Distribution and behavioral ecology of manatees in Southeastern Georgia

DISTRIBUTION AND BEHAVIORAL ECOLOGY
OF MANATEES IN SOUTHEASTERN GEORGIA

Barbara J. Zoodsma

December 1991

Department of Wildlife and Range Sciences
University of Florida

Cooperative Agreement
#14-16-009-1544, RWO #65

Final Report to
U.S. Fish and Wildlife Service

ACKNOWLEDGEMENTS

I would like to thank Drs. Stephen Cofer-Shabica and Susan Bratton of the

National Park Service Coop Unit at the University of Georgia and the US Navy for

funding this study. I am also grateful to Stephen for supporting me during the analysis

and write-up portion of the study.

My sincerest thanks go to the personnel of the U.S. Fish and Wildlife Service

Sirenia Project: Dr. T. O'Shea, Dr. L. Lefebvre, C. Beck, R. Bonde, J. Reid, and H.

Kochman. Without these individuals I would not have been able to successfully conduct

this study. I will always work towards attaining their level of professionalism and

dedication. I also thank them for teaching me the meaning of "Vaudeville."

I am especially grateful to my advisor and chairperson, Dr. Lynn Lefebvre. Lynn

provided guidance and spent many hours in discussion with me. I feel privileged to have

worked with her. I also thank the other members of my committee, Drs. Thomas O'Shea

and Melvin Sunquist, for their insightful comments. Howard Kochman and Steve Linda

provided valuable support with statistical analyses.

Many thanks are due the people who valuably assisted me with radio-tracking in

the unforgiving sun of southeast Georgia: Sue Straut, Andy Bergeron, Tom Engelsma,

Ken Niewenhouse, C. Adams, Pam Meyer, and Charlotte Fries. I give special thanks

ii

to Sue Straut for her energy level and sense of humor during my last summer of radio-

tracking. I am also grateful to the many volunteers who risked being forever lost in

black, oozing mud to assist with manatee captures.

Several other individuals and agencies provided assistance with this study. The

National Park Service personnel at Cumberland Island National Seashore allowed me to

stay in the island research dormitory and provided me with some logistical support -

particularly K. Kacer, D. Koontz, C. Davis, C. Brown, R. Robards, Z. Kirkland, and

C. Fries. Georgia Department of Natural Resources produced the first poster in Georgia

informing the public of radio-tagged manatees and forwarded public reports of radio-

tagged manatees. I thank the security personnel at Kings Bay Naval Submarine Base for

their interest in manatees and this project and for their numerous reports of tagged

manatee locations. Jim Valade of the Port of Fernandina Beach Construction Project

assisted with captures and radio tracking. I am also sincerely grateful to my skillful pilot,

the late Dr. William Fosmire.

Container Corporation of America (CCA), Fernandina Beach, FL, and Gilman

Paper Company, St. Marys, GA, kindly allowed me access to their warm-water discharge

areas. Additionally, CCA allowed us to capture manatees in their warm-water discharge

and provided us with much welcomed hot coffee and food on those chilly, early

mornings.

Finally, I wish to acknowledge the meaningful support and encouragement of my

special friends: P. Meyer, C. Fries, and T. Larsen. I thank Pam and Charlotte who

often donated their days off to assist me with this study.

iii

TABLE OF CONTENTS

page
ACKNOWLEDGEMENTS ................................. ii

LIST OF TABLES ...................................... vii

LIST OF FIGURES .................................... viii

LIST OF ABBREVIATIONS ............................... xi

ABSTRACT ............................................. xii

INTRODUCTION ......................................... 1
Sirenians and Their Distribution .......................... 1
Florida M anatees ................................... 1
Activity Patterns and Feeding Ecology of
Florida M anatees ............................... 2
Florida Manatees in Georgia ............................ 7

OBJECTIVES. ......................................... 10

HYPOTHESES ON MANATEE ACTIVITY. .................. .... 11

STUDY AREA .......................................... 13

M ETHODS ........................................... 27
Radio Telemetry ................................... 27
Behavioral Observations .............................. 32
Data Analysis ................................... 34
Food Habits .................................... 35
Aerial Surveys .................................... 37

RESULTS ........................................... 43
Manatees Radio-Tracked .............................. 43
Areas of Greatest Use by Manatees ........ ............... 47
Diurnal Time-Activity Patterns ......................... 52
Food Habits and Feeding Ecology ........................ 60

DISCUSSION ........................................ 101
Seasonal Patterns .................................. 101
Areas of Greatest Use by Manatees ...................... 105
Activity Patterns and Time Budget ....................... 108
Food Habits .................................... 114
Dredging in Cumberland Sound ................ ........ 115
Management Implications ............................ 116

CONCLUSIONS ................. ...................... 123

APPENDIX A. Definitions of behavioral categories used in activity pattern
analyses ...................................... 125

APPENDIX B. ID numbers, sex, body length (cm) and capture dates of manatees
captured in Fernandina Beach, FL, 1987-1989 ................ 127

APPENDIX C. ID numbers, sex, and body length (cm) of manatees that migrated
to southeastern Georgia from Brevard Co., FL. ................ 128

APPENDIX D. Summary of Southeastern Georgia manatee radio-tracking. 129

APPENDIX E. Salinity values of some selected areas frequented by radio-tagged
manatees in southeastern Georgia ....................... 133

APPENDIX F. Frequency distribution of the number of Spartina-feeding bouts
by length of observation time (h), for manatees radio-tracked during cold
and warm seasons .................................. 134

APPENDIX G. Mean length (min) of Spartina-feeding sessions by observation
time (h) for radio-tracked manatees during cold and warm seasons .. 135

APPENDIX H. Frequency distribution of the number of algal-feeding bouts
observed by length of observation time (h) for radio-tracked manatees in
the warm season ................................... 136

APPENDIX I. Mean length (min) of algal-feeding sessions by observation time
(h) for radio-tracked manatees during the warm season ........... 137

APPENDIX J. Frequency distribution of the number of resting bouts observed by
length of observation time (h) for radio-tracked manatees during the cold
and warm seasons .................................. 138

APPENDIX K. Mean length (min) of resting sessions by observation time (h) for
radio-tracked manatees during cold and warm seasons ............ 139

APPENDIX L. Frequency distribution of the number of cavorting bouts observed
by length of observation time (h) for radio-tracked manatees during cold
and warm seasons (h) ............................... 140

APPENDIX M. Mean length (min) of cavorting sessions by observation time (h)
for radio-tracked manatees during cold and warm seasons .......... 141

APPENDIX N. Frequency distribution of the number of moving bouts observed
by length of observation time (h) for radio-tracked manatees during cold
and warm seasons .................................. 142

APPENDIX O. Mean length (min) of moving sessions by observation time (h) for
radio-tracked manatees during cold and warm seasons ............ 143

APPENDIX P. Collection dates, locations and I.D. numbers of manatees
recovered dead from the Cumberland Sound vicinity. Only stomach
samples were analyzed for this study ..................... 144

LITERATURE CITED ................................... 145

LIST OF TABLES

Table pge

1. Number of manatees observed during each aerial survey conducted over
Cumberland Sound and adjacent waters in 1988 (left side) and 1989 (right
side). Yearly means and their standard deviations are given at the
bottom ..... ..... ....... .................... .... 81

2. Mean session durations (min) of four major activity categories and
probability values associated with tests of differences in mean activity
session durations between cold and warm seasons . . ..... 90

3. Mean session durations of two subcategories of moving and probability
values associated with tests of differences in mean activity session
durations between cold and warm seasons. The subcategory "travelling"
represents directed movements by manatees and "milling" represents
manatee movements that appeared to be aimless . . . ..... 90

4. Forage items and their mean percentages in manatee stomach content
samples (n=15). Stomach samples were taken from dead manatees
recovered in Cumberland Sound and Georgia between 1982 and 1990, and
were examined using the microhistological technique of Hurst and Beck
(1988) .. .. .. ... .. .. .. .. .. .. .. .. .. .. .. .. 99

LIST OF FIGURES

Figures pag

1. Map illustrating some general topographic features in the Cumberland
Sound, GA, vicinity and the location of Satilla River. Cumberland Sound
extends from northern Amelia Island, FL, to the southern terminus of
Crooked River, GA .................................. 18

2. Map illustrating the location of Kings Bay Naval Submarine Base and
other aquatic features in the Cumberland Sound, GA vicinity ...... 20

3. Map illustrating the warm-water discharge locations of Gilman Paper
Company and Container Corporation of America and other terrestrial
features and locations in the Cumberland Sound, GA, vicinity ....... 22

4. Map illustrating the locations of Gilman Paper Company's (GPC) warm-
water discharge on North River, Camden Co., GA, and Container
Corporation of America's (CCA) warm-water discharge on Amelia River,
Nassau Co., FL ................................... 24

5. Illustration of the pond-like discharge area of Container Corporation of
America. The rocky breakwater and mud flats are submerged at high
tide. (Drawing is not to scale.) ........................ 26

6. Diagram illustrating layout of nets and lines during all manatee captures
in the Container Corporation of America warm-water discharge area from
1987 to 1989 .................................... 40

7. Map illustrating manatee aerial survey routes in the Cumberland Sound
region. Note: survey routes depicted by a solid line indicate routes that
were flown 1987-1989; the dotted line indicates an extension to the route
that was added in 1989. (S=survey starting point; E=survey ending
point.) . . . . . . . . . 42

LIST OF FIGURES (con't)

8. Month of each year that radio-tagged manatees were documented to
migrate between southeastern Georgia and Brevard Co., FL: a) Spring
and early summer migrations from Brevard Co., FL, to southeastern
Georgia; b) Late summer and fall migrations from southeastern Georgia
to Brevard Co., FL. ............................... 64

9. Locations of radio-tracked manatees in the Cumberland Sound region
during the cold season in 1989 ................... ....... 66

10. 1989 cold-season locations of manatee TNC-05 .............. 68

11. Map illustrating warm-season site fidelity of TBC-06 from 1987 to
1988 . . . . .. . . . . . .. 70

12. Map illustrating 1989 warm-season locations of TBC-06. These locations
are similar to it's warm-season locations in 1987 and 1988 (Fig. 11) 72

13. Map illustrating warm-season site fidelity of TBC-23 from 1988 to
1989 . . . . . . . . . . 74

14. Map illustrating warm-season locations of TNC-01 in 1987 and 1988 76

15. Digitized locations of radio-tracked manatees in the Cumberland Sound
region during the warm season in 1989 . . . ...... ....... 78

16. Areas in Kings Bay used frequently by radio-tagged manatees . 80

17. Locations of manatee sightings during aerial surveys conducted in summer
1988 and spring and summer 1989 in Cumberland Sound . ... 83

18. Locations of boats sighted during manatee aerial surveys in June, July,
and August, 1989 .................................. 85

19. Frequency distribution of the durations of manatee observation periods (h)
for cold and warm seasons ............................. 86

20. Total number of observation periods recorded for each of the eight
manatees included in the behavioral analyses: a) Manatees radio-tracked
in the cold season; b) Manatees radio-tracked in the warm season . 87

LIST OF FIGURES (con't)

21. Percentages of all observation time (cold and warm seasons pooled) that
radio-tracked manatees exhibited one of four major activities during low,
mid-, or high tide: a) Feeding on Spartina; b) Resting; c) Cavorting; d)
M oving ....................................... 89

22. Percentage of all cold-season observation time that radio-tracked manatees
exhibited one of four major behaviors during low, mid-, or high tide: a)
Feeding on Spartina; b) Resting; c) Cavorting; d) Moving . ... 92

23. The ANOVA-adjusted mean proportion of time that manatees engaged in
each major behavior during each tide stage in the cold season . ... 93

24. The mean, minimum, and maximum duration of activity sessions for
radio-tracked manatees during the cold season. (FS =Feeding on Spartina,
R=Resting, C=Cavorting, M=Moving) . . . .... ....... 94

25. Percentage of all warm-season observation time that radio-tracked
manatees exhibited one of four major behaviors during low, mid-, or high
tide: a) Feeding on Spartina; b) Feeding on alga; c) Resting; d)
Cavorting; e) Moving ............ . . . ..... 96

26. The mean, minimum, and maximum duration of activity sessions for
manatees radio-tracked during the warm season. (FS=Feeding on
Spartina, FA =Feeding on Alga, R =Resting, C =Cavorting,
M =M oving) . . . . . . . . . 97

27. The ANOVA-adjusted mean proportion of time that manatees engaged in
each major behavior during each tide stage in the warm season ....... 98

28. Relative frequencies of food species found in stomach contents of dead
manatees recovered from Cumberland Sound and southeastern Georgia
(n=15) from 1982 to 1990 (S=Spartina, UG=Unknown Grass,
G=Gracilaria, U=Ulva, E=Enteromorpha, S/R= Stem/Rhizome,
I=Invertebrates, O=Other): a) Comparison between male and female
manatees; b) Comparison between season in which manatee was recovered
(warm or cold) ........ ................... ....... 100

29. Map illustrating how Navy vessels travelling the main channel in Kings
Bay go through the manatee travel corridor between two of their preferred
sites: the fresh water seep area (A), and the shallow cove-like area near
Crab Island (B) .................................... 122

LIST OF ABBREVIATIONS

BPPC ....... Brunswick Pulp and Paper Company
CCA .......... Container Corporation of America
CINS ........ Cumberland Island National Seashore
FLDNR .... Florida Department of Natural Resources
GADNR Georgia Department of Natural Resources
GLM ................. General Linear Model
GPC ................ Gilman Paper Company
KBNSB ........ Kings Bay Naval Submarine Base
MGD ................. Million gallons per day
NPS .................. National Park Service
USCGS ......... U.S. Coast and Geodetic Survey
UF ................... University of Florida
UGA .................. University of Georgia
UHF .................. Ultra-high frequency
USFWS .......... U.S. Fish and Wildlife Service
VHF .................. Very-high frequency

DISTRIBUTION AND BEHAVIORAL ECOLOGY
OF MANATEES IN SOUTHEAST GEORGIA

By

Barbara J. Zoodsma

October, 1991

Although sighting reports indicate that manatees are not uncommon in southeastern

Georgia, which represents the northern limit of their typical range, little was known about their

distribution and ecology in this region. A 3-year radio-tracking study was conducted from spring

1987 through summer 1990 to investigate the ecology of manatees along the southeastern coast

of Georgia, and to evaluate the potential effects of dredging on manatees in Cumberland Sound

and Kings Bay, GA. Fourteen manatees were radio-tracked for varying amounts of time during

the study: 8 manatees were captured and radio-tagged in Fernandina Beach, FL; 5 manatees that

were radio-tagged in Brevard Co., FL, migrated to Georgia during the warm season; one

manatee was radio-tagged in Kings Bay, GA, in 1989. Weekly aerial surveys were conducted

from May through September, 1988, and April through August, 1989, to determine manatee

distribution in Cumberland Sound and Kings Bay. Microhistological analyses were performed

on the stomach contents of 15 dead manatees that were recovered in Georgia in order to provide

further information on food habits.

Georgia salt marshes are an important manatee habitat. Some of the manatees that winter

in Florida return to Georgia in repeated years during the warm season. Manatees feed primarily

xii

Georgia salt marshes are an important manatee habitat. Some of the manatees that winter

in Florida return to Georgia in repeated years during the warm season. Manatees feed primarily

on Spartina alterniflora, which is accessible to manatees only during high tide, and the

macroalgae Ulva spp. and Gracilaria spp. Manatee activity patterns are rhythmic in response to

tidal cycles: manatees feed on Spartina most often at high tide, and rest or cavort at low and

mid-tide in both cold and warm seasons. Manatees spend a significantly greater proportion of

time resting in a man-made warm-water source at low and mid-tide than at high tide in the cold

season, presumably to conserve energy. Aerial surveys as well as radio-tracking indicated that

Cumberland Sound, Kings Bay Naval Submarine Base, Tiger Island Marsh, and the Fernandina

Beach waterfront are some of the areas in and near Georgia that manatees frequent in the warm

season. The Satilla and Altamaha Rivers, north of Cumberland Sound, were also utilized by

several radio-tagged manatees. There was a negligible amount of use of the eastern side side of

Cumberland Island (i.e., ocean) by these "marine mammals." Potential negative effects of

dredging on manatees can be minimized by providing for manatee watches on dredges, requiring

support vessels to move at slow speeds, avoiding dredging in manatee feeding areas, and

monitoring the effects of dredging on the saltmarsh shoreline and benthic macrophytes,

particularly in known manatee feeding areas.

INTRODUCTION

Sirenians and Their Distribution

Manatees belong to the mammalian Order Sirenia, which is represented by 4

living species: the dugong (Dugong dugon) and three manatees (Trichechus spp.).

Sirenians (sea cows) are large aquatic herbivores that are tropical to subtropical in

distribution. The dugong is found in tropical and subtropical coastal waters of the Indian

and western Pacific Oceans. Amazonian manatees (T. inunguis) are found in the Amazon

River Basin, and West African manatees (T. senegalensis) are found in coastal waters and

major rivers of tropical West Africa (Husar 1977, Thorback and Jenkins 1982). Two

subspecies of the West Indian manatee are currently recognized: T. m. manatus found

in the Greater Antilles, eastern Mexico, Central America, and northern South America;

and T.m. latirostris (Florida manatees) found in the southeastern United States (Domning

and Hayek 1986).

Florida Manatees

Florida manatees are the only Sirenian species to range into temperate regions

(Irvine 1983). Florida and extreme southern Georgia are the northern limits of the

manatee's year-round range in the U.S., but even in Florida manatees must seek warm

1

2
water in winter. Low metabolic rates and high thermal conductance leave manatees

susceptible to cold-related mortality. The behavior of manatees seeking warm-water

refuges from the cold is unique to Florida manatees in comparison with other sirenians

(Irvine 1983). Florida manatees are concentrated primarily in warm waters of southern

Florida during cold weather, but in spring and summer, as coastal waters warm, the

Florida manatee may move as far north as North Carolina on the Atlantic coast and as

far west as Louisiana on the Gulf coast (Hartman 1974, Rathbun et al. 1982).

Florida manatees are listed as endangered and are protected under two pieces of

federal legislation: the U.S. Marine Mammal Protection Act of 1972, and the

Endangered Species Act of 1973. Additionally, manatees have been protected in the state

of Florida as early as the eighteenth century; protection was increased more recently

under the Florida Manatee Sanctuary Act of 1978. The West Indian manatee is legally

protected worldwide and is listed as vulnerable in the IUCN Red Data Book.

Activity Patterns and Feeding Ecology of Florida Manatees

Manatees occur in freshwater, estuarine, and marine environments. The feeding

ecology of manatees in freshwater habitats has been studied in several Florida locations:

Crystal River (Hartman 1979), Blue Spring and the St. Johns River (Powell and Waldron

1978, Hartman 1979, and Bengtson 1981), Turkey Creek (Tiedemann 1980), and Blue

Lagoon in Dade Co. (Reynolds 1981). Hartman (1979) studied manatees in Crystal River

and found no predictable sequence in activities nor a difference in frequencies of behavior

3
relative to time of day. Manatees in Turkey Creek, Brevard Co., also reportedly did not

demonstrate a pattern in their diurnal activities (Tiedemann 1980).

Manatees appear to respond behaviorally to miscellaneous human and

environmental stimuli. Reynolds (1981) speculated that manatees in Blue Lagoon Lake,

Miami, FL, refrained from "body-surfing," "cruising," and "follow[ing]-the-leader"

during mid-day to avoid periods of heaviest boat traffic. The feeding forays of the

overwintering population of manatees in Blue Spring, Volusia Co., FL, coincided with

warmest ambient water temperatures (Powell and Waldron 1978, Bengtson 1981). This

particular diel cycle, probably structured for maximum energy conservation, broke down

in warm weather (Bengtson 1981). Manatees in Honduras have been reported by local

residents to have become nocturnal in response to hunting pressure (Rathbun et al. 1983).

Manatee distribution in Kings Bay, Crystal River, FL, was influenced by water

temperature and the amount of boating activity in the area (Kochman et al. 1985,

Buckingham 1990, Rathbun et al. 1990). Water temperature and diurnal boating activity

in Kings Bay may explain why a radio-tagged manatee travelled at night down Crystal

River and then returned to Kings Bay by dawn the next day (Powell and Rathbun 1984).

Observations of dugongs suggest that they also shift their activity patterns in

response to human and environmental pressures. Sander (1980) reported that dugongs

in New Guinea feed at night near villages where they were hunted. Dugongs in

Shoalwater Bay, Australia, fed inshore from approximately 2 hours before until 2 hours

after high tide when high water levels afforded them access to rich seagrass beds in

4

shallow areas (Anderson and Birtles 1978). The dugongs restricted their offshore feeding

excursions to periods of low tide when shallow water prevented access to inshore areas.

Hartman (1979) also reported the strong influence of tides on manatee movements

in Florida. Channels that served as thoroughfares at high tide could often be inaccessible

at low tide. In addition, at high tide manatees could feed on shoal margins and bank

growth that were otherwise unreachable. In "the turbid Intracoastal Waterway of

northeastern Florida where bank grasses are the staple food of manatees, high tides dictate

the hours during which animals may feed" (Hartman 1979:38).

Manatees in Guyana preferred submerged, floating, and emergent vegetation in

that order (Allsopp 1969), and manatees in Florida appear to have the same preferences

(Hartman 1979, Domning 1980, and Reynolds 1981). Submerged, freshwater vegetation

that is eaten by Florida manatees includes: Hydrilla verticillata, Myriophyllum spicatum,

Elodea densa, and Ceratophyllum demersum (coontail) (Hartman 1979, Tiedemann 1980,

Bengtson 1981). Floating vegetation consumed by manatees includes: Eichhornia

crassipes (water hyacinth) and Pistia stratiotes (water lettuce). Alternanthera

philoxeroides (alligatorweed), and Lemna spp. (duckweed) are also occasionally taken

(Hartman 1979, Tiedemann 1980, Bengtson 1981). Bengtson (1981) reported that

manatees in the upper St. Johns River fed on floating mats of grasses (Panicum and

Paspalum).

Little information is available on the feeding ecology of manatees in estuarine or

marine environments (Campbell and Irvine 1977; Packard 1981, 1984; Lefebvre and

Powell 1990). The most important food species for manatees in salt water habitats

5
throughout much of Florida are seagrasses: manatee grass (Syringodium filiforme),

shoalgrass (Halodule wrightii), turtlegrass (Thalassia testudium) (Hartman 1979; Packard

1981, 1984; Ledder 1986; C. Beck, USFWS, personal communication), and

widgeongrass (Rupia maritima) (Powell and Rathbun 1984, Rathbun et al. 1990).

Packard (1984) studied manatee impacts on seagrass beds in Jupiter Sound on the east

coast of Florida and found that in addition to grazing above-ground biomass, manatees

rooted for rhizome material. Starch is concentrated in seagrass rhizomes during the

winter (Dawes and Lawrence 1983).

Manatees will occasionally feed on emergent or bank-growing vegetation,

particularly in turbid waterways where seagrasses are absent (Reynolds 1977, Powell and

Waldron 1978, Hartman 1979, Tiedemann 1980, Bengtson 1981). Manatees have been

seen feeding on Typha latifolia, Distichlis spicata, Panicum purpurascens, Panicum sp.,

Paspalum vaginatum, Vigna reopens, and Phragmites (Hartman 1979, Tiedemann 1980,

Reynolds 1981). The stomach contents of 12 manatees recovered dead in Duval Co., FL,

were analyzed and the data suggest that Spartina is an important food item for manatees

in northeastern Florida (L. A. Hurst and G. B. Rathbun, in litt.). Manatees have been

observed feeding on Spartina alterniflora in Admiral's Cove, Palm Beach Co., in the St.

Johns River, Duval Co., in Cumberland Sound, and near St. Simons Island, GA

(Hardisky 1979, Baugh et al. 1989).

Marine algae have been described as being of little significance in manatee diets

unless vascular vegetation is unavailable (Hartman 1979). Several algal species were

found in stomach samples from dead manatees recovered in Brevard Co., FL (C. Beck,

6

USFWS, personal communication). A few stomach samples contained a large percentage

of Gracilaria; however, although algae were common in stomach samples, their

percentages of total contents were generally low (C. Beck, personal communication),

which may indicate that algae are ingested incidentally when seagrasses are consumed.

I observed manatees feeding on Ulva sp. growing on the floating docks at

Cumberland Island National Seashore (CINS) (Zoodsma and Bratton 1987). Manatees

in the Alafia River near Tampa Bay, FL, were observed feeding on Ulva sp., and

Gracilaria sp. was abundant in the vicinity of the feeding manatees (Lewis et al. 1984).

Seagrasses have essentially disappeared from this area (Lewis et al. 1984). Tiedemann

(1980) observed manatees in Turkey Creek feeding on epilithophytic (growing on objects

such as piers, pilings, barges, etc.) algae in 5% of feeding observations. Feeding on

epilithophytic algae increased to 30% of feeding observations after water hyacinth, the

preferred food item, had been chemically removed from the creek.

It is believed that manatees feed 6 to 8 hours per day (Bertram and Bertram 1964,

Hartman 1979). Etheridge et al. (1985) radio-tracked three manatees in Crystal River

and the longest they observed any of the three manatees feeding was 6 hours and 10

minutes. Manatees in the upper St. Johns River spent a mean time of approximately 5.1

hours per day feeding (Bengtson 1983). Manatee feeding activity may vary seasonally

with fluxes in forage nutrient levels and energy demands on the manatees themselves

(Best 1981, Bengtson 1983, Etheridge et al. 1985).

Manatees observed during cold weather fed in discrete sessions (Hartman 1979,

Bengtson 1981). Bengtson (1981:87) reported that "while out in the cold water, manatees

7

traveled directly to the feeding site, ate for a prolonged period, and returned rapidly to

Blue Spring's warm water." Manatees in Crystal River fed in sessions that normally

lasted "between 30 and 90 min but continued for more than 2 hours if animals were

exceptionally hungry" (Hartman 1979:55). Reynolds (1981) found that manatees in Blue

Lagoon fed in sessions ranging from several seconds to over 45 min. Manatees in the

St. Johns River fed more in winter than in spring, perhaps because energy demands were

greater in winter or because forage quality varied seasonally (Bengtson 1981).

Florida Manatees in Georgia

Florida manatees frequent the coastal waters of Georgia. Hartman (1974)

conducted interviews and aerial surveys that included the Georgia coast south of

Brunswick (Glynn Co.), and Rathbun et al. (1982) analyzed sighting reports from

Georgia between the early 1900s to June 1981. Beeler and O'Shea (1988) recently

updated the 120 Georgia records in the latter sources with 51 records obtained between

April 1974 and August 1986. Aerial surveys were conducted in August 1980 over the

region between the Crooked River (Camden Co., GA) and Fernandina Beach (Nassau

Co., FL) (Valade 1980). Results from these studies indicate that manatees are present

along the Georgia coast primarily during summer months, and in early autumn and

spring. Manatees move south to Florida in the winter, although some may remain at

warm water effluents: one in the North River, Camden Co., GA, and one in Fernandina

8
Beach, FL. Some of the same individuals that winter in Florida may return to Georgia

each summer (Rathbun et al. 1982).

Some aspects of the ecology of manatees in Florida have been well documented,

but little is known about manatees elsewhere. Coastal habitats encountered by manatees

in southeastern Georgia differ considerably from those encountered by manatees further

south in Florida. Seagrasses, the primary food of manatees in saltwater habitats

throughout most of Florida, are absent from the turbid coastal waters of southeastern

Georgia, and vascular plant forage is limited to emergent shoreline vegetation. Unlike

Florida, where seagrasses are generally available to manatees at all times, shoreline

vegetation in Georgia estuaries are more accessible to manatees during the elevated water

levels accompanying high tide. Theoretically, manatees might be expected to "link" their

feeding activity with the tidal cycle. This feeding phenomena has never before been

studied and little is known about it.

The purpose of the present study was to increase understanding of manatee

distribution, activity patterns, and feeding ecology in coastal southeastern Georgia. This

region is of particular interest because it is the northern limit of the Florida manatee's

typical range; because it is an estuarine habitat that is strongly influenced by tides; and

because only one vascular plant (Spartina spp., primarily S. alterniflora) is highly

available as manatee forage. This situation provides an opportunity to study manatee

behavior in the presence of strong environmental cues (temperature and food availability).

Results from this study should also provide insight to manatee behavior in situations

9

influenced by more subtle and complex environmental factors, and should allow better

decision-making for manatee conservation in this area.

OBJECTIVES

The major objectives of this study are:

1) Investigate manatee activity patterns relative to the tidal cycle and cold and

warm seasons.

2) Identify areas in southeastern Georgia that manatees frequently used in

cold and warm seasons.

3) Determine the food habits of manatees in coastal southeastern Georgia.

4) Use the information from this study to examine the potential influences of

dredging in Cumberland Sound on manatees in that area.

HYPOTHESES ON MANATEE ACTIVITY

The major null hypotheses of this study include:

1) Ho: The proportion of time in which manatees engage in four behaviors

(feeding, resting, cavorting, travelling) is similar during all tide stages, and

in both cold and warm seasons.

2) Ho: Behavioral session lengths are similar in both seasons.

3) Ho: Manatees consume submerged macrophytic or epilithophytic algae, and

emergent, vascular vegetation with equal frequency, both in cold and warm

seasons.

I anticipated the rejection of all three null hypotheses. I predicted tide and season

to influence the activity session length and proportion of time that manatees engaged in

at least some behaviors, particularly feeding.

Because Spartina alterniflora is most available to manatees at high tide, I predicted

manatees to spend a proportionately greater amount of time feeding on Spartina near high

tide, and, consequently rest and cavort near low and mid-tide.

I predicted season to influence activity since Georgia is outside of the Florida

manatee's traditional winter range. In the cold season, I predicted little cavorting and

12

more resting (both in proportion of time and session length) at warm water sources as an

energy saving measure. I predicted manatees would cavort more during the warm season

when energy loss due to cold water temperatures was no longer a threat to manatees.

Finally, since algae are both benthic and epilithophytic and may be available to

manatees for greater amounts of time (i.e., low and mid-tide in addition to high tide), I

predicted manatees to consume algae more frequently at low or mid-tide, and Spartina

more frequently at high tide.

STUDY AREA

The Cumberland Sound region extends approximately 18 km north from Amelia

Island, Nassau Co., FL, to the southern terminus of Crooked River, Camden Co., GA

(Fig. 1). A number of rivers flow into the region, including the Amelia River

(Intracoastal Waterway) from the south; St. Marys River, a "blackwater" river draining

the Okefenokee Swamp; Crooked River, and Cumberland River (Intracoastal Waterway)

from the north. The North River flows into St. Marys River roughly 4 km east of where

St. Marys River flows into the sound. Cumberland Sound's major inlet and outlet is the

St. Marys Entrance located between Cumberland Island, GA, to the north and Amelia

Island, FL, to the south. Kings Bay Naval Submarine Base (KBNSB) is situated in

Cumberland Sound just west of Cumberland Island (Fig. 2). Areas north of Cumberland

Sound include the Satilla River and White Oak Creek, a tributary of the Satilla, both in

Camden Co. (Fig. 1), and the Altamaha River near Brunswick, Glynn Co., GA.

Coastal southeastern Georgia experiences twice daily tidal cycles with mean

amplitudes of 1.71 to 2.16 m (mean spring amplitudes = 2.10 to 2.45 m) and mean

currents of 0.3 to 2.7 knots (USCGS 1990). Water visibility is minimal as a result of

strong tidal currents stirring up the silty, muddy bottom substrates and tannic acids that

are washed in from area marshes and woodlands. Consequently, except for benthic

14

marine algae that grow in shallow areas, submerged vegetation is lacking. The area is

warm temperate to subtropical in climate.

Vast salt marshes dominated by smooth cordgrass, Spartina alterniflora, typify the

area (unless indicated otherwise, Spartina hereafter refers to Spartina alterniflora). "Tidal

marshes are formed in conjunction with barrier island development" and occupy a 4- to

6-mile band between the barrier islands and the mainland (Johnson et al. 1974). Coastal

Georgia supports 2,856 ha of regularly flooded salt marsh (Johnson et al. 1974).

Three growth forms of Spartina occur: a tall form (to 2 m) found along the banks

of meandering estuarine creeks and rivers; a short form (to 0.5 m) growing in more

elevated marsh areas; and a medium form (to 1 m) growing between the two sites (Teal

1958, Johnson et al. 1974). The reason for this zonation is believed to be the result of

environmental differences such as nitrate availability (Anderson and Treshow 1980),

increased soluble salt concentrations (Linthurst and Seneca 1981), and water drainage

patterns (Mendelssohn and Seneca 1980, Wiegert et al. 1983). There is no genetic

difference between the 3 forms of Spartina (Shea et al. 1975). Stands of Spartina are

flooded at high tide but at low tide they are often found exposed several feet above water

level.

Manatees periodically use two industrial warm-water effluents in Cumberland

Sound during winter and early spring: the Gilman Paper Company (GPC) warm-water

discharge at St. Marys, Camden Co., GA, and the Container Corporation of America

(CCA) warm-water discharge in Fernandina Beach, Nassau Co., FL (Fig. 3). The GPC

warm-water discharge is situated on North River (Fig. 4). River depth near the discharge

15

varies from 0 to 2.4 m at mean low tide. Water is discharged directly into the river at

a mean rate of 39.5 million gallons/day (mgd) and at an average temperature of 32.2 C

(D. Rumptz, GPC, personal communication).

The CCA warm-water discharge is located on Amelia River, Fernandina Beach,

FL (Fig. 4). Water depth near the discharge is approximately 1 m at mean low tide.

Water is discharged into a pond-like area (Fig. 5) at a mean year-round rate of 14.5 mgd

and a mean (winter) temperature of 29.4 to 32.2 C (P. Hagan, CCA, personal

communication). Water depth inside the pond area varies from 0 to 2 m at mean low

water. In winter and spring, as many as seven manatees at a time will frequently utilize

the warm-water effluent at Container Corporation of America (CCA) in Fernandina

Beach, FL (personal observation). Employees of GPC and CCA claim that manatees

have used their respective warm water discharges during winter for "several" years.

Human use of Cumberland Sound and coastal Georgia has historically been

extensive. First records indicate that Timucuan and Guale Indians were occupants of

Cumberland Sound and adjacent areas prior to European settlement (Torres 1977).

Spaniards occupied most of coastal Georgia and northern Florida from 1562 to 1732;

during that time Europeans travelled through the area investigating Spanish settlements.

From 1732 to 1785 Spanish and English settlers fought for land bordering Spanish

Florida and the Georgia border (Torres 1977). Silk was a major but not profitable crop

and by 1770, rice and indigo were established as profitable crops. A number of large

lumber and cotton plantations grew during the Plantation Era from 1785 to 1880. As a

result of the Civil War (1861 to 1865), Union forces overran and controlled nearly the

16

entire Georgia coast despite confederate fortresses. From 1880 to the present, private and

public land holdings were established, including Cumberland Island National Seashore in

1972 (Torres 1977).

Coastal areas are under increasing development pressure. St. Marys, in particular,

is experiencing a growth boom as a result of the construction of KBNSB. Coincidentally,

local residents report an observable increase in boat traffic within Cumberland Sound over

the past few years .

The Intracoastal Waterway (ICW) in Cumberland Sound has undergone extensive

dredging since the 1950s as part of the area's military development (Anonymous 1985).

The U.S. Army began purchasing land to build a military ocean terminal at Kings Bay

in 1954 (Anonymous 1985). In 1978 Kings Bay was transferred to the U.S. Navy. The

area served as a naval submarine support base until 1981 when it assumed its present

function as a 12,000+ ha nuclear submarine base that is restricted to public boat and

vehicle traffic. At one time the ICW was proposed to be relocated closer to the western

shore of Cumberland Island (i.e. Dungeness Cut) "to accommodate anticipated increases

in military and commercial traffic from expansion of the Kings Bay Naval facility" (S.V.

Cofer-Shabica, in litt.).

Fig. 1. Map illustrating some general topographic features in the Cumberland Sound,
GA, vicinity and the location of Satilla River. Cumberland Sound extends from northern
Amelia Island, FL, to the southern terminus of Crooked River, GA.

blll-rlc.L

Fig. 2. Map illustrating the location of Kings Bay Naval Submarine Base and other
aquatic features in the Cumberland Sound, GA, vicinity (... = Intracoastal Waterway,
=uplands, = saltmarsh, i: .= sand or mudflat).

Crooked River -

Marianna r
SCreek

-N-
.. I --:" -

" .. "- Kings B;

....: .... -'.: f. .:- ,.

:
n-~
r
.I---:r; _. ..,

---
i. I
-i
.-' .--

a

av,

.-. .

ATLANTIC
OCEAN

North Ri

St. ars Rive
/
:. St. Marvs River

ver .. -

*

Tiger
SBasin

Lanceford Creek --

Dungeness Cut

<.

St. Marys
Entrance

St. Joseph Creek

SC -Egans Creek

. Amelia
: River -

Bells River

*I

Cumberland River

SOld House Creek

.- .- ..- -.

- ".
~-
-:- '~"

''
;"

.

-

i-

Fig. 3. Map illustrating the warm-water discharge locations of Gilman Paper Company
and Container Corporation of America and other terrestrial features and locations in the
Cumberland Sound, GA, vicinity.

-Crab Isla

.7 .. -%

- :: GEORGIA

: Gilman Paper
Company Discharge .'

S.-

SFLORIDA
...:- .: FLORIDA

nd

i : : '

.Drum Point
Island

-.

Little Tiger
Island

'' '~' ~'-''7 ~' '- `I
I
r_ -I -- ~~ ~-'~~ .' -.``.'
-I ..
'-"'
a
`
'''
_--~
'- "'
?.
;
~---
r
-;r

7w, -,7 f .:I;: -

?Table Point-

S. CUMBERLAND
ISLAND

Stafford :- -:
/ Island ,

". i
-N-

Greyfield Inn

Sea Camp Dock

Dungeness Dock

... _--- ---- -

Container Corp.
Discharge

.44
Fernandina Waterfront

AMELIA
ISLAND

Fig. 4. Map illustrating the locations of Gilman Paper Company's (GPC) warm-water
discharge on North River, Camden Co., GA, and Container Corporation of America's
warm-water discharge on Amelia River, Nassau Co., FL.

ANd
*1

Fig. 5. Illustration of the pond-like discharge area of Container Corporation of America.
The rocky breakwater and mud flats are submerged at high tide. (Drawing is not to
scale.)

Amelia River

SWater exits pond-like area
into the Amelia River
through narrow channel
0

I- z t -

Discharge Wall

-f:___::__-_::fl~r:3t::

METHODS

Radio Telemetry

Captures. Manatees were captured and radio tagged on 11 March 1987, 25

February 1988, and 18 January and 16 March 1990 in the CCA warm-water discharge

area. Manatees are very tolerant of capture and handling (O'Shea et al. 1985). Near low

tide, manatees leave the shallow discharge area through a narrow channel to the deeper

water of Amelia River. Two 10-cm2-mesh nylon nets were positioned just after sunrise

at mid-outgoing tide to intercept manatees leaving the discharge area through the narrow

channel. Capture methods generally followed those described by Bengtson (1981), but

are described below in detail. Capture teams consisted of a core of USFWS Sirenia

Project personnel (3 to 5) who led plans and supervised the team and up to eight

additional volunteers from agencies such as the National Park Service (NPS), U.S. Navy,

Georgia Department of Natural Resources (GADNR), Florida Department of Natural

Resources (FLDNR), University of Florida (UF), Jacksonville University, and the Port

of Fernandina Manatee Project. Capture, radio tagging, and tracking of manatees were

carried out under Federal USFWS Permit No. PRT-684532 issued to the Sirenia Project.

An 18-m-long, 7-m-deep primary net was stretched from the rocky breakwater on

the west side of the channel to a wooden piling on the east side of the channel. A

secondary net (25 m long and 4 m deep) was used as a backup capture net and was

27

28
deployed approximately 20 m downstream from the primary net (Fig. 6). The secondary

net stretched from the breakwater to the opposite mudbank by a 60-m, 0.6-cm braided

nylon line anchored to a railroad track 53 m east of the channel mouth. Lead and float

lines on both nets ensured that the nets stretched from the bottom substrate to the water's

surface while the net itself formed a "bag." "Pull" lines attached to the eastern ends of

both nets were held by capture team members and were drawn to collapse the net around

the manatee(s) (Fig. 6). The anchor line was released from the piling or railroad track

when collapsing the net, and later could be pulled rapidly back to the piling or railroad

track to re-deploy the net for additional manatee captures.

Entrapped manatees were pulled slowly to the breakwater and restrained in water

10 to 50 cm deep. Three criteria were used to determine which manatees were suitable

for radio tagging: 1) minimum total body length of 250 cm, 2) not a nursing calf, and

3) tail without deep propeller cuts along the margin. Manatees that met these criteria

were fitted with VHF radio telemetry assemblies. All captured individuals were

photographed for the manatee scar catalog maintained at the U.S. Fish and Wildlife

Service (USFWS) Sirenia Project laboratory in Gainesville, FL, and sketches were made

for quick field identification if the transmitters were lost.

Transmitter assemblies. Radio transmitter assemblies were designed by the

USFWS Sirenia Project of the National Ecology Research Center and the Denver Wildlife

Research Center Electronics Laboratory (Rathbun et al. 1987). Assemblies included a

rubber-coated peduncle belt, a nylon tether about 1 cm in diameter and 2 m in length,

and a floating transmitter package. Peduncle belts were adjusted to custom fit each

29

manatee. Tethers coupled the transmitter package to the peduncle belt and were designed

to break off if they became snagged as a safety precaution for the manatees. One

transmitter was coupled to a tether with a swivel snap and carried into the field for

opportunistic attachment to the belt of a previously radio-tagged manatee whose

transmitter had become detached.

Very high frequency (VHF) radio transmitters were deployed on manatees

captured at the CCA warm-water discharge and two manatees that were radio-tagged

opportunistically in Kings Bay, GA. Expected battery life of the VHF transmitter was

up to 2 years; signal reception was generally 8 to 11 km from the ground and 30 to 50

km from the air. Manatees that migrated into the study area from Florida, had been

tagged by the USFWS Sirenia Project and were equipped with either standard VHF radio

transmitters, or with UHF (ultra high frequency) satellite-monitored transmitters. The

satellite transmitter package also contained a VHF transmitter, however, the range of this

VHF signal was only approximately 3 km. No satellite-monitored UHF data are included

in this study.

Tracking. Radio-tagged manatees were generally located at least three times a

week with no more than 4 days passing between visual locations. Radio tracking was

conducted by aircraft, vehicle, and boat. Manatees were most frequently radio-located

by vehicle in 1987; however, in 1988 and 1989 manatees were typically located and

followed in a 4.5-m Boston Whaler equipped with a 40-hp Johnson outboard motor.

Tagged manatees were radio-tracked along the east coast of Georgia and Florida

through cooperative efforts by the University of Georgia (UGA), UF, USFWS, NPS,

30

FLDNR, and GADNR. The USFWS Sirenia Project coordinated all east-coast manatee

radio-tracking studies and notified participants of "missing" manatees or manatees on the

move. Radio tracking summary sheets were also sent to all participants, keeping them

abreast of tagged manatee locations.

Radio tracking was conducted by aircraft when a tagged manatee had not been

radio located for 4 consecutive days despite concerted ground efforts by cooperating

individuals and agencies. A Cessna 172 aircraft fitted with a Yagi antenna below each

wing strut was used for aerial tracking and a portable switch box permitted the

"activation" of either antenna individually or both antennas simultaneously. Headphones

virtually eliminated background noise from the aircraft and allowed any incoming signals

from manatee transmitters to be heard audibly. Aerial tracking procedures followed those

outlined by Gilmer et al. (1981).

Tagged manatees were tracked from land to generally locate an animal prior to

boat launching, when weather conditions were unsuitable for boating, or when a boat was

unavailable. A pick-up truck with a single Yagi antenna was most often used for land

tracking. The antenna was mounted on a 2-m wooden mast in a roof rack designed by

the USFWS Sirenia Project, which elevated and allowed for 360-degree antenna rotation

while the vehicle was moving or stopped. Vehicle tracking was conducted continuously

along roads adjacent to waterways used by radio-tagged manatees. Tracking stations were

established on Cumberland Island and the mainland for triangulation of compass bearings.

Tracking stations were treeless and preferably elevated locations where an unobstructed

radio signal could be received from the water. I visited appropriate tracking stations

31

consecutively for no less than 15 min each. When a signal was received, a compass

bearing was taken and recorded along with date, time, station location, and signal

strength; the same procedure was then followed within 15 to 20 min (ideally) at one and

occasionally two additional stations located within 5 km.

A single Yagi antenna mounted on a hand-held, 2-m wooden mast, a receiver, and

headphones or an external speaker were used to track manatees by boat. Tagged

manatees were located by reducing first the volume and then the RF gain while following

the direction of the radio signal. Binoculars or the naked eye were used to search for

transmitters or to spot manatee transmitters floating on the surface of the water. The

outboard motor was turned off as soon as possible to avoid disrupting manatees, and

whenever possible, I drifted with tidal currents to an acceptable viewing range.

After locating a radio tag, attachment of the transmitter to the manatee was

verified, the transmitter condition was assessed, and behavioral observations were

recorded. I attempted to minimize possible manatee disturbance while following

manatees by paddling with a canoe paddle, poling, drifting with tidal currents, or

advancing the boat forward by quietly gripping and pulling on shore vegetation. Data

collected included: date, time of day, tide level, air and water temperature, water

salinity, shore and benthic vegetation types, vegetation type consumed by the focal

animal, and presence or absence of other manatees. For all modes of radio tracking,

tagged manatee locations were plotted on copies of USGS 7.5-second topographic

32

quadrangles. It was difficult to approach manatees by boat at night without disturbing

them. Consequently, only a few nighttime observations were attempted.

Behavioral Observations

Manatee behavior was noted and recorded during the 1987 and 1988 field seasons,

but not at regular intervals over continuous observation periods. Manatee behavior was

identified and recorded every 5 min during observation periods in the 1989 field season

using a method similar to that of Altmann's (1974) focal-animal sampling. When the

focal manatee could not be seen at exactly the beginning of a 5-min interval, the last

behavior observed was used to estimate the behavior. Behavioral categories and

definitions followed Bengtson (1981), who radio-tracked manatees in the St. Johns River,

FL, near Blue Spring. Additionally, one activity type was included in this study that was

not included by Bengtson (1981): "intermittent feeding" was recorded when a manatee

was alternating between feeding and moving short distances to new feeding sites along

the shore. As in Bengtson's study, all activities were grouped into four main behavioral

types: 1) moving, 2) cavorting, 3) feeding, and 4) resting (Appendix A). "Cavorting"

referred to any obvious social encounters such as nuzzling, grabbing, or chasing

(Bengtson 1981).

Because water conditions were often very turbid, indirect cues were occasionally

used to determine manatee behaviors. Respiration rates, for example, were important in

discerning between manatees that were resting, idling, or feeding on benthic marine

algae. Resting manatees surfaced to breathe typically no more than once every 4 min,

33

idling manatees approximately once every 2 to 4 min, and feeding manatees

approximately once every 1 to 2 min. The head-raising posture of manatees that surfaced

to breathe was also used to distinguish between manatees that were milling about and

those that were feeding on marine algae. Manatees that were feeding on benthic algae

typically lifted their heads higher out of the water and often exposed their mouth parts

while breathing; manatees that were milling merely lifted their nostrils clear of the water

to breathe and rarely exposed any other part of their head unless surface wave chop was

present. Radio signals were rarely used to discern manatee behavior; however, distinctive

radio signals of 1 to 3 pulses alternating with 2- to 3- min periods of silence were

occasionally used to identify a travelling manatee.

All manatee locations were digitized on a Calcomp 9000 digitizer, stored in a

TRS-80 Radio Shack microcomputer, and later transferred to a Packard Bell

microcomputer. Activity data were entered and stored in the Packard Bell computer and

could be cross-referenced with the location data set by date, time, and manatee ID. Raw

data from these files are stored at the USFWS Sirenia Project.

Data Analysis

Within each continuous observation period, discrete "activity bouts and sessions"

were defined. A "bout" is defined as a continuous period during which a manatee's

recorded behavior fell into only one of five major behavioral categories: feeding on

Spartina, feeding on algae, resting, travelling, or cavorting. An "activity session" began

34

with at least 10 continuous minutes of one behavior and is defined as one or more bouts

of the same major behavior separated by less than 10 min of any other major behavior.

"Activity sessions" are similar to Altmann's (1974) definition of behavioral "states" that

"have appreciable durations." To help examine the possible influence of survey effort

on observed behavior, the frequency of bouts and the length of bouts and sessions were

plotted against observation time, for each combination of season and behavior. Smoothed

lines were fitted to the data with a spline routine (SM interpolation) in SAS/GRAPH

software (SAS Institute Inc., 1987a).

The session lengths of each major activity were summed for each manatee and the

proportion of time spent in each behavior was determined. Differences in the proportion

of time spent in the five major behaviors were tested with a weighted analysis of variance

(ANOVA) in a split-plot design, using the SAS General Linear Models (GLM) procedure

(SAS Institute Inc. 1987b). The ANOVA was weighted based on the total observation

time for each manatee. There were two levels of the main plot effect (season): January

through April, when manatees were located near warm-water discharges (the cold

season); May through August, when ambient water temperatures were considerably

warmer and manatees were no longer visiting warm-water discharges (the warm season).

Because no manatee was tracked in both seasons, individual manatees were nested within

season. Only manatees tracked in 1989 were used in the analysis: four tracked in the

35

cold season (TNC-05, TNC-06, TNC-07, and TNC-08) and four tracked in the warm

season (TBC-06, S-18, TBC-23, and TGA-01).

There were three levels of the subplot effect (tide stage): high, mid, and low.

High tide was considered to be the interval from 2 h before until 2 h after the high-tide

time given in the published Tide Tables (USCGS 1989). Low tide was considered to be

the interval from 2 h before until 2 h after the predicted low-tide time, and mid tide was

the approximately 2-h interval between high and low tide. I referred to the National

Oceanic and Atmospheric Administration (NOAA) Subordinate Station (USCGS 1989)

that was closest to a manatee's location for high and low tide predictions. However,

some areas that I radio-tracked manatees in were relatively remote, and were more than

three miles away from the nearest NOAA Subordinate Station. In these instances, I

interpolated my own high and low tide prediction based on published predictions for

NOAA Subordinate Stations that were nearest to the manatee's location (USCGS 1989).

The effects of season (fixed effect), tide (fixed effect), their interaction, and manatee ID

(random effect) were tested. Adjusted mean proportions from the ANOVA model were

also calculated using SAS GLM for each behavior, by season and tide stage. Differences

in activity session lengths between seasons were tested in a weighted, one-way ANOVA

using the SAS GLM procedure (SAS Institute Inc. 1987b).

Food Habits

Vegetation that was consumed by tagged manatees was identified by direct

observation whenever possible. If vegetation was not positively identified in the field,

36

specimens were collected for a laboratory reference collection. Dr. Joseph Richardson,

Savannah State College, Savannah, GA, assisted with the identification of marine algae.

To supplement these observations, samples of stomach contents from 15 manatees

that were recovered dead in Georgia and the Cumberland Sound region between 22 June

1982 and 26 February 1990 were identified and quantified in the laboratory using a

modified microscope point technique. This technique is described in detail by Hurst and

Beck (1988) but will be briefly outlined here.

Stomach contents were washed on a 30-mesh screen under running water. Five

arbitrary subsets or "grabs" were extracted from the washed ingesta with forceps, and

each grab was spread evenly over a microscope slide. Epidermal cells were removed

from the plant fragments by boiling ingesta in Hertwig's solution, and cleared with 70%

alcohol. Cover slips were placed over the prepared material.

Prepared slides were place on a graduated mechanical stage and examined under

a microscope at 100x power. For each field, plant fragments were identified at five

predetermined locations where lines from the eyepiece micrometer grid intercepted plant

material. Twenty fields were analyzed for each subsample (slide). Plant fragments were

identified at a total of 500 microscope points per stomach sample. Galt et al. (1982)

reported that a similar microscope point technique using 400 points per sample achieved

a 90% probability (a = 0.05) for species constituting 30 to 60% of the sample weight

37

and a 90% probability (a = 0.2) for species constituting 10 to 30% of the sample

weight.

Identification of plants was possible by referring to Taylor (1960), Godfrey and

Wooten (1979, 1981), and Hurst and Beck (1988). The reference slide and herbarium

collection at the USFWS Sirenia Project laboratory and mounted algal specimens

collected from Cumberland Island National Seashore, Camden Co., GA, and provided

by J. Richardson were also used extensively.

Aerial Surveys

Approximately weekly aerial surveys were conducted during summers 1988 and

1989. The region encompassed in aerial surveys extended from Fernandina Beach, FL,

northward to Table Point, Cumberland Island, and west to St. Marys, GA, and Kings

Bay Naval Submarine Base (Fig. 7). All aerial surveys were conducted in a Cessna 172

aircraft flown by a pilot who was familiar with manatee aerial survey procedures.

Surveys were flown along preplanned survey routes at altitudes of 150 to 170 m and

airspeeds of 128 to 150 km/h. Surveys generally followed the guidelines of Irvine (1982)

and lasted approximately 110 min. Data collected during aerial surveys included: date,

air temperature, wind speed, water surface conditions, water clarity, and number and

locations of manatees, Atlantic bottlenose dolphins (Tursiops truncatus), and loggerhead

sea turtles (Caretta caretta). Aerial surveys were not conducted under sea-surface

conditions of Beaufort 3 or more. When manatees were spotted, the primary observer,

sitting in the front passenger seat, informed the pilot of the sighting. The pilot then

circled tightly over the manatee(s) and the primary observer, wearing polarized sun

38

glasses to reduce the glare, would determine the number and behavior of the observed

manatees. Binoculars were often used to identify the manatee behavior.

A secondary observer participated in the 1989 surveys and was seated on the

plane's passenger side in the back seat. The secondary observer recorded dolphin and

sea turtle locations and, beginning in June, plotted boat types and their locations within

the survey area. The primary observer assisted in locating sea turtles and dolphins but

the secondary observer did not assist the primary observer in locating manatees.

Fig. 6. Diagram illustrating layout of nets and lines during all manatee captures in the
Container Corporation of America warm-water discharge area from 1987 to 1989.

Amelia River

Secondary
Net ,

Fig. 7. Map illustrating manatee aerial survey routes in the Cumberland Sound region.
Note: survey routes depicted by a solid line indicate routes that were flown 1987-1989;
the dotted line indicates an extension to the route that was added in 1989. (S= survey
starting point; E=survey ending point.)

r

.;.

'\V
:--

4 --. p-..
r -;-A .
\ .. :-.r. --
\--

X^-^.'^ss
'L- ^ ^ -f-

^ .- "s '^ -" -'=* -'~"

..-;?.-' --_ -.-.-;; :i

;-.-.. :-F- : .*,: ^; -:.

'&;^iC.^

\^--ri ^*iN^]0^?^^

-*' .- '" '.-;:?: ~ "r "' '"' i*^

~''-'

-;
;

RESULTS

Manatees Radio Tracked

Captured and tagged in Fernandina Beach. Five female and three male manatees

were captured and radio-tagged in Fernandina Beach, FL: TNC-01 and -02 (a male and

a female) in 1987, TNC-03 and -04 (2 females with accompanying calves) in 1988, and

TNC-05 through -08 (2 females, one of which was accompanied by a small dependent

calf, and 2 males) in 1989. Only manatee TNC-01 was radio-tracked in more than one

calendar year (1987 and 1988). Manatees that were radio-tagged in Fernandina Beach

ranged from 250 to 340 cm in total body length (Appendix B). Seven other manatees

were captured but not tagged because they were either too small (n=6) or had a tail that

was too mutilated (n= 1).

Other tagged manatees. Four female and one male manatee that were radio-tagged

in Brevard Co., FL, migrated to southeastern Georgia for the warm season from Brevard

Co. in one or more calendar years of this study: TBC-06 (male) in all 3 years; TBC-08

(with calf) in 1987; TBC-19 in 1988; TBC-23 in 1988 and 1989; and TBC-15 in 1989.

Radio-tagged manatees that migrated to southeastern Georgia for the warm season ranged

from 260 to 325 cm in total body length (Appendix C).

One adult female manatee (accompanied by a large dependent calf) was radio-

tagged without capture in July 1989 in Kings Bay Naval Submarine Base, GA. Manatee

43

44

TGA-01's body length was never measured. Manatee TBC-23 was retagged in the same

area with the safety transmitter a few days earlier.

In all, 14 manatees were radio-tracked in the study area for varying amounts of

time. Only manatees that were radio-tracked in 1989 (n = 8) were included in the

behavioral analyses.

One manatee died during the course of this study and two others died shortly after

the study's completion. Manatee TBC-08 was recovered (transmitter missing but still

wearing her peduncle belt) from the St. Johns River, Duval Co., FL, in September 1987;

her small calf was missing. The cause of TBC-08's death could not be determined.

Manatee TNC-08 was recovered near St. Marys, GA, in February 1990; its cause of

death was unknown. Manatee TGA-01 was killed by a Navy tugboat in KBNSB, GA,

in May 1990 (FLDNR, unpub. data).

Tag loss. Tag detachment and/or failure were major, but unavoidable, inhibiting

factors in this study. Tethered, floating transmitters were necessarily vulnerable to

detachment as a safety precaution for the manatees (Rathbun et al. 1987). Five of the

eight manatees that were radio-tagged in Fernandina Beach either lost their transmitters

or their tags stopped transmitting radio signals within 2.5 months of their respective

tagging dates; seven of the eight manatees could not be radio-located within

approximately 5 months of being tagged.

Tethered transmitters were pulled off three manatees by fishermen in the Altamaha

River. Other transmitters were found floating, lying on mud flats, or were found in

stands of Spartina (n=4); had malfunctioned (n= 1); or the transmitter's fate was never

45

determined (n=4). Transmitters that were recovered typically were still attached to the

tether and the tether had snapped at the weak link. Manatee TNC-05's peduncle belt

broke at the buckle.

Seasonal migratory trends. Two of the eight manatees that were radio-tagged in

Fernandina Beach were eventually located in Brevard Co. (approximately 270 km south

of Cumberland Sound, GA); 8 out of the 14 study animals were located in both Brevard

Co., FL, and southeastern Georgia. Four manatees that were radio-tracked in Brevard

Co., FL, during the cold season were located in Georgia for two or more warm seasons

including those that returned to southeastern Georgia in spring 1990 (USFWS, unpub.

data). Additionally, four free-ranging manatees observed in KBNSB were photo-

identified through the USFWS Manatee Individual Identification System as animals that

overwintered in Brevard Co.; two of these were documented in KBNSB in two successive

summers. Manatee TNC-01 ventured as far south as Miami during the cold season.

Those radio-tagged manatees that migrated to southeastern Georgia would typically

arrive in spring and early summer, around April and May (Fig. 8a). April was the

earliest month that a tagged manatee was documented in southeastern Georgia after

having overwintered in Brevard Co.; however, manatee TBC-23, an individual that

ventured to southeastern Georgia for the warm season and typically overwintered in

Brevard Co., was spotted at the GPC warm-water discharge in March 1989.

Unfortunately, she was without a transmitter for winter 1988-1989 so it is uncertain if

46

she overwintered in Brevard Co. and, indeed, migrated to Georgia in March. The latest

Brevard Co. arrival to the study area was TBC-23 in July 1988 (Fig 8a).

The beginning of southward migrations in late summer and autumn is not as well

documented as spring arrivals because a number of manatees lost their tags while in

Georgia over the warm season (e.g., S-12, S-18, TBC-06, and TBC-08). Typically, late

summer and autumn marked the beginning of the manatees' southward migrations from

southeastern Georgia to Brevard Co. (Fig. 8b). However, in 1988 two manatees lingered

in the Cumberland Sound region until late autumn and winter: TBC-06 remained until

early December and TBC-23 remained until at least mid-November (Fig. 8b) when her

transmitter was recovered near KBNSB. The earliest month that a tagged manatee left

the southeastern Georgia area for the year (i.e., did not subsequently return to Georgia

in the same calendar year) was June, when manatee TNC-01 began her 1988 trip

southward.

Manatee TNC-01 was unique in that she traveled from southeastern Georgia to

Brevard Co. twice in summer 1987. In September 1987, she made her final trip from

southeastern Georgia to Brevard Co. where she spent the winter. In 1988, manatee TNC-

01 traveled to Georgia in April, but remained only for a few months before returning to

Brevard Co.. Other manatees, such as TBC-06 and -23, were very predictable in that

they made one trip north to Georgia in spring and one trip south to Brevard Co. in late

47

summer or autumn (Fig. 8a and b). Appendix D summarizes the movements of manatees

radio-tracked during this study.

Areas of Greatest Use by Manatees

Cold season. Tagged manatees aggregated in two locations during the cold

season: the warm-water discharges of CCA in Fernandina Beach and GPC in St. Marys

(Figs. 3 through 6; Fig. 9). At least 13 manatees have been observed simultaneously at

the GPC discharge and at least 7 manatees have been observed at one time in the CCA

discharge. Manatee TNC-04 was observed in the Brunswick Pulp and Paper Company

(BPPC) warm-water discharge in Brunswick, GA, during spring 1988. Manatees moved

readily between the GPC and the CCA warm-water discharges during the cold season,

a swimming distance of approximately 16.5 km. Five of the eight manatees that were

captured and radio-tagged at the CCA discharge were also observed at the GPC discharge

during the cold season. Additionally, at least two other free-ranging manatees,

identifiable by their scar patterns, were observed at both the CCA and GPC discharges.

It is unlikely that manatees travel the 55 + km between Cumberland Sound and Brunswick

during very cold weather, although, manatee TNC-04 traveled from the CCA discharge

to the BPPC discharge in spring when ambient water temperatures had risen above 190C.

Manatees in Fernandina Beach left the CCA discharge to feed nearby and were

located in such areas as Egans Creek, the margins of Little Tiger Island, St. Joseph

Creek, and the vicinity of the confluence of Lanceford Creek, Bells River, and Amelia

River. The greatest distance from the CCA discharge at which I observed a manatee

48

feeding was approximately 4 km in the mouth of Lanceford Creek. Manatees were often

found in small, dendritic creeks of a main creek, such as Egans or Lanceford creeks, or

rivers.

Manatees that used GPC discharge remained primarily in North River to feed.

They, too, were often found in the dendritic creeks of North River. Manatee TNC-05,

for example, repeatedly fed in a small creek approximately 2 to 2.5 km north of the

discharge (Fig. 10). Manatees TNC-05, -06, and at least two other manatees were

observed feeding in a dendritic creek just north of New Point Peter Road. Manatees

TNC-05 and -02 were documented feeding in small dendritic creeks of North River.

Manatees were also observed feeding along marshes bordering North River, and a 1989

survey of North River for the presence of manatee grazing evidence indicated that the

majority of feeding occurred north of the GPC warm-water discharge (Zoodsma, unpub.

data). Water temperatures in the Cumberland Sound region were recorded as low as

14.50 C during the cold season.

Radio-tagged manatees dispersed from the industrial warm-water effluents in

spring when ambient air and water temperatures reached above 190C: TNC-03 and -04

dispersed from the GPC warm-water discharge when ambient water temperatures had

risen to 190C, TNC-07 left the GPC warm-water discharge for the last time when

ambient water temperatures were between 17.5 and 220C, TNC-08 dispersed from the

CCA warm-water discharge when ambient water temperatures had risen to at least

19.5 C, and TNC-05 and her calf left the GPC for the final time when ambient water

temperatures were between 26 and 280C. Manatees either moved further north in

49

Georgia or south to Florida. Manatees TNC-02, -03, and -04 moved as far north as the

Brunswick, Glynn Co., vicinity; and TNC-05 and -07 eventually moved south to St.

Johns River in Jacksonville, FL.

Warm Season. The locations of radio-tagged manatees were dispersed throughout

Cumberland Sound and as far north as Glynn Co., including Dents Creek, Altamaha

River, and the Brunswick waterfront; however, the Cumberland Sound and Satilla

River/White Oak Creek regions were used most often. Three manatees moved from

Brevard Co. to southeastern Georgia, primarily KBNSB, during more than one warm

season: TBC-06 in three consecutive years, and TNC-01 and TBC-23 in two consecutive

years. All three manatees demonstrated remarkable site fidelity (Figs. 11, 12, 13, and

14).

Six of seven manatees that were radio-tracked in southeastern Georgia during

warm weather were located in KBNSB at one time or another. Manatee TBC-19 may

also have been in KBNSB; however, her transmitter malfunctioned, making it difficult

to track her movements. Manatees such as TBC-06, TBC-23, TGA-01, and S-18 seemed

to localize in Kings Bay and left only during short jaunts to the adjacent Marianna Creek

and Crooked River marsh systems, usually to feed (Fig. 15). Manatees that utilized the

KBNSB area typically moved no more than 3 or 4 km a day. Two locations within

KBNSB were frequently used by manatees: 1) the area near the Port Services T-Pier

where an artesian well had previously been plugged, and 2) a shallow, cove-like area near

50

Crab Island. A third location, near the Explosives Handling Wharf (EHW), was used

often but not as frequently as the other two locations (Fig. 16).

Manatees also frequented the Fernandina Beach waterfront area. Fernandina

Harbour Marina was visited most often by radio-tagged manatees, but I also documented

tagged manatees at the sewage discharge just south of the Fernandina Beach waterfront.

Manatees TBC-06 and TGA-01 were often radio-located and observed near the

northern shore of Little Tiger Island east of Tiger Creek. Manatees TNC-04 and -06

were also observed in this area, but less frequently than TBC-06 and TGA-01. Manatee

TBC-06 was also radio-located in Tiger Basin.

Other areas in the Cumberland Sound region that were frequently used by tagged

manatees include Drum Point Island, the two Cumberland Island NPS Docks, the marsh

islands opposite of Greyfield Inn on Cumberland Island, and the marsh complex near

Stafford Island, Flood Island, and Oldhouse Creek.

Satilla River and White Oak Creek were used by four manatees: TBC-06, TBC-

23, TNC-07, and TBC-19. Manatees TBC-06 and -23 were often observed with other,

untagged manatees in this region. Boats were rarely observed during radio-tracking

excursions in the Satilla River and White Oak Creek region. Water salinity values for

areas where manatees were radio-tracked are found in Appendix E.

Aerial Surveys. Thirty-seven aerial surveys were conducted from May to

September, 1988, and from April to August, 1989. Total survey time was 18.3 and 36.5

h in 1988 and 1989, respectively. Manatees were spotted during every survey except for

three in 1988. The number of manatees observed per survey ranged from 0 to 10 in 1988

51

and from 2 to 11 in 1989 (Table 1). A mean of 3.33 manatees were seen per hour of

survey time in 1988 and in 1989 a mean of 4.14 manatees were seen per hour of survey

time.

Most of the manatee sightings during late spring and summer were in KBNSB and

Cumberland Sound between Crooked River, the southern end of Cumberland Island, and

northeastern Tiger Island (Fig. 17). Manatees were seen near the GPC warm-water

discharge in April 1989. Few manatees were seen between Crooked River and Table

Point, Cumberland Island (Fig. 17).

Boats were recorded during 12 surveys in summer 1989 and were observed

throughout the survey zone (Fig. 18). Several boats were observed in Amelia River

anchored off Fernandina Harbour Marina and large vessels frequently anchored just off

Seacamp Dock, Cumberland Island National Seashore (CINS). Fishing boats were seen

throughout Cumberland Sound but primarily by the jetties, the southern end of CINS, the

northern end of Amelia Island, and Drum Point Island (Fig. 18).

The Cumberland Oueen, a large boat that transports visitors to and from

Cumberland Island, was seen during 9 of the 12 surveys when boat locations were being

recorded. On one occasion I was observing two manatees travelling north along the

western side of CINS when they were passed by the Queen as they approached Dungeness

Dock. The manatees appeared to be startled by the ferry because they stopped suddenly.

I could not make out what they were doing as they stirred up a considerable amount of

mud. After approximately 30 to 60 seconds the Queen had docked and the manatees

continued their northward journey. The two manatees swam under Dungeness Dock as

52

approximately 20 to 30 visitors were disembarking from the Cumberland Queen. There

was no indication that the receiving park ranger, ferry captain, or any of the visitors had

noticed the manatees as they swam directly beneath the dock.

Diurnal Time-Activity Patterns

Relationship of observation period to frequency of activity bouts and length of

bouts and sessions. Ninety-one observation periods were included in the behavioral

analyses: 39 in the cold and 52 in the warm season. Observation periods ranged from

1 to 7.92 h (mean = 4.38) in the cold season and from 1 to 9.25 h (mean = 3.74) in the

warm season (Fig. 19). The number of observation periods per manatee varied greatly,

ranging from 3 (TNC-08) to 30 (TBC-06) (Figs. 20a. and b.).

There was a positive correlation between the observed number of Spartina-feeding

bouts and the length of observation periods, especially during the warm season. The

majority of observation periods in both seasons contained 1 to 3 Spartina feeding bouts

(Appendix F). Mean duration of Spartina-feeding sessions also increased with length of

observation period, but only during the cold season (Appendix G).

The number of observed algal-feeding bouts was positively correlated with

observation period length in the warm season (Appendix H). The majority of warm-

season observation periods contained 1 to 2 algal feeding bouts, but 3 to 4 bouts were not

uncommon in longer observation periods (Appendix H). Mean algal-feeding session

durations in the warm season remained nearly constant over varying lengths of

observation periods (Appendix I). The relationship of the length of observation period

53

to algal-feeding bout frequency and session length was not plotted for cold season because

of the small sample size (n =3).

There was a positive correlation between the number of resting bouts observed and

observation period length in both seasons (Appendix J). The maximum number of resting

bouts in any warm season observation period was eight; the majority of observation

periods contained 1 to 2 bouts of resting. The maximum number of resting bouts in any

cold season observation period was nine; the majority of the cold-season observation

periods contained 1 to 5 resting bouts (Appendix J). The four longest resting bouts were

observed in observation periods of less than 4 h. Resting session lengths appeared to be

unaffected by observation-period durations (Appendix K).

Cavorting bout frequencies per observation period increased with observation

period length in both seasons (Appendix L). Most observation periods (both seasons)

contained 1 to 3 cavorting bouts. An apparent inverse relationship existed between mean

cavorting session lengths and observation-period durations in the warm season; however,

mean cavorting session lengths in the cold season appeared to increase with observation

time during the cold season up to 4 h, and then decrease (Appendix M).

The number of moving bouts increased with an increase in observation period

duration (Appendix N). Observation periods contained from 1 to 8 bouts of moving,

however the majority of observation periods contained 1 to 6 moving bouts. Mean

54

moving session lengths increased with observation time during the cold season but

decreased with observation time over 4 h during the warm season (Appendix O).

Activity over both seasons. The overwhelming majority of my observations of

manatees feeding on Spartina occurred during high tide. In contrast, very few of my

observations of manatee Spartina-feeding occurred during low tide (Fig. 21a). The

overall effect of tide stage (seasons pooled together) had a significant effect on the

proportion of time that manatees fed on Spartina (P = 0.0003). Manatees fed on

Spartina for a significantly greater proportion of time during high tide than during low

and mid-tides ( = 0.0001 and 0.0263, respectively), and during mid-tide than during

low tide (P = 0.0029).

Manatees rested or cavorted most often during low and mid-tides. I observed

manatees resting most often during mid-tide (Fig. 21b). Tide stage was found to

significantly affect the proportion of time that manatees rested (P = 0.0005). Manatees

rested for a significantly greater proportion of time during low and mid-tides than high

tide (P = 0.0002 and 0.0004, respectively); there was no significant difference between

low and mid-tide.

The proportion of time that manatees cavorted was also affected significantly by

tide stage (P = 0.0001). I most often observed manatees cavorting during low tide (Fig.

21c). Manatees cavorted for a significantly greater proportion of time during low tide

55

than mid- or high tides (P = 0.0002 and 0.0001, respectively), and during mid-tide than

during high tide (P = 0.0320).

Moving was most often observed during mid- and high tide (Fig. 21d). Overall,

however, tide stage did not have a significant effect on the proportion of time that

manatees moved (P = 0.3509).

Season, adjusted for the effect of tide stage, also influenced manatee activity.

Manatees rested for a significantly greater proportion of time during the cold season than

they did during the warm season (P = 0.0016). Manatees fed on algae (P = 0.0131) for

a significantly greater proportion of time during the warm season than they did in the

cold season. The number of algal-feeding observations in the cold season was very low

(n=3). Season did not significantly effect the proportion of time that manatees fed on

Spartina (P = 0.7451), moved (P = 0.9251), or cavorted (P = 0.1313). There were

no significant interactions between season and tide stage for any behavioral category

except resting (P = 0.0242), although it approached significance for Spartina feeding (P

= 0.0614).

The mean duration of Spartina-feeding sessions was significantly longer during the

cold season (99.6 min) than the warm season (46.1 min) (P = 0.0153). Mean session

lengths for resting, cavorting, and moving were not significantly affected by season

(Table 2). However, when the subcategories of "moving" ("milling" and "travelling")

56

were tested separately, milling sessions were found to be significantly longer during the

warm season than during the cold season (P = 0.0358) (Table 3).

Activity in the cold season. Manatee activity in the Cumberland Sound region

during the cold season pivoted around the warm-water discharge areas of GPC and CCA.

Manatees remained in warm-water discharge areas resting or cavorting, or fed in nearby

areas. Three of the four manatees that were radio-tracked in cold weather during 1989

(TNC-05, -06, and -07) and all manatees observed during the 1987 and 1988 cold seasons

(n=3) fed exclusively on Spartina during observation periods. Only manatee TNC-08

fed on benthic marine macroalgae during cold season observation periods.

Manatees fed on Spartina primarily at high tide: 76.3% of all Spartina-feeding

observations (all bouts combined) in the cold season occurred at high tide and only 23.7%

occurred at mid-tide (Fig. 22a). I never observed manatees feeding on Spartina at low

tide in the cold season.

The adjusted mean proportion of time that manatees spend in each major

behavioral category during each tide stage was determined from an ANOVA (Fig. 23).

Manatees spent a significantly greater proportion of total observation time feeding on

Spartina at high tide than at low or mid-tide (P = 0.0003 and 0.0009, respectively).

There was no significant difference between the proportion of time that manatees spent

feeding on Spartina at mid- and low tides (P = 0.1239). Feeding sessions averaged 99.6

min in length (Fig. 24). The longest feeding session, 235 min, was recorded while

57

observing manatee TNC-05 (a 340-cm female with a small dependent calf) feeding on

Spartina in North River.

Resting was observed more than any other behavior. The largest percentage of

resting observations occurred during mid-tide (Fig. 22b). Manatees rested significantly

more at low and mid-tides than at high tide (P = 0.0017 and 0.0001, respectively).

There was no significant difference between the proportion of time that manatees spent

resting at low tide and mid-tides (P = 0.6400). Manatees rested on the bottom as well

as at the surface, and the mean duration of a resting session was 58 min (Fig. 24).

During the cold season, cavorting was observed less frequently than any other

behavior. A major portion of the total observations of cavorting occurred during low tide

(Fig. 22c). Manatees cavorted for a significantly greater proportion of time at low tide

than at mid- or high tide (P = 0.0013 and 0.0008, respectively). There was no

significant difference between cavorting at mid- and high tides (P = 0.6169).

Cavorting sessions averaged 28.2 min with a maximum of 120 min during the cold season

(Fig. 24).

Moving was observed during all three tidal stages with the majority of

observations of moving occurring at high and mid-tides (Fig. 22d). The proportion of

time that manatees moved during each tide stage, however, was not significantly different

(P = 0.9216, 0.7305, and 0.7569 for low and mid-, low and high, and mid- and high

tide, respectively). Moving sessions averaged 35.8 min with a maximum of 140 min

(Fig. 24). It appeared that most moving occurred as manatees traveled to and from their

feeding locations; however, the proportion of time spent in "travelling," a subcategory

58

of "moving," did not differ significantly among tide stages (P = 0.3554, 0.1597, and

0.5058 for low vs. mid-, low vs. high, and mid- vs. high tides, respectively).

Adjusted mean proportion of time that manatees spend in each behavior by tide

stage during the cold season are shown in Figure 23. The activities of resting and

Spartina feeding are inversely related, as are cavorting and feeding to a lesser degree

(Fig. 23). Manatees rested at low and mid-tides and fed at high tide. There was no

obvious tide-related moving pattern. Algal feeding was not included in the model since

only a few cold season observation periods contained that behavioral observation.

Consequently, values in each tide stage do not equal 100%.

Activity in the warm season. Manatees were observed feeding on Spartina mostly

at high and mid-tide during the warm season. Very few of my observations of Spartina

feeding occurred at low tides (Fig. 25a). Manatees fed on Spartina for a significantly

greater proportion of time at high tide than at low tide (P = 0.0154). There was no

significant difference in the proportion of time that manatees spent feeding on Spartina

at mid- and low tides (although it approached significance: P = 0.0564), and there was

no difference between mid- and high tides (P = 0.4209). The mean duration of feeding

sessions during the warm season was 46.1 min with a maximum of 135 min (Fig. 26).

Manatees also fed a great deal on macroalgae during the warm season. Algal

feeding was observed nearly equally among the three tide stages (Fig. 25b). There was

no difference in the proportion of time that manatees spent feeding on algae between low

59

and mid-tides ( = 0.6195), low and high tides (P = 0.8230), or mid- and high tides

(P = 0.4358).

Resting was frequently observed during mid- and low tides (Fig. 25c). The

proportion of time that manatees rested was significantly greater during low tide than

during high tide ( = 0.0456). There was no difference between low and mid-tides (

= 0.2358) and mid- and high tides (P = 0.2869). The mean duration of resting sessions

during the warm season was 43.8 min (Fig. 26).

Manatees were consistently observed cavorting during low and mid-tides (Fig.

25d). Manatees cavorted for a significantly greater proportion of time during low and

mid-tides than during high tide (P = 0.0002 and P = 0.0217, respectively), and they

also cavorted for a significantly greater proportion of time during low tide than during

mid-tide (P = 0.0143). The mean duration of cavorting sessions during the warm season

was 39.4 min (Fig. 26).

Manatees were observed moving mostly during high and mid-tides, but manatees

shifted locations during low tide as well (Fig. 25e). The ANOVA indicated that tide

stage had no significant effect on the proportion of time that manatees moved, although

the difference between low and high tides approached significance at P = 0.0924. The

mean duration of the warm-season moving sessions was 39.3 min (Fig. 26).

When "travelling" was tested separately from the other subcategories in moving,

the proportion of time that manatees traveled was not found to be significantly different

between low and mid- (P = 0.3107), mid- and high (P = 0.3530), or low and high tides

60

(P = 0.0814). The mean travelling session duration was 39.4 min. "Milling" sessions

averaged 29.8 min during the warm season.

The ANOVA-adjusted mean proportion of time that manatees spend in each major

behavioral category during each tide stage in the warm season is shown in Figure 27.

Inverse relationships exist between Spartina feeding at high tide and resting and cavorting

at low tide. Algal feeding occurred in all three tide stages (Fig. 27).

Food Habits and Feeding Ecology

Feeding observations. Manatee consumption of Spartina alterniflora was observed

most frequently; however, during the warm season manatees also fed on marine algae.

Ulva sp. and Gracilaria sp. were occasionally seen protruding from the mouths of feeding

manatees as they surfaced to breathe. Manatee TNC-08 was the only manatee observed

feeding on algae during cold weather. Manatee TBC-06 and another manatee were

observed feeding on Pontederia sp. in a dendritic creek of Satilla River on a few

occasions. Leaves and stalks of the Pontederia were consumed by the feeding manatees.

Other plant species were available in the Satilla River but I never saw manatees consume

them. One manatee in the CCA discharge was observed with a cannonball jellyfish

(Stomolophus meleagris) in its mouth. I am not sure whether the manatee was actually

consuming the jellyfish or merely "playing" with it.

Manatee feeding ranged from intermittent feeding on vegetation while milling or

cavorting, to uninterrupted intensive feeding along the water's edge. Consumption of the

tall form of Spartina (Ornes and Kaplan 1989), which was most accessible to manatees,

61

occurred as manatees moved slowly along marsh edges. They were never observed to

feed for an extended period in one localized area and eventually "tunnel" directly into the

vegetation, as reported by Hartman (1979) and Reynolds (1981) for other areas.

Manatees would take Spartina from just below, at, or just above the water's

surface. From my boat I often saw an individual Spartina plant quiver or tremble before

it disappeared below the surface. I suspected that a submerged manatee was grasping the

lower portions of the plant with its mouth and the plant was pulled under as it was

consumed. I never observed manatees using their pectoral flippers to knock plants down

or push plants into their mouths. When manatees took Spartina at the water's surface,

I saw the tops of their heads and snouts as the plants were consumed.

Occasionally manatees were observed to support themselves along the shoreline

using their pectoral flippers, as they grazed Spartina shoots from above the water's

surface. Manatees that fed in this way sounded similar to what one would hear if

listening to a cow or horse grazing: the shearing of vegetation followed by a distinct

munching sound as the vegetation was ground by molars. Shoots that had been grazed

intensively were 5 to 15 cm tall. Leaves were often cropped back to within 2 to 5 cm

of the plant stems.

Manatees fed alone as well as in groups. Feeding manatees often submerged

quietly when a boat approached and would resume feeding when the boat had passed by.

On one occasion feeding manatees (including TNC-05) were startled by a large shrimp

62

boat as it reversed its engines in order to dock. The manatees thrashed violently, causing

a succession of tremendous splashes, and moved quickly into deeper water.

Stomach contents. Microhistological analyses (Hurst and Beck 1988) of stomach

contents taken from dead manatees that were recovered from southeastern Georgia and

the Cumberland Sound region supported feeding observations made while radio-tracking

manatees (manatee IDs and collection locations and dates appear in Appendix P).

Spartina was found in all stomach samples and also formed the highest mean percentage

of contents (Table 4). There was a low frequency of Gracilaria in manatees examined

(2 of 15); however, it constituted a high percentage of content when it occurred (64 and

68%). Ulva was also relatively low in frequency (4 of 15 specimens), but sometimes

high in percentage content: 40% in one stomach sample and 50% in another. In

specimens with low percentages of Gracilaria in stomach contents, Gracilaria and Ulva

may have been present but identified as "unknown alga." Either stem or rhizome

fragments were found in every sample, but they were low in percent composition.

Female and male diets were similar (Fig. 28a). Manatees that were recovered during

warm weather had a higher percentage of Spartina in their stomachs than did manatees

collected during cold weather (Fig. 28b).

Fig. 8. Month of each year that radio-tagged manatees were documented to migrate
between southeastern Georgia and Brevard Co., FL: a) Spring and early summer
migrations from Brevard Co., FL, to southeastern Georgia; b) Late summer and autumn
migrations from southeastern Georgia to Brevard Co., FL.

64

a. Spring and Early Summer

TNC- 0 188

TGA-01

TBC-23 188

TBC-08 1

2 TBC-08 1988, 1989 1987

8-18 1989

S-12 1988
0
Mar Apr May Jun Jul
Month

b. Late Summer and Autumn

TNC-01 1987

TGA-01 1989

TBC-23 1 989

0
TBC-OB 1987 1988

S-18 8

S-12

Aug Sep Oct Nov Dec
Month

Fig. 9. Locations of radio-tracked manatees in the Cumberland Sound region during the
cold season in 1989.

Z r7

=:7

+._ --:b
4. +

be

Fig. 10. 1989 cold-season locations of manatee TNC-05.

7-. -- 7- =- Z _7=*:-.-_: -- ^ F -:-: -;-

^^:-r+ ^^y
Z..
--Iim--m_
77I: H

~~~~~~~~~~~~~~ ........ -II im . IIl M

+ ~71
=77

=7 +

+7

Fig. 11. Map illustrating warm-season site fidelity of TBC-06 from 1987 to 1988
(e=1987, +=1988).

4:

7L-----:- ,r-l.

97 0
++_

+ z7 0
.-77I.
-~~IA ? I

S L --

4.-- -

-4.--r. .

4.- ---
1------ -.r-

+4 7.rr.-~~''-

0- -- i---
0 ~ -~-------

0* ---=--

*c -
4 0~~~;=;; ~ c

+0
0.- ___ --

Fig. 12. Map illustrating 1989 warm-season locations of TBC-06. These locations are
similar to it's warm-season locations in 1987 and 1988 (Fig. 11).

? '.- ---_-_- _.
S-- -
7'_ _-.7

_.. :z.-7

i. :-=_- =7:

Z 2 -. 5 -: :- .-:: 2 - 7- : --.r.' -2 -

77--." .7_- :-.7 -_-:7'- : .

-. :- -Z-: 2 "-- ---' .. .. .. --- .-----_---

-_:..: _- --7_: -:.

-.- --: -: =- _-- 7- --- -= .-.: -.- .-7 -

---- ----~

a_

z--

7 -

Fig. 13. Map illustrating warm-season site fidelity of TBC-23 from 1988 to 1989
(* = 1988, + = 1989).

--- S

.. .-. ..
:: .. -.. --: ii ._. -_-

-. ;- ---

--.-_:r_:__:-

r~

:-' .--`--~------_-

L-

z~-:

1_
i__.

I :I

-_ .I

-- '-~I`-:-- --

:-

------
-- ----
-r
--------
-----
-- ----
-r_~.-~~r~- -----r-: r_
-;- ---- ----
-- -
~--.-----=r._.~
-------- -
------~-~-..._~ ~r;

--_.

4- --- : --
-- .--- .
.- .

.- ---------.
-. .. -- .._-- : : -

--.

Fig. 14. Map illustrating locations of TNC-01 in 1987 and 1988 (+ = 1987, = 1988).

Z------. a-7---7= 4

::7.

=Z- .- i+

4. -

-. .. -. .. .
:

.-7..Z

4 ----- _

+ .--

+I. -.
+..2",'
+--:

7-7-

Fig. 15. Digitized locations of radio-tracked manatees in the Cumberland Sound region
during the warm season in 1989.

-7-
;:__~____L_.--I_.-_

.. .IM l 4 o --._ =-.-

--.--_ -----:- _-_ .-

..- :7'-i.. ... .- .1 -

-_-- -: .
--L.:--------== _-:. Y-_----' _-- : -
.-~. -.:_--.-.:---_---___ .-- :.-_- _I=T--:--.!. _-_ :..-..:-.

f -._- :-= -- : ::
_I ~~... -.=:I .:. ..-.>-~~ .._ _:_.

. . ..... ... .. ... .
-:..-_ __.;_- -:_::. -= ~ _-_:.._- -_ -- =-''-.-_

. I- . . . .. .

_. _. .. ..- .? "
:. : .. sZ - z=:- _._ _-- .- :. .

* -

m

4+:

7+7

---+

+ -2-f-_~-

Fig. 16. Areas in Kings Bay used frequently by radio-tagged manatees. Note: areas are
indicated by (///).

Table 1. Number of manatees observed during each aerial survey conducted over
Cumberland Sound and adjacent waters in 1988 (left side) and 1989 (right side). Yearly
means and their standard deviations are given at the bottom.

NO. OF NO. OF
DATE MANATEES DATE MANATEES

20 May 1988 2 06 Apr 1989 9
03 Jun 1988 1 13 Apr 1989 9
13 Jun 1988 4 19 Apr 1989 11
22 Jun 1988 9 26 Apr 1989 10
29 Jun 1988 0 03 May 1989 7
12 Jul 1988 10 10 May 1989 9
19 Jul 1988 6 18 May 1989 7
29 Jul 1988 0 24 May 1989 9
02 Aug 1988 1 31 May 1989 4
30 Aug 1988 8 07 Jun 1989 6
02 Sep 1988 1 14 Jun 1989 11
09 Sep 1988 0 21 Jun 1989 9
16 Sep 1988 3 28 Jun 1989 8
23 Sep 1988 10 05 Jul 1989 5
30 Sep 1988 6 14 Jul 1989 2
19 Jul 1989 6
28 Jul 1989 5
02 Aug 1989 7
09 Aug 1989 3
16 Aug 1989 4
23 Aug 1989 8
30 Aug 1989 2

X 4.1 X 6.9
SD 3.9 SD 2.7

Fig. 17. Locations of manatee sightings during aerial surveys conducted in summer 1988
and spring and summer 1989 in Cumberland Sound.

-'-~-
rs

-~

---~

-=

`1.

7

----.----
i --s
_~__
;---~-~ --
r ----
--- ---~ ----lr:-,rr~~
.-r
.- c-

-- .--;-
..---. -.. s.;-
'---. I --~-r--
,r- S--~T;

~-,r --~;--1-----.--
~~C-----~-
--1-I~ -~"
T.- ._: *~:
Cc

i--
i_,
'" i- .r-
------
-----~,;-
---- --~- ~-
-~--~-~3;~
.-----=
~---- ~T~;~-~-~=~------~=~-::.
: -c-.. j,----
;e ----` --
r_-.---.--
n "'
5--=
I~ i--
----'----~- --= "-;-

--- ---~ ---
1----'=_,. n -

-~---~-~~: ~
"'S
-- -~
--~L-- r=------~-~ jT-~ --
---~--1.
--
... 5=
s-c.~ ;-~----
-s,-- -------
---s'-C
TTT- ;r---~c~-T
5i
-----r
--,.-
-c ----
-- -~--i 7
"-' -I-~

~
-jr
i -"

-.-- --~=
----
--
----:--
:;e
-1.. i--~

0

0 16

--r. --I-
----

:--c
:---- ..--
---
--
---.-i--
-----II~--~
---
-------- ;--1----;--
c~- ;.-
-- --'- ;---;
--~--- ---
----

----J- --- _~---r
-- --

---- --I ---1
---~~
~~...~ -------- -_.---1 --:--
.-- ---- :
---
--

--
"----
-'
C-

r.
--r-

~-=.-~
:: I- '~ -'---
---r,--l
~-- _s::
~' :c- -----.:---
-.-- --
c----
.---

--
---

--

Fig. 18. Locations of boats sighted during manatee aerial surveys in June, July, and
August, 1989.

+

'I

t .

_+ .. +-
- -i

::. : ":: -- .+ :.* -

.. *:: .: :! : : -.-: ... +

4

+ +-

+

.11+
#c

LJIII

+ q .

A t

+4+
- *::-I 4. .

* ....

+
+ -.

,+ ,CI

:^^)& i

~T.j~

I,

+

" :

.-- ; : :-- : ;_' ~

r,:i

r
,

i

.'Z "

::
i-

0
- 12-

S10-

4
S8--

I .
O i' '.

2-

0
1-<2 2-<3 3-<4 4-<5 5-<6 6-<7 7-<8 =>8
Duration (h)

SCold Season EJWarm Season

Fig. 19. Frequency distribution of the durations of manatee observation
periods (h) for cold and warm seasons.

87

a. Cold Season

TNC-05 TNC-06 TNC-07 TNC-08
Manatee ID

b. Warm Season

TBC.06 TBC-23 TGA-01 S-18
Manatee ID

Fig. 20. Total number of observation periods recorded for each of
the eight manatees included in the behavioral analyses: a)
Manatees radio-tracked in the cold season; b) Manatees radio-
tracked in the warm season.

30

25-
'1

0
' 20-

15-
o
'0 10-

z

MILLISECOND	CLASS.METHOD	MESSAGE
0	sobekcm_page_globals.constructor
0	sobekcm_page_globals.constructor	Application State validated or built
0	sobekcm_database.verify_item_lookup_object
0	sobekcm_page_globals.constructor	Navigation Object created from URI query string
0	sobekcm_database.verify_item_lookup_object
0	sobekcm_page_globals.display_item	Retrieving item or group information
0	sobekcm_page_globals.get_entire_collection_hierarchy	Retrieving hierarchy information
0	sobekcm_assistant.get_entire_collection_hierarchy
0	cached_data_manager.retrieve_item_aggregation
0	cached_data_manager.retrieve_item_aggregation	Found item aggregation on local cache
0	item_aggregation_builder.get_item_aggregation	Found 'all' item aggregation in cache
0	system.web.ui.page.page_load (ufdc.page_load)
0	sobekcm_page_globals.constructor.on_page_load
0	html_echo_mainwriter.add_style_references	Adding style references to HTML
0	html_echo_mainwriter.add_text_to_page	Reading the text from the file and echoing back to the output stream
194	html_echo_mainwriter.add_text_to_page	Finished reading and writing the file