor consumed a wide variety of benthic organisms, and appear to be opportunistic generalists that feed upon prey items that are readily available. In the present study, bivalves represented only 12.6% of the total IRI. These findings differ from previously published information that Rhinoptera species have a restricted diet consisting mainly of bivalves (James, 1962; Smith and Merriner, 1985; Blaylock, 1992). Hard prey items (i.e. sand dollar Mellita sp., oyster Crassostrea virginica ) were among identified prey types, confirming that R. bonasus can be durophagous in Charlotte Harbor, but these items did not dominate their diet. Diversity (H ) values represented the wide range of prey types identified across all R. bonasus stomachs examined (Table 2-2, Figure 2-4). Other dietary diversity values for Rhinoptera species are lacking; however, our values for R. bonasus are similar to those calculated for the bat ray Myliobatis californica (overall H = 0.73) (Gray et al., 1997). Evenness (E) values for both M. californica and R. bonasus were consistently low, indicating that the proportions of possible prey items within individual stomachs are not equally distributed (Krebs, 1999). Overall, R. bonasus stomachs contained a large array of prey types, but individual stomachs were typically dominated by a single prey item. At least 40% of all stomachs consisted of a single prey species (or other lowest possible taxonomic group), suggesting that R. bonasus are selectively foraging in areas of high prey density, and consequently encountering and consuming dense patches of various prey species. Rhinoptera bonasus selectively chose habitat patches of highest prey
25 density when feeding on bay scallops ( Argopecten irradians concentricus ) in North Carolina (Peterson et al., 2001). Similarly, foraging eagle rays Myliobatis tenuicaudatus in New Zealand dramatically increased their foraging behavior above a certain prey density threshold (Hines et al., 1997). Low evenness values for M. californica were suggested to be attributed to selective feeding and/or the relative abundances of prey items (Gray et al. 1997, p. 236). Stomach volume (SV) for R. bonasus generally increased with ray disc width, which has also been noted for R. bonasus in Chesapeake Bay (Blaylock, 1992). Similar to SV, prey size also showed a general increase with R. bonasus disc width. Neither of these results was particularly surprising, as larger animals are usually capable of consuming higher quantities and larger sizes of prey. For example, adult R. bonasus in the Chesapeake Bay region ate larger bivalves than juveniles (Smith and Merriner, 1985), and larger R. bonasus in North Carolina were capable of eating adult bay scallops while smaller, juvenile rays were not (Peterson et al., 2001). A similar pattern was noted for M. californica in Humboldt Bay with larger prey items being consumed as ray size increased (Gray et al., 1997). Although SV increased with ray size, R. bonasus stomach fullness (SF) did not show any relationship to disc width. This implies that although the absolute volume of food increases with size, the rays consume approximately the same relative amount of food resulting in similar fullness. However, fullness indices must be taken with caution as these values were qualitatively assessed. Dietary overlap was high between both size and sex groups of R. bonasus (CH Imm. vs.Mat. = 0 87; CH M vs.F = 0.84). Rhinoptera shoals usually form based on size (James,
26 1962; Blaylock, 1989; Smith and Merriner, 1987), although sexual segregation has also been noted (Smith and Merriner, 1987) Mixed-sex shoals are more often observed in Charlotte Harbor (Collins, unpublished data), so diets were expected to be similar between sexes of each size group. Previously published studies have found differences between adult and juvenile diets, as larger rays are more capable of consuming bigger and/or deeper buried prey (Smith and Merriner, 1985; Peterson et al., 2001). However, the high dietary overlap between mature and immature R. bonasus detected in this study was not surprising given the accessibility and small size of the predominant prey groups. Perhaps ray size does not play a major role in prey selection in Charlotte Harbor since there is such a wide array of easily captured benthic prey available (Estevez, 1986). Both immature and mature rays were found to consume bivalves, but they were less important in mature diets (17.0% for immature rays vs. 5.7 % for mature rays). Although it appears that mature R. bonasus do not feed on as many bivalves as immature rays, it is possible that prey manipulation could be the reason for this deficit. For example, adults may be more proficient at rejecting shell fragments (Smith and Merriner, 1985), resulting in consumption of only visceral masses which could be digested more rapidly and escape detection. Sasko (2000) documented winnowing behavior in feeding R. bonasus, demonstrating that they are capable of rejecting shell fragments and other debris. The majority of bivalves represented in both immature and mature stomachs were small and thin-shelled, indicating that these are the preferred bivalve prey types, but it is possible that larger, thicker shelled bivalves that necessitate fragmentation and shell rejection were not identifiable.
27 Seasonal dietary overlap was high for all season comparisons except those involving winter, which had low overlap values when compared to summer and spring samples, and medium overlap values compared to fall samples. Interestingly, R. bonasus captured during winter months also displayed the greatest dietary diversity (mean H = 1.10) (Table 5). Seasonal differences in diet could be explained by fluctuations in invertebrate fauna based on temperature, freshwater input and corresponding nutrient levels (Rubec et al., 1999; Arnold et al., 2000). Floridas rainy season typically occurs between June and October, and is characterized by lower salinities within estuaries (Estevez et al., 1984; Donaldson, 1985). The diversity of organisms within an estuary typically decreases as salinity decreases (Day et al., 1989). Estevez (1986) found that species richness of Charlotte Harbors benthic community was highest over winter months when rainfall is less, and lowest during the summer season. The differences detected in R. bonasus stomach contents between seasons may simply reflect the relative abundances of prey items. If R. bonasus is feeding within the estuary throughout the year, then adapting to changing dynamics of available prey would be a good strategy. By eating the most common or available organisms, the opportunistic foraging behavior of cownose rays is suited to handle changing prey densities and availability that may take place within estuaries. Examination of stomach contents demonstrated the importance of suction feeding for R. bonasus Pectinaria gouldii is a tube dwelling polychaete that builds fragile sand cones in the sediment (Uebelacker and Johnson, 1984), and is abundant within Charlotte Harbor (Donaldson, 1985; Estevez, 1986.). Intact, whole cones up to 30 mm long were found in many stomachs. The fact that crushed echinoderms and bivalves were
28 sometimes found in the same stomachs, in addition to complete P. gouldii sand tubes, suggests that R. bonasus are capable of distinguishing different prey types and alternating between suction only and a combination of suction and crushing. This has also been observed for another benthic feeding elasmobranch, the nurse shark ( Ginglymostoma cirratum ) (Mattot et al., in press). Prey type probably influences this as much as prey size. Hard organisms like sand dollars and thick shelled bivalves can be fractured not only to ease swallowing but also to increase digestion rate while thin shelled and softer prey (like A. papyrium cumaceans and polychaetes) can be swallowed whole. The thinshelled paper mussel A. papyrium was the predominant bivalve consumed by both immature and mature R. bonasus in Charlotte Harbor. Similar to Smith and Merriners (1985) findings, whole and fractured valves of A. papyrium were present within stomachs collected from Charlotte Harbor. However, there was little evidence of individuals preying on larger, thicker shelled mollusks. This could be due to preferential feeding on A. papyrium or reflect the likelihood of encountering this particular prey type. Pectinaria gouldii, the most dominant polychaete observed in their diets, are also common in Charlotte Harbor (Donaldson, 1985; Estevez, 1986), and although there are no reports detailing the relative abundance of cumaceans within the area, it is known that they do occur in Charlotte Harbor in high numbers (Anamari Boyes, pers. comm.). All species representing the three dominant prey groups are abundant benthic invertebrates common within estuaries of the Gulf of Mexico (Uebelacker and Johnson, 1984; Culter, 1986; Anamari Boyes and Jay Leverone, pers. comm.) and together comprised over 94% of the total numerical composition of prey identified in this study. It appears that R. bonasus are feeding upon the most abundant and commonly encountered prey items within Charlotte
29 Harbor and the fact that the dominant polychaete P. gouldii generally occurs in unvegetated habitat (Uebelacker and Johnson, 1984; J. Leverone, pers. comm.) indicates that cownose rays are not feeding exclusively in grass beds in Charlotte Harbor. The shoaling tendency of R. bonasus should increase their likelihood of encountering these widely distributed abundant prey sources. One of the benefits of shoaling is increased foraging efficiency (Pitcher and Parrish, 1993). Traveling in a group should allow for a larger search area and therefore increased prey detection capability. Shoals of R. bonasus have been observed feeding synchronously (Orth, 1975; Sasko, 2000), and the results of this study demonstrate that members of a shoal have more similar diets to each other than to members of another shoal. Blaylock (1993) calculated the daily ration for R. bonasus in the Chesapeake Bay to be ~3% of an individuals body weight. This would mean that daily intake should range between 150 270 grams for cownose rays between 5 and 9 kg. Cownose rays have also been documented to consume 1.5 liters of clams in one day (Schwartz, 1990). Matern et al. (2000) suggested that high energy output is required for the consumption of large bivalves and deep burrowing invertebrates by M. californica The benefit to consuming small items like epibenthic cumaceans and shallow infaunal polychaetes is that capture presumably takes less effort than deep bivalve excavation. It is notable that none of the dominant prey found in this study are deep infauna, suggesting that deep feeding pits are not necessary for R. bonasus in Charlotte Harbor. There may be benefits to consuming easily accessible, abundant small items vs. less abundant, larger items. Feeding on swarming or highly abundant species may reduce the amount of energy spent on prey capture.
30 It has been noted that foraging cownose rays may also play a beneficial, nondestructive role in structuring benthic communities by aerating benthic environments (Blaylock, 1993). Pit formation may extend the feeding activities of other organisms, such as filter feeding clams and worms (Cross and Cunan, 2000). Orth (1994) suggested that feeding cownose rays may actually enhance the dispersal of eelgrass by dislodging reproductive shoots. Cownose ray foraging can influence community structure by exposing normally obscured prey to other predators such as sharks, drum, jacks, cobia and menhaden, which associate with cownose ray schools (Rogers, 1990; Thrush et al., 1991).
31 Chapter 3: Movement Patterns and Residence Introduction Atlantic cownose rays ( Rhinoptera bonasus Mitchill 1815) are believed to be highly migratory, transient and seasonal residents of coastal and inshore waters. They exhibit shoaling and schooling behavior, and have at times been observed traveling in groups numbering thousands of individuals (Clark, 1963; Blaylock, 1989; Rogers et al., 1990). Although large shoals are often observed traveling in coastal and inshore waters (Clark, 1963; Blaylock, 1989, 1992; Smith and Merriner, 1985) little data are available concerning the residence and movement patterns of individual R. bonasus. Schwartz (1990) suggested that R. bonasus in the Gulf of Mexico make up a separate population from those in the western Atlantic, and that each population displays a unique migration pattern in response to seasonal changes in water temperature. The Atlantic population is believed to move north/south along the coast from New England to South America. This hypothesis is based on results from a tag-recapture study where cownose rays tagged in Chesapeake Bay were recaptured as far south as Venezuela and Brazil (Schwartz, 1965). The population in the Gulf of Mexico is hypothesized to travel clockwise from the Yucatan peninsula to Florida. Beyond the data of Schwartz (1965) there are no quantitative data available to support this hypothesis. Alternately, it has been suggested that R. bonasus may simply move offshore to warmer water during winter (Smith and Merriner, 1987) rather than partake in extensive fall and spring migrations. In either case,
32 cownose rays are believed to leave shallow estuaries and bays during the winter. Currently, there are no available data on movement patterns for cownose rays in the Gulf of Mexico. Large schools of foraging R. bonasus have been implicated in commercial shellfish predation as well as seagrass habitat destruction along the eastern coast of the United States (Orth, 1975; Merriner and Smith, 1979; Peterson et al., 2001). However, the lack of data regarding cownose ray movement patterns makes it difficult to define their behavior and residence within coastal estuaries. Determining and interpreting the movement patterns of cownose rays would be beneficial to understanding their life history, use of coastal habitats and role within the community. The most common method for monitoring marine animal movements is acoustic telemetry, which is becoming more widespread as technology advances and availability increases (Voegeli, 2001). Most previous elasmobranch marine telemetry studies consisted of manually following individuals fitted with transmitters and typically provided short term (<72 hours) or intermittent behavioral data (Sciarotta and Nelson, 1977; Standora and Nelson, 1977; Klimley and Nelson, 1984; Gruber et al., 1988). While manual telemetry studies are useful in defining fine scale habitat use, they do not provide the continuous data for individuals over extended periods (weeks months) that are necessary for a more complete understanding of residency and movement patterns. The advent of passive acoustic monitoring technology has allowed the movements of multiple animals to be continuously tracked over extended time frames (Klimley et al., 1988; Meyer et al., 2000; Vogeli, 2001; Heupel and Hueter, 2001; Heupel et al., 2004) allowing
33 predictions of long-term patterns, home ranges and response of individuals to environmental factors (Heupel et al., 2004). Although sharks were among the first subjects of marine telemetry studies (Nelson and Johnson, 1970; Sciarotta and Nelson, 1977; Klimley and Nelson, 1984; Nelson, 1990; Morrissey and Gruber, 1993), batoids (skates and rays) have been relatively overlooked. As with sharks, most tracking data that are available for rays consist of short-term or intermittent manual tracking data (Blaylock, 1992, 1993; Gilliam and Sullivan, 1993; Silliman and Gruber, 1999; Matern et al., 2000; Cartamil et al., 2003). Continuous, long-term tracking data are currently limited for ray species. The objectives of this study were to investigate R. bonasus residency, movement patterns and habitat use using passive acoustic telemetry to better understand the ecological significance of this species within an estuary on the southwest coast of Florida.
34 Methods Study area This research was conducted in Pine Island Sound, located in lower Charlotte Harbor (Figure 3-1). The shore of Pine Island Sound is largely undeveloped including state and federally protected areas (National Wildlife Refuge). The aquatic habitat within the sound varies from shallow sand and mud flats to seagrass beds and deep channels. Depths within the sound range from zero to ten meters with a tidal range of 0.7 to 1.8 meters. Pine Island Sound experiences significant freshwater input from the Caloosahatchee River that caused salinity to vary widely (15-37 ppt) over the course of the study. Temperature within the sound ranged between 20 33 C over the course of the study period (Heupel, unpublished data). Acoustic array An array of acoustic hydrophones (Vemco VR2 receivers) deployed within Pine Island Sound from April to December of 2003 and 2004 was used to track movements of R. bonasus within the study site. The main array system within Pine Island Sound (PIS) consisted of 40 hydrophones deployed in open water areas (Figure 1). The total monitoring area within PIS was approximately 184 km2. Data were downloaded from hydrophones every 2-3 weeks over the course of the study period.
35 Figure 3-1. Map of Pine Island Sound and its location along the Gulf coast of Florida. Hatched areas indicate seagrass beds and bathymetry contour indicates depths > 4 meters. Triangles designate hydrophone station locations. Study species Cownose rays were collected in 200-m (11 or 30 cm stretch mesh) or 400-m (11 cm stretch mesh) entanglement nets. Individuals were sexed, weighed to the nearest 0.1 kilogram, and measured to the nearest centimeter (straight and curved disc width and length). Rays were defined as mature (>70 cm SDW) or immature (<70 cm) based upon degree of clasper calcification in males or minimum size at pregnancy in females within
36 Charlotte Harbor (Collins, unpublished data). Sizes at maturity for R. bonasus within Charlotte Harbor differed from those established for cownose rays in Chesapeake Bay by Smith and Merriner (1986), who recognized maturity at >90 and >84 cm for females and males, respectively. This variation may be due to latitudinal or population differences. Sizes conservatively agreed with those established for R. bonasus in the northern Gulf of Mexico (approximately 65 cm disc width) (Neer and Thompson, 2005). Prior to release rays were fitted with acoustic transmitters measuring 8 x 28 mm (Vemco V8, Vemco Ltd.) that had an expected battery life of 250 days and operated on 69 kHz. Transmitters were coded to allow individual identification and were set to pulse randomly once every 90-180 seconds. Random repeat rates allowed multiple individuals to be monitored simultaneously within a given area without continuous signal overlap. Detection distances for V8 transmitters within Pine Island Sound were tested within the region and determined to be 450 m (average) with a maximum detection distance of 800 m (Collins unpublished data). Wax coated transmitters were externally attached to rays by a cinch tag (Floy Tags, Seattle, WA) inserted through the spiracular cartilage. All rays were released in good condition within 500 meters of their capture location. To determine effects of transmitter attachment and retention time, four R. bonasus were housed together in a circular saltwater tank (8,700 liters) at MML for 16 months. Three of the captive rays were fitted with dummy transmitters, equal in all physical aspects to field transmitters. One captive ray was not tagged and acted as a control. All four rays were monitored for changes in physical appearance or behavior. Ray swimming and feeding behaviors were qualitatively assessed daily, and each individual was photographed, weighed and measured monthly.
37 Data Analysis The residence of cownose rays within Pine Island Sound was examined by determining the number of days individuals were present within the study site. Individuals were considered present when a minimum of two detections were recorded for that individual within a single day. Daily presence data were analyzed to determine the number of consecutive days that an individual was resident (continuous presence), as well as the total number of days that it was detected within the study area. The number of days present (both continuous and total) within the study area was tested for differences based on sex or maturity stage using t -tests. The number of detections per hour was assessed for each tracked individual over its total monitoring period to define any diel differences in detection patterns. Each detection recorded for an individual was assigned to one of 24 bins based upon the hour of the detection. Assuming an equal distribution of detections over a 24 hour period, Chisquare tests were performed to determine whether the observed proportion of detections differed significantly from an expected even distribution. Significant differences from expected values would show individuals were more frequently detected at specific times of the day. This difference in detections could reflect diel changes in activity pattern or habitat use (e.g. movement into shallow regions where detection was difficult). To determine whether detection numbers were higher during day or nighttime hours, hourly detections were divided into day or night for each individual. Daytime fell between the hours of 06:00 and 18:59, and nighttime fell between 19:00 and 05:59. Day and night total detections were summed for each individual and compared using the Chi-square test
38 to determine whether an individual had significantly more detections during the daytime or nighttime hours. Similar to examination of detections per hour, the number of detections at various tidal stages (in 20 cm increments of tidal height) was examined to define any effects of tidal variation on individual detections. Tide data were obtained for Galt Island, Pine Island Sound, using the program Tides and Currents (MAKER). Chi-square tests were then used to compare the frequency of detections per tidal stage with the frequency of all tidal heights during the periods individuals were monitored. Statistically significant differences in the distribution of detections would indicate that an individual was detected more frequently at specific tidal heights. Detection data from R. bonasus were processed using a custom-written FORTRAN program (Simpfendorfer et al., 2002). This program used a mean position algorithm to calculate position estimates, or center of activity locations, for monitored individuals every thirty minutes. Simpfendorfer et al. (2002) compared real time location data from small sharks to mean position estimates and calculated an error of approximately 225 meters using this method. Processed data were used for all of the following analyses to define individual locations and movements through time. Calculated center of activity locations through time were used to calculate home range and activity space. Minimum convex polygons (MCP) and kernel utilization distributions (95% and 50% KUD) (Worton, 1989) were calculated for each individual using the Animal Movement Analyst Extension (AMAE) in ArcView 3.2 GIS (Hooge and Eichenlaub, 2000). Minimum convex polygons were used to demonstrate the extent of an animals range over a given period, while KUDs illustrated the utilization of space
39 within that range (Worton,1987). Changes in the extent of an individuals range over time were assessed by examining daily, 3-day, 7-day, 30-day and total MCPs. To detect changes in the utilization of the study site over time, KUDs were also examined over 3day, 7-day, 30-day and total monitoring periods. Daily KUDs were not calculated because most rays did not have enough detections over a single day to calculate an accurate utilization distribution. Unpaired t -tests were used to determine if there were sex or size differences in home ranges within each time period (e.g. 3, 7, 30 or all days) for both estimators. To determine whether the size of an animals range depended upon the number of days it was present, total MCPs and KUDs were compared to length of residency using linear regression. Activity space areas were also compared to straight disc width (SDW) using linear regression to determine the relationship between animal size and home range size. Daytime and nighttime KUDs were calculated for each examined time period to determine whether individuals exhibited diel changes in home range size or distribution. Paired ttests were used to detect differences in size of day vs. night activity spaces. Day and night KUDs were then compared between immature and mature R. bonasus using unpaired ttests to determine whether maturity had an effect on the size of day and night utilization distributions. To determine the extent of seagrass habitat utilization, total 50% KUDs (representing core areas of use over entire monitoring period) were overlaid on a map of the areas known seagrass habitat using ArcView. Total 50% KUDs were quantitatively examined for overlap by calculating the area of seagrass overlap and determining the percentage of the KUD that it represented.
40 The distance between consecutive centers of activity was calculated using standard geometric formulae. Distances between consecutive 30 minute centers of activity locations were used to examine linearity of movement, and define the net movement of an individual between detections. This analysis helped determine if individuals tended to remain within a confined location (small distances between consecutive locations) or moved large distances over time. Large distances between consecutive centers of activity would suggest rays were either moving very rapidly or were traveling in shallow water where detection was limited or impossible. Calculated distances were used to define the minimum distance traveled between consecutive centers of activity, and indicate whether rays were generally stationary or mobile over short periods. Linear regression was used to detect relationships between ray size and distance moved, and unpaired t -tests were used to detect size or sex differences in mean distances moved over time. Statistical tests were performed using Statistica (1999 edition) and Sigma Plot (version 9.0).
41 Results Effect of transmitter attachment No significant tagging effects were observed in three captive R. bonasus (2 mature, 1 immature) over 16 months in captivity. Transmitters were retained in captive rays for periods of 156, 187 and 457 days. Tagging sites displayed some necrosis (Figure 3-2) but transmitters were not dislodged and no negative effects on animal feeding or health were observed. Swimming and feeding behavior were considered normal based on field observations and in comparison to the control individual. All captive rays demonstrated growth and weight gain over 16 months (Appendix A). The control animal died after three months and was not included in growth analyses. Results from captive individuals suggest that wild caught individuals should retain transmitters for at least several months and not suffer any significant damage from transmitter attachment. Figure 3-2. Method of attachment and site for external transmitter. Pictures illustrate captive ray on day 1 (a) and day 300 (b) after transmitter attachment. a. b.
42 Sampling results Twenty one R. bonasus (15 male, 6 female) were monitored using passive acoustic telemetry between July 2003 and November 2004 within Pine Island Sound. Four R. bonasus were monitored in 2003 and 17 were monitored in 2004. Monitored R. bonasus ranged in size from 49 90 cm straight disc width (SDW) and 12 of the 21 rays were mature (8 male, 4 female). Of the 9 immature rays monitored, 7 were male and 2 were female. Residence time Residence time was highly variable among individuals and revealed no distinct seasonal pattern (Figure 3-3). Although some rays were resident for extended periods of time others only remained within the study site for brief periods. In 2003, all R. bonasus were tagged in July. One left the study area in July, two left in August, and one remained through October. In 2004, rays were captured in June, July, August and October, and left the study area in June, July, August, October and November. One ray tagged in July 2004 (no. 266) and another tagged in October (no. 512) were still in the estuary upon equipment removal from Pine Island Sound at the end of November 2004. Two R. bonasus fitted with transmitters in 2004 were within the study area for less than two days and were excluded from all further analyses.
43 Figure 3-3. Daily presence of Rhinoptera bonasus monitored within Pine Island Sound between June and November of 2003 and 2004. 2003 animals (no. 95 98) are listed at the top and separated by a gap in the Y-axis. Total monitoring periods ranged between one and 102 days (Figure 3-4). The mean total monitoring period for all R. bonasus over both years was 32 days. Total monitoring periods varied between years from 7 to 78 days in 2003 (mean=26) and 1 to 102 days in 2004 (mean=33). Females remained within the monitoring area for significantly longer periods than males (male mean = 27.7, female mean = 78.0, t -test, p=0.043). There were no significant differences between the total residence time of mature (mean = 50.7) and immature rays (mean = 30.6) ( t -test, p=0.398). Periods of continuous presence ranged from 1 to 34 days in 2003 (mean = 8) and 1 to 29 days in
44 2004 (mean = 3). However, most periods of continuous presence were between one and three days (Figure 3-4). Figure 3-4. Frequency histogram of residency times for Rhinoptera bonasus monitored in Pine Island Sound: (a) total monitoring period, and (b) periods of continuous presence.
45 Activity Patterns The variability in detection rates between individuals indicates that there were no consistent behavioral patterns that clearly demonstrate a pattern in the distribution of the numbers of hourly detections. Nineteen individuals (those present for more than two days) were tested for differences in the number of hourly detections. Observed presence by hour was significantly different from expected for 18 individuals (Chi-square test, p < 0.05; Table 3-1). Four juvenile rays that were tagged together in 2004 had very similar hourly detections, with substantially more detections from 06:00 to 12:00 than at other times of the day (Figure 3-5) The remaining 15 individuals did not reveal similar hourly detection patterns (Figure 3-5). Six individuals had higher detection numbers during the day, seven had higher numbers of nighttime detections, and seven showed no pattern (Table 3-1). Analysis of detections by tidal height showed no significant differences between observed and expected detection rates for any of the monitored R. bonasus (Table 3-1). This result also indicates that hydrophones were capable of recording equally well at all stages of the tide.
46 Table 3-1. Results for Chi2 analysis of the proportion of detections for R. bonasus within Pine Island Sound by hour, time of day (daytime vs. nighttime detections), and tidal height. Detections were expected to be equally distributed over all hours and tidal stages. P values < 0.05 indicate significant differences. (--) indicates there were not enough detections to test for differences. Hourly differences Day/Night differences Highest number of detections Tidal stage differences Ray ID p value p value Time of day p value 95 0.00000 0.00000 Night 0.19000 96 0.00000 0.00000 Night 0.87717 97 0.00000 0.60200 No pattern 0.94476 98 0.00004 0.12419 No pattern 0.99933 262 0.00000 0.01069 Night 0.90000 263 0.00000 0.16767 No pattern 0.78658 264 0.00000 0.00024 Night -265 0.00000 0.46894 No pattern 0.05020 266 0.00018 0.22490 No pattern 0.89325 268 0.00000 0.00000 Day 0.20587 269 0.00000 0.00000 Day 0.28716 270 0.00000 0.00000 Day 0.14818 271 0.05422 0.07343 No pattern 0.95257 273 0.00000 0.00000 Day 0.34372 508 0.00000 0.00017 Day 0.89053 509 0.00000 0.00000 Day 0.99908 511 0.00000 0.00001 Night 0.28688 512 0.00000 0.00000 Night 0.82930 514 0.00025 0.07184 No pattern 0.76041
47 Figure 3-5. Hourly detection patterns. Similar patterns in hourly detections were observed for four immature R. bonasus tagged together (a) while no distinct pattern in hourly detections emerged for the remaining 17 tracked rays (b). Home range analyses MCPs Daily MCPs for all rays ranged between 0.001 and 25.78 km2, with a mean of 5.25 km2. Most daily MCPs were less than 5 km2 for both mature and immature
48 individuals (Figure 3-6), and did not differ significantly between size groups (p = 0.7018). Mature rays displayed significantly larger 3-day MCP areas and significantly smaller 30-day MCP areas (Table 3-2) (Figure 3-7). There were no significant differences between mature and immature R. bonasus 7-day or total MCP areas (Table 3-2) (Figure 3-7). Although a consistent significant relationship with state of maturity was not detected, largest MCP sizes were observed in individuals greater than 60 cm SDW (Figure 3-8). MCP sizes were generally larger for mature animals over all examined time frames, except the 30-day period. This discrepancy was most likely due to the small number of animals that were present within the study area for 30 consecutive days and two large 30 day ranges for immature individuals. Figure 3-6. Frequency histogram representing the size of daily MCPs for immature and mature R. bonasus. No significant differences were detected between mature and immature daily MCPs ( ttest; p = 0.7018). Total MCPs for all R. bonasus ranged between 0.81 and 71.78 km2 (mean=22.01 km2). Total MCPs showed a significant increase with longer residence (r2 = 0.6315, slope = 0.2339, p = 0.0419). Although females remained within the study site for significantly
49 longer periods than males, no significant sex differences were detected in MCP size over any time frame (Table 3-3). Table 3-2. Ttest probability values ( p ) comparing activity space sizes between immature and mature R. bonasus over 3 day, 7day, 30 day and total monitoring periods. P values < 0.05 indicate significant differences. All significant values showed that mature individuals used larger areas than immature except for the 30 day MCP comparison (in italics). Monitoring period MCP p value KUD (95%) p value KUD (50%) p value 3 day 0.013115** 0.000373** 0.004505** 7 day 0.308216 0.047843* 0.016526* 30 day 0.002652** 0.451095 0.643910 Total 0.779804 0.002296** 0.000941** Figure 3-7. Mean MCP areas for immature (closed circles) and mature (open circles) R. bonasus over all monitoring periods (3 day, 7 day, 30 day and total).
50 Figure 3-8. MCP areas over 3 day (closed circles), 7 day (open circles), 30 day (triangles) and total (stars) monitoring periods compared to R. bonasus straight disc width (SDW) (cm). Dashed line divides immature (<70 cm) from mature (>70 cm) individuals. Table 3-3. T -test probability values ( p ) of t -tests comparing activity space sizes between male and female R. bonasus over 3 day, 7 day and total monitoring periods. P values < 0.05 indicate significant differences. Differences between 30 day monitoring periods could not be tested because only 1 female was present within the study area for 30 consecutive days. Monitoring period MCP p value KUD (95%) p value KUD (50%) p value 3 day 0.7288 0.2007 0.0602 7 day 0.2221 0.8014 0.0500 Total 0.1870 0.4097 0.0010**
51KUDs Mature R. bonasus had significantly larger KUDs (95 and 50%) than immature rays over all time frames (3-day, 7-day and total) except the 30-day period, which showed no significant difference (Figure 3-9; Table 3-2). Similar to MCP, largest KUDs occurred for individuals with disc widths greater than 60 cm (Figure 3-10). Total 95% KUDs ranged between 0.18 and 62.44 km2 (mean=22.63 km2), while total 50% KUDs were significantly smaller and ranged between 0.09 and 9.68 km2 (mean=3.33 km2) (Figure 3-11). Both 95% and 50% KUDs significantly increased with ray disc width (r2 = 0.84, p = 0.0046 and r2 = 0.88, p = 0.0002 for 95% and 50% KUDs, respectively) (Figure 3-11). Total KUD areas also increased with total monitoring periods (for 95% KUD, r2 = 0.67, slope = 0.152, p = 0.0084; for 50% KUD, r2 = 0.61, slope = 0.0281, p = 0.0266). No significant sex differences were detected for either the 95 or 50% KUD areas over the 3 day or 7 day time frame (Table 3-3). Differences between 30 day KUD areas could not be calculated because only one female R. bonasus was present for 30 consecutive days. No difference was detected between sexes for 95% KUDs or for the 3 day and 7 day 50% KUDs, but males had significantly larger total 50% KUD areas than females (Table 3-3). No significant differences were detected between the sizes of day and night KUDs for immature or mature rays over any time frame (Table 3-4). However, mature rays had significantly larger nighttime 95% KUDs than immature rays over all time frames except the 30 day period (Table 3-5). Nighttime 50% KUDs did not differ significantly between size classes.
52 Figure 3-9. Mean 95% (a) and 50% (b) KUD areas for immature (closed circles) and mature (open circles) R. bonasus over all monitoring periods (3 day, 7 day, 30 day and total).
53 Figure 3-10. 95% (a) and 50% (b) kernel utilization distribution (KUD) areas over 3 day (closed circles), 7 day (open circles), 30 day (closed triangles), and total (open triangles) monitoring periods vs. R. bonasus straight disc width (SDW) (cm). Dashed line divides immature (<70 cm) from mature (>70 cm) individuals.
54 Figure 3-11. Total 95% (closed circles) and 50% (open circles) KUD areas compared to R.. bonasus straight disc width (SDW) (cm). Both KUD areas get significantly larger as SDW increases: For 95% KUD, r2 = 0.84; p = 0.0046 and for 50% KUD, r2 = 0.88, p = 0.0002. Table 3-4 Ttest probability values ( p ) of t -tests comparing day and night KUD areas for immature and mature individuals. P values < 0.05 indicate significant differences. Monitoring Period KUD Immature p value Mature p value 3 day 95% 0.07108 0.63983 50% 0.11769 0.46212 7 day 95% 0.36495 0.95589 50% 0.41674 0.88990 30 day 95% 0.80050 0.50366 50% 0.55939 0.66154 Total 95% 0.73829 0.1806 50% 0.72024 0.5426
55Table 3-5. Ttest probability values ( p ) of t -tests comparing daytime and nighttime 95% and 50% KUDs between immature and mature animals. P -values < 0.05 indicate significant differences. 3 day 7 day 30 day Total Day (95%) 0.29891 0.42331 0.51579 0.21303 Day (50%) 0.23751 0.36200 0.90713 0.10499 Night (95%) 0.04160* 0.03315* 0.96272 0.00682** Night (50%) 0.14699 0.07037 0.82106 0.05058 Seagrass habitat overlap Examination of core areas (50% KUDs) showed variability in overlap with seagrass areas. Four rays had 50% KUDs that overlapped seagrass habitat by more than 10% (range: 12% 96%), six displayed an overlap of <5%, and nine exhibited 50% KUDs that did not overlap seagrass habitat at all (Figure 3-12). a. b.c. Figure 3-12. Examples of 50% KUD overlap of seagrass habitat. No overlap, (a), <5% overlap (b) and >10% overlap (c). Hatched areas represent seagrass beds and solid black circular shapes represent 50% KUD areas.
56 Distance between consecutive centers of activity Analysis of the distances between 30 minute centers of activity revealed that the majority of locations were within close proximity to consecutive locations. Movement distances ranged from 0 to 13 km, but mean distances per individual ranged between 0.1 and 0.7 km (Figure 3-13). Mature R. bonasus displayed significantly larger distances between consecutive centers of activity (0.50 km) than immature rays (0.26 km) ( t -test, p = 0.006), and distances generally increased with disc width (Figure 3-13; r2 = 0.92, slope = 0.02, p = 0.006). Movements over large distances were not common, with the majority of rays traveling between 0 2 km within a 30 minute period (Figure 3-14). Figure 3-13. Mean distance moved over 30-minute intervals vs. ray straight disc width (SDW) (cm).
57 Figure 3-14. Frequency histogram of mean distance moved over 30-minute intervals for immature (black bars) and mature (gray bars) R. bonasus.
58 Discussion Acoustic monitoring of R. bonasus within Pine Island Sound revealed that the residence and movement patterns of this species are complex and variable. Rays were detected within the estuary for varying lengths of time, but detection of individuals was usually not contiguous. Continuous presence data showed that most animals were not monitored for more than two consecutive days at a time. This suggests intermittent and transient use of the monitored area. Loss of contact with individuals could be the result of movement out of the study area or movement into shallow water regions where detection was improbable. Utilization of intertidal and subtidal shallow flats has been well documented for many ray species (Smith and Merriner, 1985; Snelson et al., 1988; Matern et al., 2000), so it is possible that individuals were not exiting the study site but simply moving into shallow water and out of detection range. Although data from most individuals suggested transient use of the study site, some remained within the study area for extended periods. Seven of 21 tracked rays were monitored for at least one month, and three rays were detected within the array for over three months. This demonstrates that some R. bonasus do remain resident within estuaries for extended lengths of time. However, the majority of R. bonasus fitted with transmitters were monitored for shorter periods, leaving the study area after days or weeks and were not detected again. These data suggest that there is no consistent pattern of residency among the individuals monitored in this study. Although there are no other continuous batoid tracking studies available for comparison, these results are similar to those from other studies of benthic feeding elasmobranchs. Bat rays ( Myliobatis californica ) in Tomales Bay, California, have been
59 documented within 1 km of their initial tagging site after periods of 106, 359, and 583 days during a tag-recapture study (Hopkins and Cech, 2003). However, recapture data were gathered from only three out of 257 tagged bat rays. Based on the low number of individuals recaptured during this study, the results presented by Hopkins and Cech (2003) are similar to those collected here where a small number of individuals were resident in the area while the remainder moved out of the region. Similarly, in a manual tracking study of the Pacific angel shark ( Squatina californica ), two out of nine tagged sharks were relocated after three days in the same home area that they were tagged (Standora and Nelson, 1977). Both of these studies suggest that individual variability in residence is not unusual among elasmobranchs. As with residency patterns, there was no distinct pattern in use of the study site when examining activity pattern based on tide stage, which contrasts with existing research on cownose rays. Blaylock (1992, 1993) attached satellite tags to six R. bonasus in Chesapeake Bay and manually tracked individuals for 4.3 13.5 hours and found that all tracked individuals moved in the same general direction as tidal flow. Smith and Merriner (1985, 1987) reported that cownose rays in Chesapeake Bay were moving into intertidal zones to feed during high tide. However, if this were occurring in Pine Island Sound, the data would have reflected an absence of detections during high tide because these shallow areas are not within detection range of the hydrophone array. Similar to the present study, Matern et al. (2000) did not observe any tidal correlations while tracking bat rays in Tomales Bay, but they did notice a distinct diel pattern that they attributed to behavioral thermoregulation.
60 Although there was not a distinct diel activity pattern detected for cownose rays in Pine Island Sound as reported for the bat ray in Tomales Bay (Matern at al. 2000), there was a general trend for lower detection rates for R. bonasus between 12:00 18:00 over both years of this study, which could not be attributed to tidal changes. This behavior may be explained by rays moving into canal systems or onto shallow seagrass areas and thus out of detection range during this period. Movement into these areas could be for feeding or other behaviors. Movement into shallow water could also serve as a means of predator avoidance by using habitats that large sharks do not enter as frequently. In similar studies, juvenile lemon sharks ( Negaprion brevirostris ) in the Bahamas were documented to select shallower, warmer water (Morrissey and Gruber, 1993) and hammerhead ( Sphyrna lewini ) pups in Hawaii aggregated during daylight hours in more turbid parts of Kaneohe Bay (Holland et al., 1993). Examination of home range sizes also showed that movement patterns and habitat use were not consistent across individuals. Minimum convex polygons can overestimate home range size (Kernohan et al., 2001), but are useful to illustrate boundaries and the extent to which an animal travels. Daily MCP areas between individuals varied greatly over the course of this study (<1 26 km2), but the majority were between 1 and 5 km2 suggesting individuals generally used a small portion of the study site within the course of a single day. This is comparable to activity space estimates for bonnethead sharks ( Sphyrna tiburo ), also tracked within Pine Island Sound, which had a mean daily activity space of 8.31 km2 (Heupel et al., manuscript). Like the bonnethead shark, R. bonasus are capable of traversing fairly large distances over the course of a 24 hour period, but tended to remain within a confined area over most days. Total MCP size significantly increased
61 with monitoring period, suggesting that individuals expand the areas utilized over time. The maximum total MCP recorded was 71.78 km2 which falls within the range of MCPs recorded for other coastal elasmobranchs: lemon sharks displayed activity spaces ranging from 9 93 km2 over intermittent tacking periods of 1-8 days (Gruber et al., 1988). Juvenile sandbar sharks ( C. plumbeus ) were tracked for periods between 2.5 and 70 hours, with resulting MCPs ranging between 1.1 and 333.9 km2 (Rechisky and Wetherbee, 2003). Mean MCP areas tended to increase with animal size, but significant differences based on maturity were only detected over the 3 day and 30 day time frames. Mature R. bonasus had significantly larger MCPs over 3 day periods and significantly smaller MCPs over 30-day periods. This result is confounding because of the correlation of MCP area with disc width. The reversal of trend for the 30 day data is likely the result of a small sample size and the large home ranges calculated for two juvenile individuals over the 30 day period. These two individuals, although nearing maturity, had much larger 30 day home ranges than any other immature or mature individual. Although female rays remained within the study area for significantly longer periods than males, there were no significant differences for MCP areas between sexes The KUD is more descriptive than the MCP, illustrating the use of space within a home range and differentiating areas of intense use from those that are only briefly occupied (Worton, 1987). Based on KUD analysis R. bonasus demonstrated 95% usage areas between 0.18 and 62.44 km2 over monitoring periods of 1 -102 days. These values are generally larger than those obtained for the more stationary, benthic myliobatiform Dasyatis lata, which exhibited 95% KUDs between 0.62 and 2.77 km2 during manual
62 tracking over 31-74 hour periods (Cartamil et al., 2003). In contrast, the values observed for R. bonasus were much smaller than juvenile sandbar sharks, which exhibited 2.8 to 315.4 km2 95% KUDs over manual tracking periods up to 70 hours (Rechisky and Wetherbee, 2003). As pelagic swimmers, myliobatid rays like R. bonasus are typically more mobile than the predominantly sedentary dasyatid rays (Bigelow and Schroeder, 1953). However, R. bonasus do spend time foraging for benthic prey in one location (Smith and Merriner, 1985; Sasko, 2000), and can rest motionless on the bottom for long periods (Collins, personal observation). This behavior may produce smaller home range sizes than for a species that is continuously pelagic such as the sandbar shark ( C. plumbeus ). Although 95% KUDs tended to expand over time, 50% KUDs appeared to be more stable. This suggested that individuals maintained a consistent small area (3.33 km2) that was used repeatedly, indicating individuals concentrated large amounts of time within these restricted core areas. Rhinoptera bonasus monitored during this study were active diurnally and nocturnally, as daytime and nighttime activity spaces were not significantly different for any individual. This differs from patterns observed in other benthic feeding elasmobranchs. Dasyatis lata showed significantly larger activity spaces during the night than during the day (Cartamil et al., 2003). Similarly, both the Pacific electric ray ( Torpedo californica ) and the Pacific angel shark have been documented as nocturnal, covering much larger distances between dusk and dawn than during daylight hours (Bray and Hixon, 1978; Standora and Nelson, 1977). Both the horn shark ( Heterodontus francisci ) and the swell shark ( Cephaloscyllium ventriosum ) displayed distinct diel
63 patterns, increasing activity levels at night (Nelson and Johnson, 1970). Mature rays had significantly larger nightime 95% KUDs than immature rays over all time periods, but total daytime 95% KUD comparison showed no difference between size groups. This indicates that mature animals are utilizing larger areas within their range during the nighttime hours, while immature animals remain within more confined spaces This behavior could be attributed to increased predation risk for smaller individuals. The increased nighttime activity space for larger individuals contributed to mature R. bonasus having larger home range sizes overall. The fact that mature R. bonasus generally used larger areas than immature rays may be explained by their increased swimming capability or by increased energetic demands and consequently larger foraging areas. Foraging area has been described as inversely related to food availability (Hamilton and Watt, 1970); thus the increased dietary requirements of larger animals could result in larger activity spaces. The difference may also be attributed to differential predation on smaller R. bonasus which may influence the distribution of immature animals as has been suggested for juvenile lemon sharks in Bimini, Bahamas (Morrissey and Gruber, 1993). Larger animals have often been reported to cover greater distances in given amounts of time than their smaller counterparts: lemon sharks utilize greater spaces as they grow (Sundstrom et al., 2001), longer leopard sharks ( Triakis semifasciata ) generally move faster than shorter ones (Ackerman et al., 2000) and bonnethead sharks ( Sphyrna tiburo ) have higher rates of movement as mature animals (Parsons, 1990). In the present study R. bonasus moved significantly larger distances over 30 minute intervals as disc width increased, and mature rays moved mean distances almost double those of immature rays (0.26 vs. 0.50 km.).
64 Examination of home range areas revealed that all rays had total MCP and 95% KUDs that overlapped seagrass habitat. However, only four rays had more than 10% of their core areas (50% KUD) that overlapped with seagrass habitat. Such an observation contrasted with existing evidence that R. bonasus utilize grass beds extensively (Orth, 1975; Peterson, 2001). The predominant use of sand and mud areas is supported by feeding data that show that the majority of R. bonasus prey in Pine Island Sound is found in unvegetated areas (Collins et al., manuscript). However, these results could be misleading and should be considered cautiously since there are regions of seagrass habitat within Pine Island Sound that are beyond detection range of the hydrophone array. Rays could be using this habitat during periods when detections are low or absent. In this study, R. bonasus was present within Pine Island Sound throughout the summer months for the duration of the project and some individuals appear to have remained beyond the end of the study period. These results demonstrate that individual R. bonasus can have prolonged residency time in estuarine systems. Although some rays were only present for brief periods, there was not a distinct seasonal departure of the tagged rays as would be expected if they were undergoing seasonal migration. Rays were still present in Pine Island Sound when the hydrophone equipment was removed at the end of November 2004, and rays were detected in another portion of the estuary in January, 2005, suggesting they were not migrating out of the area for the winter even though water temperatures had declined to 16 C (Heupel, unpublished data). In addition, R. bonasus were tracked in the Caloosahatchee River (adjacent to Pine Island Sound) January May of 2004 and 2005 as part of another project (Collins, unpublished data), and eighteen R. bonasus were also captured in north Charlotte Harbor between November
65 2003 and March 2004 (Collins et al., manuscript). Along the eastern coast of the United States, seasonal fall departure and summer arrival of R. bonasus has been reported (Schwartz, 1965, 1990; Smith and Merriner, 1985, 1986; Blaylock, 1993), and it was assumed that cownose rays within Charlotte Harbor would follow a similar seasonal pattern. The current data do not support this hypothesis. Hopkins and Cech (2003) noted that the departure of bat rays from coastal California during winter months was not as pronounced in more southern counties where water temperatures were warmer. Neer et al. (2005, manuscript) reported encountering cownose rays in the northeast Gulf of Mexico at temperatures between 15.5 33.6 C. Based on these results it seems possible that temperatures in Charlotte Harbor do not get low enough to trigger seasonal migrations, and that individual movements may be related to other factors (e.g. prey availability, mating or predator avoidance). This concept is supported by reports of a resident population of cownose rays in Venezuela (Smith and Merriner, 1987) and year-round presence of R. bonasus in coastal North Carolina (Smith, 1907). It is possible that R. bonasus inhabiting the southwest coast of Florida, where water temperatures rarely drop below 15 C, simply move offshore to warmer, deeper water when necessary and can easily return to inland waters when temperatures increase. Although migratory movements, and movements outside the study area, could not be identified in the present study, data are available to show the extent of an individuals movement during the course of a day via home range estimates and consecutive movement locations. The distances between centers of activity of tracked cownose rays in Pine Island Sound ranged between 0 and 13 km per 30 minute period, and daily MCPs ranged between 0.001 25.78 km2. The average traveling speed for R. bonasus has been
66 estimated between 3.7 5.6 km/hour (Blaylock, 1992) and 7.4 9.3 km/hour (Smith, 1980). Smith (1980) reported R. bonasus in the South Atlantic Bight capable of traveling 14.6 23.2 km in one day. As highly mobile, pelagic swimmers, R. bonasus have the capability to traverse large distances over relatively short lengths of time, but the current data show that they also spent long periods within a small area. The extended presence of R. bonasus in Pine Island Sound suggests that the estuary provides adequate resources to support their dietary and home range requirements.
67 Chapter 4: General Conclusions All previous reports regarding the diet and movement patterns of R. bonasus have indicated that this species is a highly migratory, transient ray that feeds almost exclusively upon bivalve prey (Schwartz, 1965, 1990; Smith and Merriner, 1985, 1986, 1987). Quantative data were lacking regarding these aspects of R. bonasus ecology within the Gulf of Mexcio. Contrasting with evidence from previous studies, Rhinoptera bonasus in Charlotte Harbor, Florida do not appear to be exclusive hard prey specialists. Bivalves represented only 12.6% of the total IRI, contrasting with existing evidence that cownose rays are stenophagous in their choice of molluscan prey. Primary prey, in decreasing order, consisted of cumaceans, the polychaete P. gouldii and the bivalve A. papyrium which are all common benthic invertebrates within Charlotte Harbor The results of this study suggest that R. bonasus are opportunistic predators that will exploit abundant prey types within their foraging area. The high dietary overlap between immature and mature R. bonasus, as well as between sexes and seasons, indicate that the same prey types remain important to all groups of R. bonasus using the estuary. Shoal mates had significantly more similar diets than non-shoal mates, supporting group feeding hypotheses and suggesting school fidelity, at least over short periods. Most of the consumed prey were not deep infauna, indicating that deep feeding pits are not necessary for foraging in this area.
68 Tracking data have shown that activity patterns of R. bonasus within Charlotte Harbor are not significantly related to time of day or tidal stage. Tagged rays did not display a distinct seasonal departure, and tracking data provide no evidence to support theories of a massive R. bonasus migration in the Gulf of Mexico. Movement patterns do not appear to be predictable for this population as a whole, but indicate that Pine Island Sound provides a suitable environment for R. bonasus to remain resident for extended periods while also serving as a transient stopping ground for numerous individuals. As highly mobile pelagic swimmers, R. bonasus are capable of traveling large distances over short time periods, but they can also spend long periods within relatively small home ranges. These results provide new insight into the feeding and movement dynamics of a population of rays along the Gulf coast of Florida. Bivalves were not the most important prey, and most organisms consumed were epifauna or shallow infauna, suggesting that cownose rays are unlikely to be responsible for commercial shellfish decline or massive modification of seagrass beds in this region. Although tracking data were produced variable results, there is no evidence to suggest a massive migration or seasonal exodus for R. bonasus utilizing Charlotte Harbor.
69 References Ackerman, J.T., Kondratieff, M.C., Matern, S.A., and Cech, J.J., Jr. (2000) Tidal influence on spatial dynamics of leopard sharks, Triakis semifasciata in Tomales Bay, California. Environmental Biology of Fishes 58, 33-43. Arendt, M.D., Olney, J.E., and Lucy, J.A. (2001). Stomach content analysis of cobia, Rachycentron canadum from lower Chesapeake Bay. Fishery Bulletin 99, 665-660. Bigelow, H.B. and Schroeder, W.C. (1953). Fishes of the Western North Atlantic. Part Two: Sawfishes, Guitarfishes, Skates, Rays and Chimaeroids. (Sears Foundation for Marine Research, Yale University: New Haven, Connecticut.) Bizzarro, J.J., Smith, W.D., Mrquez-Faras, J.F., and Hueter, R.E. (2005). Reproductive biology and artisinal fisheries for the golden cownose ray ( Rhinoptera steindachneri ) in the northern Mexican Pacific. Environmental Biology of Fishes. In press. Blaylock, R.A. (1989). A massive school of cownose rays, Rhinoptera bonasus (Rhinopteridae), in lower Chesapeake Bay, Virginia. Copeia 3, 744-748.
70 Blaylock, R.A. (1992). Distribution, abundance and behavior of the cownose ray, Rhinoptera bonasus (Mitchill 1815), in lower Chesapeake Bay. PhD Dissertation, College of William and Mary, Virginia, USA. Blaylock, R.A. (1993). Distribution and abundance of the cownose ray, Rhinoptera bonasus in lower Chesapeake Bay. Estuaries 16, 255-263. Bray, R.N., and Hixon, M.A. (1978). Night-shocker: Predatory behavior of the Pacific electric ray ( Torpedo californica ). Science 200,.333-334. Cartamil, D.P., Vaudo, J.J., Lowe, C.G., Wetherbee, B.M., and Holland, K.M. (2003). Diel movement patterns of the Hawaiian stingray, Dasyatis lata : implications for ecological interactions between sympatric elasmobranch species. Marine Biology 142, 841-847. Choat, J.H. (1982). Fish feeding and the structure of benthic communities in temperate waters. Annual Review of Ecology and Systematics 13, 423-449. Clark, E. (1963). Massive aggregations of large rays and sharks in and near Sarasota, Florida. Zoologica 48, 61-64.
71 Corts, E. (1997). A critical review of methods of studying fish feeding based on analysis of stomach contents: application to elasmobranch fishes. Canadian Journal of Fisheries and Aquatic Sciences 54, 726-738. Cross, R.E. and Curran, M.C.. (2000). Effects of feeding pit formation by rays on an intertidal meiobenthic community. Estuarine, Coastal and Shelf Science 51, 293-298. Culter, J.K. (1986). Manual for identification of marine invertebrates: A guide to some common macroinvertebrates of the Big Bend region, Tampa Bay, Florida. United States Environmental Protection Agency, Contract # CA-8-3159-A. Day, J.W. Jr., Hall, C.A.S., Kemp, W.M., and Yanez-Arancibia, A. (1989) Estuarine Ecology. John Wiley and Sons, Inc. New York.558p. Donaldson, J. (1985). A small scale benthic study of the Myaka River estuary. B.S. thesis. New College of the University of South Florida. Tampa, Florida. Dulvy, N.K., Metcalfe, D.J., Glanville, J., Pawson, G.M., and Reynolds, D.J. (2000). Fishery stability, local extinctions and shifts in community structure of skates. Conservation Biology 14, 283-293. Estevez, E.D. (1981). Charlotte Harbor estuarine ecosystem complex, a summary of scientific information. Mote Marine Laboratory Review Series 3, 1077 p.
72 Estevez, E.D. (1986). Infaunal macroinvertebrates of the Charlotte Harbor estuarine system and surrounding inshore waters, Florida. United States Geological Survey, Water Resources Investigations Report No. 85-4260, Tallahassee, Florida. Gelsleichter, J., Musick, J.A., and Nichols, S. (1999). Food habits of the smooth dogfish, Mustelus canis dusky shark, Carcharhinus obscurus Atlantic sharpnose shark, Rhizoprionodon terraenovae and the sand tiger, Carcharias Taurus from the northwest Atlantic Ocean. Environmental Biology of Fishes 54, 205-217. Gilliam, D.S., and Sullivan, K.M. (1993). Diet and feeding habits of the southern stingray, Dasyatis americana in the central Bahamas. Bulletin of Marine Science 52, 1007-1013. Grant, J. (1983). The relative magnitude of biological and physical sediment reworking in an intertidal community. Journal of Marine Research 41, 673-689. Gray, A.E., Mulligan, T.J. and Hannah, R.W. (1997). Food habits, occurrence and population structure of the bat ray, Myliobatis californica in Humboldt Bay, California. Environmental Biology of Fishes. 49, 227-238.
73 Gruber S.H., Nelson, D.R., and Morrissey, J.F. (1988). Patterns of activity and space utilization of lemon sharks, Negaprion brevirostris in a shallow Bahamian lagoon. Bulletin of Marine Science 43, 61-76. Hamilton, W.J., III, and Watt, K.E.F. (1970). Refuging. Annual Review of Ecology and Systematics 1, 263-286. Heupel, M.R. and Hueter, R.E. (2001). Use of a remote acoustic telemetry system to monitor shark movements in a coastal nursery area. In: J.R. Sibert and J.L. Nielsen (eds.) Electronic Tagging and Tracking in Marine Fisheries. Kluwer Academic Publishers, Netherlands, pp 217-236. Heupel, M.R., Simpfendorfer, C.A. and Hueter, R.E. (2004). Estimation of shark home ranges using passive monitoring techniques. Environmental Biology of Fishes 71, 135142. Heupel, M.R., Simpfendorfer, C.A., Barker-Collins, A., and Tyminski, J.P. (in review). Residency and movement patterns of bonnethead sharks Sphyrna tiburo in a large Florida estuary. Hines, A.H., Whitlatch, R.B., Thrush, S.F., Hewitt, J.E., Cummings, V.J., Dayton, P.K., and Legendre, P. (1997). Nonlinear foraging response of a large marine predator to
74 benthic prey: eagle ray pits and bivalves in a New Zealand sandflat. Journal of Experimental Marine Biology and Ecology. 216, 191-210. Hoese, H.D., and Moore, R.H. (1977). Fishes of the Gulf of Mexico. (Texas A&M University Press: College Station, TX, USA.) Horn, H.S. (1966). Measurement of overlap in comparative ecological studies. American Naturalist 100, 419-424. Holland, K.N., Wetherbee, B.M., Peterson, J.D. and Lowe, C.G. (1993). Movements and distribution of hammerhead shark pups on their natal grounds. Copeia 1993 (2), 495-502. Hooge, P.N., and Eichenlaub, W.M. (2000). Animal movements extension to ArcView. Alaska Biological Center, U.S. Geological Survey, Anchorage. Hopkins, T.E., and Cech, J.J., Jr. (2003). The influence of environmental variables on the distribution and abundance of three elasmobanchs in Tomales Bay, California. Environmental Biology of Fishes 66, 279-291. Hovel, K.A., and Lipcius, R.N. (2001). Habitat fragmentation in a seagrass landscape: patch size and complexity control blue crab survival. Ecology 82, 1814-1829.
75 Hyslop, E.J. (1980). Stomach content analysis-a review of methods and their application. Journal of Fisheries Biology 17, 411-429. James, P.S.B.R. (1962). Observations on shoals of the Javanese cownose ray Rhinoptera javanica Mller and Henle from the Gulf of Mannar, with additional notes on the species. Journal of the Marine Biological Association of India 4, 217-223. Kernohan, B.J., Gitzen, R.A., and Millspaugh, J.J. Analysis of Animal Space and Movements. In Radio Tracking and Animal Movements. (Eds. Millsapugh, J.J and Marzluff, J.M.) pp. 126-168. (Academic Press, CA.) Klimley, A.P., Butler, S.B., Nelson, D.R., and Stull, A.T. (1988). Diel movements of scalloped hammerhead sharks, Sphyrna lewinii Griffith and Smith, to and from a seamount in the Gulf of California. Journal of Fish Biology 33, 751-761. Klimley, A.P. and Nelson, D.R. (1984). Diel movement patterns of the scalloped hammerhead shark ( Sphyrna lewini ) in relation to El Bajo Espiritu Santo: a refuging central-position social system. Behavioral Ecology and Sociobiology 15, 45-54. Kraeuter, J.N., and Castagna, M. (1980). Effects of large predators on the field culture of the hard clam Mercenaria mercenaria. Fishery Bulletin 78, 538-540.
76 Krebs, C.J. (1999). Ecological Methodology. 2nd Edition. (Addison-Wesley: Menlo Park, California.) Kyne, P.M., and Bennett, M.B. (2002). Diet of the eastern shovelnose ray, Aptychotrema rostrata (Shaw & Nodder, 1974), from Moreton Bay, Queensland, Australia. Marine and Freshwater Research 53, 679-686. Livingston, R.J. (1984). Trophic response of fishes to habitat variability in coastal seagrass systems. Ecology 65, 1258-1275. Mrquez-Faras, F.J. (2002). The artisinal ray fishery in the Gulf of California: development, fisheries research, and management issues. IUCN Shark Specialist Group. Shark News 14, 1-5. Matern, S.A., Cech, J.J., Jr., and Hopkins, T.E. (2000). Diel movements of bat rays, Myliobatis californica in Tomales Bay, California: evidence for behavioral thermoregulation? Environmental Biology of Fishes 58, 173-182. Mattot, M.P., Motta, P.J. and Hueter, R.E. (2005). Modulation in feeding kinematics and motor patterns of the nurse shark Ginglymostoma cirratum. Environmental Biology of Fishes. (in press).
77 Merriner, J.V., and Smith, J.W. (1979). A report to the oyster industry of Virginia on the biology and management of the cownose ray ( Rhinoptera bonasus, Mitchill) in lower Chesapeake Bay. Virgina Institute of Marine Science, College of William and Mary, Special Report in Applied Marine Science and Ocean Engineering 216. 33 p. Meyer, C.G., Holland, K.N., Wetherbee, B.M. and Lowe, C.G. (2000). Movement patterns, habitat utilization, home range size and site fidelity of whitesaddle goatfish, Paripeneus porphyreus in a marine reserve. Environmental Biology of Fishes 59, 235242. Morrissey, J.F. and Gruber, S.H. (1993). Home range of juvenile lemon sharks, Negaprion brevirostris Copeia 1974, 425-434. Morte, M.S., Redn, M.J., and Sanz-Brau, A. (2002). Diet of Phycis blennoides (Gadidae) in relation to fish size and season in the western Mediterranean (Spain). Marine Ecology 23, 141-155. Musick, J.A. (ed.) (1999). Life in the Slow Lane: Ecology and Conservation of Longlived Marine Animals. American Fisheries Society Symposium 23, 260 pp. Neer, J. A., Carlson, J.K., and Thompson, B.A. (2005). Oxygen consumption of the cownose ray, Rhinoptera bonasus (Mitchill 1815): the effect of temperature and salinity. Submitted.
78 Neer, J.A., and Thompson, B.A. (2005). Life history of the cownose ray, Rhinoptera bonasus in the northern Gulf of Mexico, with comments on geographic variability in life history traits. Environmental Biology of Fishes 73, 321-331. Nelson, D.R. (1990). Telemetry studies of sharks: A review, with applications in resource management. In: Pratt, H.L., Jr., Gruber, S.H., and Taniuchi, T. (eds.) Elasmobranchs as Living Resources: Advances in the Biology, Ecology, Systematics and the Status of Fisheries. NOAA Technical Report NMFS 90, pp. 239-256. Nelson, D.R. and Johnson, R.H. (1970). Diel activity rhythms in the nocturnal, bottomdwelling sharks, Heterodontus francisci and Cephaloscyllium ventriosum Copeia 1970 (4), 732739. Orth, R.J. (1975). Destruction of eelgrass, Zostera marina by the cownose ray, Rhinoptera bonasus in the Chesapeake Bay. Chesapeake Science 16, 205-208. Parsons, G.R. (1990). Metabolism and swimming efficiency of the bonnethead shark, Sphyrna tiburo. Marine Biology 104, 363-367. Peterson, C.H., Fodrie, F.J., Summerson, H.C. and Powers, S.P. (2001). Site-specific and density-dependent extinction of prey by schooling rays: generation of a population sink in top-quality habitat for bay scallops. Oecologia 129, 349-356.
79 Pierce, R.H., Brown, R.C., Van Vleet, E.S., and Joyce, R.M. (1986). Hydrocarbon contamination from coastal development. Organic Marine Geochemistry 305, 229-246. Pierce, R.H., Wetzel, D.L., and Estevez, E.D. (2004). Charlotte Harbor initiative: assessing the ecological health of southwest Floridas Charlotte Harbor estuary. Ecotoxicology 13, 275-284. Pinkas, L.M., Oliphant, S., and Iverson, I.L.K. (1971). Food habits of albacore, bluefin tuna and bonito in Californian waters. California Fish and Game. 152, 1-105. Pitcher, T.J. and Parrish, J.K. (1993). Functions of shoaling behavior in teleosts. In Behavior of Teleost Fishes. (Ed. Pitcher, T.J.) pp 363-439. (Chapman and Hall: London.) Rechisky, E.L., and Wetherbee, B.M. (2003). Short-term movements of juvenile sandbar sharks, Carcharhinus plumbeus on their nursery grounds in Delaware Bay. Environmental Biology of Fishes 68, 113-128. Rogers, C., Roden, C., Lohoefener, R., Mullin, K., and Hoggard, W. (1990). Behavior, distribution, and relative abundance of cownose ray schools Rhinoptera bonasus in the northern Gulf of Mexico. Northeast Gulf Science 11, 69-76.
80 Sasko, D.E. (2000). The prey capture behavior of the Atlantic cownose ray, Rhinoptera bonasus. MS Thesis, University of South Florida, Florida, USA. Schwartz, F.J. (1964). Effects of winter water conditions on fifteen species of captive marine fishes. American Midland Naturalist 71 (2), 434-444. Schwartz, F.J. (1965). Inter-American migrations and systematics of the western Atlantic cownose ray, Rhinoptera bonasus In: Association of Island Marine Laboratories of the Caribbean, Sixth Meeting. p.1. (Isla Margarita, Venezuela.) Schwartz, F.J. (1990). Mass migratory congregations and movements of several species of cownose rays, genus Rhinoptera: A world-wide review. The Journal of the Elisha Mitchell Scientific Society 106, 10-13. Sciarotta, T.C. and Nelson, D.R. (1977). Diel behavior of the blue shark Prionace glauca near Santa Catalina Island. California Fishery Bulletin 75, 519-528. Silberhorn, G.M., Dewing, S., and Mason, P.A. (1996). Production of reproductive shoots, vegetative shoots, and seeds in populations of Ruppia maritima L. from the Chesapeake Bay, Virginia. Wetlands 16, 232-239. Silliman, W., and Gruber, S.H. (1999). Behavioral biology of the spotted eagle ray Aetobatus narnari. Bahamas Journal of Science 7, 13-20.
81 Simpfendorfer, C.A., Heupel, M.R. and Hueter, R.E. (2002). Estimation of short-term centers of activity from an array of omnidirectional hydrophones and its use in studying animal movements. Canadian Journal of Fisheries and Aquatic Sciences 59, 23-32. Smith, H.M. (1907). The Fishes of North Carolina. North Carolina Geological and. Economic Survey 2. 47p. Smith, J.W. (1980). The life history of the cownose ray, Rhinoptera bonasus (Mitchill 1815), in lower Chesapeake Bay, with notes on the management of this species. M.A. thesis. School of Marine Science, College of William and Mary. Williamsburg, VA. 150p. Smith, J.W., and Merriner, J.V. (1985). Food habits and feeding behavior of the cownose ray, Rhinoptera bonasus in lower Chesapeake Bay. Estuaries 8, 305-310. Smith, J.W., and Merriner, J.V. (1986). Observations on the reproductive biology of the cownose ray, Rhinoptera bonasus in Chesapeake Bay. Fishery Bulletin 84, 871-877. Smith, J.W. and Merriner, J.V. (1987). Age and growth, movements and distribution of the cownose ray, Rhinoptera bonasus in Chesapeake Bay. Estuaries 10, 153-164.
82 Snelson, F.F., and Williams, S.E. (1981). Notes on the occurrence, distribution, and biology of elasmobranch fishes in the Indian River Lagoon System, Florida. Estuaries 4, 110-120. Snelson F.F., Williams-Hooper, S.E., and Schmid, T.H. (1988). Reproduction and ecology of the Atlantic stingray, Dasyatis sabina in Florida coastal lagoons. Copeia 1988, 729-739. Standora, E.A. and Nelson, D.R. (1977). A telemetric study of the behavior of freeswimming angel sharks Squatina californica Bulletin of Southern California Academy of Sciences 76, 193-201. Summers, A.P. (2000). Stiffening the stingray skeleton an investigation of durophagy in myliobatid stingrays (Chondrichthyes, Batoidea, Myliobatidae). Journal of Morphology 243, 113-126. Summers, A.P. (2003). A novel fibrocartilaginous tendon from an elasmobranch fish ( Rhinoptera bonasus ). Cell and Tissue Research 312, 221-227. Sundstrom, L.F., Gruber, S.H., Clermont, S.M., Correia, J.P.S., de Marignac, J.R.C., Morrissey, J.F., Lowrance, C.R., Thomassen, L., and Oliveira, M.T. (2001). Review of elasmobranch behavioral studies using ultrasonic telemetry with special reference to the lemon shark, Negaprion brevirostris around Bimini Islands. In T.C. Tricas and S.H.
83 Gruber (eds.) The Behavior and Sensory Biology of Elasmobranch fishes: an Anthology in Memory of Donald Richard Nelson, Kluwer Academic Publishers, the Netherlands. Tiller, R.E., Glude, J.B., and Stringer, L.D. (1952). Hard clam fishery of the Atlantic coast. Commercial Fisheries Review 14, 1-25. Thrush, S.F., R.D. Pridmore, J.E. Hewitt and V.J. Cummings. (1991). Impact of ray feeding disturbances on sandflat macrobenthos: do communities dominated by polychaetes or shellfish respond differently? Marine Ecology Progress Series 69, 245252. Trent, L., Parshley, D.E., and Carlson, J.K. Catch and bycatch in the shark drift gillnet fishery off Georgia and East Florida. Marine Fisheries Review 59 (1), 19. Uebelacker, J.M. and Johnson, P.G. (1984). (eds.)Final Report to the Minerals Management Service, contract # 14-12-001-29091. Barry A. Vittor and Associates, Inc. Mobile, Alabama. 7 Vols. Valentine, J.F., Heck, K.L., Jr., Harper, P., and Beck, M. (1994). Effects of bioturbation in controlling seagrass ( Thalassia testudinum Banks ex Knig) abundance: evidence from field enclosures and observations in the Northern Gulf of Mexico. Journal of Experimental Marine Biology and Ecology 178, 181-192.
84 Voegeli, F.A., Smale, M.J., Webber, D.M., Andrade, Y., and ODor, R.K. (2001). Ultrasonic telemetry, tracking and automated monitoring technology for sharks. Environmental Biology of Fishes 60, 267-281. Wang, J.C.S., and Raney, E.C. (1971). Distribution and fluctuations in the fish fauna of the Charlotte Harbor Estuary, Florida. Mote Marine Laboratory Contributions. 112, 1-21. Worton, B.J. (1987). A review of models of home range for animal movement. Ecological Modelling 38, 277-298. Worton, B.J. (1989). Kernel methods for estimating the utilization distribution in homerange studies. Ecology 70, 164-168.
86Appendix A. Appendix A. Length weight data for three captive R. bonasus used in transmitter attachment and retention studies. Individuals were maintained at Mote Marine Laboratory between July 2003 and October 2004. SDW (cm) Weight (kg) Tag retention Sex start end start end (days) F 76 77 6.1 6.6 457 F 67 72.5 3.4 5.3 187 M 72 74 4.6 5 156
87Appendix B. Appendix B. Minimum convex polygon (MCP) and kernel utilization distributions (95 and 50% KUD) in kilometers2 for all tracked individuals over 3 day, 7 day, 30 day (1 month) and total monitoring periods. 3 day 7 day 30 day Total Ray ID MCP 95% KUD 50% KUD MCP 95% KUD 50% KUD MCP 95% KUD 50% KUD MCP 95% KUD 50% KUD 95 8.65 9.83 1.00 43.9741.713.45 55.6346.023.80 55.63 43.005.50 96 3.96 12.82 3.07 10.6911.481.89 31.64 21.793.12 97 8.24 14.83 1.82 10.07 9.22 1.02 98 20.24 11.49 1.14 23.574.97 0.63 39.83 16.301.48 262 22.77 62.44 6.96 22.77 62.446.96 263 16.21 53.30 16.96 17.5243.487.35 27.9946.359.07 27.99 46.359.07 264 6.88 23.325.04 265 15.41 36.98 6.37 31.5462.966.67 34.528.35 1.66 39.83 20.943.13 266 16.20 33.224.50 268 0.01 0.00 0.00 0.02 0.01 0.08 0.22 0.05 269 0.06 0.01 0.00 3.93 0.27 0.09 4.02 0.67 0.13 270 0.00 0.00 0.65 0.09 0.03 0.81 0.18 0.05 271 2.10 6.58 0.97 2.10 6.58 0.97 273 0.35 0.03 0.01 1.03 0.16 0.05 1.10 0.37 0.09 508 15.19 29.86 4.89 15.19 29.864.89 509 12.71 41.479.68 511 2.19 1.35 0.33 16.378.35 1.03 59.4133.091.84 71.78 20.381.65 512 24.03 16.86 1.94 31.2819.351.86 31.6825.112.51 31.77 25.112.51 514 19.83 29.80 3.69 27.72 28.563.44