• TABLE OF CONTENTS
HIDE
 Cover
 Dedication
 Table of Contents
 Frontispiece
 Introduction
 Florida manatee population biology:...
 Aerial surveys of manatees: A summary...
 Assessment of trends in sizes of...
 Fixed-width aerial transects for...
 Aerial survey as a technique for...
 The life history, pattern of breeding,...
 Age and seasonality in spermatogenesis...
 Age and reproduction in female...
 An automated photo-identification...
 Reproduction in free-ranging Florida...
 Reproduction and early-age survival...
 Reproduction and mortality of radio-tagged...
 Reproduction of the West Indian...
 Estimation of survival of adult...
 Trends and patterns in mortality...
 Analysis of watercraft-related...
 Integration of manatee life-history...
 Population biology of the Florida...
 Endpiece: Counties of Florida














Title: Population biology of the Florida manatee
CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00000121/00001
 Material Information
Title: Population biology of the Florida manatee
Series Title: Population biology of the Florida manatee
Physical Description: Book
Creator: O'Shea, Thomas J.
 Record Information
Bibliographic ID: UF00000121
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: ltqf - AAA0276
ltuf - AKQ7506
alephbibnum - 002069255
 Related Items
Other version: Alternate version (PALMM)
PALMM Version

Table of Contents
    Cover
        Page i
        Page ii
    Dedication
        Page iii
        Page iv
    Table of Contents
        Page v
    Frontispiece
        Page vi
    Introduction
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
    Florida manatee population biology: Research progress, infrastructure, and applications for conservation and management
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
    Aerial surveys of manatees: A summary and progress report
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
    Assessment of trends in sizes of manatee populations at several Florida aggregation sites
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
    Fixed-width aerial transects for determining dugong population sizes and distribution patterns
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
    Aerial survey as a technique for estimating trends in manatee population size-problems and prospects
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
    The life history, pattern of breeding, and population dynamics of the dugong
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
    Age and seasonality in spermatogenesis of Florida manatees
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
    Age and reproduction in female Florida manatees
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
    An automated photo-identification catalog for studies of the life history of the Florida manatee
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
    Reproduction in free-ranging Florida manatees
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
        Page 152
        Page 153
        Page 154
        Page 155
        Page 156
    Reproduction and early-age survival of manatees at blue spring, upper St. Johns River, Florida
        Page 157
        Page 158
        Page 159
        Page 160
        Page 161
        Page 162
        Page 163
        Page 164
        Page 165
        Page 166
        Page 167
        Page 168
        Page 169
        Page 170
    Reproduction and mortality of radio-tagged and recognizable manatees on the Atlantic Coast of Florida
        Page 171
        Page 172
        Page 173
        Page 174
        Page 175
        Page 176
        Page 177
        Page 178
        Page 179
        Page 180
        Page 181
        Page 182
        Page 183
        Page 184
        Page 185
        Page 186
        Page 187
        Page 188
        Page 189
        Page 190
        Page 191
    Reproduction of the West Indian manatee in captivity
        Page 192
        Page 193
    Estimation of survival of adult Florida manatees in the Crystal River, at Blue Spring, and on the Atlantic Coast
        Page 194
        Page 195
        Page 196
        Page 197
        Page 198
        Page 199
        Page 200
        Page 201
        Page 202
        Page 203
        Page 204
        Page 205
        Page 206
        Page 207
        Page 208
        Page 209
        Page 210
        Page 211
        Page 212
        Page 213
        Page 214
        Page 215
        Page 216
        Page 217
        Page 218
        Page 219
        Page 220
        Page 221
        Page 222
    Trends and patterns in mortality of manatees in Florida, 1974-1992
        Page 223
        Page 224
        Page 225
        Page 226
        Page 227
        Page 228
        Page 229
        Page 230
        Page 231
        Page 232
        Page 233
        Page 234
        Page 235
        Page 236
        Page 237
        Page 238
        Page 239
        Page 240
        Page 241
        Page 242
        Page 243
        Page 244
        Page 245
        Page 246
        Page 247
        Page 248
        Page 249
        Page 250
        Page 251
        Page 252
        Page 253
        Page 254
        Page 255
        Page 256
        Page 257
        Page 258
    Analysis of watercraft-related mortality of manatees in Florida, 1979-1991
        Page 259
        Page 260
        Page 261
        Page 262
        Page 263
        Page 264
        Page 265
        Page 266
        Page 267
        Page 268
    Integration of manatee life-history data and population modeling
        Page 269
        Page 270
        Page 271
        Page 272
        Page 273
        Page 274
        Page 275
        Page 276
        Page 277
        Page 278
        Page 279
    Population biology of the Florida manatee: An overview
        Page 280
        Page 281
        Page 282
        Page 283
        Page 284
        Page 285
        Page 286
        Page 287
        Page 288
    Endpiece: Counties of Florida
        Page 289
Full Text




















U.S. DEPARTMENT OF THE INTERIOR
NATIONAL BIOLOGICAL SERVICE
WASHINGTON, D.C. 20240


INFORMATION AND TECHNOLOGY REPORT 1
AUGUST 1995







POPULATION BIOLOGY OF

THE FLORIDA MANATEE



Edited by



Thomas J. O'Shea,

Bruce B. Ackerman,

and

H. Franklin Percival


UNIVERSITY OF FLORIDA LIBRARIES








Dedication


Dedicated to colleagues in some 50 developing nations of the world who strive to maintain manatees in their faunal
heritage and to the memory of Tammy Dominguez and Amaury Villalba who lost their lives during an aerial survey of
manatees on the southern coast of the Dominican Republic, January 1995.










Contents

Page
Dedication .. .. . . .. . .. .. .. .. . . . .. .. ... .. . . . . iii
Frontispiece ................... ...... .................... .. .......... vi
Introduction, Thomas J. O'Shea, Bruce B. Ackerman, and H. Franklin Percival . . . . . . .....
Florida Manatee Population Biology: Research Progress, Infrastructure, and Applications for
Conservation and Management, John E. Reynolds, III ...... . . . . . . . . . . 6
Aerial Surveys of Manatees: A Summary and Progress Report, Bruce B. Ackerman . . . . . . ... 13
Assessment of Trends in Sizes of Manatee Populations at Several Florida Aggregation Sites,
Robert A. Garrott, Bruce B. Ackerman, John R. Cary, Dennis M. Heisey, John E. Reynolds, III,
and J. Ross Wilcox ................. . ................... ............34
Fixed-width Aerial Transects for Determining Dugong Population Sizes and Distribution Patterns,
Helene Marsh ................... .................. ............. 56
Aerial Survey as a Technique for Estimating Trends in Manatee Population Size-Problems and
Prospects, Lynn W. Lefebvre, Bruce B. Ackerman, Kenneth M. Portier, and Kenneth H. Pollock . . ... 63
The Life History, Pattern of Breeding, and Population Dynamics of the Dugong, Helene Marsh . . ... 75
Age and Seasonality in Spermatogenesis of Florida Manatees, Patricia Hernandez, John E.
Reynolds, III, Helene Marsh, and Miriam Marmontel ................... . . . . ..84
Age and Reproduction in Female Florida Manatees, Miriam Marmontel . . . . . . . ... . 98
An Automated Photo-identification Catalog for Studies of the Life History of the Florida Manatee,
CathyA. Beck and James P. Reid ....................................... ..120
Reproduction in Free-ranging Florida Manatees, Galen B. Rathbun, James P. Reid, Robert K.
Bonde, and James A. Powell .................. ........................ 135
Reproduction and Early-age Survival of Manatees at Blue Spring, Upper St. Johns River, Florida,
Thomas J. O'Shea and Wayne C. Hartley ................... ................. 157
Reproduction and Mortality of Radio-tagged and Recognizable Manatees on the Atlantic Coast of
Florida, James P. Reid, Robert K. Bonde, and Thomas J. O'Shea . . . . . . . . . . .. ....171
Reproduction of the West Indian Manatee in Captivity, Daniel K. Odell, Gregory D. Bossart,
Mark T. Lowe, and Thomas D. Hopkins (Abstract only) . . . . . . . ..... ......... 192
Estimation of Survival of Adult Florida Manatees in the Crystal River, at Blue Spring, and on the
Atlantic Coast, Thomas J. O'Shea and Catherine A. Langtimm . . . . . . . . . .... 194
Trends and Patterns in Mortality of Manatees in Florida, 1974-1992, Bruce B. Ackerman, Scott D.
Wright, Robert K. Bonde, Daniel K. Odell, and Donna J. Banowetz . . . . . . . . . . ... 223
Analysis of Watercraft-related Mortality of Manatees in Florida, 1979-1991, Scott D. Wright,
Bruce B. Ackerman, Robert K. Bonde, Cathy A. Beck, and Donna J. Banowetz . . . . . . . .. 259
Integration of Manatee Life-history Data and Population Modeling, L. L. Eberhardt and
ThomasJ. O'Shea ........... ...... ... .............. ........ .269
Population Biology of the Florida Manatee: An Overview, Thomas J. O'Shea and
Bruce B. Ackerman ......... . . . . . . ................... ........ 280
Endpiece ................ ... ................... ................ 289











































Frontispiece: Female Florida manatee (Trichechus manatus latirostris) with nursing calf Photo by P. M. Rose.










Introduction


by


Thomas J. O'Shea1

U.S. Fish and Wildlife Service
National Ecology Research Center
Sirenia Project
412 N.E. 16th Avenue, Room 250
Gainesville, Florida 32601


Bruce B. Ackerman

Florida Department of Environmental Protection
Florida Marine Research Institute
100 Eighth Avenue S.E.
St. Petersburg, Florida 33701


and


H. Franklin Percival2

U.S. Fish and Wildlife Service
Florida Cooperative Fish and Wildlife Research Unit
University of Florida
P.O. Box 110450
Gainesville, Florida 32611


Florida Manatees

The Florida manatee (Trichechus manatus latirostris)
is a unique element of the U.S. fauna. It is a distinct
subspecies of the West Indian manatee (Domning and
Hayek 1986) and one of the largest inshore mammals of
the continent, reaching weights to 1,650 kg (Rathbun et al.
1990). Annual migratory circuits of some individuals
through the intracoastal waterways of the Atlantic Coast
are 1,700 km round trips at seasonal travel rates as high as
50 km/day (*3Reid and O'Shea 1989; Reid et al. 1991),
resulting in one of the longest remaining intact mammalian
migrations in the eastern United States. Manatees are the


SPresent address: National Biological Service, Midcontinent Ecological
Science Center, 4512 McMurry Avenue, Fort Collins, Colo. 80525.
2 Now with the National Biological Service, same address.
3 An asterisk denotes unpublished material.


only living North American representatives of the small
mammalian Order Sirenia and are therefore the only em-
bodiment of the unique suite of biological features that
define the distinctive adaptive syndrome at the ordinal
level of the taxonomic hierarchy. Features of this adaptive
syndrome can be directly related to the lifestyle of aquatic
herbivory (O'Shea and Reep 1990). In manatees, these
features include unique aspects of morphology (pachyos-
tosis; horizontal, unilobular lungs and diaphragm; indeter-
minate numbers of molars undergoing continuous replace-
ment; dexterous forelimbs and prehensile lips; and a
hind-gut as long or longer than 30 m), physiology (in
particular an unusually low metabolic rate and a high
thermal conductance that lead to energetic stresses in
winter, ameliorated by migrations and aggregations in
warm-water refugia), and behavioral ecology (lack of a
rigid social organization, seasonal migrations, absence of
strong circadian rhythms). Hartman (1979) and Reynolds






2 INFORMATION AND TECHNOLOGY REPORT 1


and Odell (1991) provided further details on distinctive
features of manatee biology. Other aspects of this adaptive
syndrome include life-history traits, such as litter size, age
at first reproduction, interbirth intervals, and longevity,
which profoundly dictate the population consequences of
human intrusion on the manatee's unique lifestyle.
Human intrusions on manatees in Florida have been
substantial. Many manatees are accidentally killed by wa-
tercraft and other anthropogenic causes; traditional migra-
tions have been preempted or modified by coastal develop-
ment and industrially heated water (where aggregations of
hundreds of animals may form and are liable to catastrophic
losses), and human activities have impinged on habitat
quality throughout the range of the Florida manatee. Never-
theless, quantification of these impacts on the population has
been elusive, and available data frequently lead to enigmas
about some of the seemingly simplest questions (O'Shea
1988). How many are there? Is the population decreasing or
increasing? What are the potential rates of population
growth? What rates of mortality can manatees sustain with-
out being driven closer to extinction? How do these compare
with existing information on population size, observed num-
bers of dead manatees, and life-history traits?
These questions have high priorities for research on
Florida manatees. Manatees are protected under the Florida
Manatee Sanctuary Act of 1978, the U.S. Endangered Spe-
cies Act of 1973, and the U.S. Marine Mammal Protection
Act of 1972 and their amendments (*U.S. Fish and Wildlife
Service 1989). Federal management is directed to improve
the population status of Florida manatees so that the subspe-
cies can be removed from the endangered-species list. The
U.S. Marine Mammal Protection Act also requires that
Florida manatees must be maintained at optimum sustain-
able population levels and that the marine ecosystems of
which they are a part also remain healthy and stable. Mana-
tees have become increasingly popular with the public be-
cause of their biological uniqueness and their ability to exist
in densely populated areas, where some individuals become
accustomed to people and provide them with some of the
most singular, inspiring wildlife encounters of their life-
times. Some of these animals, however, also bear inhumane
and gruesome wounds from boat propellers, and images of
manatees killed by these means are given substantial media
attention, leading to deepening public concern.


Information on Manatee
Population Biology

Increasing public concern and requirements of legisla-
tion notwithstanding, the answers that can be provided to
questions about manatee population biology are only as
good as the quality of available information. Recognizing
this need for a new assessment of information about manatee


population dynamics, the U.S. Fish and Wildlife Service and
the Florida Department of Natural Resources held a techni-
cal workshop on manatee population biology at the Austin
Cary Forest of the University of Florida during 4-6 February
1992.4
This was the first workshop on the topic since a meeting
in Orlando in 1978 (*Brownell and Rails 1981). At the 1978
meeting, several recommendations were made for further
research, and conclusions were drawn about certain aspects
of manatee population biology, but these conclusions were
tentative because of the scant amount of then-available data.
Also in 1978, the U.S. Fish and Wildlife Service organized
the Sirenia Project at the Gainesville Field Station of the
National Fish and Wildlife Laboratory.5 One of the major
objectives for the Sirenia Project was long-term research to
expand databases on manatee life history, population size,
and mortality. Subsequently, in the early 1980's, the service
and the U.S. Marine Mammal Commission (through the
Florida Cooperative Fish and Wildlife Research Unit) em-
barked on a series of studies to address other issues in
manatee population biology, particularly aerial survey, cen-
sus, estimation of trends, and development of a preliminary
population model. These studies resulted in a series of
publications and reports, including those by Eberhardt
(*1982), Packard and Mulholland (*1983), Packard and
Nichols (*1983), Packard (*1984, *1985a, *1985b), and
Packard et al. (1985, 1986). However, conclusions of many
of these studies remained preliminary because of the large
degree of uncertainty about manatee population attributes.
Although databases had moved from being scant to modest
by the mid-1980's, continued long-term research and addi-
tional experimentation were necessary to improve certainty.
Population research was significantly augmented in 1985
when the state of Florida initiated manatee research at what
is now the Florida Marine Research Institute. This program
encompasses several areas of study in manatee population
biology, including aerial surveys, research with telemetry,
habitat studies, and the statewide recovery of carcasses and
analyses of mortality data.

4 In 1993, the research functions of the U.S. Fish and Wildlife Service
were transferred to the National Biological Service, and the Florida
Department of Natural Resources was merged with the Florida
Department of Environmental Regulation to form the Department of
Environmental Protection.
5 The National Fish and Wildlife Laboratory of the U.S. Fish and Wildlife
Service was responsible for all of the service's marine mammal research
under the Marine Mammal Protection Act and began federal manatee
research in 1973 prior to formal establishment of the Sirenia Project in
1978. The Sirenia Project was subsequently administered by the Denver
Wildlife Research Center of the service after the Denver Wildlife
Research Center and the National Fish and Wildlife Laboratory were
merged. The National Ecology Research Center of the service later
administered the Sirenia Project after the transfer of many other Denver
Wildlife Research Center programs to the U.S. Department of
Agriculture. Currently the Sirenia Project is conducted by the
Southeastern Biological Science Center of the National Biological
Service.






INTRODUCTION 3


Because the Sirenia Project, the Florida Marine Re-
search Institute, and the Florida Cooperative Fish and
Wildlife Research Unit have been involved in much of the
past and present research about manatee population biol-
ogy, sponsorship of the 1992 workshop by these institu-
tions was appropriate. However, significant progress in
understanding manatee population biology has also been
due to the contributions of many other individuals and
organizations. In addition to findings of federal and state
agencies, the workshop also included contributions by
oceanaria, industry, and the academic community. Much
of the research on manatee population biology presented
at the workshop was called for in the revised Florida
Manatee Recovery Plan (*U.S. Fish and Wildlife Service
1989).


Objectives of Workshop
Participants

The goals for the 1992 workshop were (1) a synthesis of
information about manatee population biology, (2) an evalu-
ation of the strengths and weaknesses of data sets and
approaches to manatee-population research, and (3) recom-
mn,ndations for future research. The primary objective for
ne revised Florida Manatee Recovery Plan (*U.S. Fish and
Wildlife Service 1989) was to change the status of the
subspecies from endangered to threatened. According to the
plan, this objective is met when viable, self-sustaining popu-
lations are maintained on Florida's Gulf and Atlantic coasts.
The criteria for a change of status encompasses four ele-
ments: the population is growing or stable, mortality factors
are controlled at acceptable levels or decreasing, habitats are
secure, and threats to habitats are controlled or decreasing.
The recovery plan also called for the development of popu-
lation models to assist in assessing progress toward meeting
the first of these criteria if methodology and data are avail-
able. One of the tasks of the workshop participants was the
assessment of existing data and research.
The duration of the workshop was 3 days. On the first
day, 14 invited papers were delivered to an audience of
approximately 75 individuals, including members of the
recovery team, researchers and managers from govern-
ment agencies, members of the academic community, and
representatives of industry and conservation groups. The
paper session was followed by a 2-day retreat of group
meetings and informal evening sessions. The group ses-
sions were devoted to aerial-survey techniques and esti-
mation of population size and growth trend, reproduction,
age structure, mortality, photo identification, estimation of
survival, and integration of life-history data and popula-
tion modeling. The groups were limited to a small number
of researchers in manatee population biology and ex-
perts in statistics and mammalian population dynamics.


Individuals were assigned to various groups and were
provided a list of topics for discussion. Each group pro-
vided a written summary and set of recommendations.
Abstracts of presentations, group reports, and lists of at-
tendees were assembled in an interim report with limited
distribution (*O'Shea et al. 1992).
The proceedings contain the peer-reviewed presented
papers, except the information on reproduction in captive
manatees (summarized by Odell et al. 1995) and the life-
history information based on field studies of the wild
population at Blue Spring (O'Shea and Hartley 1995).
Development of the latter paper was a recommendation of
the workshop participants. Several papers include results
through calendar year 1992 or later, but most papers report
findings through 1991. Most papers benefitted from com-
ments by participants in the 1992 workshop and were
modified with feedback in response to their verbal presen-
tation. The contribution on population modeling (Eber-
hardt and O'Shea 1995) was completed after results of
related studies became available in manuscript form. Par-
ticipants in the 1992 working group sessions were: B. B.
Ackerman, Florida Marine Research Institute; C. A. Beck,
Sirenia Project; I. E. Beeler, Florida Marine Research
Institute; R. K. Bonde, Sirenia Project; D. P. DeMaster,
National Marine Fisheries Service, La Jolla, California;
L. L. Eberhardt, Battelle Northwest, Kennewick, Wash-
ington; C. W. Fowler, National Marine Fisheries Service,
Seattle, Washington; R. A. Garrott, University of Wiscon-
sin; T. Gerrodette, National Marine Fisheries Service,
La Jolla, California; S. R. Humphrey, University of Flor-
ida; H. I. Kochman, Sirenia Project; C. A. Langtimm,
Sirenia Project; L. W. Lefebvre, Sirenia Project; M. Mar-
montel, University of Florida; H. Marsh, James Cook Uni-
versity, Townsville, Australia; J. D. Nichols, Patuxent
Wildlife Research Center, Laurel, Maryland; D. K. Odell,
Sea World of Florida, Orlando; T. J. O'Shea, Sirenia
Project; K. H. Pollock, North Carolina State University,
Raleigh; K. M. Portier, University of Florida; G. B. Rath-
bun, National Biological Service, San Simeon, California;
J. P. Reid, Sirenia Project; J. E. Reynolds, III, Eckerd
College,St. Petersburg, Florida and U.S. Marine Mammal
Commission, Washington, D.C.; P. M. Rose, Florida De-
partment of Environmental Protection, Tallahassee; B. L.
Weigle, Florida Marine Research Institute; and S. D.
Wright, Florida Marine Research Institute.
Articles in the current volume provide the most inten-
sive analyses of original data and overview perspectives
yet available on the topic of the population biology of the
Florida manatee. Contents are arranged to provide an
overview of the numerous programs and institutions in
manatee research and related conservation in Florida to the
time of the workshop, information about manatee aerial-
survey techniques and estimation of population size and






4 INFORMATION AND TECHNOLOGY REPORT 1


trend, background information on research approaches
and current understanding of the population biology of the
closely related dugong (Dugong dugon, that for years
provided the best standard and proxy for evaluating
sirenian population biology), anatomical studies of mana-
tee age-related reproduction, techniques in assembling
photo-identification records for life-history studies in the
field, longitudinal field studies of manatee reproduction,
field studies of survival, analyses of manatee mortality
databases, and integration of manatee population and life-
history data with a classical approach to population dy-
namics modeling. The volume is concluded with an over-
view, including summaries of recommendations from the
workshop and a brief synopsis of pertinent work that has
been published elsewhere since the workshop.



Acknowledgments

In addition to working group participants, we thank the
numerous individuals who contributed to the success of
the 1992 workshop: R. L. Brownell, Jr., L. Buckle,
C. Campbell, R. Carthy, D. Easton, B. Fesler, J. Grear,
R. J. Hofman, W. Jones, M. Klaips, D. Laist, P. Lefebvre,
M. Mangel, J. Serino, T. D. Smith, M. Suarez, R. O.
Turner, Jr., J. R. Twiss, Jr., P. A. Underwood, J. R. Wil-
cox, and K. Wood. We also thank the many peer reviewers
for helpful comments on each of the manuscripts. The
workshop was administered in part through Research
Work Order No. 88 issued by the National Ecology Re-
search Center to the Florida Cooperative Fish and Wildlife
Research Unit at the University of Florida. Logistics and
administration were eased by participation of the Florida
Wildlife Federation. The use of Mace Lodge at the Austin
Cary Forest was provided by the School of Forest Re-
sources and Conservation, University of Florida. We are
especially grateful to our families for tolerating our ab-
sences while assembling and editing the material for this
volume. Assistance with the preparation of final versions
of the manuscripts was provided by B. Coen and
D. Medellin. J. Shoemaker, D. Crawford, and J. Goehring
assisted with the final preparation of the figures and the
cover art.



Cited References6

*Brownell, R. L., Jr., and K. Rails, editors. 1981. The West Indian
manatee in Florida. Proceedings of a workshop held in Or-
lando, Florida, 27-29 March 1978. Florida Department of
Natural Resources, Tallahassee. 154 pp.


6 An asterisk denotes unpublished material.


Domning, D. P., and L. C. Hayek. 1986. Interspecific and intras-
pecific morphological variation in manatees (Sirenia:
Trichechus). Marine Mammal Science 2:87-144.
*Eberhardt, L. L. 1982. Censusing manatees. Manatee Popula-
tion Research Report 1, Florida Cooperative Fish and
Wildlife Research Unit, University of Florida, Gainesville.
18 pp.
Eberhardt, L. L., and T. J. O'Shea. 1995. Integration of manatee
life-history data and population modeling. Pages 269-279 in
T. J. O'Shea, B. B. Ackerman, and H. F. Percival, editors.
Population biology of the Florida manatee. National Biologi-
cal Service Information and Technology Report 1.
Hartman, D. S. 1979. Ecology and behavior of the mana-
tee (Trichechus manatus) in Florida. American Society of
Mammalogists, Special Publication 5. Lawrence, Kans.
153 pp.
Odell, D. K., G. D. Bossart, M. T. Lowe, and T. D. Hopkins.
1995. Reproduction of the West Indian manatee in captivity.
Pages 192-193 in T. J. O'Shea, B. B. Ackerman, and H. F.
Percival, editors. Population biology of the Florida manatee.
National Biological Service Information and Technology
Report 1.
O'Shea, T. J. 1988. The past, present, and future of manatees in
the southeastern United States: realities, misunderstandings
and enigmas. Pages 184-204 in R. R. Odom, K. A. Riddleber-
ger, and J. C. Ozier, editors. Proceedings of the Third South-
eastern Nongame and Endangered Wildlife Symposium.
Georgia Department of Natural Resources, Game and Fish
Division, Social Circle.
*O'Shea, T. J., B. B. Ackerman, and H. F. Percival, editors. 1992.
Interim report of the technical workshop on manatee popula-
tion biology. Manatee Population Research Report 10, Flor-
ida Cooperative Fish and Wildlife Research Unit, University
of Florida, Gainesville. 83 pp.
O'Shea, T. J., and W. C. Hartley. 1995. Reproduction and
early-age survival of manatees at Blue Spring, upper St. Johns
River, Florida. Pages 157-170 in T. J. O'Shea, B. B. Acker-
man, and H. F. Percival, editors. Population biology of the
Florida manatee. National Biological Service Information
and Technology Report 1.
O'Shea, T. J., and R. L. Reep. 1990. Encephalization quotients
and life-history traits in the Sirenia. Journal of Mammalogy
71:534-543.
*Packard, J. M. 1984. Review of manatee marking techniques.
Manatee Population Research Report 6, Florida Cooperative
Fish and Wildlife Research Unit, University of Florida,
Gainesville. 29 pp.
*Packard, J. M. 1985a. Development of manatee aerial survey
techniques. Manatee Population Research Report 7, Florida
Cooperative Fish and Wildlife Research Unit, University of
Florida, Gainesville. 68 pp.
*Packard, J. M. 1985b. Preliminary assessment of uncertainty
involved in modeling manatee populations. Manatee Popu-
lation Research Report 9, Florida Cooperative Fish and
Wildlife Research Unit, University of Florida, Gainesville.
19 pp.
*Packard, J. M., and R. Mulholland. 1983. Analysis of manatee
aerial surveys: a compilation and preliminary analysis of win-
ter aerial surveys conducted in Florida between 1977 and 1982.






INTRODUCTION 5


Manatee Population Research Report 2, Florida Cooperative
Fish and Wildlife Research Unit, University of Florida,
Gainesville. 119 pp.
*Packard, J. M., and J. D. Nichols. 1983. A preliminary analysis
of sample sizes required for mark-recovery and mark-re-
sighting studies of manatees (Trichechus manatus) in
Florida. Manatee Population Research Report 4, Florida
Cooperative Fish and Wildlife Research Unit, University of
Florida, Gainesville. 14 pp.
Packard, J. M., D. B. Siniff, and J. A. Cornell. 1986. Use of
replicate counts to improve indices of trends in manatee
abundance. Wildlife Society Bulletin 14:265-275.
Packard, J. M., R. C. Summers, andL. B. Barnes. 1985. Variation
of visibility bias during aerial surveys of manatees. Journal
of Wildlife Management 49:347-351.
Rathbun, G. B., J. P. Reid, and G. Carowan. 1990. Distribution
and movement patterns of manatees (Trichechus manatus) in


northwestern peninsular Florida. Florida Marine Research
Publications 48:1-38.
*Reid, J. P., and T. J. O'Shea. 1989. Three years operational use
of satellite telemetry on Florida manatees: tag improvements
based on challenges from the field. Pages 217-232 in Proceed-
ings of the 1989 North American Argos Users Conference.
Service Argos, Inc., Landover, Md. 361 pp.
Reid, J. P., G. B. Rathbun, and J. R. Wilcox. 1991. Distribution
patterns of individually identifiable West Indian manatees
(Trichechus manatus) in Florida. Marine Mammal Science
7:180-190.
Reynolds, J. E., III, and D. K. Odell. 1991. Manatees and du-
gongs. Facts On File, Inc., New York. 192 pp.
*U.S. Fish and Wildlife Service. 1989. Florida manatee
(Trichechus manatus latirostris) Recovery Plan. Prepared
by the Florida Manatee Recovery Team for the U.S. Fish and
Wildlife Service, Atlanta, Ga. 98 pp.






6 INFORMATION AND TECHNOLOGY REPORT 1


Florida Manatee Population Biology:
Research Progress, Infrastructure, and Applications for

Conservation and Management



by


John E. Reynolds, III1

Marine Mammal Commission
1825 Connecticut Avenue N. W.
Washington, D.C. 20009


Abstract. Twenty-five years ago, a single scientist focused his research on Florida manatees
(Trichechus manatus latirostris). Today, millions of dollars are spent annually on research and
management of this endangered species. In the meantime, several databases with valuable information
about manatee population biology, including some particularly valuable long-term databases, were
created. Cooperation and interaction among various individuals, agencies, and institutions have been
great strengths of the overall research and the management of manatees in Florida. The presentation
and evaluation of the collectively gathered information about the population biology of manatees in
this workshop are therefore appropriate. The recovery of the subspecies is not assured; with some
justification, it has been portrayed in grim terms. I hope that this workshop provides new insights, new
syntheses, and new directions for research and management and that these, in turn, establish new hope
for the recovery of Florida manatees.
Key words: Manatee, Trichechus, Florida, population biology, recovery plan, trends.


The first and, to my knowledge, only workshop that
focused on population biology of manatees in Florida
was held almost exactly 14 years ago in March 1978
(2*Brownell and Ralls 1981). The attending individuals
and represented institutions and agencies at the work-
shop were different from those of today's workshop. In
fact, fewer than a half dozen individuals appear on both
rosters, and only two institutions, the U.S. Fish and
Wildlife Service and Sea World, provided repre-
sentatives to both workshops.
The players but not the objectives changed with time.
The objectives, reflected in the revised Florida Manatee
Recovery Plan (*U.S. Fish and Wildlife Service 1989),
include assessment of the causes of manatee mortality,
population size and distribution, habitat-use patterns,
reproduction, behavior, life history, and application of
available data to the management of the subspecies. I
provide a brief historical context for manatee research




1 Biology and Marine Science Departments, Eckerd College, 4200 54th
Avenue South, St. Petersburg, Fla. 33711.
2 An asterisk denotes unpublished material.


and management, describe the current status of activi-
ties, and speculate about the future.



History and Organization of
Manatee Research

Modern manatee research can be traced to Daniel S.
Hartman who began to study manatees in the Crystal
River in the mid-1960's. In 1974, he prophetically stated
that "the manatee is protected...but its survival is still
threatened by the propellers of power boats...and habitat
alteration. Recommendations for the manatee's conser-
vation include the regulation of boat speeds on strategic
bodies of water and the establishment of sanctuaries..."
(*Hartman 1974:239). His insightful observations and
advice for manatee protection and his descriptions of
manatee behavior (Hartman 1979) still stand.
The year 1974 marked the beginnings of multifaceted
manatee research at the University of Miami and at the
Gainesville Field Station of the U.S. Fish and Wildlife
Service, a program that in 1978 was recognized as the
Sirenia Project. The 1978 workshop mentioned earlier,






JOHN E. REYNOLDS, i 7


in fact, helped catalyze the reorganization of the Sirenia
Project.
Research by the U.S. Fish and Wildlife Service and
University of Miami personnel has been on behavior and
ecology, causes of death, distribution and abundance,
functional anatomy, food habits, movement patterns,
acoustics, growth, and recently, age determination. Most
long-term databases, which are essential for population
assessment, were begun by scientists of the U.S. Fish and
Wildlife Service and the University of Miami and were
covered in the 1978 workshop (*Brownell and Rails
1981). The state of Florida demonstrated increased in-
volvement in studies of manatees with the passage of the
Florida Manatee Sanctuary Act in 1978.
By 1984, the Florida Department of Natural Resources
(now the Florida Department of Environmental Protection)
began to assume a more active role in manatee management
and various types of research, including the operation of the
carcass-salvage network in mid-1985. Note that 1990
amendments of certain state statutes, including the Florida
Manatee Sanctuary Act, gave the department more author-
ity, more funding, and more positions, such that the depart-
ment has a larger budget and more people for manatee
research and management than other agencies or organiza-
tions in manatee conservation.
Over time, several other players joined the conserva-
tion, research, or education programs on manatees. Many
organizations and individuals are involved in research,
management, education, or public awareness efforts on
manatees in Florida (Tables 1-4). These organizations
generally work well together.
Cooperation among the various agencies has been dem-
onstrated in many ways. Recently the development and
adoption of a revised Florida Manatee Recovery Plan
(signed by heads or representatives of 13 agencies, conser-
vation organizations, and industry groups) and the mul-
tiagency support of proposed regulations by the Florida
Department of Natural Resources to the Florida Governor
and Cabinet have provided good examples of cooperation
and shared goals (*U.S. Fish and Wildlife Service 1989).
The history of manatee conservation can be divided into
three eras. From the late 1960's until 1980, various activities
got underway, setting the stage for cooperative, focused
conservation. In 1980, with the direction of the Marine
Mammal Commission, the first Manatee Recovery Plan was
adopted and a coordinator was hired (*U.S. Fish and Wild-
life Service 1980); from that date until 1989, joint studies
increased in which the state began to participate in 1984.
Finally, in 1989, the revised Recovery Plan was adopted
(*U.S. Fish and Wildlife Service 1989); it focused even
better on specific research or management. The transition to
each new stage, marked by increased research and manage-
ment, was promoted and guided by the Marine Mammal


Commission, the agency with oversight of all marine mam-
mal conservation in this country (*Marine Mammal Com-
mission 1980, 1981, 1991). Since the late 1970's, the com-
mission has been concerned about the status of manatees and
manatee management in Florida and, as a result, has devoted
time and funding to research and recovery of the manatee.
In fact, the bulk of the commission's 1988 research budget
was spent on manatees (*Marine Mammal Commission
1991). In 1980, 1987, and 1992, the commission held its
annual meeting in Florida to focus on manatee issues.
All the individual and collective efforts to date advanced
our knowledge about manatees. In the 1970's, popular and
technical publications may well have given the impression
that the common name of the animal was not simply "mana-
tee" but "the poorly studied manatee." That situation
changed. Unanswered questions remain, but important da-
tabases exist. For example, several long-term manatee data-
bases were initiated at least 15 years ago by the Sirenia
Project, by Daniel Odell at the University of Miami, and by
the Florida Power and Light Company. They include the
causes and locations of mortality; distribution and abun-
dance in winter; numbers of calves in aggregations in winter;
reproduction and behavior of distinctly marked manatees;
and movement patterns and high use at sites on the eastern
coast of Florida, the Crystal River, Blue Spring, Fort Myers,
and in Tampa Bay. Recent studies provided valuable insight
into manatee population biology and conservation; papers
in this volume describe many of the newer databases. For
this particular workshop on manatee population biology, it
is important to note the pioneering work of J. M. Packard,
who compared a variety of survey techniques to determine
their value in assessing manatee abundance and trends
(*Packard and Mulholland 1983; Packard et al. 1986).


Applications of Manatee
Research to Conservation and
Management

The existing databases and technology facilitate studies
to determine the following:
1) the number of manatees that died where, when, and
how to highlight specific areas and key counties where
manatee abundance or mortality is highest,
2) locations of warm-water refugia and other high-use
areas where protection of manatees is especially im-
portant,
3) approximate age and reproductive status of manatees,
4) the number of calves in warm-water refugia in winter,
and
5) characteristics of basic life-history events such as re-
productive intervals, age at first reproduction, and sur-
vival.







8 INFORMATION AND TECHNOLOGY REPORT 1


Table 1. Summary of research or research-related activities conducted now or in the recent past by the federal
government and the state of Florida on manatees (Trichechus manatus latirostris; updated from Reynolds and
Gluckman 1988).


Agencv


Topics


U.S. Fish and Wildlife Service (under
National Biological Service as of 1993),
Sirenia Project







Florida Cooperative Fish and Wildlife
Research Unit (National Biological Service)





Florida Department of Natural Resources
(under Department of Environmental
Protection as of 1993)



U.S. Army Corps of Engineers (Jacksonville, Florida
District)

National Aeronautics and Space Administration

U.S. Marine Mammal Commission



National Marine Fisheries Service
National Park Service

U.S. Navy
Chassahowitzka National Wildlife Refuge


Life-history studies
Carcass salvage
Development and application of telemetry
Behavior
Ecosystems
Aerial surveys
Scar catalog
Technical assistance
Age determination
Human recreational activities (Crystal River)
Research and management plans (Crystal River)
Aerial surveys (Caloosahatchee and Crystal Rivers)
Aerial surveys/human activities (Northeastern Florida)
Radio tracking (Cumberland Sound)
Description of habitat types and manatee distribution (Crystal River)
Carcass salvage
Aerial surveys
Geographic information systems
Telemetry (central, western Florida)
Ecosystem studies (Hobe Sound)
Human activities (boat use patterns)
Overview of manatee status
Aerial surveys (eastern Florida)
Telemetry support (eastern Florida)
Aerial surveys (Banana River)
Ecosystem studies (Banana River)
Constituting/convening the Manatee Technical Advisory Council
Studies to enhance protection and management of manatees
Research and management plans (Crystal River)
Food sources and feeding habits (Hobe Sound)
Ecosystems studies (Hobe Sound)
Aerial surveys (Cumberland Island, Everglades National Park)
Telemetry applications (Cumberland Sound)
Telemetry applications (Cumberland Sound)
Aerial surveys (Citrus County)


The databases also provide much additional informa-
tion. The manatee entered a new era when the term "poorly
studied" was no longer an apt description.
Good biological data permit more and better manage-
ment. The U.S. Fish and Wildlife Service continues to use
the best available data to formulate endangered species act
jeopardy opinions. To date, more opinions (99 between
1984 and November 1991) were issued for manatees than
for all other endangered species in this country combined
(R. Turner, U.S. Fish and Wildlife Service, personal com-
munication). Similarly, the state continues to work with


local governments to implement local management for the
adequate protection of the manatee and its habitats. To
date, plans that regulate boat speeds and boat access in
areas known to be frequented by manatees were approved
for six counties by the state; three other plans will probably
be approved soon. These plans should reduce fatal and
non-fatal collisions between watercraft and manatees. As
mandated by Florida's Growth Management Act of 1985,
local regulations for watercraft will be incorporated into
management to control the effects of human population
growth in local areas (*Reynolds and Gluckman 1988).


Aizencv






JOHN E. REYNOLDS, m 9


Table 2. Research conducted or sponsored by private organizations or by scientists (updated from Reynolds and
Gluckman 1988).


Organization or individual


Topic


Florida Power & Light Company



Save the Manatee Club
Sea World (D. Odell, others)





Miami Seaquarium (G. Bossart)
J. Reynolds (Eckerd College)



D. Domning (Howard University)



G. Patton (Mote Marine Laboratory)


J. Morris (Florida Institute of Technology)


E. Gerstein (Florida Atlantic University)
J. and M. Provancha (Bionetics Corp.)


Aerial surveys
Scar catalog
Applications of telemetry
Water temperatures
Aerial surveys, support to various other projects including international efforts
Clinical parameters
Carcass salvage
Morphometrics
Parasites
Aerial surveys
Behavior and ecology
Clinical parameters, immunology
Carcass salvage
Aerial surveys
Functional anatomy
Behavior and ecology
Anatomy
Paleontology and systematics
Feeding ecology
Bibliography and index
Aerial surveys
Means of detecting manatees
Acoustics
Ecology and behavior
Analysis of vocalizations
Nutrition
Acoustics
Aerial surveys
Habitat assessment


Public support is important for the passing of regula-
tions to protect manatees or their habitat. The Save the
Manatee Club now has more than 30,000 members
(J. Vallee, Save the Manatee Club, personal communi-
cation) and more than 130,000 special state of Florida
auto tags depicting a manatee were sold (B. Weigle,
Florida Marine Research Institute, personal communica-
tion).
The biological data on manatees increased, efforts to
conserve manatees have broad support and should re-
duce manatee deaths, public support and interest in
manatees are widespread, and several agencies and insti-
tutions work hard to protect manatees. Nonetheless, I
consider that manatee recovery and protection are, in
fact, in jeopardy for four fundamental reasons: (1) the
poorly managed human population growth in Florida, (2)
insufficient funds for acquisition of habitat and enforce-
ment of regulations, (3) the increasing strength and ef-


fectiveness of opposition, and (4) the size of the problem.
A better understanding of manatee population biology
would permit a measure of progress in the face of these
limitations.
In recent years, the human population of Florida in-
creased by nearly 1,000 new residents a day (Anony-
mous 1987). About 90% of Florida's residents live
within 16 km of the coast, concentrating impacts of
human activities on coastal habitats. Human activities
continue to kill and injure manatees; the number of
manatee deaths attributable to human factors grows with
the human population.
A conference titled Managing Cumulative Effects in
Florida Wetlands (*Estevez et al. 1986a, 1986b) re-
vealed that in the early 1970's the average annual loss of
wetlands was approximately 29,160 ha. The loss of Flor-
ida wetlands continues today at a reduced rate. However,
cumulative effects of development and other human






10 INFORMATION AND TECHNOLOGY REPORT 1


Table 3. Primary state and federal agencies that manage Florida manatees (Trichechus manatus latirostris) or manatee
habitat, and examples of duties (Reynolds and Gluckman 1988).


Agency


Duties


Florida Department of Community
Affairs
Florida Department of
Environmental Regulation
Florida Department of
Natural Resources

Florida Game and Freshwater
Fish Commission
Marine Mammal Commission
U.S. Army Corps of Engineers

U.S. Fish and Wildlife Service


Approval of local growth management plans; approval of major developments
with regional impact.
Administers permits for dredging and filling of wetlands; designates aquatic preserves
as Outstanding Florida Waterways
Develop local management plans for Florida manatees; enforce state laws
affecting manatees and their habitats; use of state-owned lands; review of permit
applications, and other regulatory activities that can enhance manatee protection
Enforcement of boat speed zones and other manatee protection activities on
inland waterways
Oversight for all federal agency activities affecting marine mammals
Federal permits for activities in coastal areas with potential impacts on manatees;
initiate Section 7 consultations
Consultations with other federal agencies on activities with potential impacts on manatees
as required under Section 7 of the Endangered Species Act; federal wildlife law enforce-
ment; review of permit applications under various federal statutes; development and
updating of Recovery Plan; coordination of a wide array of management activities


Table 4. Organizations involved in education and awareness programs about Florida manatees (Trichechus manatus
latirostris) in Florida. A brief statement of efforts to date is included. More complete information about education
and awareness is in Reynolds and Gluckman (1988).
Organization Efforts
Colleges and Universities Classes; public lectures
Florida Department of Brochures, bumper stickers, posters, public lectures, etc.
Natural Resources
Florida Power & Light Company Booklets, bumper stickers, Educator's Guide, films, public workshops
Marine Mammal Commission Funding support for development of educational materials and for Sirenews;
distribution of annual reports and other literature
Marine Zoological Parks Tours, efforts focused on children and teachers, educational displays
Port Everglades Authority Bumper stickers, brochures, public tours, and lectures
Save the Manatee Club Newsletter, public service announcements, signs, press releases, in-service programs for
teachers, brochures
Tampa Electric Company Manatee viewing platform, lectures, displays
U.S. Fish and Wildlife Service Interpretive/Education Center literature distribution


activities have been devastating. In Tampa Bay, for
example, Lewis (*1986) estimated that more than 80%
of the seagrass community was destroyed by human
activities. The status of the manatee is not unique but
symptomatic of the status of all coastal and estuarine
resources in Florida. Efforts focused entirely on manatee
protection are doomed to failure if erosion of habitat
extent and quality is allowed to continue.
Important solutions to manatees' woes and to
coastal degradation include the acquisition of habitat


and subsequent creation of reserves and further develop-
ment and enforcement of regulations to protect habitat
and manatees. Where true sanctuaries (i.e., no entry
areas, such as at the Crystal River, the Banana River, and
the Tampa Electric Company's Apollo Beach plant)
were created and incorporate adequate resources for
manatees, manatee use increased profoundly.
Lack of money is a problem for habitat acquisition
and enforcement. The state's Conservation and Recrea-
tional Lands Acquisition List includes properties valued






JOHN E. REYNOLDS, m 11


at more than $900 million, but the program has an annual
budget of about $50 million (*Reynolds and Gluckman
1988). These are large sums but perhaps not large enough
to secure sufficient habitat for Florida wildlife. Funding
has been insufficient to hire adequate enforcement staff.
An additional 250 field officers (the number needed in
1988; *Reynolds and Gluckman 1988) for the Florida
Marine Patrol would cost more than $11 million a year.
A state that is currently implementing significant cut-
backs in education and other programs to control its
budget will probably not make these expenditures.
Damaging the protection of manatees are reports in
various media that basically state that humans cause little
damage to manatees and manatee habitat. Some articles
simply contain false information. Others take advantage
of the fact that precise population data on size and trends
are unavailable. Scientists may be reluctant to speculate
on insufficient data; many journalists are not so reticent.
Therefore, the public does not know what to believe. A
solution is to continue research to provide more defini-
tive population data.
Some people in Florida are openly antagonistic to-
ward regulations that restrict human activities. The atti-
tude has been publicly expressed: waterways belong to
humans and, if manatees and other creatures cannot exist
where people want to boat or ski, that is unfortunate. The
boating industry, among others, has a strong and effec-
tive lobby in a state with 750,000 registered boats. How-
ever, a survey by scientists (*Parker 1989) at the Florida
State University revealed that, in reality, most boaters
support protection of manatees.
The magnitude of the problem is immense. Almost
any human activity in any inshore body of water or
wetland in Florida can and does harm manatees or their
habitat. Significant progress in regulating human activi-
ties to protect wildlife and habitat in the last 20 years is
a tribute to those who have worked in this area. Whereas
some other marine mammal-human interactions may be
limited to a small number of activities or particular
locations, human-manatee interactions cover all activi-
ties along Florida's lengthy waterways (*Marine Mam-
mal Commission 1991). People who are unfamiliar with
Florida may be surprised to learn that Florida and Geor-
gia have more coastline (2,333 km) than California,
Oregon, and Washington combined (2,080 km; Anony-
mous 1991).
The current research and management must continue
and expand. In particular, the valuable long-term data-
bases must be maintained. Enforcement must increase.
Habitat protection with regulations or acquisition must
continue. Education must expand. But currently the
question remains: Will all this be enough? I worry that it
may not. I also worry that by the time we have enough


pieces of the puzzle in place and then evaluate their
effects in terms of habitat and manatee recovery, years
will have passed. If our efforts do not work, those years
of continued high mortality and continued habitat de-
struction may be irreversible.
This workshop and its proceedings take advantage of
the degree of interagency cooperation and the existence
of certain data to attempt to better understand Florida
manatee population biology. Let us hope that the effort
results not only in increased scientific insight but also in
contributions to better management and conservation of
this endangered species.


Acknowledgments

The following people read and commented on drafts
of this presentation: R. J. Hofman, D. W. Laist, L. W.
Lefebvre, D. K. Odell, T. J. O'Shea, J. R. Twiss, Jr., and
B. L. Weigle. I am grateful for their assistance and in-
sight.


Cited References3

Anonymous. 1987. Keys to Florida's future. Winning in a
competitive world. Final report of the State Comprehensive
Planning Committee. Florida Department of Community
Affairs, Tallahassee. 46 pp.
Anonymous. 1991. Information please almanac. Forty-fourth
edition. Houghton Mifflin Company, Boston, Mass. 992 pp.
*Brownell, R. L., Jr., and K. Rails, editors. 1981. The West Indian
manatee in Florida. Proceedings of a workshop held in Or-
lando, Florida, 27-29 March 1978. Florida Department of
Natural Resources, Tallahassee. 154 pp.
*Estevez, E. D, J. Miller, J. Morris, and R. Hammen. 1986a.
Proceedings of the conference: managing cumulative effects
in Florida wetlands. October 1985, Sarasota, Fla. New Col-
lege Environmental Studies Program Publication 37,
Omnipress, Madison, Wis. 326 pp. + appendix.
*Estevez, E. D., J. Miller, J. Morris, and R. Hammen. 1986b.
Executive summary of the conference: managing cumulative
effects in Florida wetlands. October 1985, Sarasota, Fla.
New College Environmental Studies Program Publication
37, Omnipress, Madison, Wis. 50 pp. + appendix.
*Hartman, D. S. 1974. Distribution, status, and conservation of
the manatee in the United States. Document PB 81-140725,
National Technical Information Service, Springfield, Va.
246 pp.
Hartman, D. S. 1979. Ecology and behavior of the manatee
(Trichechus manatus) in Florida. Special Publication 5,
American Society of Mammalogists. Lawrence, Kans.
153 pp.
*Lewis, R. R., III. 1986. Marine wetland loss in Tampa Bay and
management/restoration recommendation. Pages 159-174 in


3 An asterisk denotes unpublished material.






12 INFORMATION AND TECHNOLOGY REPORT 1


E. D. Estevez, et al., editors. Proceedings of the conference:
managing cumulative effects in Florida wetlands. October
1985, Sarasota, Fla. New College Environmental Studies Pro-
gram Publication 37, Omnipress, Madison, Wis.
*Marine Mammal Commission. 1980. Annual report of the
Marine Mammal Commission, calendar year 1979. A report
to Congress. Washington, D.C.
*Marine Mammal Commission. 1981. Annual report of the
Marine Mammal Commission, calendar year 1980. A report
to Congress. Washington, D.C.
*Marine Mammal Commission. 1991. Annual report of the
Marine Mammal Commission, calendar year 1990. A report
to Congress. Washington, D.C.
*Packard, J. M., and R. Mulholland. 1983. Analysis of mana-
tee aerial surveys. Prepared for U.S. Fish and Wildlife Serv-
ice, Cooperative Agreement 14-16-0009-1544. 32 pp.
Packard, J. M., D. B. Siniff, and J. A. Cornell. 1986. Use of
replicate counts to improve indices of trends in manatee
abundance. Wildlife Society Bulletin 14:165-175.


*Parker, S. L. 1989. Report on a survey of Florida's licensed boat
owners for the Save the Manatee Committee. Survey Research
Laboratory, Florida State University, Tallahassee. 52 pp.
*Reynolds, J. E., III, and C. J. Gluckman. 1988. Protection of
the West Indian manatee (Trichechus manatus) in Florida.
Document PB 88-222922, National Technical Information
Service, Springfield, Va. 85 pp.
Reynolds, J. E., III, and K. D. Haddad, editors. 1990. Report of
the workshop on geographic information systems as an aid to
managing habitat for West Indian manatees in Florida and
Georgia. Florida Marine Research Publication 49, St. Peters-
burg. 98 pp.
*U.S. Fish and Wildlife Service. 1980. West Indian manatee
recovery plan. Prepared by the U.S. Fish and Wildlife Serv-
ice in cooperation with the Recovery Team. 27 pp.
*U.S. Fish and Wildlife Service. 1989. Florida manatee
(Trichechus manatus latirostris) recovery plan. Prepared by
the Florida Manatee Recovery Team for the U.S. Fish and
Wildlife Service, Atlanta, Ga. 98 pp.






BRUCE B. ACKERMAN 13


Aerial Surveys of Manatees:
A Summary and Progress Report



by


Bruce B. Ackerman


Florida Department of Environmental Protection
Florida Marine Research Institute
100 Eighth Avenue S.E.
St. Petersburg, Florida 33701


Abstract. Aerial surveys are used to document the distribution and relative abundance of Florida
manatees (Trichechus manatus latirostris) and to assess population trends. Recent research included
aerial surveys by various agencies to determine the distributions of manatees in 10 areas of Florida.
In most of these studies, twice-monthly flights were made for at least 2 years. Surveys of distributions
have now been made in all areas of the state that are extensively used by manatees. The resulting data
have been used for the protection of manatees. Various groups conducted counts of manatees at
aggregation sites in winter at selected power plants and at the Crystal River and Blue Spring. These
counts have been used to assess population-size trends. Based on mandates by the state legislature, a
2-day, synoptic aerial survey was made to obtain a single annual high count by maximizing survey
effort under optimal conditions. These surveys followed two major cold fronts each winter in 1991
and 1992. On 17-18 January 1992, a high count of 1,856 manatees was made (907 on the eastern coast,
949 on the western coast, 8.7% calves). Although not statistical estimates, these counts provide new
information about the minimum size of the population. The higher synoptic-survey counts are not proof
of an increase of the population through time but are consistent with increases in long-term counts in
some areas of the state, including aerial counts at the Crystal River, ground counts at Blue Spring, and
counts of aggregations at power plants adjusted for temperature covariates. Current research on aerial
surveys is focused on new techniques to improve estimates of population size and trend.
Key words: Aerial surveys, Florida manatee, synoptic survey, trends, Trichechus manatus latirostris,
warm-water refuges.


Aerial surveys to count and map the distribution of
Florida manatees (Trichechus manatus latirostris) have
been used since 1967 (Hartman 1979). Various methods
to survey manatees were used by subsequent researchers
(* Hartman 1974; Irvine and Campbell 1978; *Rose and
McCutcheon 1980; Irvine 1982; Irvine et al. 1982; Shane
1983; Kinnaird 1985; *Packard 1985; Reynolds and Wil-
cox 1985, 1986, 1994; Packard et al. 1986). Aerial sur-
veys are useful and cost-effective for counting manatees
and for mapping manatee distribution and seem to be the
only method with which large numbers of manatees in
large areas can be counted. Distribution data have been
used extensively for the protection and management of
manatees. However, aerial surveys have significant


1 An asterisk denotes unpublished material.


drawbacks for obtaining precise population-size esti-
mates (*Eberhardt 1982; Packard et al. *1984, 1985;
Lefebvre et al. 1995). Manatees are difficult to detect
and, once seen, are often difficult to count accurately
(Packard et al. 1985, 1986). Therefore, aerial counts are
generally assumed to be too low. Lefebvre et al. (1995)
discussed the theory and the problems of aerial surveys
of manatees, particularly for the estimation of population
sizes and trends.
My objectives were to provide descriptions of recent and
ongoing aerial surveys of manatee distribution, surveys of
manatee aggregations in winter, and synoptic surveys. De-
scriptions of each of these survey categories include state-
ments of survey objectives, summaries of employed proce-
dures, results, and inherent problems and limitations. I also
provided information on population size and trend based on






14 INFORMATION AND TECHNOLOGY REPORT 1


aerial surveys and gave brief descriptions of new research
to improve survey techniques.
This paper is largely a summary of all recent informa-
tion on aerial surveys of manatees and interpretation of
results of these surveys, particularly as they relate to
estimation of manatee distribution, population size, and
trend. Considerable resources are expended on this topic
by several organizations, but many of the results appear in
technical documents, agency files, or other unpublished
sources. I examined these sources and tabulated descrip-
tions to provide a comprehensive overview of aerial sur-
vey activities, focusing on work carried out since 1986.
Studies prior to 1986 were previously reviewed by
Packard (*1985), Beeler and O'Shea (*1988), and O'Shea
(1988).
Reviews of methodology employed in previous studies
are included under the heading of procedures in descriptions
of surveys of manatee distribution and surveys of manatee
aggregations in winter. I present original information, in-
cluding the first published description of the objectives,
procedures, and results of synoptic surveys. I also analyzed
data on trends in counts of manatees in the Crystal River and
at Blue Spring with regression techniques. Details on meth-
ods for obtaining and analyzing original information are
provided in the sections devoted to these topics.


Recent and Ongoing Aerial
Surveys

Surveys of Distribution

Objectives
The purpose of surveys of distribution (extended-area
surveys, *Packard 1985) is to document the spatial dis-
tribution and seasonal habitat use of manatees. Although
these surveys do not give accurate population-size esti-
mates, they provide a minimum estimate of the number
of animals in an area on a given day. Abundance is
usually considered to be relative because it is believed
(or hoped) to include a roughly constant proportion of
the animals. Surveys over a large area are made repeat-
edly during 1 or more years and provide data on seasonal
and yearly changes of relative abundance. Data from
these surveys have been useful for management because
they reveal areas of high seasonal usage and support the
protection of manatees (e.g., restrictions on boat speeds
or conditions for development).

Procedures
The following is a summary of procedures by the
Florida Department of Environmental Protection; the
procedures are similar to those of other agencies. The


procedures typically followed protocols established by
earlier studies such as those by Shane (1983), Kinnaird
(1985), Provancha and Provancha (1988), and Rathbun
et al. (1990). Some studies may vary from these guide-
lines, but for each study, procedures are kept as consis-
tent as possible to maximize comparability of counts.
Most aerial surveys are conducted with a Cessna 172
or with a similar small, high-winged, 4-seat airplane with
good downward visibility (Irvine 1982). Small helicopters
have also been used but are more expensive (Rathbun
1988). Helicopters are useful for surveys in urban or
residential areas and in congested airspaces such as near
major airports. Pilots experienced in low-altitude, slow-
speed, circling flight are used. One or two experienced
observers are usually seated on the right side of the aircraft;
the door is attached and the window is open. Observers
wear polarized sunglasses to reduce glare. The primary
(most experienced) observer has a minimum of 30 hours
experience in aerial surveys of manatees, has detailed
knowledge of the survey area, and sits in the right front
seat. A secondary observer is not required for most survey
areas but, if used, usually sits in the right rear seat. How-
ever, if wide expanses of shallow water are covered, the
second observer can view from the left rear seat to help
cover the area more effectively (Irvine 1982; Shane 1983;
*Packard et al. 1984). A higher proportion of manatees is
seen with two experienced observers than with one.
Flights are usually at an altitude of 150 m and at an air
speed of 130 km/h. When manatees are seen, the airplane
slows and circles the area clockwise until the observer is
reasonably sure that an accurate count was made (i.e., until
repetitive counts become consistent). Manatees may be
spotted by any observer or by the pilot, but manatees are
officially counted and mapped only when confirmed by the
primary observer (*Packard et al. 1984). Counts are more
consistent when the same observers are used each time.
Surveys follow a standardized flight path and are de-
signed to cover the most probable manatee habitats in an
area, as described by Irvine (1982), Shane (1983), Packard
(*1985), and Rathbun et al. (1990). The route is marked in
advance on National Oceanic and Atmospheric Admini-
stration 1:40,000 navigation charts or U.S. Geological
Survey 1:24,000 topographic maps, and observations are
written on the maps. Since 1992, the Florida Department
of Environmental Protection has used a portable global
positioning system unit (Trimble Pathfinder Basic Plus,
Trimble Navigation, Sunnyvale, California) to accurately
store information on the position of sightings and the flight
path. Routes may include coastal areas, major rivers and
estuaries (usually to depths of 3 m) and their tributaries,
and freshwater and saltwater canals. Surveys are intensi-
fied over aggregation sites in winter, areas within 500 m
of shore, offshore areas (shoals) that are shallower than






BRUCE B. ACKERMAN 15


2 m, areas with aquatic vegetation, freshwater sources, and
areas in which manatees have been sighted historically.
High-density concentrations of manatees such as those
at power plants are surveyed intensively (intensive-area
method; *Packard 1985). Each small area is circled clock-
wise at least twice before the aircraft moves on. This
technique takes more time but gives a higher probability
of detecting manatees, including manatees that rest on the
bottom and must rise for a breath while the aircraft is
passing over. Surveys in areas of low manatee density are
less intensive (extended-area method; *Packard 1985).
Only one pass is made over each area, which reduces the
probability of seeing all manatees but allows sampling in
a larger area.
Wide expanses of shallow water (e.g., Indian River,
Whitewater Bay) have been covered in a series of transects
that were 0.8 km apart (Odell 1979; Shane 1983). Wide
expanses of deeper water, such as Tampa Bay, are only
covered along the shoreline and around spoil islands
(*Reynolds et al. 1991). This allows coverage of a larger
area in a given amount of time but decreases the prob-
ability of detecting all the manatees.
Recent Florida Department of Environmental Protec-
tion surveys were typically conducted twice per month for
2 or more years. The unpublished protocols are similar to
those of other agencies (Shane 1983; Provancha and
Provancha 1988; Rathbun et al. 1990). Data about each
survey flight are recorded on standardized forms and in-
clude date; start and end time; observer and pilot names;
and aircraft type, speed, and altitude. Weather and water
conditions recorded for each segment of the flight include
wind speed and direction, air temperature, percentage of
clouds, water clarity (depth to which a manatee can be
seen), and water-surface conditions. A scale of water-sur-
face conditions was adapted from the Beaufort Scale
(Woolf 1977:98): (0) smooth like glass; (1) ripples with
appearance of scales, no foam crests; (2) small wavelets,
crests of glossy appearance, not breaking, no whitecaps;
and (3) large wavelets, crests beginning to break, scattered
whitecaps. Flights are canceled at conditions rated 3 or
higher. The best visibility below the water surface occurs
in smooth, clear water in the presence of few clouds and a
bright sun. Flying conditions are best in the presence of
little wind and no fog or precipitation.
Data about each observed group are recorded on maps
and include the number of adult and calf manatees and their
locations and behavior. Calves are defined as animals
closely associated with an adult but less than about half the
adult's length (Irvine and Campbell 1978; Irvine 1982).
Behavior categories include resting (motionless manatees),
traveling (swimming manatees), feeding (recognized by the
presence of a manatee in a vegetated area and a nearby plume
of suspended sediment), and cavorting (group of manatees


rolling, splashing, or swimming in tight circles). Most mana-
tees are seen close to the flight path but not directly under
the aircraft. Splashes, surface wakes, mud trails, and mud
plumes may draw the observer's attention to more distant
manatees (Irvine 1982). Manatees in aggregations or in clear
water are easiest to find. In winter, aggregations occur at or
near warm-water refuges and are often accompanied by
large amounts of stirred-up mud. Photographs can confirm
counts of groups in clear water but are not of much value for
large groups in turbid water.

Results

More than 30 studies were made between 1984 and
1993 (Table 1). Studies through 1986 were reviewed by
Beeler and O'Shea (*1988). Surveys have been conducted
in most areas of the state since that time. In most studies,
twice-monthly surveys were conducted year-round for 2
or more years. Results of many studies have not yet been
published (see Table 1 for unpublished sources). Exam-
ples of long-term or extensive studies follow.
On the western coast of Florida, the U.S. Fish and Wild-
life Service conducted surveys in Lee County during 1984-
85 (R. K. Frohlich, Florida Department of Environmental
Protection, unpublished data). Subsequent surveys by the
department were of manatees in Charlotte, Lee, and Collier
counties (Florida Department of Environmental Protection,
unpublished data). The Mote Marine Laboratory conducted
surveys in Manatee, Sarasota, and Charlotte counties from
1985 to the present (*Kadel and Patton 1992). Eckerd Col-
lege and the department conducted surveys in Tampa Bay
from 1987 to 1994 (*Reynolds et al. 1991; Eckerd College
and Florida Department of Environmental Protection, un-
published data). A series of studies of distribution were
conducted in northwestern Florida beginning in 1967, cov-
ering aggregation sites in winter and warm-season habitats
(Charlotte, Dixie, and Levy counties; Powell and Rathbun
1984; Kochman et al. 1985; Rathbun et al. 1990; Chassa-
howitzka National Wildlife Refuge, unpublished data).
On the eastern coast of Florida, four teams from differ-
ent agencies conducted simultaneous counts twice-
monthly for 1 year during 1986 in five adjacent counties
from Volusia to Martin counties (B. L. Weigle, Florida
Department of Environmental Protection, St. Petersburg,
Florida, and R. K. Bonde, National Biological Service,
Gainesville, Florida, unpublished data). These coordi-
nated surveys provided information about seasonal migra-
tions of manatees. Provancha and Provancha (1988,
*1989) conducted surveys of manatees in the Banana
River from 1984 to the present, expanding the database
provided by surveys conducted during 1978-80 by Shane
(1983). These surveys revealed high use by manatees of
the Banana River, especially during spring migration.






16 INFORMATION AND TECHNOLOGY REPORT 1


Table 1. Recent aerial surveys of distribution of the Florida manatee (Trichechus manatus latirostris).

Area Dates Citations

Eastern coast


Southeastern Georgia (Camden County,
Cumberland Sound, warm seasons)
Southeastern Georgia (Camden,
Glynn, Mclntosh counties)
Nassau County
Duval County

Duval, Clay, St. Johns counties
(lower St. Johns River)
Duval County (lower
St. Johns River)
St. Johns, Clay, Putnam counties
(middle St. Johns River)
Nassau, Duval, St. Johns, Flagler,
Volusia counties (ICW, coast)
St. Johns, Flagler, Volusia counties
(ICW, coast)
Volusia County (Tomoka River)

Brevard County
Brevard County (Banana River,
warm seasons)

Interagency cooperative aerial survey
Volusia County (Halifax River,
Tomoka River, Mosquito Lagoon)
Brevard County (northern Indian River,
Banana River, Banana Creek)
Brevard County (southern Indian River)
Indian River, St. Lucie counties

St. Lucie, Martin counties

St. Lucie, Martin counties

Palm Beach County
Broward, northern Dade counties

Broward County
Dade County (Biscayne Bay)
Dade County (County-wide)
Western coast
Citrus, Levy, Dixie counties
Winter (Crystal River, Homosassa River)
Summer (coast and rivers)


Northern Manatee County

Pinellas, Hillsborough, northern
Manatee counties


Southern Manatee, Sarasota,
northern Charlotte counties


May 1988-Aug 1989

Jun 1989-May 1990

Oct 1986-Oct 1988
May 1988-Apr 1990

Jul 1982-Jun 1983

May 1993-May 1994

Jun 1985-Jun 1986

Jul 1982-Jun 1983

Mar 1991-Nov 1993

May 1985-Dec 1985

Jan 1978-Feb 1980
Jun 1984-Apr 1986;
Feb 1987-ongoing



Dec 1985-Jan 1987



Jan 1986-Jan 1987
Jun 1985-Dec 1987

Jan 1986-Jan 1987

Aug 1990-Jun 1993

Aug 1990-Jun 1993
Jan 1988-Mar 1990

Nov 1991-Jun 1993
Jul 1974-Jun 1975
Jun 1989-ongoing


1967-ongoing




Apr 1985-Dec 1986

Nov 1987-May 1994


Jan 1985-ongoing


Zoodsma (1991)

*aValade (1990)

Zoodsma (1991)
City of Jacksonville, Florida,
unpublished report
Kinnaird (1985)

Florida Department of Environmental
Protection, unpublished data
*CH2M Hill (1986)

Kinnaird (1985)

Florida Department of Environmental
Protection, unpublished data
Florida Department of Environmental
Protection, unpublished data
U.S. Fish and Wildlife Service, Shane (1983)
Provancha and Provancha (1988, *1989);
National Aeronautic and Space
Administration, unpublished data


U.S. Fish and Wildlife Service,
unpublished data


Brevard County, unpublished data
Florida Department of Environmental
Protection, unpublished data
U.S. Fish and Wildlife Service,
unpublished data
Florida Department of Environmental
Protection, unpublished data
Palm Beach County, unpublished data
Florida Department of Environmental
Protection, unpublished data
Broward County, unpublished data
*Odell (1976)
Dade County, unpublished data


Hartman (1979); *Powell (1981);
Powell and Rathbun (1984);
Kochman et al. (1985); Rathbun et al.
(1990); Chassahowitzka National
Wildlife Refuge, unpublished data
Florida Department of Environmental
Protection, unpublished data
Reynolds et al. (1991); Florida Department
of Environmental Protection and Eckerd
College, unpublished data
Kadel and Patton (1992)






BRUCE B. ACKERMAN 17


Table 1. Continued.
Area Dates Citations
Charlotte County Jan 1987-Dec 1988 Florida Department of Environmental
Protection, unpublished report


Lee County
Core area
Hendry Creek
Imperial River
Deep Lagoon
Collier County
North Collier
Naples area
Wiggins Pass area
Marco Island area
Everglades City
Ochopee
Port of the Islands
Ten Thousand Islands
(includes Everglades City, Ochopee,
Port of the Islands)
Everglades National Park
Everglades National Park
Everglades National Park


Jan 1984-Dec 1985
May 1988-Dec 1988
Mar 1987-Feb 1988
Jul 1986-Feb 1988

Feb 1987-Feb 1988
Jan 1986-Jan 1987
Feb 1987-Sep 1987
Jan 1989-Dec 1990
Jan 1986-Jan 1987
Mar 1987-Feb 1988
Jan 1986-Dec 1990
Jan 1991-Nov 1993


Sep 1973-Jun 1976
Dec 1979-Sep 1981
Mar 1990-Mar 1993


Florida Department of Environmental
Protection, unpublished report



Florida Department of Environmental
Protection, unpublished report








Odell (1979)
Everglades National Park, unpublished report
*Snow (1992)


a An asterisk denotes unpublished material.


In 1992, 10 surveys of distribution were in progress (3
by the Florida Department of Environmental Protection, 3
by other agencies, and 4 jointly by the department and
other agencies; Table 1). The department conducted sur-
veys in 1992 in St. Johns, Flagler, and Volusia counties
(unpublished data); St. Lucie and Martin counties (unpub-
lished data); Tampa Bay (*Reynolds et al. 1991; Florida
Department of Environmental Protection and Eckerd Col-
lege, unpublished data); and the Ten Thousand Islands in
Collier County (unpublished data). Surveys were con-
ducted in 1992 by the Broward County Office of Planning
(D. Burgess, unpublished data), Palm Beach County De-
partment of Environmental Resources (D. Carson, unpub-
lished data), and the Everglades National Park (*Snow
1992), each co-sponsored by the Florida Department of
Environmental Protection. Surveys were also conducted
in Dade County by the Dade County Department of Envi-
ronmental Resources Management (S. Markley, unpub-
lished data), in the Crystal River area by the Chassa-
howitzka National Wildlife Refuge (J. Kleen, unpublished
data), and in the Banana River by the National Aeronautic
and Space Administration (Provancha and Provancha
1988, *1989, unpublished data).
These surveys provided a detailed, up-to-date cover-
age of all counties in the typical range of manatees. An
extensive database of counts and distribution maps now
exists for most of the state (Table 1; Fig. 1). These data
were needed to support intensive manatee protection


proposed in 1989 to regulate watercraft speeds in 13
counties in Florida (*Florida Department of Natural Re-
sources 1989). Sighting data from all surveys conducted
by the Florida Department of Environmental Protection
and numerous other groups were digitized and entered
into the department's Marine Resources Geographic In-
formation System (O'Shea and Kochman 1990; Weigle
and Haddad 1990). Maps were created that display
manatee sightings from aerial surveys and locations
where dead manatees were reported, locations of mana-
tees tracked by telemetry, shorelines, aquatic habitats,
shoreline development, boat ramps, sources of fresh and
warm water, and water depths. The Marine Resources
Geographic Information System has been used exten-
sively by the department to develop protection of mana-
tees in concentration areas and to plan watercraft and
shoreline developments (Weigle and Haddad 1990;
B. Ackerman and K. Clifton, in preparation).

Problems and Limitations

Surveys provide valuable data on the seasonal distri-
bution and abundance of manatees and have now been
conducted in all major areas in the typical range of
Florida manatees around the state. They also provide
detailed information on habitat use and have been used
extensively to define areas that require legal protection.
However, a major criticism of surveys of distribution is
that they usually do not provide accurate or statistical






18 INFORMATION AND TECHNOLOGY REPORT 1


Fig. 1. Counties (shaded) in
Florida where aerial surveys
ofmranate (Trihecharma-
natuarostria)distribution
wre conduct.


estimates of the number of manatees present (*Eberhardt
1982; *Packard 1985; Lefebvre et al. 1995). Effort var-
ies by flight and by area. Visibility biases are probably
not equal among various habitats. Therefore, surveys of
distribution do not provide good population-size esti-
mates, especially where visibility bias is large.
Surveys by the Florida Department of Environmental
Protection and others were usually conducted twice
monthly for 2 years. This provided a sample of the mana-
tees' seasonal distribution; however, a determination of
whether this is a sufficient sample has not been possible.
Twice-monthly surveys for 2 years seem more adequate
than a smaller sample (less frequent or shorter duration
surveys) for documenting manatee distribution. This
schedule at least compensates for short-term weather
changes between months and between consecutive
years. But longer-term studies are needed to monitor
changes in numbers or shifts in habitat use (Provancha
and Provancha 1988, *1989; Rathbun et al. 1990;
*Reynolds et al. 1991; *Kadel and Patton 1992). Lim-
ited resources require trade-offs between surveys in


many study areas fora short time each or surveys in few
areas for longer periods.
Because of resource limitations, in 1992 the Florida
Department of Environmental Protection began shifting its
emphasis from these surveys to the improvement of tech-
niques. However, this temporary moratorium on new sur-
veys of distribution by the department is not without costs.
Long-term monitoring is probably needed to assess
changes in populations and habitat use. At least, additional
distribution data are probably needed on a rotating sched-
ule, perhaps every 5 years, to update data used for the
protection of manatees. Many existing data are now older
than 5 years, and additional distribution information may
soon be required.

Surveys ofManateeAggregations in
Winter

In winter, Florida manatees are forced to travel to warm
water because of the low water temperatures in much of
the state (-Lefebvre and Frohlich 1986; Reid et al. 1991;


fm \_
I W--<






BRUCE B. ACKERMAN 19


Ackerman et al. 1995; Reid et al. 1995). Manatees migrate
either far to the south (e.g., Dade, Monroe, and Collier
counties), to a few natural springs (principally the Crystal
River and Blue Spring), or to industrial warm-water efflu-
ents (Fig. 2). Aggregations in these areas allow the count-
ing of large numbers of manatees from the air with rela-
tively short, concentrated efforts. Some long-term studies
provided information on trends in the sizes of these aggre-
gations. Best results should be obtained from a regional
population that aggregates in one small area where clear
water allows accurate counting.

Objectives

Surveys of manatee aggregations in winter serve to
determine the changing numbers of the animals at warm-
water sites. Surveys may also be useful for assessing
trends in counts. These objectives are primarily useful
for measuring progress toward long-term recovery goals.
However, these surveys are also valuable for manage-
ment, such as defining boundaries of seasonal sanctuar-
ies to protect manatee aggregations in winter. Intensive
counts were made at aggregation sites in winter as early
as 1967 (Hartman 1979). These counts allow the eco-
nomical counting of a high proportion of all animals in
a large region. Long-term studies have been made at the
Crystal River, at Blue Spring, and at power plants on the
eastern coast and near Fort Myers. These surveys were


initially used to determine the number of manatees near
aggregation sites throughout the year but later were
focused on counts only in winter.

Procedures

Surveys of manatees at power plants on the eastern
and southwestern coasts of Florida have been conducted
from 1977 to the present (*Rose and McCutcheon 1980;
*Raymond 1981; *McGehee 1982; Reynolds and Wil-
cox 1985, 1986, 1994; *Reynolds 1993, 1994). These
surveys were a part of larger research to determine the
year-round distribution and abundance of manatees. Sur-
veys were funded by the Florida Power and Light Com-
pany. Surveys took place at six major power-generating
stations, including five Florida Power and Light Com-
pany plants and one Orlando Utility Commission plant
(Fig. 2) and their adjacent waters. Manatees were sur-
veyed at several other sites, including smaller power
plants at Fort Pierce and Vero Beach and in the Hobe
Sound area. Areas within about 8 km of each power plant
were included in surveys of manatees (*Rose and
McCutcheon 1980; Reynolds and Wilcox 1994) with
intensive-area methods (*Packard 1985). Water clarity,
general visibility, air traffic problems, and attractiveness
to manatees varied among sites.
Details of survey timing and methodology were provided
by Rose and McCutcheon (*1980), *Raymond (1981),


Spring
Indian River
wCape Canaveral

Vero Beach
Harbor Branch
Fort Pierce
Hobe Sound
Riviera Beach


Fig. 2. Warmwater sites in
Florida that manatees
(Trichechus manatus lati-
rostris) frequented in
winter. Symbols indicate
natural springs (trian-
gles) and power plants
(circles) that were in-
cluded in surveys by the
Florida Power and Light
Company in winter.


Fort Lauderdale
Port Everglades


/:


Crystal


Fort MerI






20 INFORMATION AND TECHNOLOGY REPORT 1


*Packard and Mulholland (1983), and Reynolds and Wilcox
(1985, 1986, 1994). Surveys of manatees at power plants
were initially 2-day surveys in each week in winter and
2/month in summer. In 1980, these were reduced to 1-day
surveys in each week in winter. By 1982 a schedule was
achieved that consisted of 4-10 flights each year, only after
winter cold fronts. Except for scheduling, the same survey
methods have been used since 1977, and the same biologist
has conducted the surveys since 1982. This is the longest
series of available counts of manatees on the eastern and
southwestern coasts of Florida.

Results
Following the suggestions of Eberhardt (* 1982), Packard
and Mulholland (*1983) conducted preliminary statistical
analyses of the survey counts during 1977-82. They at-
tempted to adjust or correct the counts at each power plant,
based on air and water temperatures (*Packard and Mulhol-
land 1983; *Packard et al. 1984). Their analyses showed
complex relations between the counts at each plant and air
and water temperatures and other environmental factors, but
counts could not be adjusted for these factors. The numbers
of manatees at each plant increased from fall to winter but
were highly variable at each plant between consecutive
flights and among years. Cold fronts caused manatees to
aggregate at certain plants, resulting in high counts. More
manatees stayed near the plants on cold days but dispersed
from the plants on warmer days to feed. More recent studies
with telemetry revealed that some manatees migrate farther
southward during the coldest part of the winter, moving
among southern plants for a few days, then migrating back
northward in spring (Reid et al. 1991, 1995).
The highest summation of counts of manatees at all
eastern coast power plants on a single date each winter
showed an upward trend (linear regression, r2= 0.39,
n = 17, P= 0.01; Fig. 3). Annual high counts at the Fort My-


78 80 82 84 86 88 90 92 94
Year
Fig. 3. Largest numbers of manatees (Trichechus manatus
latirostris) in aggregation areas at power plants on the eastern
coast of Florida, 1977-1994. Data were collected during
aerial surveys in winter (Reynolds and Wilcox 1994 and
sources therein). Data from 10 power plants were combined.


ers plant were too variable to show a trend (r2 = 0.01, n = 17,
P = 0.78; Fig. 4). However, without correcting for short-
term and long-term temperature effects on counts, the annual
high counts did not reveal convincing trends (Figs. 3 and 4).
Garrott et al. (1995) improved trend information from
data from the surveys at power plants during 1982-91 by
developing statistical models that adjust the counts based
on short-term and long-term air and water temperature
patterns. The adjusted counts in 1982-91 at the power
plants on the Atlantic Coast of Florida significantly
increased when corrected for temperature. This sug-
gested but did not prove that the actual size of the
Atlantic Coast manatee population also increased. The
adjusted counts at the Fort Myers plant did not show a
significant trend.

Problems and Limitations

The percentage of manatees that is in aggregations in
winter to be counted at any given time and the percentage
of animals that are actually observed are not known (*Eber-
hardt 1982; *Packard 1985; Lefebvre et al. 1995). These
counts are the only long-term data from the eastern-coast
and southwestern aggregation sites, and considerable effort
is justifiable to develop an index to past and future popula-
tion-size trends from these data (*Packard 1985; Packard
et al. *1984, 1986, 1989). An increase in knowledge of
factors that influence these counts is important. Analysis of
counts with telemetry data on locations and behavior of
manatees by the area (Reid et al. 1995) and air and water
temperatures may provide further information for correcting
for possible biases in these counts.

Synoptic Survey

The synoptic survey is designed to obtain statewide

counts of manatees in all winter habitats at one time. De-


78 80 82 84 86 88 90 92 94
Year
Fig. 4. Largest numbers of manatees (Trichechus manatus
latirostris) at the Fort Myers power plant, Lee County,
southwestern Florida, 1977-1994. Data were collected
during aerial surveys in winter (Reynolds and Wilcox 1994
and sources therein).


r2 = 0.01
p = 0.78






BRUCE B. ACKERMAN 21


scriptions of the synoptic survey have not been published.
This procedure combines some features of the intensive
counts of manatees in aggregations in winter and extended-
area surveys of manatee distribution. The goal is to obtain
the highest, presumably most accurate count, which then
serves as a new baseline to evaluate indexes from other
surveys. Plans were initially made by the Florida Depart-
ment of Environmental Protection to conduct a synoptic
survey in 1989. Surveys were not made until 1991, however,
because of the lack of weather patterns suitable for synoptic
surveys in 1989 or 1990. In 1990, the Florida State Legisla-
ture mandated "an impartial scientific benchmark census of
the manatee population to be conducted annually" by the
Florida Department of Environmental Protection (Florida
Statute 370.12.5a). The subsequent synoptic surveys were
made to comply with this mandate.

Procedures
The term synoptic was used to designate comprehensive
coverage of a large area at one time. Plans were made to
cover the entire potential winter range of manatees in Florida
and in southeastern Georgia in 2 days (Fig. 1). Most routes
followed those of recent surveys, principally over rivers,
estuaries, bays, streams, and canals along most of the coast-
line of peninsular Florida. Most aircraft were Cessna 172's,
although in a few areas manatees were surveyed from heli-
copters. Ground-based observers searched Blue Spring and
two industrial sites in southeastern Georgia where vertical
visibility was poor. Planes flew at an altitude of 150 m and
at a speed of 130 km/h. Transects were used to cover some
wider areas. These methods were flexible and were altered
in some cases to accommodate local conditions. Routes
were planned and teams were on standby for 2 months each
winter. Each count was planned to follow two cold fronts or
a prolonged cold period in January or February when mana-
tees were concentrated at warm-water aggregation sites and
when the winter migration was probably completed. The
desired weather pattern usually was forecast with only 2-3
days notice, which made planning difficult.
Manatees on the eastern coast of Florida were surveyed
on the first day and on the western coast on the second day.
Surveys required nearly simultaneous counts by biologists
from numerous agencies. Whenever possible, manatees in


adjacent areas were simultaneously surveyed to minimize
the effects on counts from manatee movements between
areas. The survey was made on 2 days because of the limited
number of available biologists and aircraft. Movement of
manatees between the eastern and western coasts was as-
sumed to be minimal (O'Shea 1988). Counts of manatees in
all areas were tallied, and possible duplicate counts were
eliminated by taking into consideration the manatees'
mapped positions and the distance the manatees may travel
in the elapsed time between adjacent surveys.
Synoptic surveys of manatees were conducted three
times, twice in 1991 and once in 1992. Following successful
surveys in 1991, the flight window in 1992 was shortened
to the period from early January to mid-February, and
planned coverage was reduced in a few areas. After 1991,
surveys conducted north of Tampa on the western coast and
north of Fort Pierce on the Atlantic Coast covered only
warm-water aggregation sites. This reduced the number of
teams and the cost. Few manatees were in the omitted areas
during the two previous surveys. Conditions on the most
recent survey in 1992 were more conducive to higher counts
of manatees than during previous synoptic surveys. A sig-
nificant cold front passed across Florida on 14 January 1992,
and a second on 16 January. Weather conditions were good;
several weeks of cool weather were followed by several days
of steadily decreasing temperatures. The survey was con-
ducted during 17-18 January; the weather was clear and
cold and winds diminished in most areas. Counts were made
on 21 survey routes. Ten aircraft flew simultaneously on the
first day, nine on the second day. Surveys were not con-
ducted during the winters of 1992-93 or 1993-94 because
of weather conditions.

Results
Two surveys were completed in 1991 (23-24 January
and 17-18 February). Counts during the first survey were
679 manatees on the eastern coast and 589 on the western
coast or a total of 1,268 manatees (8.6% calves; Table 2). A
second survey was conducted during 17-18 February 1991.
Weather had been warm for almost a month, and a strong
cold front passed on 15 February. A total of 1,465 manatees
(8.8% calves) was seen, 813 on the eastern coast and 652 on
the western coast. This was about 20% more than in any


Table 2. Counts of Florida manatees (Trichechus manatus latirostris) from synoptic aerial surveys on the eastern and
western coasts of Florida, 1991-1992.
Percent Number of
Date Eastern Western Total calves Teams Observers
23-24 Jan 1991 679 589 1,268 8.6 27 32
17-18 Feb 1991 813 652 1,465 8.8 27 32
17-18 Jan 1992 907 949 1,856 8.7 21 28






22 INFORMATION AND TECHNOLOGY REPORT 1


previous statewide count (Table 2). During both surveys,
almost all of the manatees seen in the northern part of the
state were at warm-water aggregation sites (Table 3). During
the second survey in the southern part of the state, about half
were away from warm water (Table 3), particularly in the
southeast.
The largest number of Florida manatees ever recorded
was seen during the synoptic survey during 17-18 January
1992. A total of 1,856 manatees (8.7% calves) was counted,
907 on the eastern coast (7.6% calves) and 949 on the
western coast (9.6% calves; Tables 2-4; Fig. 5).

Problems and Limitations of the Synoptic
Survey
The synoptic surveys provided new information on the
minimum size of the manatee population in Florida: 1,856
manatees in January 1992. Until that time, only 1,200 were
known to exist. The greatest value of the surveys may be as
a snapshot of the whole state population at once, reducing
the possibility of movements among areas between adjacent
count segments. However, these counts seem to be highly
variable among surveys, depending on weather conditions.
Weather conditions will never be optimal in all areas at once,
and counts in some areas may be maximized under opposite
weather conditions. In several areas, higher counts were
obtained on dates other than during synoptic surveys. Opti-
mal conditions are difficult to predict, and the best condi-
tions in each winter are easily missed. Surveys were not
conducted during some winters because of inadequate
weather conditions. The results are not statistical estimates
of population size and will probably not provide estimates
of population-size trends. They are more costly than many
other surveys because of the required large number of
aircraft and biologists and the large amounts of coordination.


Population Size and Size Trends

Trends Based on Surveys of
Aggregations in Winter

Crystal River Area
Aerial surveys of manatees have been conducted in the
Crystal River region in winter since 1967, except during the
three winters from 1970-71 to 1972-73. Surveys covered


the Crystal and Homosassa rivers and surrounding areas
(Hartman 1979; *Powell 1981; Powell and Rathbun 1984;
Kochman et al. 1985; Rathbun et al. 1990; Chassahowitzka
National Wildlife Refuge, unpublished data). These are the
only aggregation sites in winter in northwestern Florida and
are characterized by clear water. This area is largely isolated
in winter from other aggregation sites; however, counts and
distribution in the Crystal River area markedly change dur-
ing and between winters. Animals leave on feeding excur-
sions, and changes in counts can be substantial from week
to week (Rathbun et al. 1990).
I used exponential regression to examine maximum aer-
ial counts in each winter from 1967 to 1994 for trends.
Maximum aerial counts each winter in the Crystal River area
were based on data from Hartman (1979), Powell (*1981),
Powell and Rathbun (1984), Rathbun et al. (1990), and
unpublished data of the Chassahowitzka National Wildlife
Refuge. Exponential regression was used because popula-
tions often change exponentially and because it allows
simpler expression of the annual percentage change in num-
bers. The formula for exponential regression is
bt
y = axe (1)
where y is the count, t is the year, and a and b are regression
coefficients (Eberhardt and Simmons 1992). This is equiva-
lent to the linear regression form
In(y) = In(a) + bt (2)
This further gives
b
annual percent change = (e 1) x 100%.(3)
Manatee counts in the Crystal River area showed a
significant upward trend with low variability (Fig. 6;
+9.7%/year; r2 = 0.93, n = 23, P < 0.001). Maximum
counts increased from 38 during 1967-68 to 292 during
1992-93. Similarly, the total number of manatees iden-
tified during each winter (Hartman 1979; *Powell 1981;
Powell and Rathbun 1984; Rathbun et al. 1990) in-
creased through time.
Moore (1951) described manatee sightings as rare in the
Crystal River area in the 1940's and earlier. Increasing
populations were noted by later researchers (Hartman 1979;
Powell and Rathbun 1984; *Beeler and O'Shea 1988;
O'Shea 1988; Rathbun et al. 1990). Life-history studies of
known individuals suggested that most of this increase could


Table 3. Percentage of Florida manatees (Trichechus manatus latirostris) at warm-water sources during synoptic
surveys in northern and southern zones of Florida, 1991.
Date Northern zone Southern zone Statewide
23-24 Jan 1991 93 7 53
17-18 Feb 1991 97 47 85






BRUCE B. ACKERMAN 23


Table 4. Synoptic aerial survey of Florida manatees (Trichechus manatus latirostris) in Florida and Georgia, 17 (eastern
and coast) 18 (western coast) January 1992, unless otherwise noted.
Total number Number of
Location of survey of manatees calves Comments


Eastern coast
Camden/Glynn counties, Georgia

Nassau/Duval counties
St. Johns/Flagler/Volusia counties
Volusia County (St. Johns River)
Blue Spring Run
Brevard County
OUC plant
Cape Canaveral plant
Indian River County
Vero Beach plant
Other sites
St. Lucie/Martin counties
Fort Pierce plant
Other sites
Palm Beach County
Riviera Beach plant
Other sites
Broward County
Port Everglades plant
Fort Lauderdale plant
Other sites
Dade County
Monroe County (Florida Keys)
Eastern total
Western coast
Citrus/Levy counties
Western Tampa Bay
Eastern Tampa Bay
Manatee/Sarasota counties
Sarasota/Charlotte counties
Charlotte/Lee counties
Lee/Collier counties
Monroe County (western Everglades)
Monroe County (eastern Everglades)
Okeechobee Waterway/Lee/Hendry/
Glades/ Okeechobee/Martin/
Palm Beach counties
Western total
Grand total
a Calf counts were not obtained at three sites.


Count from shore at four
industrial plants
Five industrial plants and vicinity
Conducted 18 Jan
Count made from canoe

Aerial counts at power plants,
warm-water sites, known use
areas, and immediate vicinity


Survey using helicopter


Survey using helicopter


Conducted 17 Jan
Conducted 17 Jan


949
1,856


result from reproduction and survival of resident manatees
(Eberhardt and O'Shea 1995), but some of the increase
probably also resulted from permanent immigration by
adults from areas farther south and recruitment of their
descendants (Powell and Rathbun 1984; O'Shea 1988;
Rathbun et al. 1990; Rathbun et al. 1995). At the Crystal


River, food supplies are abundant, unlike at most other
winter aggregation sites (Hartman 1979; O'Shea 1988), and
increases in manatee numbers in the region have been
attributed in part to the introduction of exotic aquatic vege-
tation in the mid-1960's (Hartman 1979; Powell and Rath-
bun 1984; O'Shea 1988).






24 INFORMATION AND TECHNOLOGY REPORT 1


kilometers
15 0 15 30


Fig. 5. Locations of 1,856 manatees (Trichechus manatus latirostris) seen during the third synoptic aerial survey in Florida and in
southeastern Georgia, 17-18 January 1992. Each symbol indicates one group.


E 200
o
0 150-
r2 = 0.93
100- 9.7 %/yr
50p- T! p < 0.-001

69 71 73 75 77 79 81 83 85 87 89 91 93
Year
Fig. 6. Trend (exponential regression) in counts of Florida
manatees (Trichechus manatus latirostris) made during aerial
surveys at winter aggregation sites in the Crystal River area,
Citrus County, 1967-1994. Highest count obtained during
each winter is shown. Data were collected by Hartman (1979),
Powell (*1981), Powell and Rathbun (1984), Rathbun et al.
(1990), and the Chassahowitzka National Wildlife Refuge
(unpublished data).


Blue Spring

Counts of manatees at the aggregation site at Blue Spring
State Park provide the only long-term trend data about
manatees in the St. Johns River. The spring run is shaded
by a tree canopy, and aerial surveys are therefore not prac-
tical. However, the clear water allows accurate counts from
canoes and shore, and all individuals are identifiable (T. J.
O'Shea, National Biological Service, Fort Collins, Colo-
rado, personal communication). Counts have been con-
ducted almost daily in winter since 1970 except during the
winters of 1972-73 and 1973-74. Analysis of the total
number of manatees identified during each winter was based
on data obtained from Hartman (1979), Powell and Waldron
(*1981), O'Shea (1988), T. J. O'Shea (National Biological
Service, Fort Collins, Colorado, personal communication),
and W. C. Hartley (Florida Park Service, Orange City,
Florida, personal communication).






BRUCE B. ACKERMAN 25


VU-
80"
70-
n 60
S50-
4 40
2 30
20
10


70 72 74 76 78 80 82 84 86 88 90 92 94
Year
Fig. 7. Trend (exponential regression) in counts (highest total
number of animals identified during each winter) of Florida
manatees (Trichechus manatus latirostris) at the winter
aggregation site in Blue Spring State Park, Volusia County,
1970-1994. Highest counts of manatees on a single day each
winter at Blue Spring are also shown. Data were collected
during surveys by canoe and snorkeling by Hartman (1979),
Powell and Waldron (*1981), O'Shea (1988) and the National
Biological Service and Florida Park Service (unpublished
data).


The number of manatees identified at Blue Spring
each winter showed an upward trend with low variability
(+8.2%/year; r2 = 0.94, n = 22, P < 0.001; Fig. 7), in-
creasing from 11 during 1970-71 to 88 during 1993-94.
Similarly, highest single-day counts each year increased
from 11 during 1970-71 to 81 during 1993-94
(+8.7%/year; r2 = 0.96, n = 19, P < 0.001).
Increases in counts at Blue Spring were discussed by
Beeler and O'Shea (*1988), O'Shea (1988), O'Shea and
Hartley (1995) and O'Shea and Langtimm (1995). Stud-
ies of known individuals showed that most of this in-
crease resulted from reproduction and survival of resi-
dent animals, but part was also from permanent
immigration of adults (O'Shea 1988). A maximum of
about one-third of the increase in counts at Blue Spring
may be due to immigration and subsequent reproduction
by the immigrants (T. J. O'Shea, National Biological
Service, Fort Collins, Colorado, personal communica-
tion). Forty-two of the 63 manatees identified at Blue
Spring during the 1990-91 winter season were present
during the first 3 years of intensive study (winters 1978-
79 through 1980-81, n = 14 animals) or were descen-
dants (n = 28) of those animals. The remaining 21 were
immigrants or their offspring. Therefore, during a 10-
year period, 67% of the net population growth was from
internal recruitment (T. J. O'Shea, National Biological
Service, Fort Collins, Colorado, personal communica-
tion). This is a minimum estimate because some indi-
viduals identified as new immigrants could have been
offspring of long-term residents that returned when older
but were no longer identifiable as such because of new


STotal identified
] Single day count



n" o +8.2 %/yr
ar2 = 0.94
7 p < 0.001


marks. O'Shea and Langtimm (1995) found high adult
survival at Blue Spring. Eberhardt and O'Shea (1995)
estimated positive population growth rates at Blue
Spring based on demographic data and discussed results
in relation to trends in counts. Manatees at Blue Spring
have been increasingly protected from direct injury from
watercraft and from disturbance by boats, swimmers,
and divers (Hartman 1979; *Powell and Waldron 1981;
*Beeler and O'Shea 1988; O'Shea 1988). This increased
protection may have encouraged more manatees to use
these aggregation sites.

Tampa Bay
Weigle et al. (*1988) summarized the highest annual
aerial counts of manatees obtained in the Tampa Bay area
from 1979 to 1986, and no counts exceeded 76 manatees.
Subsequently, counts of manatees during surveys during
the winters of 1987-94 were as high as 190 (24 January
1994; *Reynolds et al. 1991; Florida Department of Envi-
ronmental Protection and Eckerd College, unpublished
data). However, major changes also occurred in warm-
water sources in the 1980's (*Weigle et al. 1988). The
Gardinier Phosphate Plant discharge into the Alafia River
ceased in 1986, and a small no-entry zone was created at
the Tampa Electric Company's Big Bend plant in 1986 and
expanded to the entire discharge canal in 1989 (*Reynolds
et al. 1991). These actions probably reduced disturbance
to manatees from boats and resulted in a shift in manatee
use from the Alafia River to the Big Bend plant during
1985-86 (*Weigle et al. 1988). The increasing levels of
protection may have encouraged manatees to immigrate
from other wintering areas, although this has not so far
been detected by studies with telemetry (*Lefebvre and
Frohlich 1986; B. L. Weigle, Florida Department of Envi-
ronmental Protection, St. Petersburg, Florida, unpublished
data) or scar catalog studies (Beck and Reid 1995; B. L.
Weigle, Florida Department of Environmental Protection,
unpublished data). The number of documented manatee
deaths is low in the Tampa Bay area (O'Shea et al. 1985;
*Reynolds et al. 1991; Ackerman et al. 1995). Manatee
sightings were apparently rare in the Tampa Bay area until
the 1950's (Moore 1951), and numbers seemingly were
low until the 1970's (Hartman *1974, 1979; Irvine and
Campbell 1978; Irvine et al. 1982; *Beeler and O'Shea
1988).

Counts at Power Plants
Trends in counts of aggregations in winter at seven
power plants on the eastern coast and at the Fort Myers
plant on the western coast were presented above and by
Garrott et al. (1995). These are the only long-term trend
data in these areas. Results from Garrott et al. (1995)
suggested that adjusted counts at the plants on the eastern






26 INFORMATION AND TECHNOLOGY REPORT 1


coast increased but showed no evidence of change at the
Fort Myers plant on the western coast. Although the num-
ber of documented deaths in both areas is relatively high
(O'Shea et al. 1985; Ackerman et al. 1995), models based
on demographic information suggested that past manatee
population growth on the eastern coast could have oc-
curred but, if so, at a low rate (Eberhardt and O'Shea
1995).

Trends during Warm Seasons in the
Banana River

Intensive surveys of manatees during warm seasons
have been made in the northern Banana River since 1977
(Shane 1983; Provancha and Provancha 1988, *1989;
National Aeronautical and Space Administration, unpub-
lished data). These are the only long-term counts during
warm seasons on the eastern coast. Counts were made
from an airplane during 1977-80 and from a helicopter
during 1980-81 and from 1984 to the present; similar
flight routes were used (Provancha and Provancha 1988).
Aerial counts increased through time; the counts were
highest in spring each year (Provancha and Provancha
1988, *1989; National Aeronautical and Space Admini-
stration, unpublished data). High counts in spring seemed
to reflect a temporary influx of manatees in transit during
spring migration (Provancha and Provancha 1988). Recent
counts in spring were as high as 200-400 manatees.
Counts in summer (June to August) north of the NASA
Causeway from 1977 to 1981 did not exceed 30 animals
(Provancha and Provancha 1988), but recent counts were
as high as 139 (J. Provancha, National Aeronautical and
Space Administration, Kennedy Space Center, Florida,
unpublished data). Provancha and Provancha (1988) sug-
gested that no other area on the eastern coast offers as
much protected suitable habitat as the northern Banana
River. Much of this area is inside the Kennedy Space
Center and has been closed to boating for many years for
security reasons. A larger area in the Merritt Island Na-
tional Wildlife Refuge (north of the Cape Canaveral Barge
Canal) was closed in 1990 for manatee protection, substan-
tially decreasing human disturbance. Provancha and
Provancha (1988) suggested two possible reasons for in-
creased counts in the northern Banana River: increases in
manatee population size and shifts in habitat use by mana-
tees into the northern Banana River because of increased
development or disturbance outside this sanctuary.

Problems and Limitations in Long-term
Survey Data

Manatee populations in all areas for which long-term
data exist seem to be increasing (eastern-coast power
plants, Crystal River, Blue Spring, Tampa Bay, Banana


River in summer) or appear stable (Fort Myers power
plant). Although these data are encouraging, they do not
encompass all areas, and trends may differ in populations
lacking long-term data. Moreover, interpretations of in-
creases in counts over time are difficult because of various
confounding factors (Reynolds and Wilcox 1994; Garrott
et al. 1995). Although the increasing counts in most areas
where manatees were surveyed suggested that populations
increased, alternative explanations exist. Manatees are
able to locate and use protected areas. Manatees may have
increased use of these areas and avoided other areas in
response to improving resources such as availability of
warm water, protection from human disturbance, and in-
creasing aquatic plants. Manatees may have become more
visible to biologists, perhaps because a higher proportion
uses aggregation sites than in the past or because of im-
proved visibility or changes in plant operations (Packard
et al. 1989; Reynolds and Wilcox 1994). In most cases,
different people conducted surveys, the survey procedures
evolved, and skills or techniques for surveying aggregated
manatees in winter may have improved. Each of these
alternative explanations could also result in the observed
upward trends of counts in some refugia in the absence of
an actual increase in the manatee population size. Al-
though no data support these alternative speculations on
the apparent increases in counts, conclusions that popula-
tions actually increased in these parts of the state could be
erroneous (Reynolds and Wilcox 1994; Eberhardt and
O'Shea 1995; Garrott et al. 1995).

Statewide Population-size Estimates

A population-size estimate of Florida manatees is de-
sirable as a baseline for estimating trends, modeling popu-
lations, and assessing the effect of observed mortality.
Statewide surveys were designed to provide such a base-
line. However, other than the synoptic surveys, only a
small number of studies included simultaneous counts of
manatees throughout the southeastern United States. Hart-
man (*1974) counted 255 manatees throughout Florida
and southeastern Georgia in summer 1973; he used one
plane during six consecutive weeks. A crude correction
factor was based on water clarity (*Hartman 1974). Hart-
man concluded from these counts and interviews of the
public that probably 800 manatees (range 750-850) were
in Florida and Georgia. Counts in summer have since been
lower than counts in the same general areas in winter
(Irvine and Campbell 1978; Rathbun et al. 1990;
*Reynolds et al. 1991). Manatees are dispersed widely in
summer in small groups at low densities in unpredictable
locations. Visibility is poor in many areas because of
turbid water and overhanging trees.
In winter 1976, Irvine and Campbell (1978)
counted 738 manatees in Florida and Georgia. Nine teams






BRUCE B. ACKERMAN 27


conducted surveys statewide in 6 days after a cold front.
Manatees in most areas were counted simultaneously in a
single day. These and other data suggested a population of
at least 800-1,000 in 1978 (*Brownell et al. 1981). A
survey was also conducted in August 1976 (Irvine and
Campbell 1978) and included parts of the Florida panhan-
dle, Georgia, and the Carolinas; 245 manatees were seen,
similar to Hartman's (* 1974) count in summer 1973. After
1976, research shifted to other topics because these mass
efforts were too costly and logistically difficult and did not
provide data for clear interpretation (*Eberhardt 1982;
G. B. Rathbun, National Biological Service, San Simeon,
California, personal communication).
High counts at power plants in January 1985 and counts
in other areas of the state led experts to revise the minimum
statewide estimate to 1,200 in 1985 (O'Shea 1988). How-
ever, surveys were not made at the same time in all areas,
and manatees in some important areas had not been
counted at all in several years. A record single-day count
of 717 manatees at selected power plants was made under
favorable counting conditions in January 1986, and a
higher count of 804 in February 1988 (Reynolds and
Wilcox 1994). In December 1989, a composite of counts
in various areas during a short time period revealed 1,240
manatees (B. B. Ackerman, Florida Department of Envi-
ronmental Protection, St. Petersburg, Florida, unpublished
data). As discussed above, the synoptic survey in January
1992 revealed a count of 1,856.
Interpretation of the results of these statewide surveys
is difficult. In 19 years the best minimum estimate in-
creased from 800 to 1,856, but these data were obtained
with survey methods that differed in several important
ways. O'Shea (1988) reviewed statewide manatee popu-
lation counts through 1985. He found no firm evidence of
a decrease or increase in manatee populations in spite of
the increase in the official minimum estimate because the
methods were without a measure of precision.
What does the record count of 1,856 in January 1992
reveal about trends in the statewide manatee population?
Perhaps not much. Even though the best minimum counts
increased from 800 in 1973 to 1,000 in 1978, to 1,200 in
1985, and to 1,856 in 1992, a basis to determine the
statistical significance does not exist (*Eberhardt 1982;
O'Shea 1988). Previous surveys were not over as large an
area, under as good conditions, as comprehensive, or in as
short a time as the 1992 synoptic survey.
This higher count does not provide evidence that the
population is no longer endangered. Rescaling to a new
baseline does not change the fact that mortality from
various anthropogenic causes is still increasing and that
these threats may be greater than the population can with-
stand (*Brownell et al. 1981; O'Shea 1988; Marmontel
1993; Ackerman et al. 1995; Eberhardt and O'Shea 1995).


Accurate assessment of the effects of anthropogenic mor-
tality on the manatee population is not yet possible. There-
fore, higher statewide counts provide no reason to relax
conservation. In the interim, however, methods of state-
wide surveys must be improved. Goals should be to use
statistical sampling methods to reduce the required effort
in a wide-scale survey, provide statistically meaningful
estimates with confidence limits, correct for counting bi-
ases, and reduce the total cost (Lefebvre et al. 1995).

Research to Improve Survey Techniques

Techniques for estimating population sizes of Florida
manatees are currently inadequate (*Eberhardt 1982;
Packard et al. 1985, 1986; O'Shea 1988; *Reynolds and
Gluckman 1988; *U.S. Fish and Wildlife Service 1989;
Lefebvre and Kochman 1991; Lefebvre et al. 1995). No
basis exists to statistically measure trends in population
size, correct for visibility errors (visibility bias), or assign
confidence levels to minimum counts (*Eberhardt 1982;
*Packard and Mulholland 1983; Packard et al. *1984,
1986; *Packard 1985; Lefebvre et al. 1995). The number
of manatees cannot be estimated from a random sample of
current surveys. The estimated number of manatees in one
subunit cannot be extrapolated to other subunits. Counts
are not corrected for visibility bias. The current survey
procedures probably lead to an underestimation of the
number of manatees and do not provide estimates of the
precision of the count (i.e., standard deviation).

Research to Determine Visibility and Absence
Bias
Visibility bias (the proportion of missed animals) is one
of the largest problems in estimating manatee population
sizes (Lefebvre et al. 1995). Visibility bias in aerial counts
of other animals was determined with known or marked
subpopulations (Eberhardt et al. 1979; Pollock and Ken-
dall 1987). The proportion of a known number of radio-
tagged animals observed during counts has also been used
to estimate bias in various species of large mammals
(Floyd et al. 1979; Gasaway et al. 1985; Packard et al.
1985, 1989; Samuel et al. 1987; Ackerman 1988). Other
researchers used known subpopulations comprising
groups monitored intensively by another method or from
the ground (*Hartman 1974; Samuel and Pollock 1981) or
known numbers of penned animals (Packard et al. 1989;
Unsworth et al. 1990). Accurate counts are needed to
obtain correction factors (Eberhardt et al. 1979; *Eber-
hardt 1982; Pollock and Kendall 1987). Correction factors
were developed for counts of other species-some terres-
trial-and were based on group size, behavior, and habitat
(Eberhardt et al. 1979; Floyd et al. 1979; Samuel and
Pollock 1981; Gasaway et al. 1985; Samuel et al. 1987;
Ackerman 1988; Marsh and Sinclair 1989b).






28 INFORMATION AND TECHNOLOGY REPORT 1


Packard et al. (1985) investigated visibility bias in sur-
veys of manatees in the St. Johns River based on known
numbers of radio-tagged and unmarked manatees near Blue
Spring. Many manatees were not seen in these turbid waters;
an estimated 38-47% were seen. The seen proportion varied
by river, lake, and creek habitats. The radio tags were often
not visible from aircraft when the peduncle was submerged.
Packard et al. (1989) investigated visibility bias in winter
near Fort Myers with telemetry. Floating transmitters used
then were not easily seen from aircraft (*Packard 1985;
Packard et al. 1989). These researchers did not determine
the environmental variables that correlate with visibility or
suitable correction factors for current surveys.
However, as recommended by Lefebvre et al. (1995),
additional assessment of visibility bias is underway. Pre-
liminary tests of visibility bias in aerial counts of manatees
were made by the Florida Department of Environmental
Protection and collaborators during 1990-92 with radio-
telemetry and are planned for the future. Analysis of data
will be made with logistic regression to identify variables
(covariates) that influence the probability of seeing a given
group of manatees. Significant variables will be used to
develop an equation (visibility model) to predict the prob-
ability of seeing groups under various environmental con-
ditions. This equation can then be converted to a visibility
correction factor (Samuel et al. 1987; Ackerman 1988;
Steinhorst and Samuel 1989) on a group-by-group basis.
Success of this approach depends on the visibility of the
tags from the air (*Packard 1985; Pollock and Kendall
1987).
A different approach to estimating the number of ani-
mals missed in surveys is a double-sampling technique
(Pollock and Kendall 1987; Marsh and Sinclair 1989b;
Marsh 1995). Paired observers view the same area, and
each counts and maps seen animals. The number of ani-
mals seen by each observer is determined, and mark-re-
capture statistics are used to estimate the number of objects
missed by both observers. This provides an estimate with
confidence limits of the total number of objects in the
surveyed area. When animals are in groups, the number of
groups is estimated and is multiplied by the average group
size to provide the total number of present animals (Marsh
and Sinclair 1989b; Marsh 1995). Correction factors are
based on seen or missed groups-not individuals-be-
cause sightings of members in a group are not considered
independent. Double-observer counts were tested during
monthly aerial surveys of manatee distributions in Tampa
Bay during 1989-92 (*Reynolds et al. 1991; B. Acker-
man, Florida Department of Environmental Protection, St.
Petersburg, Florida, unpublished data) and will be made in
the future.
Telemetry data can also be used to investigate absence
bias in determining whether manatees are present in the


survey area as expected (Packard et al. *1984, 1989;
Marsh and Sinclair 1989b; Lefebvre et al. 1995; Marsh
1995). For example, locations of radio-tagged animals can
be used to estimate the proportion of all animals available
to be counted at power plants on a given day.

Use of Novel Approaches
Lefebvre et al. (1995) suggested that standardized sur-
veys during warm seasons should be tested as supplements
to counts in winter. Short-term weather patterns probably
affect the density of manatees less in summer than in
winter. Therefore, counts during warm seasons with a
standardized procedure and during a short time period
should provide more consistent data for determining an-
nual indices and population-size trends. Strip transects
have been used for population-size estimates of dugongs
(Dugong dugong; Marsh and Sinclair 1989a, 1989b;
Marsh 1995) and other species (Eberhardt et al. 1979;
Pollock and Kendall 1987; Barlow et al. 1988; Graham
and Bell 1989). Strip transects were used for counting
bottlenosed dolphins (Tursiops truncatus) and manatees in
the Indian and Banana rivers, (Leatherwood 1979), but
manatee counts were incidental and were not used for
calculating estimates. Transects are most suitable in large,
homogenous areas. The Florida Department of Environ-
mental Protection and the National Biological Service
recently selected three areas-Charlotte Harbor, Ten
Thousand Islands, and the Banana River-to test a strip
transect methodology for counting manatees in warm sea-
sons. This work was initiated in 1992 and is in progress.
Ongoing research also includes the use of other novel
approaches. Recent test flights were conducted to com-
pare counts from an airship and from an airplane, namely
a Cessna 172 and the Airship Shamu. The airship is
provided by Sea World of Florida and has a length of
58 m and a capacity for two pilots and five passengers.
The flights were made over three Tampa Bay power
plants and adjacent areas on 12 December 1990 and on
22 and 25 January 1992. Additional flights were made in
1993. The airship was flown along discharge canals and
transects across adjacent water bodies. Observers looked
out of large, open windows on either side of the airship.
Altitudes were usually 150 m but also 20-300 m.
Ground speed was 0-15 km/h. Because of the stability
and slow speed of the airship, observers were able to use
binoculars, long telephoto lenses on still cameras, and
video cameras. Preliminary flights demonstrated the use-
fulness of an airship for counting manatees, although
counts from the two aircraft were similar. Manatees had
less reaction to the airship than to an airplane. This
allowed investigators to observe manatees without re-
peatedly circling in a small plane or disturbing the ani-
mals with the noise and turbulence from a helicopter.






BRUCE B. ACKERMAN 29


Large airships may not be as effective for surveys of
manatee distributions because they are not as maneuver-
able as an airplane. However, comparisons of the two
aircraft must be made to determine the strengths and
weaknesses of each kind of aircraft for various tasks. I
recommend additional testing of airships. Smaller, more
maneuverable and more affordable airships may soon be
available for extended observation.
Other technologies are probably useful for aerial sur-
veys of manatees. High-resolution aerial video may be
used to record sightings on long transects for later view-
ing and counting and simultaneous documentation of
water conditions (Sidle and Ziewitz 1990). Computer-
image analysis may be available for detecting and count-
ing manatees on videotapes and for measuring body
lengths (Ratnaswamy and Winn 1993) and quantifying
visibility conditions. Use of global positioning systems
will improve accuracy of sighting locations and will
accurately record the flight path to document the exact
areas covered during flights. Advances in sonar may
allow accurate detection of manatees in some small areas
where aerial counting of the manatees is difficult; counts
on the ground could provide verification of counts from
aircraft. Sensitive time-depth recorders (Goodyear 1993)
may be used to document when manatees are at the
surface and to improve correction factors. Previously
classified, military remote-sensing technology may
eventually be used to detect and count marine mammals
with various platforms (military satellites, high-altitude
reconnaissance planes, or unoccupied military drones
with video or real-time artificial intelligence algorithms)
in large areas.


Conclusions

Recommendations for improvements in aerial survey
methods periodically have been made (*Eberhardt 1982;
*Packard 1985; Packard et al. 1986; Lefebvre et al.
1995). Improvements were made, but progress has been
slow. Most survey methods are the same as 10-15 years
ago. Surveys of distribution have been made in almost
all areas of Florida that manatees use substantially.
These data are in demand for developing protection of
manatees, particularly for planning boat traffic regula-
tions and coastal development. Long-term, regular moni-
toring of manatees may be necessary to update distribu-
tion data and to reassess manatee protection needs.
The January 1992 synoptic survey revealed more
manatees than had ever been recorded. However, the
method is not adequate to track statewide trends, and a
more standardized method is needed. Counts from future
surveys will probably be as variable as those from the
three surveys in 1991 and 1992. Estimated population


sizes based on surveys at aggregation sites in winter
increased in several areas of the state where long-term
studies were made, but the trends in population sizes in
other areas are unclear. Proposed techniques for moni-
toring trends must include tests of their statistical power
to detect small changes (Gerrodette 1987; Taylor and
Gerrodette 1993).
Improvement of aerial-survey techniques is a high
priority in manatee-population research. Statistically
valid techniques are needed. Development of correction
procedures for visibility bias is continuing. These new
techniques probably require advanced statistical analy-
ses, more observers in the aircraft, more intensive survey
effort, and more funds. Assessment of trends may require
more replication of surveys within years.
Recent advances in electronic equipment and comput-
ers will make the data more usable. Examples include use
of a global positioning system to more accurately record
the locations and geographic information system mapping
techniques and sighting density maps to display the results.
Small airships may improve surveys.


Acknowledgments

Most aerial surveys by the Florida Department of
Environmental Protection were funded by the Florida
Department of Environmental Protection Save the
Manatee Trust Fund with sales of Save the Manatee auto
license tags. Additional funding was provided by the
Save the Manatee Club. Many biologists from the Flor-
ida Department of Environmental Protection and other
agencies, universities, and private research laboratories
participated in these surveys during many years. Various
agencies generously provided data from aerial surveys
for the Florida Department of Environmental Protection
Marine Resources Geographic Information System.
The synoptic surveys were conducted with the coop-
eration of biologists from many agencies and groups.
Synoptic-survey flight expenses were paid by participat-
ing agencies or by the Florida Department of Environ-
mental Protection. Additional funding was provided by
R. Boone of Chevron, Inc., the Save the Manatee Club,
and the Florida Power and Light Company. The follow-
ing agencies and individuals participated during 1991 or
1992: the Florida Department of Environmental Protec-
tion (K. Abbott, B. Ackerman, B. Beeler, M. Clemons,
K. Curtin, K. Frohlich, W. Hartley, M. Justice, J. Ken-
ner, C. Knox, D. LaGrua, R. Mezich, M. Morris,
P. Nabor, J. Serino, S. Tyson, J. Valade, B. Weigle,
P. Wells, A. Welty, and E. Westerman), the St. Johns
River Water Management District (J. Higman), the
Georgia Department of Natural Resources (M. Harris),
Brevard County (R. Day and J. Snodgrass), Broward






30 INFORMATION AND TECHNOLOGY REPORT 1


County (D. Burgess and D. Stone), Dade County
(S. Markley and K. Mayo), Palm Beach County (D. Car-
son), the U.S. Fish and Wildlife Service
(F. Drauszewski, J. Kleen, S. Marcus, and R. Quarles),
the National Biological Service (R. Bonde, L. Lefebvre,
and J. Reid), Everglades National Park (S. Snow), the
Kennedy Space Center (J. and M. Provancha), Eckerd
College (P. Boland and J. Reynolds), the Mote Marine
Laboratory (J. Gorzelany, J. Kadel, and G. Patton), and
the Save the Manatee Club (P. Thompson). Additional
Florida Department of Environmental Protection and
National Biological Service staff located radio-tagged
manatees on the days of the surveys, entered data into
computer databases, and produced geographic informa-
tion system maps. I thank the pilots and crew of Sea
World's Airship Shamu for making those flights possi-
ble, especially pilot A. Judd and coordinators F. Ken-
nelly and W. Waier. N. Berill, K. Clifton, and L. Ward-
Geiger of the department entered, digitized, and proofed
the aerial survey sightings and produced the maps.
K. Clifton, A. Huff, T. O'Shea, J. Provancha,
J. Reynolds, and the Florida Marine Research Institute
editorial staff provided helpful comments on the manu-
script.



Cited References2

Ackerman, B. B. 1988. Visibility bias of mule deer aerial census
procedures in southeast Idaho. Ph.D. dissertation, University
of Idaho, Moscow. 106 pp.
Ackerman, B. B., S. D. Wright, R. K. Bonde, D. K. Odell, and
D. J. Banowetz. 1995. Trends and patterns in mortality of
manatees in Florida, 1974-1992. Pages 223-258 in T. J.
O'Shea, B. B. Ackerman, and H. F. Percival, editors. Popu-
lation biology of the Florida manatee. National Biological
Service Information and Technology Report 1.
Barlow, J., C. W. Oliver, T. D. Jackson, and B. L. Taylor. 1988.
Harbor porpoise, Phocoena phocoena, abundance estimation
for California, Oregon, and Washington: II. Aerial surveys.
Fishery Bulletin 86:433-444.
Beck, C. A., and J. P. Reid. 1995. An automated photo-identifi-
cation catalog for studies of the life history of the Florida
manatee. Pages 120-134 in T. J. O'Shea, B. B. Ackerman,
and H. F. Percival, editors. Population biology of the Florida
manatee. National Biological Service Information and Tech-
nology Report 1.
*Beeler, 1. E., andT. J. O'Shea. 1988. Distribution and mortality
of the West Indian manatee (Trichechus manatus) in the
southeastern United States: a compilation and review of re-
cent information. Document Number PB 88-207980/AS, Na-
tional Technical Information Service, Springfield, Va.
613 pp.


2 An asterisk denotes unpublished material.


*Brownell, R. L., Jr., K. Rails, and R. R. Reeves. 1981. Report
of the West Indian manatee workshop. Pages 3-16 in *R. L.
Brownell, Jr., and K. Rails, editors. 1981. The West Indian
manatee in Florida. Proceedings of a workshop held in Or-
lando, Florida, 27-29 March 1978. Florida Department of
Natural Resources, Tallahassee. 154 pp.
CH2M Hill. 1986. Aerial census of manatees in the vicinity of
Six Mile Creek, in the lower middle St. Johns River, Florida.
CH2M Hill, Gainesville, Fla. Report for St. Johns Harbour,
Inc. Unnumbered.
*Eberhardt, L. L. 1982. Censusing manatees. Manatee Popula-
tion Research Report I, Florida Cooperative Fish and Wild-
life Research Unit, University of Florida, Gainesville. 18 pp.
Eberhardt, L. L., D. G. Chapman, and J. R. Gilbert. 1979. A
review of marine mammal census methods. Wildlife Mono-
graphs 63. 46 pp.
Eberhardt, L. L., and T. J. O'Shea. 1995. Integration of manatee
life-history data and population modeling. Pages 269-279 in
T. J. O'Shea, B. B. Ackerman, and H. F. Percival, editors.
Population biology of the Florida manatee. National Biologi-
cal Service Information and Technology Report 1.
Eberhardt, L. L., and M. A. Simmons. 1992. Assessing rates of
increase from trend data. Journal of Wildlife Management
56:603-610.
*Florida Department of Natural Resources. 1989. Recommenda-
tions to improve boating safety and manatee protection for
Florida waterways. Report to the Governor and Cabinet. Final
report. 20 pp.
Floyd, T. J., L. D. Mech, and M. E. Nelson. 1979. An improved
method of censusing deer in deciduous-coniferous forests.
Journal of Wildlife Management 43:258-261.
Garrott, R. A., B. B. Ackerman, J. R. Cary, D. M. Heisey, J. E.
Reynolds, III, and J. R. Wilcox. 1995. Assessment of trends
in sizes of manatee populations at several Florida aggregation
sites. Pages 34-55 in T. J. O'Shea, B. B. Ackerman, and H.
F. Percival, editors. Population biology of the Florida mana-
tee. National Biological Service Information and Technology
Report 1.
Gasaway, W. C., S. D. Dubois, and S. J. Harbo. 1985. Biases in
aerial transect surveys for moose during May and June.
Journal of Wildlife Management 49:777-784.
Gerrodette, T. 1987. A power analysis for detecting trends.
Ecology 68:1364-1372.
Goodyear, J. D. 1993. A sonic/radio tag for monitoring dive
depths and underwater movements of whales. Journal of
Wildlife Management 57:502-513.
Graham, A., and R. Bell. 1989. Investigating observer bias in
aerial survey by simultaneous double-counts. Journal of
Wildlife Management 53:1009-1016.
*Hartman, D. S. 1974. Distribution, status, and conservation of
the manatee in the United States. Document Number PB
81-140725, National Technical Information Service, Spring-
field, Va. 246 pp.
Hartman, D. S. 1979. Ecology and behavior of the manatee
(Trichechus manatus) in Florida. American Society of Mam-
malogy Special Publication 5. 153 pp.






BRUCE B. ACKERMAN 31


Irvine, A. B. 1982. West Indian manatee. Pages 241-242 in D. E.
Davis, editor. Handbook of census methods for terrestrial
vertebrates. CRC Press, Boca Raton, Fla.
Irvine, A. B., J. E. Caffin, and H. I. Kochman. 1982. Aerial
surveys for manatees and dolphins in western peninsular
Florida. Fishery Bulletin 80:621-630.
Irvine, A. B., and H. W. Campbell. 1978. Aerial census of the
West Indian manatee, Trichechus manatus, in the southeast-
ern United States. Journal of Mammalogy 59:613-617.
*Kadel, J. J., and G. W. Patton. 1992. Aerial studies of the West
Indian manatee (Trichechus manatus) on the west coast of
Florida from 1985-90: a comprehensive six year study.
Mote Marine Laboratory Technical Report 246. Report to
Sarasota County Natural Resources Department. 39 pp. +
appendixes.
Kinnaird, M. F. 1985. Aerial census of manatees in northeastern
Florida. Biological Conservation 32:59-79.
Kochman, H. I., G. B. Rathbun, and J. A. Powell. 1985. Tempo-
ral and spatial distribution of manatees in Kings Bay, Crystal
River, Florida. Journal of Wildlife Management 49:921-924.
Leatherwood, S. 1979. Aerial survey of the bottlenosed dolphin,
Tursiops truncatus, and the West Indian manatee, Trichechus
manatus, in the Indian and Banana rivers, Florida. Fishery
Bulletin 77:47-59.
Lefebvre, L. W., B. B. Ackerman, K. M. Portier, and K. H. Pol-
lock. 1995. Aerial survey as a technique for estimating trends
in manatee population size-problems and prospects. Pages
63-74 in T. J. O'Shea, B. B. Ackerman, and H. F. Percival,
editors. Population biology of the Florida manatee. National
Biological Service Information and Technology Report 1.
*Lefebvre, L. W., and R. K. Frohlich. 1986. Movements of ra-
dio-tagged manatees in southwest Florida, January 1985-
March 1986. U.S. Fish and Wildlife Service, unpublished
report. 87 pp.
Lefebvre, L. W., and H. I. Kochman. 1991. An evaluation of aerial
survey replicate count methodology to determine trends in
manatee abundance. Wildlife Society Bulletin 19:298-309.
Marmontel, M. 1993. Age determination and population biology
of the Florida manatee, Trichechus manatus latirostris. Ph.D.
dissertation, University of Florida, Gainesville. 408 pp.
Marsh, H. 1995. Fixed-width aerial transects for determining
dugong population sizes and distribution patterns. Pages
56-62 in T. J. O'Shea, B. B. Ackerman, and H. F. Percival,
editors. Population biology of the Florida manatee. National
Biological Service Information and Technology Report 1.
Marsh, H., and D. F. Sinclair. 1989a. An experimental evaluation
of dugong and sea turtle aerial survey techniques. Australian
Wildlife Research 16:639-650.
Marsh, H., and D. F. Sinclair. 1989b. Correcting for visibility
bias in strip transect aerial surveys of aquatic fauna. Journal
of Wildlife Management 53:1017-1024.
*McGehee, M. A. 1982. Manatees (Trichechus manatus): abun-
dance and distribution in and around several Florida power
plant effluents during the winter of 1981-1982. Final report
prepared for the Florida Power and Light Company, Contract
31534-86419. 67 pp.
Moore, J. C. 1951. The range of the Florida manatee. Quarterly
Journal of the Florida Academy of Science 14:1-19.


*Odell, D. K. 1976. Distribution and abundance of marine
mammals in South Florida: Preliminary results. Pages 203-
212 in A. Thorhaug and A. Volker, editors. Biscayne Bay:
past/present/future. Proceedings of a symposium, 2-3 April
1976. Special Report 5. University of Miami, Sea Grant
Program.
Odell, D. K. 1979. Distribution and abundance of marine mam-
mals in the waters of the Everglades National Park.
Pages 673-678 in R. M. Linn, editor. Proceedings of the First
Conference on Scientific Research in the National Parks, New
Orleans, Louisiana, 9-12 November 1976. U.S. Department
of the Interior, National Park Service, Transactions and Pro-
ceedings Series 5, Volume 1.
O'Shea, T. J. 1988. The past, present, and future of manatees in
the southeastern United States: realities, misunderstandings,
and enigmas. Pages 184-204 in R. R. Odom, K. A. Riddle-
berger, and J. C. Ozier, editors. Proceedings of the Third
Southeastern Nongame and Endangered Wildlife Sympo-
sium, Georgia Department of Natural Resources, Social
Circle.
O' Shea, T. J., C. A. Beck, R. K. Bonde, H. I. Kochman, and D. K.
Odell. 1985. An analysis of manatee mortality patterns in
Florida, 1976-81. Journal of Wildlife Management 49:1-11.
O'Shea, T. J., and W. C. Hartley. 1995. Reproduction and early-
age survival of manatees at Blue Spring, St. Johns River,
Florida. Pages 157-170 in T. J. O'Shea, B. B. Ackerman, and
H. F. Percival, editors. Population biology of the Florida mana-
tee. National Biological Service Information and Technology
Report 1.
O'Shea, T. J., and H. I. Kochman. 1990. Florida manatees:
Distribution, geographically referenced data sets, and eco-
logical and behavioral aspects of habitat use. Pages 11-22
in J. E. Reynolds, III and K. Haddad, editors. Report of the
workshop on geographic information systems as an aid to
managing habitat for West Indian manatees in Florida and
Georgia. Florida Marine Research Publications 49.
O'Shea, T. J., and C. A. Langtimm. 1995. Estimation of survival
of adult Florida manatees in the Crystal River, at Blue Spring,
and on the Atlantic Coast. Pages 194-222 in T. J. O'Shea,
B. B. Ackerman, and H. F. Percival, editors. Population biol-
ogy of the Florida manatee. National Biological Service In-
formation and Technology Report 1.
*Packard, J. M. 1985. Development of manatee aerial survey
techniques. Manatee Population Research Report 7, Techni-
cal Report 8-7. Florida Cooperative Fish and Wildlife Re-
search Unit, University of Florida, Gainesville. 68 pp.
*Packard, J. M., R. K. Frohlich, J. E. Reynolds, II1, and J. R.
Wilcox. 1984. Factors influencing indices of manatee abun-
dance of manatees in the Fort Myers region, winter 1983/84.
Manatee Population Research Report 5, Technical Report
8-5. Florida Cooperative Fish and Wildlife Research Unit,
University of Florida, Gainesville. 63 pp.
Packard, J. M., R. K. Frohlich, J. E. Reynolds, III, and J. R.
Wilcox. 1989. Manatee response to interruption of a thermal
effluent. Journal of Wildlife Management 53:692-700.
*Packard, J. M., and R. Mulholland. 1983. Analysis of manatee
aerial surveys: a compilation and preliminary analysis of
winter aerial surveys conducted in Florida between 1977 and
1982. Manatee Population Research Report 2, Technical Re-






32 INFORMATION AND TECHNOLOGY REPORT 1


port 8-2. Florida Cooperative Fish and Wildlife Research
Unit, University of Florida, Gainesville. 119 pp.
Packard, J. M., D. B. Siniff, and J. A. Cornell. 1986. Use of
replicate counts to improve indices of trends in manatee
abundance. Wildlife Society Bulletin 14:265-275.
Packard, J. M., R. C. Summers, and L. B. Barnes. 1985. Vari-
ation in visibility bias during aerial surveys of manatees.
Journal of Wildlife Management 49:347-351.
Pollock, K. H., and W. L. Kendall. 1987. Visibility bias in aerial
surveys: a review of estimation procedures. Journal of Wild-
life Management 51:502-510.
*Powell, J. A. 1981. The manatee population in Crystal River,
Citrus County, Florida. Pages 33-40 in *R. L. Brownell, Jr.
and K. Rails, editors. The West Indian manatee in Florida.
Proceedings of a workshop held in Orlando, Fla., 27-29
March 1978. Florida Department of Natural Resources, Tal-
lahassee. 154 pp.
Powell, J. A., and G. B. Rathbun. 1984. Distribution and abun-
dance of manatees along the northern coast of the Gulf of
Mexico. Northeast Gulf Science 7:1-28.
*Powell, J. A., and J. C. Waldron. 1981. The manatee population
in Blue Spring, Volusia County, Florida. Pages 41-51 in
*R. L. Brownell, Jr. and K. Rails, editors. The West Indian
manatee in Florida. Proceedings of a workshop held in Or-
lando, Fla., 27-29 March 1978. Florida Department of Natu-
ral Resources, Tallahassee. 154 pp.
Provancha, J. A., and M. J. Provancha. 1988. Long-term trends
in abundance and distribution of manatees (Trichechus ma-
natus) in the northern Banana River, Brevard County, Flor-
ida. Marine Mammal Science 4:323-338.
*Provancha, J. A., and M. J. Provancha. 1989. Summary of 1987
and 1988 manatee aerial surveys at Kennedy Space Center.
National Aeronautical and Space Administration Technical
Memorandum 102783. 17 pp.
Rathbun, G. B. 1988. Fixed-wing airplane versus helicopter sur-
veys of manatees Trichechus manatus. Marine Mammal Sci-
ence 4:71-75.
Rathbun, G. B., J. P. Reid, R. K. Bonde, and J. A. Powell. 1995.
Reproduction in free-ranging Florida manatees. Pages 135-
156 in T. J. O'Shea, B. B. Ackerman, and H. F. Percival,
editors. Population biology of the Florida manatee. National
Biological Service Information and Technology Report 1.
Rathbun, G. B., J. P. Reid, and G. Carowan. 1990. Distribution
and movement patterns of manatees (Trichechus manatus) in
northwestern peninsular Florida. Florida Marine Research
Publications 48:1-33.
Ratnaswamy, M. J., and H. E. Winn. 1993. Photogrammetric
estimates of allometry and calf production in fin whales,
Balaenoptera physalus. Journal of Mammalogy 74:323-330.
*Raymond, P. W. 1981. Manatees (Trichechus manatus): abun-
dance and distribution in and around several Florida power
plant effluents during the winter of 1981-82. Final report
prepared for the Florida Power and Light Company, Contract
31534-81511. 62 pp.
Reid, J. P., R. K. Bonde, and T. J. O'Shea. 1995. Reproduction
and mortality of radio-tagged and recognizable manatees on
the Atlantic coast of Florida. Pages 171-191 in T. J. O'Shea,
B. B. Ackerman, and H. F. Percival, editors. Population biol-


ogy of the Florida manatee. National Biological Service
Information and Technology Report 1.
Reid, J. P., G. B. Rathbun, and J. R. Wilcox. 1991. Distribution
patterns of individually identifiable West Indian manatees
(Trichechus manatus) in Florida. Marine Mammal Science
7:180-190.
*Reynolds, J. E., III. 1993. Distribution and abundance of Florida
manatees (Trichechus manatus latirostris) around selected
power plants following winter cold fronts: 1992-1993. Re-
port for Florida Power and Light Company, Order B91135-
00073. 36 pp.
*Reynolds, J. E., III. 1994. Distribution and abundance of Florida
manatees (Trichechus manatus latirostris) around selected
power plants following winter cold fronts: 1993-1994. Re-
port for Florida Power and Light Company, Order B91135-
00139. 37 pp.
*Reynolds, J. E., III, B. B. Ackerman, I. E. Beeler, B. L. Weigle,
and P. F. Houhoulis. 1991. Assessment and management of
manatees (Trichechus manatus) in Tampa Bay. Pages 289-
306 in S. F. Treat and P. A. Clark, editors. Proceedings of the
Second Tampa Bay Area Scientific Information Symposium,
27 February-I March 1991, Tampa, Fla.
*Reynolds, J. E., III, and C. J. Gluckman. 1988. Protection of the
West Indian manatee in Florida. Document Number PB 88-
222922, National Technical Information Service, Spring-
field, Va. 85 pp.
Reynolds, J. E., III, and J. R. Wilcox. 1985. Abundance of West
Indian manatees (Trichechus manatus) around selected Flor-
ida power plants following winter cold fronts: 1982-1983.
Bulletin of Marine Science 36:413-422.
Reynolds, J. E., III, and J. R. Wilcox. 1986. Distribution and
abundance of the West Indian manatees (Trichechus manatus)
around selected Florida power plants following winter cold
fronts: 1984-1985. Biological Conservation 38:103-113.
Reynolds, J. E., III, and J. R. Wilcox. 1994. Observations of
Florida manatees (Trichechus manatus latirostris) around
selected power plants in winter. Marine Mammal Science
10:163-177.
*Rose, P. M., and S. P. McCutcheon. 1980. Manatees
(Trichechus manatus) abundance and distribution in and
around several Florida power plant effluents. Final report
prepared for the Florida Power and Light Company, Contract
31534-86626 128 pp.
Samuel, M. D., E. O. Garton, M. W. Schlegel, and R. G. Carson.
1987. Visibility bias during aerial surveys of elk in northcen-
tral Idaho. Journal of Wildlife Management 51:622-630.
Samuel, M. D., and K. H. Pollock. 1981. Correction of visibility
bias in aerial surveys where animals occur in groups. Journal
of Wildlife Management 45:993-997.
Shane, S. H. 1983. Abundance, distribution, and movements of
manatees in Brevard County, Florida. Bulletin of Marine
Science 33:1-9.
Sidle, J. G., andJ. W. Ziewitz. 1990. Use of aerial videography in
wildlife habitat studies. Wildlife Society Bulletin 18:56-62.
*Snow, R. W. 1992. The distribution and relative abundance of
the Florida manatee in Everglades National Park. Interim
report, March 1991 to February 1992. South Florida Research
Center, Everglades National Park. Unnumbered.






BRUCE B. ACKERMAN 33


Steinhorst, R. K., and M. D. Samuel. 1989. Sightability adjust-
ment methods for aerial surveys of wildlife populations.
Biometrics 45:415-425.
Taylor, B. L., and T. Gerrodette. 1993. The uses of statistical
power in conservation biology: the vaquita and northern
spotted owl. Conservation Biology 7:489-500.
Unsworth, J. W., L. Kuck, and E. O. Garton. 1990. Elk sightabil-
ity model validation at the National Bison Range, Montana.
Wildlife Society Bulletin 18:113-115.
*U.S. Fish and Wildlife Service. 1989. Florida manatee
(Trichechus manatus latirostris) recovery plan. U.S. Fish and
Wildlife Service, Atlanta, Ga. 98 pp.
*Valade, J. A. 1990. Coastal Georgia manatee assessment. Final
report for Georgia Department of Natural Resources, Con-
tract 462-511-960966. Unnumbered.
Weigle, B. L., and K. D. Haddad. 1990. Applications of the
Florida Department of Natural Resources' Marine Re-
sources Geographic Information System to manatee biology


and management. Pages 23-27 in J. E. Reynolds, III and
K. D. Haddad, editors. Report of the workshop on geographic
information systems as an aid to managing habitat for West
Indian manatees in Florida and Georgia. Florida Marine Re-
search Publications 49:23-27.
*Weigle, B. L., J. E. Reynolds, III, G. W. Patton, and J. R.
Wilcox. 1988. Manatee (Trichechus manatus) winter use of
warm water discharges in Tampa Bay. Pages 153-164 in
K. Mahadevan, R. K. Evans, P. Behrens, T. Biffar, and
L. Olsen, editors. Proceedings Southeastern Workshop
on Aquatic Ecological Effects of Power Generation, 3-5
December 1986, Sarasota, Fla. Mote Marine Lab. Report
124.
Woolf, H. B., editor. 1977. Webster's New Collegiate Dictionary.
G. & C. Merriam Company, Springfield, Mass. 1,536 pp.
Zoodsma, B. J. 1991. Distribution and behavioral ecology of
manatees in southern Georgia. M.S. thesis, University of
Florida, Gainesville. 202 pp.






34 INFORMATION AND TECHNOLOGY REPORT 1


Assessment of Trends in Sizes of Manatee Populations at Several

Florida Aggregation Sites



by


Robert A. Garrott

University of Wisconsin
Department of Wildlife Ecology
Madison, Wisconsin 53706


Bruce B. Ackerman

Florida Department of Environmental Protection
Florida Marine Research Institute
100 Eighth Avenue S.E.
St. Petersburg, Florida 33701


John R. Cary

University of Wisconsin
Department of Wildlife Ecology
Madison, Wisconsin 53706


Dennis M. Heisey

University of Wisconsin
Madison Academic Computing Center
Madison, Wisconsin 53706


John E. Reynolds, HI

Eckerd College
Department of Biology
St. Petersburg, Florida 33701


and


J. Ross Wilcox

Florida Power and Light Company
Environmental Affairs
P.O. Box 088801
North Palm Beach, Florida 33408






ROBERT A. GARROTT ET AL 35


Abstract. Temporal trends in sizes of Florida manatee (Trichechus manatus latirostris) populations
were assessed with counts from aerial surveys at nine aggregation sites in winter. Manatees that winter
along the eastern coast of Florida are considered one population; therefore, counts from all survey sites
in eastern Florida were pooled. Results from counts at Fort Myers in southwestern Florida were treated
separately. Simple log-linear models were used to identify physical covariates that explained a
significant amount of variability in the counts adjusted to yearly means. Then the counts were adjusted
for these covariates and, if present, a test for temporal pattern in the year coefficients was made.
Covariates included survey conditions, short-term (1- to 3-day) air and water temperatures at each
survey site, and a series of time-lagged air and water temperature variables. National Oceanic and
Atmospheric Administration degree-day summations for periods of 5-40 days prior to the survey were
used. Most temperature variables contributed significantly to the model. Based on the correlation
coefficient, however, the best general models for counts of manatees in eastern Florida and at Fort
Myers incorporated a short- and long-term summation of degree-days. Approximately 50% of the
variation in counts could be explained with this simple model. An analysis of temporal trends in the
temperature-adjusted counts suggested that population sizes in eastern Florida increased from 1982-83
to 1990-91, whereas population sizes at Fort Myers remained stable. Interpretation of these results is
problematic because corroborating evidence that counts at winter aggregation sites are reliable
indicators of manatee population size does not yet exist.
Key words: Aerial survey, manatee, power plants, trends, Trichechus manatus latirostris.


Increases in the number of recovered dead Florida
manatees (Trichechus manatus latirostris) throughout
Florida in recent years have been interpreted as evidence
of increasing mortality rates (Ackerman et al. 1995). Be-
cause manatees are long-lived mammals with low rates of
reproduction, any significant increase in mortality may
lead to a decline in the population (O'Shea 1988). No
available indices of population-size trends have been
proven reliable. However, an extensive database has been
collected by aerial surveys during the past 14 years. Re-
peated aerial surveys in winter to count aggregations of
manatees in the warm-water effluents at power plants have
been conducted since 1977 (* Rose and McCutcheon
1980, *Raymond 1981; *McGehee 1982; *Reynolds
1983, 1985, 1991; Reynolds and Wilcox 1985, 1986).
These data may provide insights into trends in manatee
numbers in several regions of Florida (Ackerman 1995;
Lefebvre et al. 1995). However, the variability of the
counts has been problematic.
Previous researchers (*Rose and McCutcheon 1980,
*Raymond 1981; *McGehee 1982; *Reynolds 1983) re-
gressed manatee counts at each major site within each year
against daily air and water temperatures. They clearly
demonstrated that the highest counts occurred at lowest
temperatures, but sample sizes in any one year were not
sufficiently large to indicate more specific patterns.
Counts did not vary solely in response to single-day tem-
peratures. Packard and Mulholland (*1983) compiled the
first 5 years of data (1977-82) and performed extensive
statistical analyses. Multiple regression was used to sepa-
rately regress counts from each site against air and water


1 An asterisk denotes unpublished material.


temperatures on the survey date and each of the two
previous days, time of year, and other weather and visibil-
ity factors. However, regressions on water temperatures
seemed to differ among years at some sites, and trends in
maximum counts and mean-adjusted counts seemed to
substantially differ among sites. These complications were
thought to be the result of large-scale movements among
sites within and between years or responses to longer-term
weather patterns. Subsequently, Reynolds and colleagues
(*Reynolds 1983, 1985, 1991; Reynolds and Wilcox 1985,
1986) regressed manatee counts at each major site each
year against 1- and 3-day mean air and water temperatures.
Although these studies provided important insights into
the factors of the variability of the counts, none of the
studies provided a means of integrating data to develop a
population index. Our objectives were the identification of
environmental variables that influenced counts, develop-
ment of a statistical model to adjust counts for these
effects, and testing of the adjusted counts for a significant
temporal trend in the index. These studies were recom-
mended in the Florida Manatee Recovery Plan (*U.S. Fish
and Wildlife Service 1989, Tasks 42 and 421) and by
Eberhardt (*1982).


Methods

Aerial Survey Methodology

Since 1977, 1- to 2-day aerial surveys have been con-
ducted to count manatees associated with warm-water
sources during cold weather. These surveys can be segre-
gated into two fundamentally different periods. During the
5-year interval from 1977 to 1982, counts were conducted






36 INFORMATION AND TECHNOLOGY REPORT 1


first year-round at regularly scheduled weekly intervals and
later weekly only in winter. Survey methodology also
evolved, and several different personnel performed the
counts (*Rose and McCutcheon 1980; *Raymond 1981;
*McGehee 1982). Beginning in winter 1982-83, the sched-
uling of surveys was changed so that flights were conducted
only immediately after cold fronts when the highest counts
were obtained. We began conducting surveys at that time
and standardized the methodology. The difference in survey
scheduling and the changes that occurred in personnel and
methodology during 1977-82 suggested that pooling data
from all years in a single analysis would be problematic.
Hence, only survey data from winter 1982-83 through
winter 1990-91 were used in this analysis.
Surveys were conducted at four major manatee aggrega-
tion sites at power plants along the eastern coast of Florida:
the Orlando Utility Commission's Indian River plant (IR)
and the Florida Power and Light Company's Cape Ca-
naveral (CC), Riviera Beach (RV), and Port Everglades (PE)
plants (Table 1; Fig. 1). Minor aggregations at three eastern-
coast plants were also surveyed: Vero Beach Municipal
Plant (VB), H. D. King Municipal Plant at Fort Pierce (FP),
and the Florida Power and Light Company's Fort Lauder-
dale plant (FL). Manatees were also counted at the Harbor
Branch Oceanographic Institute (HB), at the Hobe Sound
National Wildlife Refuge (HS), and along the Intracoastal
Waterway (IW) between the plants, although none of these
had a source of warm water (Fig. 1). These sites are believed
to include most areas along the eastern coast where manatees
are known to aggregate during winter. The Indian River and
Cape Canaveral power plants are only 3 km apart, and
counts at these two plants were pooled. Surveys were also


conducted at a single power plant on the southwestern coast
of Florida, the Florida Power and Light Company's Fort
Myers (FM) plant, believed to be the only warm-water site
in the area where manatees aggregate. In winter, manatees
are more isolated there than at eastern-coast sites.
Aerial counts were conducted by continuously circling
manatee groups detected in each area until a maximum count
was obtained. The primary observer was the same individual
during all surveys. Each survey area extended about 8 km
from one of the major plants. Recorded data included the
total number of observed manatees, the number of calves,
and a subjective ranking of survey conditions
(SURVCOND, combining wind conditions, water turbidity,
and surface conditions) ranging from 1 (excellent) to 5 (very
poor; Table 2; *Reynolds 1983). A more detailed descrip-
tion of the survey procedures appears in Reynolds and
Wilcox (1986), and discussions of survey limitations are
presented elsewhere (Ackerman 1995; Lefebvre et al.
1995).
Because temperature may strongly influence the number
of manatees counted on surveys, several temperature vari-
ables were recorded during surveys (Table 2; Hartman
1979; *Rose and McCutcheon 1980; *Packard and Mulhol-
land 1983). National Oceanic and Atmospheric Administra-
tion air temperatures were obtained from sites nearest the
four major power plants (National Oceanic and Atmos-
pheric Administration 1982-91). Air temperatures were the
average of the maximum and minimum temperatures on the
survey day and average temperature on the survey day and
on the two previous days. Water temperatures were meas-
ured at the power-plant intakes and were representative of
ambient temperatures in the area. The water temperature


Table 1. Florida manatee (Trichechus manatus latirostris) aggregation sites in winter.
Site Site
number identifier Site name City County
Eastern coast
1 IR OUCa Indian River plant Titusville Brevard
2 CC FPLb Cape Canaveral plant Titusville Brevard
3 VB Vero Beach Municipal plant Vero Beach Indian River
4 FP H. D. King Municipal plant Ft. Pierce St. Lucie
5 HB Harbor Branch Oceanographic Institute Ft. Pierce St. Lucie
6 HS Hobe Sound National Wildlife Refuge Jupiter Martin
(feeding area)
7 RV FPL Riviera Beach plant West Palm Beach Palm Beach
8 PE FPL Port Everglades plant Ft. Lauderdale Broward
9 FL FPL Ft. Lauderdale plant Ft. Lauderdale Broward
10 IW Intracoastal Waterway (Titusville to Ft. Lauderdale)
Western coast
11 FM FPL Ft. Myers plant Ft. Myers Lee
a OUC = Orlando Utility Commission.
b FPL = Florida Power and Light Company.






ROBERT A. GARROTT ET AL. 37


Indian River
Cape Canaveral Fig. 1. Aggregation sites of Florida mana-
tees (Trichechus manatus latirostris)
Vero Beach where counts were made by aerial sur-
H arbor Branch veys in winters, 1982-1983 to 1990-
Fort Pierce 1991.
Hobe Sound
Riviera Beach
Fort Myers Fort Lauderdale
Port Everglades













Table 2. Definitions of variables.
Variable name Definition
ATEMP Mean of high and low air temperature oC at a given site on survey date
ATEMP3 Mean air temperature o C at a given site, averaged over survey date and 2 previous days
NATEMP Same as ATEMP, at northern site (CC)
NATEMP3 Same as ATEMP3, at northern site (CC)
SATEMP Same as ATEMP, averaged over two southern sites (RV, PE)
SATEMP3 Same as ATEMP3, averaged over two southern sites (RV, PE)
COUNT Total number of manatees counted at a site on a given date, including adults and calves
DD5, ..., DD40 Cumulative heating degree-days F. Cumulative degrees that daily mean air
temperature is below 65* F (18.30 C), summed for 5,..., 40 days previous
to survey date at each site. Calculated for Titusville and Fort Myers
TOTDD Cumulative heating degree-days o F, summed over entire year
DD10 DD 10 (1-10 previous days)
DD10S DD 10 (1-10 previous days) squared
DX30 DD 11-30 (11-30 previous days; DD30 with DD10 removed)
DX30S DD 11-30 squared (11-30 previous days)
SURVCOND Subjective evaluation of survey conditions for each major site on day of survey, based on water turbidity,
surface glare and chop, and wind conditions (1 = excellent or very good, 2 = good, 3 = fair, 4 = poor,
5 = very poor)
NSUVCOND Same as SURVCOND, at northern site (CC)
SSUVCOND Same as SURVCOND, averaged over two southern sites (RV, PE)
WTEMP Intake water temperature C, at a given site, on survey date
WTEMP3 Intake water temperature C, at a given site, averaged over survey date and 2 previous days
NWTEMP Same as WTEMP, at northern site (CC)
NWTEMP3 Same as WTEMP3, at northern site (CC)
SWTEMP Same as WTEMP, averaged over two southern sites (RV, PE)
SWTEMP3 Same as WTEMP3, averaged over two southern sites (RV, PE)
YEAR Year of survey, 83 indicates November 1982-March 1983






38 INFORMATION AND TECHNOLOGY REPORT 1


indices were the average temperature on the survey day and
the average temperature on the survey day and on the two
previous days. Complete air and water temperature variables
were available only at three of the eastern-coast power plants
(Cape Canaveral, Riviera Beach, and Port Everglades) and
the Fort Myers power plant in southwestern Florida. Each
of the eastern Florida sites is close to a second survey area
(Indian River plant, Hobe Sound, and Ft. Lauderdale plant).
Thus, the temperatures generally also reflected conditions at
nearby sites.
The air temperature variables provided a measure of the
severity of the cold fronts that instigated each survey,
whereas water temperature variables provided a more direct
measure of the immediate thermal conditions that manatees
were experiencing. Cumulative heating degree-days were
calculated with National Oceanic and Atmospheric Admini-
stration (1982-91) air temperatures (original measurements
in o F) at Titusville (near Cape Canaveral) and Fort Myers
(Table 2). Heating degree-days were used as another meas-
ure of the winter severity. Heating degree-days are the
cumulative total of degrees that the daily average air tem-
perature is below 18.30 C each day and summed over all
days of the year (primarily occurs October-March). The
National Oceanic and Atmospheric Administration (1982-
91) provides degree-days on a monthly and yearly basis as
an index to the severity of each winter. However, degree-day
data were recalculated here to create a series of time-lag
variables that represent temperature conditions on various
numbers of days prior to each aerial survey. Eight time-lag
temperature variables were calculated by summing the de-
gree-days (daily number of degrees below 18.30 C) over the
5-40 day intervals prior to each survey (5-day increments,
DD5, DD10, ..., and DD40). Occasionally missing National
Oceanic and Atmospheric Administration temperatures
were replaced with temperatures from the most appropriate
nearby weather station.

General Approach

A population index is a statistic that correlates with the
true population level (Lancia et al. 1994). The higher the
correlation is, the more reliable is the index. Extraneous
sources of variation may weaken the correlation to the point
that the index is no longer useful. However, if the causes of
the extraneous sources of variation are largely known, it may
be possible to adjust for these causes, and some of the
reliability of the index may be recovered. Thus, an index that
poorly correlates with the population level may still be
useful if a high partial correlation with the true population
level can be achieved after appropriate adjustment for co-
variates.
Adjusting for covariates involves risks. The primary goal
is to find a model of the index for which the covariates
explain a large portion of the variability of the index. After


adjustment for these covariates, the remaining or residual
variability in the index should be caused by variations in the
true population level. However, this may not be the case.
The greater the proportion of variability in the index that is
due to covariates, the greater are the consequences of model
misspecification. Undocumented lurking variables also re-
main a problem. Furthermore, if the covariates also correlate
with the true population level, adjustment may weaken the
reliability of the index. Only a comparison with accurate
population data allows a true test of the reliability of the
index (Lancia et al. 1994).
With these cautions in mind, we developed an adjusted
aerial count index for manatees. We initially pursued a
descriptive, statistical model of the count data rather than a
more biological, model-based approach because we wanted
as simple and assumption-free a model as possible. We used
a semi-parametric approach, imposing no model structure
by year; yearly levels were nonparametrically modeled. A
nonparametric approach ensures that an assumed model
does not force the appearance of a temporal pattern (e.g., a
linear trend) that does not exist. We worked primarily with
the logarithm of counts, which is a natural scale for many
rate processes such as population growth (Lancia et al.
1994).
The proportional-rates model we developed is similar to
the semi-parametric Cox proportional-hazards model (Cox
and Oakes 1984). We assumed that, if temperatures and
survey conditions were always average, the expected num-
ber of manatees that is counted in any given place (one site
or several pooled sites) at time t is M(t). We assumed that
M(t) remains constant during a year (at least during the
counting period), represented as Mi, where i = year. Mi
should correlate with the true manatee population level and
must be assumed. We let C(t, X(t)) be the expected number
of manatees counted at time t, given conditions X(t), where
X(t) is a row vector ofcovariate values at time t. This can be
modeled as follows:
C(t, X(t)) = MiR(t, X(t))
The function R(.) is a rate function (range 0 to o) that takes
the value 1 when the conditions are at their averages. When
R(.) is greater than 1, conditions favored high counts, and
vice versa. R(.) must be nonnegative. A convenient form for
R(t, X(t)) is R(t) = exp(x(t)'B), where x(t) is a row vector of
the covariates centered at their long-term means and B is a
row vector of regression parameters. Thus, the model be-
comes
C(t, X(t)) = Mi exp(x(t)'B) or
InC(t, X(t)) = InMi + x(t)'B,
a simple log-linear model. The Mi's are the baseline annual
population index levels.
Before attempting to fit a given model to the count data,
we conducted a variety of exploratory investigations of the






ROBERT A. GARROTT ET AL. 39


covariate data in an attempt to understand the variations in
these variables and their potential relations to each other and
to the counts. Because temperature data at the two southern
power plants (RV, PE) were similar, these data were aver-
aged into a single southern site temperature variable to
contrast with the pooled northern site (IR, CC) data.
The sampling design is such that it seems difficult to
accurately model the anticipated error structure. The
timing of counts is erratic and based on subjective factors
(*Reynolds 1983). Some surveys were conducted soon
after others, whereas others were more temporally iso-
lated. This no doubt results in varying degrees of tempo-
ral auto-correlation. The fact that the timing of the sur-
veys was based on the values of covariates that may enter
into the model is also problematic. Despite these prob-
lems, we believed that statistical modeling was useful for
exploring some features of this data set. We tried a wide
variety of models but report only on those we believe are
the most valid. We caution that the analysis is purely
exploratory and that the data are from observations; we
use hypothesis- testing statistics (P-values) only as rela-
tive measures of model merit.
Trends in the population size may be detected by
treating data from each site separately, adjusting counts
with site-specific covariates, and looking for trends at
each site (*Packard and Mulholland 1983). This rela-
tively rigorous strategy allows modeling different rela-
tions of populations at each site and provides the most
detailed use of the data. Data from individually identifi-
able animals, however, indicate that manatees move
among the sites in each winter (Reid et al. 1991, 1995).
Another confounding factor is the possibility that factors
far from each survey site may influence the number of
manatees.
Manatee aggregations on the eastern coast of Florida
in winter are considered a single population that is rela-
tively closed to emigration or immigration (Ackerman
1995). For our analysis, therefore, we had to determine
whether the count data provided an indication of trend in
the population size of manatees of the entire eastern coast
of Florida. During each survey, counts were conducted
at the major aggregation sites along the eastern coast.
Counts were conducted immediately after cold fronts
because it is believed that almost all manatees in the
population are concentrated at these warm-water refuges
because of thermal stress. The structure of the surveys
and the regional dynamics of the population suggested
modeling the counts with a generalized model that
summed counts from all sites in each survey.
The strategy of our analysis was the identification of the
physical covariates that significantly contributed to the vari-
ability of the natural logarithm of the total number of mana-
tees counted at a site on a given date (COUNT in Table 2)


after adjusting to the yearly means, the adjustment of counts
for these covariates, and the determination of any temporal
pattern in the year coefficients. The covariates included the
1- and 3-day air and water temperatures and survey condi-
tions from the northern power plants (sites IR and CC;
NATEMP, NATEMP3, NWTEMP, NWTEMP3, NSU-
VCOND). The means of the same variables from the two
major southern power plants (sites RV and PE) were also
considered (SATEMP, SATEMP3, SWTEMP,
SWTEMP3, SSUVCOND). In addition, nine time-lag vari-
ables with degree-days from the northern site summed over
5- to 40-day intervals (DD5, DDIO,..., DD40) and the total
number of degree-days (TOTDD) in winter were consid-
ered. Model selection was performed by manually conduct-
ing a series of forward stepwise regressions to identify the
best model. The first step was to include the categorical
variable YEAR for the baseline parameter Mi. In Step 2, we
constructed models for each of the 19 possible covariates
and their quadratic terms and performed an F test to deter-
mine whether a covariate significantly contributed to the
simple model. The covariate that contributed the most (i.e.,
had the highest correlation coefficient) was added to the
model. In further steps, we continued to search for the next
most important physical covariate, only stopping when no
further covariates could be identified at the P < 0.05 level.


Results


Eastern Florida Aggregation Sites

Count Data
Each winter, 4-10 surveys were conducted between
December and early March (Appendix A). Three surveys,
when counts were not made at all sites (19 January 1983,
23 January 1989, 11 February 1989), and the first survey
(14 December 1982), when only a single manatee was ob-
served at a warm-water aggregation site, were excluded
from the analysis (Appendix A). Principal aggregation areas
included the combined IR (Site 1) and CC (Site 2) power
plants in the northern part of the study area and the RV (Site
7) and PE (Site 8) plants in southern Florida (Table 1;
Fig. 1). Counts at these sites were highly variable within and
among years (Fig. 2). The high variability of the counts was
also present when counts from each survey at all survey sites
were pooled (Fig. 3). Seasonal patterns in winter counts
within years were not obvious. Survey conditions were
usually recorded only at sites CC, Hobe Sound (Site 6, HS),
RV, PE, and Fort Lauderdale (Site 9, FL). Survey conditions
were intermediate during most flights; only 12% of surveys
in all years were conducted under excellent (1) conditions
and only 6% under very poor conditions (5; Tables 3 and 4).
Survey conditions were similar among sites.







40 INFORMATION AND TECHNOLOGY REPORT 1


Cu, -
100 -
0



300
200

> 100
0


3U . .
:iuu -- -- -- -- -- -
200 -
100 -
0 L.
Q ti --i -- -- -i

"In .. .


200
S100-
0 -


Year


300
200
S100
0 41



300

200 I

100

0 0


300
200

a 100




300

200 -
-J
LL 100-
0 ..L -




Year


Fig. 2. Frequency distribution of number of manatees (Trichechus manatus latirostris) counted at survey sites along the eastern coast
of Florida during aerial surveys in winter 1982-1991. Data are presented in standard box plots. Individual survey counts are
shown as solid squares; extreme values are plotted as an asterisk or open circle. For each sample, the bottom and top edges of
the box are located at the sample 25th and 75th percentiles. The center horizontal line is drawn at the sample median. The central
vertical lines extend from the 5th to the 95th percentile.


, Ye

Year


Fig. 3. Summary statistics on number of manatees (Trichechus
manatus latirostris) counted at seven aggregation sites along
the eastern coast of Florida during each winter, 1982-1991.
Counts at all sites were made on the same day from a fixed-
wing aircraft; 4-10 surveys were conducted each winter. Data
are presented in standard box plots (see Fig. 2 caption).


SI I I I ) I


- -_ 0 0 a- - -






ROBERT A. GARROTT ET AL. 41


Table 3. Survey conditions during aerial counts at seven Florida manatee (Trichechus manatus latirostris) aggregation
sites along the Florida coast, by site, during winters 1982-1983 to 1990-1991. Survey-condition scores were a
subjective evaluation of visibility based on water turbidity, surface glare and chop, and wind conditions (*Reynolds
1983). A score of 1 indicated excellent conditions and a score of 5, very poor conditions. Counts are identified in
each category and are followed by the percentage of the row total in parentheses.
Survey conditions
1 2 3 4 5
Site Excellent Good Fair Poor Very poor Total
IR 1 2 (20.0) 4 (40.0) 4 (40.0) 10
CC 2 2 (3.4) 13 (22.4) 23 (39.7) 16(27.6) 4 (6.9) 58
HS 6 1 (1.8) 17 (29.8) 23 (40.4) 16(28.1) 57
RV 7 19 (32.8) 19 (32.8) 9(15.5) 3 (5.2) 8 (13.8) 58
PE 8 12(21.4) 27 (48.2) 7 (12.5) 5 (8.9) 5 (8.9) 56
FL 9 13(24.1) 27(50.0) 14(25.9) 54
FM 11 1(1.7) 6(10.2) 26(44.1) 24 (40.7) 2(3.4) 59
Total 35 (9.9) 97(27.6) 119(33.8) 82(23.3) 19(5.4) 352


Temperature
Although we examined 1-day and 3-day air and water
temperature variables, the patterns and relations of both
were similar, and we present in detail only the 3-day
variables (ATEMP3, WTEMP3). Median 3-day water
temperatures were approximately 5-6 C warmer than
median air temperatures in the southern and northern
areas (Fig. 4). Air and water temperatures were cooler at
the northern site than at the southern site (median air
temperature 4.40 C cooler, Fig. 5; water 5.80 C cooler,
Fig. 6). Air temperatures were more variable than water
temperatures (Fig. 4), and the range of temperatures dur-
ing each survey in winter was similar in the northern
and southern areas (Fig. 5). Water temperatures at the


southern sites usually varied by no more than 2-30 C
during surveys in winter, whereas water temperatures at
the northern site usually varied 5-70 C (Fig. 6).
Manatees seem to be thermally stressed at water tem-
peratures below 180 C. Hartman (1979) found that mana-
tees aggregated at the Crystal River and Blue Spring
warm-water sites when the daily low air temperatures were
below 100 C; the number of manatees was largest at tem-
peratures below 0 C. Manatees used the refuges when
water temperatures outside the refuges were below 16-
180 C but occasionally left the refuges to feed when out-
side waters were as cool as 11-13.5 C. Irvine (1983)
found that manatees must increase their metabolic rates in
water below 20 C but did not increase metabolism at air


Table 4. Survey conditions during aerial counts at seven Florida manatee (Trichechus manatus latirostris) aggregation
sites along the Florida coast in each winter 1982-1983 to 1990-1991. Survey-condition scores were a subjective
evaluation of visibility based on water turbidity, surface glare and chop, and wind conditions (*Reynolds 1983). A
score of 1 indicated excellent conditions and a score of 5, very poor conditions. Counts are identified in each category
and are followed by the percentage of the row total in parentheses.
Survey conditions
1 2 3 4 5
Year Excellent Good Fair Poor Very poor Total
1982-83 2(5.0) 21(52.5) 6(15.0) 10 (25.0) 1(2.5) 40
1983-84 4(13.3) 13 (43.3) 12(40.0) 1(3.3) 30
1984-85 6 (10.0) 14 (23.3) 27 (45.0) 11 (18.3) 2 (3.3) 60
1985-86 7(14.6) 6(12.5) 21(43.8) 13(27.1) 1(2.1) 48
1986-87 3 (10.0) 9 (30.0) 12(40.0) 5 (16.7) 1(3.3) 30
1987-88 6(12.5) 15(31.3) 11(22.9) 16(33.3) 48
1988-89 7(25.9) 3(11.1) 12(44.4) 5(18.5) 27
1989-90 9(22.0) 11(26.8) 13(31.7) 8(19.5) 41
1990-91 7 (25.0) 7 (25.0) 8 (28.6) 6 (21.4) 28
Total 35(9.9) 97(27.6) 119(33.8) 82(23.3) 19(5.4) 352






42 INFORMATION AND TECHNOLOGY REPORT 1


ATEMP3

30

20

10


WTEMP3


21.
LU CO


-I--M
0 X
z


Dec Jan Feb


Dec Jan Feb


Fig. 4. Comparison of 3-day mean air (ATEMP3) and water
(WTEMP3) temperatures recorded at northern and southern
survey sites along the eastern coast of Florida, 1982-1991.
Data are presented in standard box plots (see Fig. 2 caption).


a.
2 L
cW CI 24

22
ox
CO


LO
0

ox
z


co


C1-



0 X
C,


Year
Fig. 5. Annual variation in 3-day mean air temperatures
(ATEMP3) recorded on each survey date at northern and
southern survey sites along the eastern coast of Florida,
1982-1991. Data are presented in standard box plots (see
Fig. 2 caption).


Year
Fig. 6. Annual variation in 3-day mean water temperatures
(WTEMP3) recorded on each survey date at northern and
southern survey sites along the eastern coast of Florida,
1982-1991. Data are presented in standard box plots (see Fig.
2 caption).


temperatures as low as 100 C. Campbell and Irvine
(*1981) reported that captive manatees exposed to water
temperatures below 16-18 C for several days experi-
enced lethargy and anorexia.
During cold fronts, water temperatures were usually
below 180 C at the northern sites but remained consider-
ably warmer in the southern area. Air temperatures dur-
ing cold fronts were generally below this reference point
in the northern and southern areas. Southern areas are
warmed by their proximity to the Florida Current. No
trends in air temperatures were apparent at the northern
or southern sites during the 9-year survey (Fig. 5); water
temperatures at the southern sites were also without
trends (Fig. 6). However, 3-day water temperatures at
the northern site suggested a general increase in tempera-
ture during 1984-85 and 1990-91 (Fig. 6).
The cumulative total number of heating degree-days in
winter (TOTDD in Table 2) recorded at Titusville during
1982-83 through 1986-87 were near the 30-year mean of
712 (Table 5). Departures from the 30-year mean were
notable during several winters. Winter 1987-88 was ab-
normally severe, and winters 1988-89 and 1990-91 were


- r 0.





o II II
1*1 1 1 1 1 1


?46
*.-


S0
-o dr CF,



dsw~w.

~



6i-+






ROBERT A. GARROTT ET AL. 43


Table 5. Number of heating degree-days at Titusville and Fort Myers, Florida, during winters 1982-1983 to 1990-1991
(National Oceanic and Atmospheric Administration 1982-1991).
Year December January February Totala
Titusville
1982-83 110 260 170 699
1983-84 184 301 200 869
1984-85 98 401 176 767
1985-86 252 232 112 768
1986-87 49 276 136 540
1987-88 173 314 304 983
1988-89 130 22 116 318
1989-90 302 118 57 541
1990-91 75 93 78 314
9-year mean 153 224 150 644
30-year mean (NOAA 1991) 712
Fort Myers
1982-83 61 133 87 347
1983-84 101 141 79 384
1984-85 38 177 77 316
1985-86 128 91 38 329
1986-87 22 116 43 198
1987-88 65 129 105 353
1988-89 70 9 73 180
1989-90 178 33 9 229
1990-91 57 24 44 146
30-year mean (NOAA 1991) 107 150 120 441
a Total degree-days represents the entire year (July-June), although almost all degree-days are accumulated from November to March.


unusually mild (Table 5). The number of heating degree-
days also increased through time similar to the northern
water temperature variable.

Model Development and Trend Analysis
Year as a categorical variable contributed signifi-
cantly (P = 0.036) to the model, accounting for 29% of
the variation in the In(COUNT) and identifying signifi-
cant differences in counts among years (Table 6). In Step
2, the correlation coefficients suggested that nearly all
temperature variables explained a significant amount of
the remaining variation. DD10, the summation of de-
gree-days of the 10-day period just prior to the survey,
was the temperature variable that explained the largest
proportion of the variation and was added to the model
(Table 6). For Step 3, we subtracted the DD10 value
from the remaining degree-day variables to exclude the
information that was incorporated into the model at
Step 2. The next iteration identified DX30, the new vari-
able summing degree-days for the 20-day period from 11
to 30 days prior to the survey, as the largest remaining
contributor to the correlation coefficient. Although the F
test indicated that DX30 contributed significantly to the
model (P < 0.001), it only accounted for an additional
13% of the variation in the In(COUNT), raising the
question whether this covariate should be added to the


model (Table 6). A plot of DX30 against the DDIO
variable already incorporated into the model, however,
showed little correlation between the two variables
(Fig. 7a). In addition, a plot of DX30 against the residu-
als from the Step 2 model suggested a general correlation
(Fig. 7b). Therefore, we chose to add DX30 as a second
temperature covariate to the model. No other variables
contributed significantly to the model, thus our best
model was
E{ln(COUNT)} = po + P1 (YEAR) + P2 (DD10) (1)
+ 13 (DD10)2 + 04 (DX30) + s5 (DX30)2
This model required estimating 13 parameters (year as
categorical variable required eight parameters; Table 7) and
accounted for 78% of the variation in the In(COUNT) data
(Table 6). Residual plots for this model indicated a relatively
good fit to the data.
Plots of the unadjusted and adjusted counts on logarith-
mic and natural scales against year, adjusted for the tempera-
ture covariates, suggested an increasing trend in the data
(Figs. 8 and 9). The model had the most significant effect on
counts that were conducted under abnormally warm condi-
tions (Fig. 10). Polynomial contrasts indicated that most of
the between-year variation could be explained by a linear
trend (P < 0.001) and that a quadratic trend was not signifi-
cant (P = 0.498).






44 INFORMATION AND TECHNOLOGY REPORT 1


Table 6. Forward step-wise regressions to select covariates to be included into a log-linear model of total counts of Florida
manatees (Trichechus manatus latirostris) along the eastern coast of Florida during winters 1982-1983 to 1990-1991. Year
was treated as a categorical variable, and linear and quadratic terms were included when testing each of the other 19 possible
covariates. Underlined values are variables that were entered into the model at each step.

Correlation coefficient (R )
Covariate Step 1 Step 2 Step 3
YEAR (categorical) 0.291
NATEMP 0.433 0.671
NATEMP3 0.456 0.662
NWTEMP 0.475 0.672
NWTEMP3 0.332 0.663
NSUVCOND 0.313 0.658
SATEMP 0.475 0.701
SATEMP3 0.536 0.696
SWTEMP 0.432 0.696
SWTEMP3 0.433 0.663
SSUVCOND 0.351 0.729
DD5 0.503
DD10 0651
DD15 0.577 0.660
DD20 0.493 0.706
DD25 0.517 0.762
DD30 0.532 0.782
DD35 0.508 0.741
DD40 0.482 0.723
TOTDD 0.369 0.699
F value 2.31 22.11 12.43
df 8,45 2,43 2,41
P 0.036 <0.001 <0.001
R2 0.29 0.65 0.782


) 50 100 150 200
DD10




I *







0 100 200 300 400
DX30


Fig. 7. Plots of manatee (Trichechus manatus latirostris) counts
pooled at all sites along the eastern coast of Florida, 1982-
1991, showing the relation between (a) DX30, the summation
of degree-days during the 20-day period from 10 to 30 days
prior to the survey, and DD10, the summation of degree-days
from 1 to 10 days prior to the survey, and (b) the residuals
from a simple log-linear model (Step 2, Table 6) and DDIO,
incorporated into the model as a covariate.


a, 400


300 -


o 2uu
x
o 100

0


b. 2


1


Ih,.? .


*
* ** *
;q, eeo
o
oe
---



,,, I






ROBERT A. GARROTT ET AL. 45


Table 7. Comparison of best models fit to counts of Florida
manatees (Trichechus manatus latirostris) at aggregation
sites on the eastern coast of Florida and at Fort Myers.
Values of year and temperature coefficients in scientific
notation.

Coefficient value
Independent variable Eastern coast Fort Myers
Constant 3.18 2.81
1982-83 -5.84 x 10-1 -7.71 x 10-1
1983-84 -7.75 x 10-1 -1.68 x 10-1
1984-85 -3.13 x 10-1 3.55 x 10-1
1985-86 -1.00x 10-2 3.32 x 10-1
1986-87 -6.00 x 10-3 1.33 x 10-1
1987-88 -1.16 x 10-1 -1.55 x 10-1
1988-89 6.44 x 10-1 -1.40 x 10-1
1989-90 1.97 x 10-1 2.71 x 10-1
1990-91 9.63 x 10-1 1.43 x l0-1
DD10 2.65 x 10-2 5.18 x 1072
DDIOS -8.40 x 105 -4.30 x 10-4
DX30 1.22 x 10-2 1.72 x 10-2
DX30S -3.68 x 10-5 -1.04 x 10-4
R2 0.78 0.68


a. 7.5

6.5


4.5

3.5

b.7.5

6.5

5.5

S4.5


- Y












Year


Fig. 8. Comparison of In(COUNT) of numbers of manatees
(Trichechus manatus latirostris) seen during aerial surveys in
winter 1982-1991, (a) before and (b) after adjustment for
temperature covariates. Data from all sites along the eastern
coast of Florida were pooled. Note the large adjustments of
counts during 1988-1989 and 1990-1991.


a. 1500


1000


0

b.1500

E
S1000
0
-0

- 500


0
















I I I I
SYear


Year


Fig. 9. Comparison of counts of manatees (Trichechus manatus
latirostris) during aerial surveys in winter 1982-1991, (a)
before and (b) after adjustment for temperature covariates.
Data from all sites along the eastern coast of Florida were
pooled. Note the large adjustments of counts during
1988-1989 and 1990-1991.


Given these results, we developed a more parsimonious
model by treating year as a continuous variable and manu-
ally performing a backwards stepwise regression to identify
parameters that could be dropped from the model. Our final
model, therefore, was
E{In(COUNT)} = Po + P1(YEAR) + P2(DDIO) (2)
+ P3(DX30) + P4(DX30)2.
This model required the estimation of five parameters
(year as continuous variable requires only one parame-
ter) and had an R2 of 0.678, which is only slightly lower
than the best model that required the estimation of 13
parameters. Again, residual plots indicated that the
model fit the data relatively well. Because the estimates
are on a logarithmic scale, the linear trend translated to
an increasing exponential (geometric) trend on the origi-
nal scale, providing no evidence that the manatee popu-
lation in eastern Florida is decreasing.

Fort Myers Aggregation Site

Numbers of Manatees and Temperature

Because manatees at the Fort Myers site were counted
during the same flights as those on the eastern coast, the











' 83
6


4
77

85
6-




7 85

4 --------------------------
7

6 87



4 i
7'-





6 89

4-


400




300-




o 200 -
0




100-




0


Year
Fig. 11. The number of manatees (Trichechus manatus
latirostris) counted in the vicinity of the Fort Myers power
plant during aerial surveys in winter 1982-1991. A total of
4-10 surveys were conducted each winter. Data are presented
in standard box plots (see Fig. 2 caption).


Fig. 10. Comparison of In(COUNT) of numbers of manatees
(Trichechus manatus latirostris) seen during aerial surveys
before and after adjustment for temperature covariates. Data
from all sites along the eastern coast of Florida during
1982-1991 were pooled. Circles connected by solid lines
represent the observed counts, and triangles connected by
dotted lines represent the counts adjusted for the temperature
covariates.



number and timing of surveys were identical (Appendix
B). A total of 58 counts was obtained between winters
1982-83 and 1990-91; however, we excluded data from
the 14 December 1982 survey because, like during the
surveys in eastern Florida, few animals were seen. The
general characteristics of the counts were also similar to
those obtained on the eastern coast of Florida, and vari-
ation within and among years was considerable
(Figs. 11-14).
The extreme outlier in these data was observed on
19 January 1985 when 338 manatees were counted in one
survey. This was approximately three times the mean num-
ber of manatees counted during nine other surveys in winter
1984-85. The 19 January survey was the only survey at Fort
Myers during the 9-year period that was judged to have been
conducted under excellent survey conditions (Appendix B;


5

4 *
0
3





Year
Fig. 12. Comparison of In(COUNT) of numbers of manatees
(Trichechus manatus latirostris) observed during aerial
surveys at the Fort Myers power plant area in winters
1982-1983 to 1990-1991, (a) before and (b) after adjustment
for temperature covariates.


46 INFORMATION AND TECHNOLOGY REPORT 1


e











7-
6
5
A


91


7
6
5
4


90
.... .... ........ ... ... . ....... .. .....







ROBERT A. GARROTT ET AL. 47


a. 350
300
250
200
8 150
100
50
0
b.350
300
250
0
0 200
S150
5' 100
50
0


Year


Fig. 13. Comparison of counts of manatees (Trichechus manatus
latirostris) during aerial surveys at the Fort Myers power
plant in winters 1982-1983 to 1990-1991, (a) before and (b)
after adjustment for temperature covariates.


Table 3). Under typical survey conditions, which were usu-
ally rated fair to poor at this site (Table 3), a relatively low
proportion of the animals may have been detected. Rela-
tively low counts were obtained during the winters 1988-89
and 1990-91 (Fig. 11). These winters were unusually milder
than previous winters, and the lowest counts in each year
were made during the warmest months (Table 5; Fig. 14).

Model Development and Trend Analysis

Because the data structure and possible covariates of
the Fort Myers manatee population were identical to those
of the manatee population on the eastern coast, we initially
fit the best model developed for this population to the Fort
Myers data. The model fit the data relatively well. How-
ever, the covariates accounted only for 68% of the vari-
ation, whereas they accounted for 78% of the variation in
counts in eastern Florida. We attempted to find a better
model by repeating the stepwise procedure to construct the
model for counts in eastern Florida. This procedure pro-
duced results similar to the analysis of the data from
eastern Florida (Table 8), indicating the best model for the
data from eastern Florida was also the best model for the
data from Fort Myers. The degree-day coefficients had the
same signs for the data from Fort Myers and eastern
Florida, and the relative magnitude of the coefficients was
also similar (Table 7).


5 8




5




-. 3
S 6 85
S 5-






1 V88
4 89


3



6 90
5-
4-

:3 89










Dec Jan Feb
Fig. 14. Comparison of In(COUNT) of numbers of manatees
(Trichechus manatus latirostris) observed during aerial
surveys at the Fort Myers power plant in winters 1982-1983
to 1990-1991 before and after adjustment for temperature
covariates. Circles connected by solid lines represent the
observed counts, and triangles connected by dotted lines
represent the counts adjusted for the temperature covariates.


Polynomial contrasts of the year coefficients indicated
that linear (P < 0.045) and quadratic temporal trends in
adjusted counts (P < 0.012) were significant; however, the
plots of the adjusted counts (Figs. 12 and 13) were not
convincing. The most significant adjustments to the counts
were those obtained during unusually warm weather (Fig.
14). Tukey multiple comparisons revealed the only signifi-
cant differences among years were between 1982-83 and
each of the years 1984-85, 1985-86, 1986-87, and 1989-
90. We concluded that no detectable trend was in the counts
of manatees in the Fort Myers area between 1983-84 and
1990-91 (although counts in 1982-83 were lower than
counts in subsequent years).
To develop a more parsimonious model, we converted
year from a categorical to a continuous variable with a linear
and quadratic term and performed a backward stepwise






48 INFORMATION AND TECHNOLOGY REPORT 1


Table 8. Forward stepwise regressions to select covariates to be included into a log-linear model of total counts of
Florida manatees (Trichechus manatus latirostris) at Fort Myers, Florida during winters 1982-1983 to 1990-1991.
Year was treated as a categorical variable, and linear and quadratic terms were included when testing each of the
other 14 possible covariates. Underlined values are variables that were entered into the model at each step.

Correlation coefficient (R2)
Covariate Step 1 Step 2 Step 3
YEAR (categorical) 33
ATEMP 0.441 0.716
ATEMP3 0.411 0.717
WTEMP 0.493 0.740
WTEMP3 0.481 0.726
SURVCOND 0.377 0.717
DD5 0.457
DD10 0.60
DD15 0.573 0.611
DD20 0.523 0.637
DD25 0.537 0.653
DD30 0.608 0.683
DD35 0.578 0.684
DD40 0.496 0.655
TOTDD 0.358 0.603
F value 3.31 14.82 5.19
df 8,48 2,46 2,44
P 0.004 <0.001 <0.010
R2 0.34 0.60 0.68


regression. This procedure resulted in dropping the linear
year and quadratic DX30 parameters from the model, pro-
ducing the model
E{ln(COUNT)} = Po + PI(YEAR)2 + P2(DD10) (3)
+ P3(DD10)2 + P4(DX30).
This model required estimating five parameters, had
an R2 of 0.501, and fit the data relatively well. Treating
year as a continuous variable resulted in a substantial
decrease in the correlation coefficient. This result and
the results of the post-hoc analysis of the year coeffi-
cients suggested that the most appropriate log-linear
model for the data from Fort Myers was the model
initially developed with the forward stepwise regression
procedure.



Discussion

Results of these analyses revealed that approximately
50% of the variation in the counts of manatees in winter
on the eastern coast of Florida and at Fort Myers in
southwestern Florida can be explained with log-linear
models that incorporate measures of temperature. The
choice of models is probably not critical because most
temperature covariates correlated with each other and
contributed significantly to the models. The best general


models we could construct for the data sets from eastern
Florida and Fort Myers incorporated the same temperature
covariates.
The models used a short-term (1-10 days) and a long-
term (11-30 days) summation of heating degree-days.
Although degree-day variables are artificial because of the
relatively arbitrary scaling to 18.30 C, they provided sev-
eral advantages over the other possible temperature co-
variates. Degree-days are calculated from standard Na-
tional Oceanic and Atmospheric Administration
temperature data that are recorded daily and are readily
available from many locations in Florida (National Oce-
anic and Atmospheric Administration 1982-91). Summa-
tion of degree-days over various time periods also allowed
the development of time-lag temperature covariates that
can provide a measure of the relative severity of winter
temperatures for extended periods prior to each survey. In
contrast, water-temperature data obtained from the power
plants were available only for the survey date and two prior
days. The aerial surveys were standardized for short-term
temperatures because they were conducted only immedi-
ately after cold fronts had moved into the area. Other
problems with power-plant water-temperature data are
that they are recorded only when the plant is operating and
the quality of the measurements was not tightly controlled
(J. R. Wilcox, Florida Power and Light Company, North
Palm Beach, personal observation). These considerations






ROBERT A. GARROTT ET AL. 49


and the comparisons of the correlation coefficients ob-
tained during the stepwise regression procedures led us to
conclude that time-lag degree-day covariates in our mod-
els were the most appropriate temperature covariates.
These results, however, were influenced by a pro-
found increase in the number of manatees counted at the
Indian River and Cape Canaveral power plants during
the late 1980's. Manatees in these areas are difficult to
count because of turbid water and surface foam from the
power-plant discharge. The increase in number of mana-
tees counted at the IR-CC sites may have resulted from
changes in counting conditions because the plants under-
went major operational changes during the mid-1980's
(J. R. Wilcox, Florida Power and Light Company, per-
sonal observation). This possibility must be further in-
vestigated.
Because a substantial proportion of the variability in
these statistical models remains unexplainable, a determina-
tion of whether an increase in the population size caused the
increase in adjusted counts on the eastern coast of Florida is
difficult. Alternate explanations can be suggested for why
the increasing counts may not accurately track trends in the
size of the population. Although the same observer con-
ducted all surveys, the observer may have become more
experienced at conducting the surveys with time. Visibility
may have increased. Changes in manatee behavior or tradi-
tions may have altered the proportion of manatees in winter
aggregations. Random variation in conditions during spe-
cific winters may have led to the appearance of a greater-
than-observed trend in the temperature-adjusted counts.
More analyses must be made. A primary goal should be
to incorporate the previous set of 1977-82 surveys (*Rose
and McCutcheon 1980; *Raymond 1981; *McGehee 1982;
*Packard and Mulholland 1983). Although those earlier
surveys were conducted year-round under a different sam-
pling scheme and surveyors recorded different variables,
data should be investigated to determine whether a suitable
subset of comparable surveys can be selected. Counts were
higher during several months (January 1979, January 1980,
January 1981) on the eastern coast of Florida than in the
years 1982-84, and their inclusion may reduce the apparent
trend from this model.


Conclusions and
Recommendations

The major impetus for this study was the determina-
tion of whether aerial survey counts of manatees at
aggregation sites in winter can provide an index to the
size of the manatee population. A consistent increase in
the number of dead manatees each year created concern
that the populations may be declining (Ackerman et al.
1995). Our analyses of the aerial counts do not provide


evidence in support of a trend of a decreasing population
size of manatees in eastern Florida or at Fort Myers.
Pooled counts of manatees at all survey sites in eastern
Florida, adjusted for temperature, increased during the
past 8 years. Several other factors may have influenced
this trend, such as changing conditions at the IR and CC
plants. In contrast, the temperature-adjusted counts from
the Fort Myers area suggested that this population re-
mained relatively stable.
The interpretation of the results of this analysis requires
caution. Several measures of temperature were identified,
each of which seemed to explain a substantial fraction of the
variation in the count data. Finding such an influential
adjustor by using the counts as a population index was good
and bad. The advantage was that much of the observed
variation in the counts can be explained and, to some extent,
adjusted. The disadvantage was that this indicated the counts
were not responding to only the underlying population levels
but also to external covariates. The better our covariate
model becomes at predicting the counts, the less residual
variability remains to predict true population fluctuations.
The more influential the covariates become, the more seri-
ous becomes the bias induced by model misspecification
(e.g., assuming a linear model instead of the appropriate
nonlinear one). Even if all year effects observed in count data
were caused by fluctuations in the true population levels, the
analysis suggests that no more than half of the variability
observed in the counts can be attributed to changes in
population levels. With the currently available data, statisti-
cal analysis alone can shed no light on the question of
whether the estimated baseline year effects are truly indica-
tive of the population-size trends. Several relatively exten-
sive databases exist on other demographic characteristics of
these manatee populations that include information on age
structure, reproduction, and survival (Eberhardt and O'Shea
1995). We caution against accepting the results of our trend
analysis unless these independent databases provide reason-
able evidence to corroborate our results. We recommend
incorporation of other population data in future attempts to
support or reject the trend suggested by this work.


Acknowledgments

Helpful reviews of the draft manuscript were provided
by K. B. Clifton, J. A. Huff, C. A. Langtimm, L. W. Le-
febvre, M. Mangel, and T. J. O'Shea. M. A. Newton and
P. M. Rose provided helpful insights. K. B. Clifton as-
sisted with the preparation of the figures. This analysis was
funded by the Save the Manatee Trust Fund of the Florida
Department of Environmental Protection. Funding for the
aerial surveys was provided by the Florida Power and
Light Company.






50 INFORMATION AND TECHNOLOGY REPORT 1


Cited References2

Ackerman, B. B. 1995. Aerial surveys of manatees: a summary
and progress report. Pages 13-33 in T. J. O'Shea, B. B. Ack-
erman, and H. F. Percival, editors. Population biology of the
Florida manatee. National Biological Service Information and
Technology Report I.
Ackerman, B. B., S. D. Wright, R. K. Bonde, D. K. Odell, and
D. J. Banowetz. 1995. Trends and patterns in mortality of
manatees in Florida, 1974-1992. Pages 223-258 in T. J.
O'Shea, B. B. Ackerman, and H. F. Percival, editors. Popula-
tion biology of the Florida manatee. National Biological Serv-
ice Information and Technology Report 1.
*Campbell, H. W., and A. B. Irvine. 1981. Manatee mortality
during the unusually cold winter of 1976-1977. Pages 86-91
in *R. L. Brownell, Jr. and K. Rails, editors. The West Indian
manatee in Florida. Proceedings of a workshop held in Or-
lando, Fla., 27-29 March 1978. Florida Department of Natural
Resources, Tallahassee. 154 pp.
Cox, D. R., and D. Oakes. 1984. Analysis of survival data.
Chapman and Hall, New York. 201 pp.
*Eberhardt, L. L. 1982. Censusing manatees. Manatee Population
Research Report 1, Florida Cooperative Fish and Wildlife
Research Unit. University of Florida, Gainesville. 18 pp.
Eberhardt, L. L., and T. J. O'Shea. 1995. Integration of manatee
life-history data and population modeling. Pages 269-279
in T. J. O'Shea, B. B. Ackerman, and H. F. Percival, editors.
Population biology of the Florida manatee. National Biologi-
cal Service Information and Technology Report 1.
Hartman, D. S. 1979. Ecology and behavior of the manatee
(Trichechus manatus) in Florida. American Society of Mam-
malogists Special Publication 5, Lawrence, Kans. 153 pp.
Irvine, A. B. 1983. Manatee metabolism and its influence on
distribution in Florida. Biological Conservation 25:315-334.
Lancia, R. A., J. D. Nichols, and K. H. Pollock. 1994. Estimating
the number of animals in wildlife populations. Pages 215-253
in T. A. Bookhout, editor. Research and management tech-
niques for wildlife and habitat. The Wildlife Society, Bethesda,
Md.
Lefebvre, L. W., B. B. Ackerman, K. M. Portier, and K. H. Pol-
lock. 1995. Aerial survey as a technique for estimating trends
in manatee population size-problems and prospects.
Pages 63-74 in T. J. O'Shea, B. B. Ackerman, and H. F. Per-
cival, editors. Population biology of the Florida manatee.
National Biological Service Information and Technology Re-
port 1.
*McGehee, M. A. 1982. Manatees (Trichechus manatus): abun-
dance and distribution in and around several Florida power
plant effluents during the winter of 1981-1982. Final report
prepared for the Florida Power and Light Co., Contract
31534-86419. 67 pp.
National Oceanic and Atmospheric Administration. 1982-91.
Climatological data, Florida. National Climatic Data Center,
Asheville, N.C. Volumes 86-95.
O'Shea, T. J. 1988. The past, present, and future of manatees in
the southeastern United States: realities, misunderstandings,


and enigmas. Pages 184-204 in R. R. Odom, K. A. Riddleber-
ger, and J. C. Ozier, editors. Proceedings of the Third South-
eastern Nongame and Endangered Wildlife Symposium,
Georgia Department of Natural Resources, Social Circle.
*Packard, J. M., and R. Mulholland. 1983. Analysis of manatee
aerial surveys: a compilation and preliminary analysis of win-
ter aerial surveys conducted in Florida between 1977 and 1982.
Manatee Population Research Report 2. Technical Report 8-2.
Florida Cooperative Fish and Wildlife Research Unit, Univer-
sity of Florida, Gainesville. 119 pp.
*Raymond, P. W. 1981. Manatees (Trichechus manatus): Abun-
dance and distribution in and around Florida power plant
effluents. Final report prepared for the Florida Power and
Light Co. Contract 31534-81511. 62 pp.
Reid, J. P., R. K. Bonde, and T. J. O'Shea. 1995. Reproduction
and mortality of radio-tagged and recognizable manatees on
the Atlantic Coast of Florida. Pages 171-191 in T. J.
O'Shea, B. B. Ackerman, and H. F. Percival, editors. Popu-
lation biology of the Florida manatee. National Biological
Service Information and Technology Report 1.
Reid, J. P., G. B. Rathbun, and J. R. Wilcox. 1991. Distribution
patterns of individually identifiable West Indian manatees
(Trichechus manatus) in Florida. Marine Mammal Science
7:180-190.
*Reynolds, J. E., III. 1983. Distribution and abundance of West
Indian manatees (Trichechus manatus) around selected Flor-
ida power plants following winter cold fronts. Final report
prepared for the Florida Power and Light Co., Contract 71527-
88609. 75 pp.
*Reynolds, J. E., III. 1985. Distribution and abundance of the
West Indian manatee (Trichechus manatus) around selected
Florida power plants following winter cold fronts: 1984-
1985. Final report prepared for the Florida Power and Light
Co., Order 01249-87331. 48 pp.
*Reynolds, J. E., III. 1991. Distribution and abundance of the
West Indian manatee (Trichechus manatus) around selected
Floridapower plants following winter cold fronts: 1990-1991.
Report for Florida Power and Light Co., Order B89806-00264.
39 pp.
Reynolds, J. E., III, and J. R. Wilcox. 1985. Abundance of West
Indian manatee Trichechus manatus around selected Florida
power plants following winter cold fronts: 1982-1983. Bul-
letin of Marine Science 36:413-422.
Reynolds, J. E., III, and J. R. Wilcox. 1986. Distribution and
abundance of the West Indian manatee Trichechus manatus
around selected Florida power plants following winter cold
fronts: 1984-85. Biological Conservation 38:103-113.
*Rose, P. M., and S. P. McCutcheon. 1980. Manatees (Trichechus
manatus) abundance and distribution in and around several
Florida power plant effluents. Final report prepared for the
Florida Power and Light Co., Contract 31534-86626. 128 pp.
*U.S. Fish and Wildlife Service. 1989. Florida manatee
(Trichechus manatus latirostris) Recovery Plan. U.S. Fish
and Wildlife Service, Atlanta, Ga. 98 pp.


2 An asterisk denotes unpublished material.





ROBERT A. GARROTT ET AL. 51


Appendix A.


Aerial counts of Florida manatees (Trichechus manatus
latirostris) along the eastern coast of Florida and
potential adjustment factors.













Survey conditions


Survey North"
date Ac A3d We W3f SCg


821214"
821220
83 113
83 116
83 119'
8325
832 8

831227
84 130
8421
8427
8431

8516
85110
85114
85119
85 122
85 127
8522
8529
85215
85218

851219
851222
851227
86112
86115
86 121
86 129
86214

87 112
87 115
87124
872 1
87211


South"
A A3 W W3 SC


15 19 21 22
13 11 16 17
7 12 18 20
12 10 17 18
10 9 15 15
12 17 18 19
10 14 18 18

0 2 12 13
11 14 20 20
5 9 16 18
6 10 17 18
9 12 17 18

8 14 18 20
11 11 17 17
7 10 13 14
13 15 16 16
-2 6 11 13
5 13 12 13
23 22 19 17
8 15 16 19
11 9 14 14
14 10 16 15

14 13 17 17
9 11 15 16
2 7 12 13
15 17 17 18
13 12 16 16
12 15 18 18
2 6 13 15
6 11 16 19

8 14 16 17
16 12 17 16
6 8 16 18
14 15 16 16
15 13 16 17


Counts by site
1 2 3 4 5 6 7 8 9 10


22.0 19.0 24.0 24.0 2.3
15.5 15.0 23.0 23.0 1.8
11.5 16.0 23.5 24.0 2.5
14.0 15.5 22.5 23.5 2.5
17.0 15.0 22.0 22.0 3.5
15.5 18.5 21.5 22.0 2.3
14.5 18.5 22.0 22.0 2.3

14.0 10.0 20.5 21.0 2.3
17.5 18.0 22.5 23.0 1.8
14.0 16.0 22.0 22.5 2.0
12.0 14.0 22.5 22.5 2.8
10.5 13.5 21.0 22.0 2.0

13.5 16.0 22.0 22.5 2.5
18.0 17.0 22.5 22.0 2.5
17.5 15.0 21.5 22.0 3.5
16.5 19.0 22.0 22.0 2.3
4.5 10.5 21.0 21.0 2.5
13.0 15.5 21.0 21.0 3.0
26.0 24.0 22.0 22.0 2.5
14.5 18.0 21.5 22.0 2.5
12.0 11.5 21.0 21.0 1.5
18.5 15.0 21.0 21.0 2.8

19.0 18.0 24.5 23.5 3.3
12.0 15.0 24.5 24.5 2.3
11.5 12.5 22.5 22.5 2.5
18.5 20.5 22.0 23.0 3.5
13.5 14.0 23.5 23.5 2.3
17.0 18.0 23.0 23.5 2.3
11.5 10.0 22.0 22.5 2.3
15.5 15.5 23.0 23.5 2.0

11.5 17.0 22.0 22.0 2.8
21.0 17.5 20.0 22.0 2.8
13.0 16.0 22.5 23.0 1.8
16.0 17.0 22.0 22.0 2.5
18.5 16.5 23.0 22.0 3.0


0 0
0 14
0 11
0 15
0 14
0 1
0 5

0 9
0 16
0 14
0 9
0 20

0 8
0 12
0 5
0 34
0 7
0 2
0 0
0 6
0 23
0 16

0 8
0 16
0 15
0 3
0 17
0 7
0 6
0 7

0 19
0 129
0 13
0 9
0 16


Cumulative degree days
S' -5i -10 -15 -20 -25 -30 -35 -40


1 0
32 10
43 39
98 56
85 0
33 29
41 41

60 12
15 16
20 35
27 15
22 30

13 20
17 30
34 62
58 144
231 109
143 234
29 28
28 23
191 190
85 116

5 17
124 109
217 172
11 57
131 133
81 158
272 185
52 35

107 160
45 86
78 59
47 182
98 134


k


41 18 18 18 18 18 18 18 41
92 48 69 69 69 69 69 69 69
202 33 69 69 69 121 161 179 179
248 78 99 115 115 135 201 225 225
299 77 131 163 166 166 208 256 276
422 29 68 112 174 252 273 289 289
450 54 66 122 151 227 281 314 317

235 90 90 114 140 140 151 166 195
563 27 65 93 148 208 288 361 418
601 65 71 113 171 223 290 355 456
674 54 137 138 181 228 280 352 427
777 26 28 35 58 145 214 241 279

192 34 34 34 34 49 117 128 142
245 71 87 87 87 87 104 180 180
295 63 137 137 137 137 140 164 231
358 63 126 200 200 200 200 203 227
424 75 170 217 266 266 266 266 274
519 95 170 265 312 361 361 361 361
559 26 106 211 284 341 401 401 401
594 35 42 106 224 279 350 420 436
680 71 117 121 185 292 345 430 488
731 85 156 172 194 256 373 436 499

101 58 58 70 77 77 77 77 77
139 57 96 102 115 115 115 115 115
220 82 138 177 183 197 197 197 197
325 16 38 83 169 232 282 282 302
360 47 73 83 170 239 289 317 330
390 19 77 91 103 170 252 308 347
482 75 108 135 168 194 221 316 371
568 40 40 67 156 190 208 254 280

151 33 88 114 128 132 138 138 151
186 54 84 138 160 166 168 173 173
244 58 62 118 152 201 221 224 226
333 56 127 147 182 215 270 296 310
400 56 67 122 194 214 249 282 336










871220
88112
88 124
88 128
88214
88217
88223
88228

881214
891219
89 123k
89 211k
89227

891210
891214
891224
891226
90115
90228

901210
91123
91 213
91 217


23.0 20.5 23.0 23.0 3.0
16.5 18.0 22.5 22.0 2.3
19.5 18.0 22.0 22.0 2.8
13.5 13.0 20.5 21.0 3.3
15.0 15.5 22.0 22.0 2.5
16.5 19.0 23.0 22.5 2.0
23.5 20.5 22.5 22.5 2.5
18.5 17.0 21.5 21.5 3.0

14.0 17.5 21.5 21.0 2.0
14.5 14.0 20.5 21.0 2.0
17.0 20.0 22.0 22.5 1.5
19.0 19.5 22.5 22.0 2.5
15.5 13.0 20.5 21.0 2.0

16.0 20.5 24.0 24.0 2.0
14.0 17.5 22.5 23.0 2.2
2.5 10.5 20.0 22.0 3.2
10.5 6.5 20.0 20.0 2.2
21.0 17.5 22.5 23.0 2.2
22.0 20.5 23.5 23.5 2.3

18.0 17.5 25.0 26.0 2.7
16.5 17.0 26.0 25.5 1.3
18.0 18.0 26.0 26.5 2.5
13.0 14.0 23.5 24.5 2.5


0 83
0 81
0 85
0 76
0 137
0 156
0 117
0 185

0 37
0 74
0 15
0 14
0 229

13 25
5 41
5 1
3 54
2 39
17 49

16 84
13 129
88 28
55 97


12 19 27
7 99 71
10 53 78
3 269 110
5 277 99
47 96 276
3 23 23
22 84 72

3 19 59
26 153 173
2 3 0
1 15 1
4 112 23

6 26 60
10 118 78
0 226 133
0 266 227
15 73 110
3 3 0

6 38 35
4 18 24
13 11 54
2 202 75


a Sites and 2.
Sites 7 and 8.
Air temperature (0 C) on day of survey.
d
Average air temperature (0 C) for 3 days prior to survey.
e Water temperature ( C) on day of survey at power plant intake.
f
Average intake water temperature (0 C) for 3 days prior to survey.
g Index to flight and sighting conditions during survey.
h
Subheadings under counts are 10 different sites surveyed.
SFrom start of winter (1 Nov) to survey date.
-5,..., -40 are starting points in days relative to the survey date.
Survey unusable for analysis.


203 65 73 88 141 151 174 181 203
353 62 114 148 150 162 219 223 260
438 44 74 138 183 212 235 236 301
521 96 127 163 230 282 308 318 320
699 71 150 158 244 305 326 386 429
740 72 145 191 219 314 346 381 449
772 25 87 160 223 238 334 378 408
843 71 95 158 231 294 308 405 449

85 34 41 76 85 85 85 85 85
142 57 91 98 133 142 142 142 142
161 14 14 14 20 20 20 20 77
187 24 24 24 33 39 39 39 45
275 81 81 81 112 112 112 114 128

115 28 60 66 88 115 115 115 115
152 50 76 104 108 130 152 152 152
254 74 102 152 178 205 209 232 254
312 125 143 185 217 253 263 281 312
444 51 65 87 132 257 275 317 349
532 38 40 42 53 57 57 78 82

53 29 32 32 34 42 53 53 53
170 40 68 74 74 74 89 89 95
219 20 28 30 40 87 110 123 123
260 41 65 70 71 90 130 158 164







54 INFORMATION AND TECHNOLOGY REPORT 1


Appendix B. Aerial counts of Florida manatees (Trichechus manatus

latirostris) near Fort Myers, Florida, and potential

adjustment factors.


Survey Survey conditions Cumulative degree days
date Aa A3b Wc W3d SCe Counts Sf -5g -10 -15 -20 -25 -30 -35 -40
821214' 18 17 24 25 4 5 18 11 11 11 11 11 11 11 18
821220 14 13 20 20 4 32 54 36 47 47 47 47 47 47 47
83 113 10 13 21 22 4 56 107 25 39 39 39 60 89 100 100
83116 12 13 18 18 4 85 134 52 56 66 66 67 108 127 127
83 119 17 15 18 18 5 53 160 43 79 90 93 93 106 142 153
8325 18 20 18 18 4 45 221 20 34 58 88 140 143 154 154
8328 13 16 18 19 4 41 235 31 35 62 75 118 154 165 168
831227 13 7 17 19 3 56 94 68 68 70 74 78 79 80 85
84130 15 16 21 23 2 122 252 14 31 40 68 90 135 169 227
8421 12 14 19 20 3 89 270 32 32 58 67 96 135 177 245
8427 10 13 19 21 3 120 303 30 61 64 91 96 120 161 209
8431 11 13 20 21 3 83 337 16 23 23 23 64 85 99 116
85 16 13 16 20 20 4 93 71 17 17 17 17 17 41 55 55
85110 17 17 19 19 3 141 91 29 37 37 37 37 37 75 75
85 114 16 13 17 18 3 80 119 31 64 64 64 64 64 66 102
85 119 17 18 18 18 1 338 133 14 45 78 78 78 78 78 80
85 122 6 9 21 20 3 136 184 54 82 104 130 130 130 130 130
85 127 13 15 21 23 3 139 231 47 101 129 151 177 177 177 177
8522 24 24 22 22 5 99 232 1 31 102 118 148 177 177 177
8529 15 16 19 21 3 70 247 16 16 38 109 124 156 185 193
85215 9 11 20 21 4 163 297 48 66 66 77 137 167 206 226
85218 19 17 21 21 3 100 309 42 68 77 77 99 176 190 221
851219 18 16 18 18 4 81 37 27 32 37 37 37 37 37 37
851222 11 13 17 17 3 126 62 27 57 57 62 62 62 62 62
851227 10 10 15 16 3 104 116 54 81 111 111 116 116 116 116
86112 18 19 19 19 3 147 138 4 10 19 66 102 133 133 138
86115 14 14 18 18 4 147 162 27 33 34 61 113 135 157 160
86121 16 18 18 19 4 248 170 8 33 36 42 55 109 136 165
86129 11 10 17 18 3 191 214 44 51 59 79 86 89 135 173
86214 15 14 18 21 4 48 243 24 24 25 73 77 81 108 114
87112 11 16 19 20 4 107 71 19 37 59 65 65 65 65 71
87115 20 17 20 20 2 140 83 31 31 61 76 77 77 77 77
87124 12 14 20 21 3 91 109 26 26 56 63 90 103 103 103
872 1 16 16 20 20 3 153 143 21 60 60 72 91 109 131 137
87211 16 15 20 20 3 111 168 25 25 46 85 85 97 115 133
871220 22 18 21 21 4 77 67 25 25 33 51 51 62 62 67
88112 15 14 19 20 4 93 118 28 37 51 51 51 76 76 93
88124 11 13 19 20 4 90 159 27 36 64 78 80 92 92 117
88 128 11 11 18 18 4 134 202 51 70 83 112 121 130 135 135
88214 13 13 19 20 4 203 272 32 62 62 111 140 145 170 184
88217 16 17 20 20 2 191 281 30 48 70 79 130 148 162 191
88223 21 18 21 21 4 40 290 9 27 50 79 79 131 157 167
88228 19 15 20 20 4 104 313 23 32 50 73 102 102 154 180
881214 13 15 21 22 3 62 34 20 22 32 34 34 34 34 34
881219 12 11 18 20 3 84 72 38 58 59 70 72 72 72 72
89123 18 18 22 23 2 21 81 3 3 3 9 9 9 9 47
89211 16 17 23 24 3 14 88 7 7 7 9 10 10 10 16
89227 15 12 20 21 3 77 153 64 64 64 73 73 73 73 75







ROBERT A. GARROTT ET AL. 55


Appendix B. Continued.

Survey Survey conditions Cumulative degree days
date Aa A3b WC W3d SCe Counts S' -58 -10 -15 -20 -25 -30 -35 -40
891210 14 19 21 21 2 113 42 10 34 35 36 42 42 42 42
891214 11 15 20 21 2 113 68 35 44 62 62 62 68 68 68
891224 2 7 17 19 3 126 133 61 65 99 109 126 126 127 133
891226 11 6 14 16 2 198 169 97 97 121 135 161 162 163 169
90115 20 15 20 20 3 134 211 23 23 25 43 139 140 163 178
90228 20 19 22 22 3 21 228 8 8 8 9 9 9 17 17

901210 14 15 21 21 2 186 37 28 35 35 35 35 37 37 37
91 123 16 17 21 21 3 26 82 10 19 23 23 23 25 25 25
91213 19 18 21 22 4 26 93 10 10 10 11 21 26 34 34
91217 13 13 21 21 3 80 120 28 38 38 38 39 48 58 62

a Air temperature ( C) on day of survey.
Average air temperature ( C) for 3 days prior to survey.
c Water temperature (o C) on day of survey at power plant intake.
dAverage intake water temperature (o C) for 3 days prior to survey.
e Index to flight and sighting conditions during survey.
From start of winter (1 Nov) to survey date.
-5, ..., -40 are starting points in days relative to the survey date.
h Survey unusable for analysis.






56 INFORMATION AND TECHNOLOGY REPORT 1


Fixed-width Aerial Transects for Determining Dugong Population

Sizes and Distribution Patterns



by


Helene Marsh


James Cook University
Department of Tropical Environment Studies and Geography
Townsville, Queensland 4811, Australia


Abstract. The fixed-width transect technique developed for surveying the dugong (Dugong dugon)
from the air at large spatial scales (tens of thousands of km2) is described and evaluated. Perception
bias (the proportion of groups that is visible in the transect, yet missed by observers) is corrected with
a modified Petersen estimate calculated for each of two teams of tandem observers, one on each side
of the aircraft. Availability bias (the proportion of animals that is unavailable to observers because of
water turbidity) is standardized by comparing the proportion of individuals at the surface during the
survey with the proportion at the surface in a clear-water area when all dugongs are potentially visible.
This fixed-width transect technique provides a standardized estimate of minimum population size and
is useful for producing density-distribution maps for monitoring trends in abundance over large spatial
scales and long time periods, and for assessing the probable impact of direct anthropogenic mortality.
However, the population'size estimated with the technique has a coefficient of variation (S.E./mean)
of 12% at best, which means detection of a low-level chronic decline in dugong abundance even at a
large spatial scale would take about one decade.
Key words: Aerial survey techniques, bias corrections, dugong, Dugong dugon.


The range of the dugong (Dugong dugon) extends
throughout the tropical and subtropical coastal and island
waters of the Indo-west Pacific from east Africa to the
Solomon Islands and Vanuatu and between 260 and 27
north and south of the equator (Nishiwaki and Marsh
1985). The distribution spans the waters of more than 40
countries. Over much of this range, dugongs are now
believed to be represented by relict populations, separated
by large areas in which they are close to extinction or
extinct (Nishiwaki and Marsh 1985) or have never oc-
curred. For most countries, however, this assessment is
based on anecdotal information and the extent of popula-
tion declines or range contractions is unknown because
quantitative information on the distribution and abun-
dance of dugongs and their habitats is unavailable.
The only quantitative information on dugong popula-
tion size is from aerial surveys. Hughes and Oxley-Oxland
(1971) demonstrated that aerial surveys were useful for
studying dugongs in Mozambique. Heinsohn et al. (1976)
were the first to survey dugongs from the air in Australia,
and most surveys have been in Australian waters. Aerial


SAn asterisk denotes unpublished material.


surveys of dugongs have also been conducted in Kenya
(*'Ligon 1976), Papua New Guinea (Ligon and Hudson
1977; Hudson 1980a, 1980b), Palau (*Brownell et al.
1981; Rathbun et al. 1988), Irian Jaya in Indonesia (*Salm
et al. 1982), the Arabian region (*Preen 1989), and Vanu-
atu (Chambers et al. 1989).
The designers of the surveys before 1983 assumed that,
because dugongs feed primarily on seagrasses (Marsh
et al. 1982), they mainly occur in coastal waters within
about 2 km of land. The survey technique was broadly
similar to the extended-area technique (sensu *Packard
1985 and Lefebvre et al. 1995) used for manatees
(Trichechus manatus latirostris). Dugongs were counted
from aircraft at altitudes of 275-300 m and about 0.8 km
from and parallel to the shore. If a large group of dugongs
was detected, a count was made while the aircraft circled.
In some studies, flights were made over additional tran-
sects where suitable habitat was known to extend farther
offshore. No corrections were made for dugongs that were
not seen by observers (e.g., because of water turbidity).
This technique proved useful for identifying areas in
which large numbers of dugongs occurred close to the
shore (see Nishiwaki and Marsh [1985] for a summary of






HELENE MARSH 57


counts in various areas). Because of its ease of implemen-
tation, the technique is still useful for identifying inshore
areas in developing countries where dugongs occur, par-
ticularly in regions where the continental shelf is narrow
(Chambers et al. 1989).
Dugongs have been sighted tens of kilometers from the
coast in large embayments (e.g., Shark Bay in Western
Australia; *Marsh et al. 1991; Marsh et al. 1994) and
where the continental shelf is broad (e.g., Torres Strait
[Marsh and Saalfeld 1988, 1991] and the northern Great
Barrier Reef lagoon [Marsh and Saalfeld 1989]). The
number of dugongs sighted during a survey over the shore-
line is an unreliable index of abundance because it depends
on the degree to which the distribution of the animals
follows the coast, which is variable, even where the con-
tinental shelf is narrow. Hence, the shoreline method is
unsuitable for tracking temporal changes in dugong abun-
dance, especially at large spatial scales.
The shoreline technique has largely been replaced by
fixed-width transect surveys (Marsh and Saalfeld 1989;
Marsh and Sinclair 1989a, 1989b) designed to provide
standardized minimum population-size estimates of du-
gongs as a basis for monitoring temporal changes in abun-
dance, for the assessment of the impact of direct anthropo-
genic mortality, and for density-distribution maps at scales
required for management by zoning. A fixed-width tran-
sect technique was adopted in preference to the line tran-
sect technique often used for dolphin surveys (Forney et al.
1991). Dugongs are generally more difficult to sight than
dolphins because they are most often seen as solitary
individuals or adult female-calf pairs in turbid water. Ac-
cordingly, we decided to use a technique with which
observers did not have to take their eyes off the water to
read an inclinometer.
Here I review the fixed-wing transect technique as a
background to the evaluation of the relevance of dugong
aerial-survey techniques for estimating manatee popula-
tion sizes or trends.


Material and Methods

Survey Procedure

Dugongs are counted on either side of an aircraft flying
at 185 km/h at an altitude of 137 m over 200-m-wide strip
transects. Altitude is maintained with the aid of a radar
altimeter. The strip transects, defined by markers on the
wing struts (see diagram in Caughley 1977), are suffi-
ciently narrow to preclude detectable variation in dugong
sightability across the transect (Marsh and Saalfeld 1990).
The survey crew comprises six people: the pilot, a
front-right survey leader, and two teams of tandem ob-
servers, one on either side of the aircraft. The survey leader


records the data with a portable computer programmed as
a data logger and timer and equipped with a printer that
produces an immediate hard copy of the data. The mid-seat
observers report their sightings to the survey leader via a
2-way intercom system connected to one track of a 2-track
tape recorder. The rear-seat observers are usually screened
from the mid-seat observers with a curtain and acoustically
isolated from the other crew members but can communi-
cate with each other. They report their sightings into the
second track of the tape recorder (Fig. 1).
All reports from observers are in standardized format:
dugong group size, number of calves, number at the sur-
face, and position of the sighting inside the transect strip.
The top (farthest from aircraft), middle, and bottom thirds
of the transect are color-marked on the wing strut to
facilitate the determination of the position of sightings
inside the strip. This information is recorded to increase
the probability of distinguishing between different, simul-
taneously reported sightings by both members of a tandem
team. Surveys are made only when the cloud cover is less
than 50% and the sea is calm ( to minimize glare off the surface of the water from a low
or mid-day sun.
After each flight, the tape record of each transect is used
to verify and edit the computer records, so that each
sighting can be coded as made by one (specified) member
or both members of a tandem observing team. The reports
of team members are different if they are unambiguously
distinct (usual situation) or if they are separated by 5 or
more seconds. Discrepancies between the reports of ob-
servers sighting the same group of dugongs are also noted.
When training a new observer (Marsh and Saalfeld
1989), I use a functional tandem team on one side of the
aircraft and one trained observer and the trainee on the
other side. During training, the intercom system is
switched so that the trainee can hear the reports of his or
her counterpart on the same side of the aircraft. This
system greatly reduces the time to train reliable ob-
servers.
Corrections are made for perception bias (dugongs that
are visible but missed by observers) and availability bias
(dugongs that are unavailable to observers because of
water turbidity) with correction factors that are calculated
separately for each survey. These corrections compensate
for fluctuations in variations in the visibility biases that are
due to changes in sea state, glare, and water turbidity
within the range of acceptable survey conditions and that
cannot be eliminated by the standardization of procedures.
They also compensate for differences between observers
and reduce the need to use the same observers for each
survey.
The correction for perception bias is based on a modi-
fied Petersen estimate calculated separately for the two






58 INFORMATION AND TECHNOLOGY REPORT 1


Survey --,
Pilot Leader

Intercom/ )
f 7 Tape Recorder
I I --

Mid-Seat Mid-Seat
observer n Observer

Screening-curtains


r
Rear-Seat
Observer

Port


r
Rear-Seat

Starboard
Starboard


Pilot
Flys/navigates aircraft
Blows whistle at start and end of transect
Uses stopwatch to check navigation
Survey Leader
Manages survey: Records aircraft height,
Landmarks, Beaufort sea state.
Standardized observations of mid-seat
observers into data-logger
Mid-Seat Observers
Communicate with each other and
survey leader during transects
Rear-Seat Observers
Communicate with each other
during transects
(All members wear polarized sunglasses)
- Track 1 Intercom/Tape Recorder
-- Track 2 Intercom/Tape Recorder


Fig. 1. The arrangement
and duties of the survey
crew during fixed-width
aerial transect surveys
of dugongs (Dugong
dugon).


teams of two observers (Marsh and Sinclair 1989a). This
correction factor requires the assumption that all groups
are equally sightable. Marsh and Sinclair (1989b) showed
that this is a reasonable assumption for the small groups
of dugongs usually observed.
The correction for availability bias (Marsh and Sin-
clair 1989a) serves to standardize the results of each
survey for the proportion of animals that was unavailable
to observers because of water turbidity during that sur-
vey. This correction rests on the untested assumption that
the proportion of dugongs at the surface is constant
within the limits of acceptable survey conditions. This
proportion (16.7%) is based on data from an aerial survey
over a clear water area in Moreton Bay near Brisbane
where all dugongs were assumed to be visible. The
proportion is not significantly different from that ob-
tained from vertical aerial photographs of dugongs in the
same area (Marsh and Sinclair 1989b). This fraction is,
however, much greater than the 1.9% suggested by An-
derson and Birtles' (1978) surface observations of diving
and surfacing dugongs in muddy water. The differences
may be due to the observation platform or spatial vari-
ation in diving by dugongs. I believe the corrections for
availability bias are probably conservative and the popu-
lation size obtained with the fixed-wing transect tech-
nique is probably underestimated.

Large Groups of Dugongs

The mean size of a group of dugongs sighted in these
surveys is between 1.3 and 2.1 animals (Marsh and
Sinclair 1989a). Groups of more than 10 dugongs are rare
(<2% of groups; Marsh and Saalfeld 1988, 1989, 1990;
*Marsh et al. 1990, 1991, 1994). When these groups are
sighted (even outside the transect), the flight course is


interrupted and the group is circled and photographed to
obtain a total count. These groups are then treated as a
separate stratum of large herds as suggested by Norton-
Griffiths (1978).

Survey Design

The survey areas ranged from about 2,000 to
40,000 km2 (Table 1). Each area was divided into blocks
(sampling strata) based on expected dugong density. The
sampling fraction in each block was proportional to
dugong density, varying from about 5% where few du-
gongs were expected to about 20% where the known
dugong density was high. The sampling fraction in the
entire survey usually averaged about 10%. Dugongs in
each block were sampled over systematically spaced
transects (the first transect was placed randomly). Be-
cause of their logistical advantages and to facilitate the
production of density-distribution maps, systematic
rather than random transects were used. The transects
were aligned across the depth contours to increase the
precision of the population-size estimates. The same
transects were used on repeated surveys of the same area
so that transects could be used as factors in the analyses.
I now use a global positioning system to locate the
aircraft on the transects.

Analysis

Because transects are variable in area, the Ratio
Method (Jolly 1969; Caughley and Grigg 1981) was used
to estimate density and population size and their associ-
ated standard errors in each block in each survey. Any
statistical bias from this method was considered incon-
sequential because of the high sampling fraction






HELENE MARSH 59


Table 1. Estimates of the number and densities of dugongs (Dugong dugon) and the associated coefficients of variation
where surveys were conducted with the fixed-width transect technique.
Population
Area estimate Density kml' Coefficients
Location (km2) S.E. S.E. of variation Reference
Shark Baya 14,240 10,146 0.71 14.6 *Marsh et al. (1991)
1,478 0.10 Marsh et al. (1994)
Exmouth Gulf- 3,397 1,964 0.58 18.5 *Marsh et al. (1991)
Ningalooa 363 0.11
North coast 28,746 13,800 0.48 19.4 Bayliss (1986) in Bayliss
Northern Territoryb 2,683 0.09 and Freeland (1989)
Western Gulf of 27,216 16,846 0.62 19.3 Bayliss and Freeland (1989)
Carpentariab 3,259 0.12
Torres Straite 30,533 12,522 0.41 11.9 *Marsh and Saalfeld
1,497 0.05 (1988 and 1991)
Northern Great 31,288 8,110 0.26 13.2 Marsh and Saalfeld (1989)
Barrier Reefc 1,073 0.03
Southern Great 39,396 3,479 0.088 13.2 Marsh and Saalfeld (1990)
Barrier Reefc 459 0.012
Southeast 9,170 2,479 0.26 14.7 *Marsh et al. (1990)
Queenslandc 365 0.04
Arabian Gulfd 34,604 7,310 0.25e 17.8 *Preen (1989)
1,300 0.045
Saudi Arabian coast 22,370 1,820 0.08e 20.9 *Preen (1989)
of Red Sea 380 0.017
a Western Australia.
b Northern Territory, Australia.
c Queensland, Australia.
d Saudi Arabia.
e Excluding zones in which too few dugongs were sighted for population-size estimates.


(Caughley and Grigg 1981). Input data were the esti-
mated number of dugongs (in groups of <10) for each
tandem team per transect, calculated with the corrections
for perception and availability biases. The resultant
standard errors were adjusted to incorporate the errors
associated with the appropriate estimates of the percep-
tion and availability correction factors and mean group
size as outlined in Marsh and Sinclair (1989a). At the end
of the analyses, the number of dugongs in groups of more
than 10 was added to the estimates of the population size
and density in the appropriate block, as outlined in Nor-
ton-Griffiths (1978).
Density diagrams, adjusted for sampling intensity,
were produced with the Arcinfo GIS package. A grid
coverage (2.5 km or 5 km2) was combined with the
coastline coverage. The corrected number of dugongs
and the transect length in each grid cell were calculated.
The density in each grid cell was then calculated as
follows:
density/km2 = corrected number of dugongs sighted in
cell/survey area in cell, where survey area = transect
length in km x transect width (i.e., 0.4 km).


Results


Distribution of Dugongs

Density distribution maps were produced for the entire
survey area (Table 1). High local densities of dugongs occur
in inshore waters sheltered from trade winds and in associa-
tion with offshore reefs and shoals in the northern Great
Barrier Reef (Marsh and Saalfeld 1989; *Marsh et al. 1993)
and Torres Strait (Marsh and Saalfeld *1988, *1991). Large
numbers of dugongs were sighted in more-than-10-m-deep
water in several areas including Shark Bay in Western
Australia (*Marsh et al. 1991; Marsh et al. 1994), Torres
Strait (Marsh and Saalfeld 1988, 1991), the northern Great
Barrier Reef region (Marsh and Saalfeld 1989; *Marsh et al.
1993), and Hervey Bay in southeastern Queensland (*Marsh
et al. 1990). The proportion of dugongs in these deeper water
areas is unknown because we lack information on the rela-
tion between diving and surfacing times at different depths.
In contrast to their essentially inshore distribution where the
continental shelf is narrow, dugongs seen in waters deeper
than about 10 m in northern Australia tend to be more than






60 INFORMATION AND TECHNOLOGY REPORT 1


10 km from land (Marsh and Saalfeld *1988, 1989, *1991;
Marsh et al. *1990, *1991, *1993, 1994).

Detection of Population Trends

Temporal changes in density have been studied with
repeated surveys of dugongs in the same area and with
analysis of variance usually with and without measures of
sea state or cloud cover at each transect as continuously
distributed covariates (Bayliss and Freeland 1989; Marsh
and Saalfeld 1989; *Marsh et al. 1993). Blocks and times
were treated as fixed factors and transects as a random factor
nested within blocks. Data for all analyses were corrected
densities/km2 based on mean group sizes and the estimated
correction factors for perception and availability bias; each
transect contributed one density per survey based on the
combined corrected counts of both tandem teams. The den-
sities were transformed (logl0x + 1) for analysis to equalize
the error variances.
The population-size estimates (Table 2; Fig. 2) are con-
sistent, especially in surveys separated by relatively short
time intervals (months). The inclusion of sea state and
cloud cover as covariates in the analyses made little dif-
ference to the results and did not alter the conclusions
(Marsh and Saalfeld 1989), suggesting that the method
was appropriate for stabilizing most biases in visibility
because of weather conditions.
Marsh and Saalfeld (1989) used Gerrodette's (1987)
power analysis technique to investigate the length of time to
detection of a hypothetical population decline of 5%/year
with acceptable levels of confidence (Type 1 and Type 2
errors at 0.05). Assuming that the precision of the popula-
tion-size estimate is 11% (which is optimistic even for
large-scale dugong surveys at the given sampling fractions;
Table 1), Marsh and Saalfeld (1989) estimated that 10 an-
nual surveys are required (i.e., 9 years to be able to detect
such a decline with 95% confidence). During this period, a
dugong population declining at 5%/year would have been


reduced to 63% of its size since the time of the first survey.
A preliminary indication of this trend could be obtained
more quickly by increasing the Type 1 and Type 2 error
rates. However, because the consequences of failing to
detect a declining trend are more serious than the conse-
quences of deciding that a declining trend is occurring when
it is not, the Type 2 error rate must be kept low. Even if the
Type 2 error rate were increased to 0.1 and the Type 1 error
rate to 0.15, eight annual surveys (7 years) are required to
detect a declining trend in the given example.


Discussion

Evaluation of the Technique

Results of the fixed-width transect technique are now
used for developing local strategies for dugong conserva-
tion. Density-distribution maps of dugongs are used for the
zoning and management of marine protected areas in north-
ern Australia, especially in the Great Barrier Reef Marine
Park. Distribution maps have also been used to produce
recommendations for the conservation and management of
dugongs in the Arabian region (*Preen et al. 1989). The
distribution of dugongs mirrors the distribution of seagrasses
in all survey areas. Indeed, the pattern of dugong sightings
has proved a reliable basis for designing recent seagrass
surveys in Torres Strait, the northern Great Barrier Reef, and
Shark Bay.
The standardized minimum population-size estimates
have been used in conjunction with a dugong population
model to assess the probable impact of direct anthropogenic
mortality of dugongs in the few cases for which a measure
of that mortality was available. For example, Smith and
Marsh (1990) concluded that the take of Aboriginal commu-
nities in Cape York was well below the sustainable yield.
In Australia, dugongs are being resurveyed at regular
intervals along fixed transects using the techniques outlined


Table 2. Comparison of the population-size estimates obtained from repeated surveys of dugongs (Dugong dugon) with
the fixed-width transect technique in the same area.
Population
Location Survey date estimate S.E. Reference
Western Gulf of August 1984 16,816 2,946 Bayliss and Freeland (1989)
Carpentariaa February 1985 16,846 3,257
Cape Bedford- November 1984 2,899 454 Marsh and Saalfeld (1989)
Cape Melvilleb November 1985 2,542 634
Campbell Point- April 1985 2,172 552 Marsh and Saalfeld (1989)
Hunter Pointb November 1985 1,938 491
Cape Bedford- November 1985 8,100 1,073 Marsh and Saalfeld (1989)
Hunter Pointb November 1990 10,742 1,579 *Marsh et al. (1993)
a Northern Territory, Australia.
b Northern Great Barrier Reef region, Queensland, Australia.






HELENE MARSH 61


o 0.30 /
/
W /
C
) 0.20 -- =0.9327
0 / 9

C 0.10

o o
0.00----------
0.00 0.10 0.20 0.30 0.40

Mean density of dugongs / block 1985

Fig. 2. The estimated density of dugongs (Dugong dugon) in each
block in the northern Great Barrier Reef region in Australia
in 1985 (Marsh and Saalfeld 1989) plotted against the
estimated numbers in the same blocks when the survey was
repeated in November 1990 (Marsh et al. 1993). The line
illustrates equal densities in the two surveys and is not the
fitted regression line represented by the r value.


in this paper. I recommended a resurvey interval of 5 years
to the management agencies. Although the expected small
population-size changes will probably not be detected in less
than a decade, a 10-year interval between surveys could
cause unwarranted delays in the management response if
numbers were declining rapidly. In addition, personnel
changes would make it difficult to guarantee continuity of
the methodology if the survey interval was much longer than
5 years.
The greatest weakness of the technique is its dependence
on the unvalidated assumption that the proportion of du-
gongs on the surface is constant. Data are urgently needed
to examine this assumption and, if it is incorrect, to develop
additional methods of compensating for the variability in the
proportion of dugongs that are not visible to observers.
The method also must be modified so that local changes
in dugong densities can be monitored, a modification that
may be relevant to the needs of surveying manatees in large
bays, lagoons, and estuaries. Theoretically, this goal can be
achieved by increasing the sampling fraction and the fre-
quency of surveys, which have yet to be empirically con-
firmed.

Applicability of the Technique to
Surveys of Manatees

Lefebvre et al. (1995) provided a corresponding review
of techniques and problems with surveys and current meth-
ods of estimation of population sizes of Florida manatees.


Manatees inhabit rivers or coastal and estuarine waters and
seemingly require access to freshwater (Hartman 1979),
whereas dugongs are strictly marine and in some areas feed
tens of kilometers offshore. Thus the spatial dimensions tend
to be more linear in manatee habitats than in dugong habitats.
The spatial design of manatee surveys must reflect these
differences, and parallel transects probably will be useful
only in large bays, estuaries, and lagoons. However, the
distinction between perception and availability bias (Marsh
and Sinclair 1989a) is relevant to manatee surveys, and the
methods developed to overcome these biases in dugong
surveys have potential application for surveys of manatees.


Acknowledgments

Surveys of dugongs in Australia were funded by the
Australian Fisheries Service, the Australian National Parks
and Wildlife Service, the Conservation Commission of the
Northern Territory, the Great Barrier Reef Marine Park
Authority, the Queensland Department of Primary Indus-
tries, and the Western Australian Department of Conserva-
tion and Land Management. B. Ackerman, L. W. Lefebvre,
T. J. O'Shea, T. Smith, and an anonymous referee made
helpful comments on the manuscript.


Cited References2

Anderson, P. K., and A. Birtles. 1978. Behavior and ecology of
the dugong, Dugong dugon (Sirenia): observations in Shoal-
water and Cleveland Bays, Queensland. Australian Wildlife
Research 5:1-23.
Bayliss, P. 1986. Factors affecting aerial surveys of marine fauna,
and their relationship to a survey of dugongs in the coastal
waters of the Northern Territory. Australian Wildlife Re-
search 13:27-37.
Bayliss, P., and W. J. Freeland. 1989. Seasonal distribution and
abundance of dugongs in the Gulf of Carpentaria. Australian
Wildlife Research 16:141-149.
*Brownell, R. L., P. K. Anderson, R. P. Owen, and K. Rails.
1981. The status of dugongs at Palau, an isolated island group.
Pages 11-23 in H. Marsh, editor. The Dugong: Proceedings
of a Seminar/Workshop Held at James Cook University,
8-13 May 1979. Townsville, Australia.
Caughley, G. 1977. Analysis of vertebrate populations. John
Wiley and Sons, London. 234 pp.
Caughley, G., and G. C. Grigg. 1981. Surveys of the distribution
and density of kangaroos in the pastoral zone in South Austra-
lia, and their bearing on the feasibility of aerial survey in large
and remote areas. Australian Wildlife Research 8:1-11.
Chambers, M. R., E. Bani, and B. E. T. Barker-Hudson. 1989.
The status of the dugong in Vanuatu. South-Pacific Regional
Environment Program Topic Review 37. 63 pp.


2 An asterisk denotes unpublished material.






62 INFORMATION AND TECHNOLOGY REPORT 1


Forney, K. A., D. A. Hanan, J. Barlow. 1991. Detecting trends
in harbor porpoise abundance from aerial surveys using
analysis of covariance. Fishery Bulletin 89:367-377.
Gerrodette, T. 1987. A power analysis for detecting trends.
Ecology 68:1364-1372.
Hartman, D. S. 1979. Ecology and behavior of the manatee
(Trichechus manatus) in Florida. Special publication, The
American Society of Mammalogists 5:1-154.
Heinsohn, G. E., A. V. Spain, and P. K. Anderson. 1976. Popula-
tions of dugongs (Mammalia: Sirenia): aerial survey over the
inshore waters of tropical Australia. Biological Conservation
9:21-23.
Hudson, B. E. T. 1980a. Dugongs in Papua New Guinea: west
New Britain. Wildlife in Papua New Guinea 80/15:1-22.
Hudson, B. E. T. 1980b. Dugongs in Papua New Guinea: Manus
Province. Wildlife in Papua New Guinea 80/16.
Hughes, G. R., andR. Oxley-Oxland. 1971. A survey ofdugong
(Dugong dugon) in and around Antonio Enes, Northern
Mozambique. Biological Conservation 3:299-301.
Jolly, G. M. 1969. Sampling methods for aerial census of wildlife
populations. East African Agriculture and Forestry Journal
34:46-49.
Lefebvre, L. W., B. B. Ackerman, K. M. Portier, and K. H.
Pollock. 1995. Aerial survey as a technique for estimating
trends in manatee population size-problems and prospects.
Pages 63-74 in T. J. O'Shea, B. B. Ackerman, and H. F.
Percival, editors. Population biology of the Florida manatee.
National Biological Service Information and Technology
Report 1.
*Ligon, S. H. 1976. Aerial survey of the dugong in Kenya.
FAO/UN Advisory Committee on Marine Resources Re-
search, Scientific Consultation on Marine Mammals, Bergen.
ACMRR/MM/SC/107. 2 pp.
Ligon, S., and B. E. T. Hudson. 1977. Aerial survey of the
dugong, Dugong dugon, in Papua New Guinea. Wildlife in
Papua New Guinea 77/17:1-5.
Marsh, H., P. W. Channells, G. E. Heinsohn, and J. Morrissey.
1982. Analysis of stomach contents of dugongs from Queens-
land. Australian Wildlife Research 9:55-67.
*Marsh, H., D. Kwan, and I. Lawler. 1993. The status of du-
gongs, sea turtles and dolphins in the northern Great Barrier
Reef region. Unpublished report to the Great Barrier Reef
Marine Park Authority.
Marsh, H., R. I. P. Prince, W. K. Saalfeld, and R. Shepherd. 1994.
The distribution and abundance of the dugong in Shark Bay.
Wildlife Research. 21:149-161.
Marsh, H., and W. K. Saalfeld. 1988. The distribution and
abundance of dugongs in the Torres Strait region. Report to
the Australian Fisheries Service, the Great Barrier Reef Ma-
rine Park Authority and the Fisheries Management Branch of
the Queensland Department of Primary Industries. 25 pp.
Marsh, H., and W. K. Saalfeld. 1989. The distribution and
abundance of dugongs in the northern Great Barrier Reef
Marine Park. Australian Wildlife Research 16:429-440.


Marsh, H., and W. K. Saalfeld. 1990. The distribution and
abundance of dugongs in the Great Barrier Reef region south
of Cape Bedford. Australian Wildlife Research 17:511-524.
*Marsh, H., and W. K. Saalfeld. 1991. The status of the dugong
in Torres Strait. Pages 187-194 in D. Lawrence and T. Cans-
field-Smith, editors. Sustainable development for traditional
inhabitants of the Torres Strait region. Great Barrier Reef
Marine Park Authority Workshop Series 10, Townsville, Aus-
tralia.
*Marsh, H., W. K. Saalfeld, and A. R. Preen. 1990. The distri-
bution and abundance of dugongs in southern Queensland.
Final report to the Queensland Department of Primary Indus-
tries, Fisheries Management Branch. 14 pp.
*Marsh, H., W. K. Saalfeld, and R. I. P. Prince. 1991. The status
of the dugong in Shark Bay and Exmouth Gulf-Ningaloo.
Report to the Western Australian Department of Conserva-
tion and Land Management. 17 pp.
Marsh, H., and D. F. Sinclair. 1989a. Correcting for visibility bias
in strip transect aerial surveys of aquatic fauna. Journal of
Wildlife Management 53:1017-1024.
Marsh, H., and D. F. Sinclair. 1989b. An experimental evaluation
of dugong and sea turtle aerial survey techniques. Australian
Wildlife Research 16:639-650.
Nishiwaki, M., and H. Marsh. 1985. The dugong. Pages 1-31
in S. H. Ridgway and R. J. Harrison, editors. Handbook of
marine mammals, volume 3. Academic Press, London.
Norton-Griffiths, M. N. 1978. Counting animals. African Wild-
life Leadership Federation Handbook 1. 139 pp.
*Packard, J. M. 1985. Development of manatee aerial survey
techniques. Manatee Population Research Report 7. Techni-
cal Report 8-7. Florida Cooperative Fish and Wildlife
Research Unit. University of Florida, Gainesville. 68 pp.
*Preen, A. R. 1989. Dugongs: technical report 10, volume 1.
Meteorology and Environmental Protection Administration,
Saudi Arabia. 200 pp.
*Preen, A. R., H. Marsh, and G. E. Heinsohn. 1989. Recommen-
dations for the conservation of the Arus Al Bahr (Dugong
dugon) in the Kingdom of Saudi Arabia and neighboring
countries, Dugongs: technical report 10, volume 2. Meteorol-
ogy and Environmental Protection Administration, Saudi Ara-
bia. 43 pp.
Rathbun, G. B., R. L. Brownell, K. Rails, and J. Engbring. 1988.
Status of dugongs in waters around Palau. Marine Mammal
Science 4:265-270.
*Salm, R., R. C. Petocz, and T. Socharlono. 1982. Survey of
coastal areas in Irian Jaya (1982). Report for the Directorate
of Nature Conservation, Directorate General of Forestry,
Republic of Indonesia.
Smith, A. J., and H. Marsh. 1990. Management of the traditional
hunting of dugongs [Dugong dugon (Muller 1776)] in the
Great Barrier Reef Marine Park. Environmental Manage-
ment.






LYNN W. LEFEBVRE ET AL. 63


Aerial Survey as a Technique for Estimating Trends in Manatee
Population Size-Problems and Prospects



by


Lynn W. Lefebvre1

U.S. Fish and Wildlife Service
National Ecology Research Center
Sirenia Project
412 N.E. 16th Avenue, Room 250
Gainesville, Florida 32601


Bruce B. Ackerman

Florida Department of Environmental Protection
Florida Marine Research Institute
100 8th Avenue S.E.
St. Petersburg, Florida 33701


Kenneth M. Portier

University of Florida
Department of Statistics
401 Rolfs Hall
Gainesville, Florida 32611


and


Kenneth H. Pollock

North Carolina State University
Department of Statistics
Box 8203
Raleigh, North Carolina 27615


1 Now with the National Biological Service, same address.






64 INFORMATION AND TECHNOLOGY REPORT 1


Abstract. Aerial surveys are the primary means of obtaining information on Florida manatee
(Trichechus manatus latirostris) distribution and abundance. Results have been used to derive
minimum estimates of population size; these are composite counts with unknown variance which
cannot be statistically compared across surveys or years. The need to establish population estimation
techniques with aerial surveys or other methods and to control or correct for the many variables that
influence counts was identified as a high priority in the Florida Manatee Recovery Plan.
Aerial-survey-based estimates of manatee abundance are biased because of visibility and sampling
problems. Visibility bias is the proportion of unobserved manatees in the survey area. Two types of
visibility bias were defined by Marsh and Sinclair: perception bias (proportion of the target species
present and visible in the survey area but not observed) and availability bias (proportion of the target
species present but not visible). A third type of bias in surveys of manatees in aggregation sites in
winter is absence bias (proportion not present in the survey area and for which the count is not
corrected). Several environmental, behavioral, and procedural factors are related to the degree of
survey bias. The current approach to compensating for variability in these factors is variable survey
effort. The greater the difficulty is in counting manatees because of water turbidity, group size, and
other factors, the longer the aircraft circles a particular site. Further investigation is needed of the
relations among survey effort, environmental variables, manatee behavior, and counts. Research is
needed to improve survey design and technique and to reduce survey bias.
Key words: Aerial survey, availability, bias, population estimate, Florida manatees, trend, visibility.


The Florida manatee (Trichechus manatus latirostris)
is a subspecies of the West Indian manatee (T. manatus),
which also occurs in coastal waters and rivers of the Gulf
coast of Mexico, the Caribbean coast of Central America,
northern South America, and the Greater Antilles. The two
other species of manatees are the Amazonian manatee
(T. inunguis) and the West African manatee (T. senegalen-
sis). Manatees, like their closest living relative, the dugong
(Dugong dugon), are tropical to subtropical in distribution
and are classified as threatened or endangered. Florida
manatees are near the northern limit of the species' range,
yet they probably represent one of the most abundant
populations of the West Indian manatee (Lefebvre et al.
1989). Aerial surveys are the primary means of obtaining
information on manatee distribution and relative abun-
dance in Florida (Ackerman 1995). Aerial survey results
have been used to derive minimum estimates of the mana-
tee population size in the state. These are actually compos-
ite counts rather than estimates and are biased by a variety
of factors that prevent complete enumeration or unbiased
extrapolation of manatees in survey areas. The composite
counts also have an unknown variance, thus, the popula-
tion-size trend (upward or downward) cannot be deter-
mined from survey results. This is hardly an optimal
situation. Managers must detect downward trends before
they become irreversible.
The two ultimate goals in aerial-survey research are the
development of indices of abundance that accurately reflect
trends and the development of population-size estimation
procedures that allow comparisons of manatee numbers at
different sites within and across years. The objectives of
managers who use the results of aerial surveys are the
determination of trends in manatee distribution over time,


the estimation of total manatee abundance, and the determi-
nation of trends in manatee abundance over time, particu-
larly the detection of any decline in manatee abundance.
Researchers and managers generally agree that current aerial
survey methods are useful for meeting the first objective but
not for the second and third objectives.
The U.S. Fish and Wildlife Service and the U.S. Marine
Mammal Commission sponsored research during 1982-85
to improve manatee population indices from aerial surveys
(*2Packard and Mulholland 1983; *Packard 1985; Packard
et al. 1985, 1986, 1989). This research revealed many
sources of bias in the surveys and indicated that methods to
reduce bias in the visibility of manatees should be explored.
The need to establish unbiased population-size estimation
techniques with aerial surveys or with other methods was
identified as a high priority in the revised Florida Manatee
Recovery Plan (*U.S. Fish and Wildlife Service 1989). A
Manatee Aerial Survey Technical Meeting was held in
December 1987 to review the current state of the art in aerial
surveys of manatees and to determine whether models of
manatee visibility can be developed, tested, and used to
correct for various sources of survey bias. We here report
the findings of the 1987 meeting and provide a report of
research since that meeting, including insights and recom-
mendations provided by the participants of the February
1992 Technical Workshop on Manatee Population Biology.
Our intended audience is technical and non-technical per-
sonnel, including managers, interested non-scientists, and
researchers. Thus we do not attempt to use detailed statistical
terms. Readers who wish to learn more about the different
types of survey techniques described here should refer to the


2 An asterisk denotes unpublished material.






LYNN W. LEFEBVRE ET AL. 65


cited literature. Results of recent and ongoing aerial surveys
of Florida manatees are described by Ackerman (1995).


Types of Aerial Surveys of
Manatees

Packard (* 1985) described three aerial survey techniques
used to assess manatee distribution and abundance: the
extended-area, intensive-search, and unit-recount tech-
niques. The extended-area technique is used to determine
manatee distribution in areas surrounding a winter refuge or
in broad areas that are distant from a winter refuge and is
usually conducted throughout the year to detect changes in
seasonal and local distribution over several years. Shane's
(1983) surveys of the Indian and Banana rivers are examples
of extended-area techniques. Intensive-search and unit-re-
count techniques are used in winter refuges, when manatees
are highly aggregated after the passage of cold fronts.

Intensive-search Technique

Irvine (1982) described the intensive-search technique.
Surveyors strive for comprehensive coverage of probable
manatee habitat in a well-defined region, usually an aggre-
gation site in winter. Manatees in Florida must move to
natural or artificial warm-water sites during winter, and
biologists have taken advantage of this situation by timing
aerial or ground surveys to coincide with periods when
manatees are probably aggregated at warm-water sites. The
use of warm-water effluents is triggered by drops in ambient
water temperature below about 200 C, and manatees may
move hundreds of kilometers to reach these refuges (O'Shea
1988; Reid et al. 1991).
Intensive-search techniques follow a general survey
route until manatees are sighted; groups are then circled until
repetitive counts are consistent. Survey effort may differ
among survey dates, so that variation from differences in
effort cannot be distinguished from variation in actual abun-
dance (*Packard 1985). Even when surveys are replicated,
variation in replicates has been ignored in favor of using the
highest count of the series. Packard (*1985) called the
obtained numbers a maximum-count index. No estimate of
error is associated with a maximum count.
Since 1977, the Florida Power and Light Company has
sponsored intensive-search surveys in winter at five of their
power plants and at several other warm-water aggregation
sites (Fig. 1; *Rose and McCutcheon 1980; *Raymond
1981; *McGehee 1982; Reynolds and Wilcox 1985, 1986;
Reynolds *1989, 1990, *1991). Biologists with the Sirenia
Project and with the Chassahowitzka National Wildlife Ref-
uge have conducted aerial surveys in winter of the aggre-
gated manatees in Kings Bay and in the Crystal River since
1978 and in the Homosassa River since 1983 (Fig. 1).


Unit-recount Technique

The unit-recount technique was the first attempt to design
aerial surveys with standardization, replication, and stratifi-
cation. The study area is divided into survey units and
stratified so that survey effort is consistent among units in
each stratum. The primary stratum includes warmer waters
(e.g., thermal plumes from power plants) where manatees
aggregate during cold periods and receives the greatest
survey effort. Manatees are counted five times in primary
units and at least once in secondary and tertiary units during
each survey (*Packard 1985).
Packard (*1985) and Packard et al. (1986) developed
the unit-recount technique to monitor annual trends in
manatee abundance at specific winter refuges. Packard
(* 1985) envisioned the development of standardized unit-
recount procedures at six major aggregation sites: Crystal
and Homosassa rivers, Tampa Bay, and Fort Myers on the
Gulf Coast of Florida; Indian River, Riviera Beach, and
Port Everglades on the Atlantic Coast (Fig. 1).
The unit-recount technique was developed in Fort Myers
during 1984-85 (Packard et al. 1986), and one of the indices
obtained with this method, the sight-resight index, was used
again in Fort Myers during the following winter (Packard
et al. 1989). However, the sight-resight index has not been
used by other survey biologists, and the unit-recount tech-
nique has only been tested at one other winter refuge, the
Crystal and Homosassa rivers (Lefebvre and Kochman
1991).

Statewide Technique

The first statewide attempt to count manatees by aerial
surveys was made in summer 1973 during a 6-week period
and resulted in an estimate of 750-850 manatees (Table 1;
*Hartman 1974). Hartman was also the first biologist to
attempt correction of raw counts based on his best guess that
90% of the manatees in clear water, 50% of those in moder-
ately turbid water, and 10% of those in highly turbid water
are visible from the air. He did not provide details of how
these correction factors were obtained. Irvine and Camp-
bell's 1976 count in winter (during a 6-day period) of 738
manatees is uncorrected (Table 1), but these researchers
assumed that they undercounted and guessed that the mana-
tee population size was between 800 and 1,000 (Irvine and
Campbell 1978). The statewide minimum estimate was
updated in 1985 to 1200 based on a composite of surveys of
manatees in aggregation sites in winter and other data
(O'Shea 1988).
The statewide synoptic technique, coordinated by the
Florida Marine Research Institute, Florida Department of
Environmental Protection, is a marathon event that com-
bines the methods of the intensive-search and extended-area
techniques. Three synoptic surveys have been conducted






66 INFORMATION AND TECHNOLOGY REPORT 1


Sebastian


Florida Manatee's Tampa
Winter Range

] High use area
* Warm water discharges
O Springs


Matlacha I


0 20 40 60 80 100
Kilometers


Fig. 1. Important warm water aggregation sites and winter range of the Florida manatee (Trichechus manatus latirostris). Aerial
surveys are conducted every year over the Crystal and Homosassa rivers and over several coastal power plants. Counts on the
ground of manatees at Blue Spring are obtained each winter.


Table 1. Statewide minimum estimates of Florida manatee
(Trichechus manatus latirostris) abundance based on
aerial surveys conducted between 1973 and 1992.
Sources of data are shown on the right.

Minimum
estimate Source

750-850 Hartman (1974)
738 Irvine and Campbell (1978)
800-1,000 *Brownell et al. (1981; no new data)
1,200 FP&La, FDNRb, and USFWSC surveys (1985)
1,268 FDNR statewide survey, 23-24 January 1991
1,465 FDNR statewide survey, 17-18 February 1991
1,856 FDNR statewide survey, 17-18 January 1992
a FP&L = Florida Power and Light Company.
b FDNR = Florida Department of Natural Resources.
c USFWS = U.S. Fish and Wildlife Service.


through January 1992, two in 1991 and one in 1992. Acker-
man (1995) described the objectives and methods of agen-
cies who made the surveys, and provided the results.
Although there is considerable doubt that the counts
(Table 1) represent the actual population-size trend, mana-
tee biologists probably agree with Eberhardt's (* 1982) con-
clusion that a substantial fraction of the population is repre-
sented in maximum aerial counts of manatees at power
plants and warm-water springs. The aggregation at Blue
Spring on the St. Johns River (Fig. 1) cannot be surveyed by
air because of overhanging branches; however, the rela-
tively small number of manatees that use this site (<90) and
the high proportion of distinctly scarred individuals
make direct enumeration by ground observation possible.
Although surveys have also been conducted at many sites in
Florida in non-winter seasons (Shane 1983; Powell and






LYNN W. LEFEBVRE ET AL. 67


Rathbun 1984; Kinnaird 1985), the relative effectiveness of
aerial surveys in winter or in other seasons for estimating
abundance or detecting trends is unknown. At the 1987
Manatee Aerial Survey Technical Meeting, the participants
assumed that aerial surveys at aggregation sites in winter
would continue to be the primary focus of research, although
much of what the participants discussed could also apply to
other sites and seasons.
Eberhardt (*1982) noted that an improved size estimate
of manatee populations would lead to some confusion and
controversy because reported abundance could increase
over the years, whereas actual numbers may be decreasing.
For example, although recent minimum estimates of mana-
tee abundance were not improved statistically, survey cov-
erage of important manatee wintering sites undoubtedly
improved, leading to an apparent increase in counts. Further-
more, even if survey methods are improved so that more
accurate, statistically valid size estimates of manatee popu-
lations can be made, a statistical detection of a decline may
never be possible because of the relatively small population
size in a survey region (Taylor and Gerrodette 1993). Esti-
mates with a precision of 12% at best were obtained with the
fixed-width transect technique in surveys of dugongs; for
example, detection of a 5%/year decline in dugong abun-
dance to a level where only 63% of the population remained
would require almost a decade (Marsh 1995).


Survey Bias


Types of Bias

The problems with aerial-survey sampling of manatee
population abundance are caused by biases in count-based
estimates. Bias is the difference between the estimated value
and the true value of a parameter (in this case, population
size). Unbiased estimators of population size do not exist in
most situations (not even of humans). Unbiased estimates


are difficult to obtain because of problems with unequal
visibility (also called observability) and sampling. A count-
based statistic, C, is related to the true population size, N, by
a proportionality constant, 3:
A
N=C/P (1)

where p is the proportion of the population represented by
the count or the probability that a given manatee will be
seen and counted (Lancia et al. 1994).
The closer the surveyor can get physically to the individu-
als that are counted, the smaller the bias in the estimator.
Thus, aerial sampling, which keeps the observer at a distance
from individuals in the population, tends to result in biased
estimators. The fact that manatees are not always at the
surface or visible when submerged obviously compounds
the problem. Two types of visibility bias in dugong and sea
turtle surveys were defined by Marsh and Sinclair (1989;
Table 2). Perception bias occurs when a manatee group (of
one or more individuals) is present and visible in the survey
area but is missed by the observerss. Availability bias is the
proportion of manatees that is present but not visible in the
survey area. Perception bias is the least problematic source
of bias; it can be tested and corrected by using at least 2
observers/survey (Marsh and Sinclair 1989). Availability
bias depends on location, environmental conditions, and
manatee behavior and is the source of bias that is most
difficult to control.
Another major source of bias in many wildlife surveys is
the necessity for less-than-complete coverage of the area of
interest (Lancia et al. 1994). Typically, sample areas are
randomly selected, and the wildlife in a fraction, a, of the
total area of interest is censused. This approach leads to the
estimator:
A A
N=N'/a (2)
A
where a is the proportion of the total covered area, N' is
the count for the covered area, and k is the estimate for the


Table 2. Types of bias that affect aerial surveys of Florida manatees (Trichechus manatus latirostris).
Visibility biasa (= sightability or observability bias): Proportion of the target species within the survey area that is not observed
(Gasaway et al. 1985; Packard et al. 1985; Pollock and Kendall 1987; Samuel et al. 1987; Lancia et al. 1994).
Perception bias (= observer bias): Proportion of groups of the target species that is visible in the transect but is not
observed (Marsh and Sinclair 1989).
Availability bias: Proportion of groups of the target species that is present on the transect but not visible (Marsh and
Sinclair 1989).
Sampling bias: Proportion of entire area of interest in which the survey actually occurs (Lancia et al. 1994).
Absence bias: Proportion of the target species that is not present in the survey area and for which the count is not corrected
(this paper).
a Authors who used these terms are listed after each definition.






68 INFORMATION AND TECHNOLOGY REPORT 1


total area. Although for many species the sampling fraction
can be determined with reasonable accuracy, manatees pre-
sent a greater challenge because of the many types of water
bodies they inhabit-from residential canals to large bays-
and the strong seasonal pattern of their movements and
habitat requirements. Most manatee biologists tend to as-
sume that their counts represent complete coverage of the
area in which the survey takes place and do not attempt to
estimate a sampling fraction. However, even the statewide
synoptic manatee surveys do not cover all potential manatee
habitat.
The statewide synoptic survey is conducted immedi-
ately after cold fronts to minimize bias caused by ab-
sence of manatees from the sample areas. Absence bias
is the proportion of manatees that are not present in the
survey area at the time of the survey. Manatees may be
in a different warm-water aggregation site or somewhere
between different sites. Absence bias affects the ex-
pected value of the estimate and is more strongly related
to temporal variability than visibility bias. It can be
controlled to some extent by timing surveys to coincide
with periods of high manatee aggregation and with
knowledge of manatee movements. Absence bias is of
considerable concern in surveys of manatees in aggrega-
tion sites in winter. Because the severity and pattern of
winter cold fronts differ across years, the manatee distri-
butions among and between aggregation sites may differ
(*Packard 1985). To obtain a statewide estimate, surveys
for manatees must take place at all major aggregation
sites in a relatively short period to avoid double-counting
(counting the same manatees twice in different sites) and
when most of the population is clumped at aggregation
sites, not dispersed between sites.

Factors That Cause Bias

Several factors are known to relate to the effective-
ness of aerial counts of manatees. These factors are in
one of three categories: environmental, behavioral, and
survey-related. Most of the environmental and behav-
ioral factors influence all three sources of bias (percep-
tion, availability, and absence) in manatee counts, mak-
ing it difficult to sort out what, where, and when bias
correction is needed (Table 3). Most survey-related fac-
tors, such as height and speed of the plane, time of day,
and survey dates, can be controlled by the sampling
protocol (Table 3). Some of the environmental factors,
such as water temperature, cloud cover, and surface
condition can be controlled to some extent by appropri-
ate timing of the survey; however, few of the factors
related to manatee behavior can be controlled through
survey methods (Table 3). All environmental and behav-
ioral factors of visibility bias can vary considerably
across survey sites, and the development of visibility


models for manatees in each major aggregation site is
essential.
The current method of compensating for variability in
environmental conditions and manatee behavior is variable
survey effort: the greater the difficulty in counting manatees
because of water turbidity or manatee submergence, the
longer the pilot is directed by the observer to circle a
particular site. This was described as the variable effort
recount (VER) method by Lefebvre and Kochman (1991).
Because this procedure has not been standardized, differ-
ences may be considerable among observers in how long
they circle individuals or groups until they are satisfied that
they obtained the best (highest) possible count. Moreover,
survey effort is not allocated objectively and consistently
across surveys; observers decide spontaneously how to al-
locate effort in each individual survey. Thus, the estimation
of the sampling fraction is difficult and differs across sur-
veys. A standardized procedure by which observers evaluate
environmental conditions also does not exist, although many
observers record a subjective ranking of sea state (Beaufort
scale), cloud cover, glare, water clarity, air temperature,
wind direction, and wind speed at the start of each survey
(Ackerman 1995).

Control of Bias

The problem of bias can be solved with three approaches.
The best is the development of a sampling design and
estimation procedure that results in unbiased estimates. One
attempts to identify sources of bias that can be controlled in
the design of the sampling scheme, for example, by replica-
tion, stratification, and proper selection of when and where
to sample (*Packard 1985). Standardization of survey pro-
cedures primarily reduces perception bias (Table 3). A
second approach is the estimation of the amount of bias in
the estimate and application of correction factors, essentially
by combining equations (1) and (2):
A
N=C/ap (3)
so that bias from problems with visibility and sampling are
corrected (Lancia et al. 1994). The difficulties with this
approach for manatees are that the degree of bias changes
with conditions, potentially requiring a large number of
correction factors, and that complete enumeration is
rarely, if ever, possible.
A third approach is the calibration of a population
index by using ratio estimation or double sampling
(Eberhardt and Simmons 1987). This approach is based
on the assumption that a relation between an index value
and the true population density is known or can be
estimated. The index value is usually obtained with
lower sampling intensity and therefore is of lower accu-
racy than the estimate of the true population density.







LYNN W. LEFEBVRE ET AL. 69


Table 3. Factors of visibility bias in aerial surveys of Florida manatees (Trichechus manatus latirostris).
Controllable
Perception Availability Absence with sampling
Factors bias bias bias protocol
Environmental factors
Water turbidity X X
Water depth X X
Distribution of food resources X X
Background of bottom X X
Water temperature X X X
Cloud cover/light X X X
Surface conditions X X X
Tides X X X
Behavioral factors
Group size X X
Propensity to aggregate X X X
Surfacing/diving patterns X X
Diurnal activities X X X
Local spatial movements X X
Seasonal spatial movements X X
Size (age, sex) X X X
Response to disturbance X X X
Survey related factors
Number of observers X
Experience of observers/pilot X
Height of plane X
Speed of plane X
Time of day X X
Length of flight X
Survey dates X X
Number of surveys X X
Flight path X X X


Manatee population sizes are generally believed to be
underestimated from aerial-survey counts (*Hartman
1974; *Eberhardt 1982; *Packard 1985), largely because
of availability bias. As a result, survey effort has been
geared to maximize counts rather than to standardize
counts. However, maximum counts of unknown accu-
racy are inferior to an index that accurately portrays the
population-size trend. More attention should be paid to
the identification of optimal survey conditions that mini-
mize visibility bias and to the development of survey
protocols that incorporate information on environmental
conditions. The geographic (aggregation sites versus
non-aggregation sites) and temporal (winter versus
warm seasons) stratification of survey effort should be
designed to optimize site-specific conditions.

Perception Bias

Counts from two independent observers can be used in a
Petersen estimate (Seber 1982; Pollock and Kendall 1987;


Marsh and Sinclair 1989) to determine observer variation
and to estimate the population size of individuals with
non-zero detection probabilities. The differences among
observers in the designation of calves and manatee activity
should also be established.
Tests of different types of flight paths should be con-
ducted. Results obtained by the variable effort recount
method could be compared with those obtained in simulta-
neous surveys in which the number and location of circles
are fixed and with surveys with straight flight lines (no
circling). Straight flight lines provide information on the
error in sighting of groups, whereas tests of circling effort
provide information on the error of counting individuals. A
stopping rule could be developed, so that observers can
standardize their circling efforts. Variance component mod-
els can be developed for repeated, fixed-effort surveys to
determine the most important factors of variability in counts,
such as survey dates, cold fronts, and survey units (Packard
et al. 1989, Lefebvre and Kochman 1991).






70 INFORMATION AND TECHNOLOGY REPORT 1


Availability Bias
The effect of environmental and behavioral factors on
counts of manatees can be tested by double-sampling strate-
gies, i.e., from counts on the ground or from a boat, radio-
telemetry, and counts from airships conducted simultane-
ously with routine aerial surveys (Ackerman 1995). The
detection probability of manatees can be estimated by the
ratio of aerial survey counts to counts on the ground, counts
from a boat, or counts from an airship or by the proportion
of radio-tagged manatees seen from the air to all tagged
manatees in the survey area. Problems with the use of
radio-tagged manatees for double-sampling experiments are
that the radio tags are variably visible, which differs from
the tagged manatee's detection probability (Packard et al.
1989; Ackerman 1995) and that an adequate sample of
tagged manatees is probable only in winter when the distri-
bution of the manatees is highly clumped. This results in
uncertainty of whether the tagged manatee was actually
sighted; one may simply know that the tag belongs to one of
the members of a group. In a strict statistical sense, estima-
tion of availability bias is dependent on the ability to estimate
the true population size by an alternative procedure. If the
estimate from the alternative procedure is consistently bi-
ased with respect to the survey estimate or count, it still may
provide information about the population-size trend.
The use of airships (Ackerman 1995) or some other
stationary observation platform to obtain close-to-exact
counts of manatees may be useful for developing models of
detection probabilities. This approach requires measuring
(or ranking) environmental and behavioral factors at various
sites (or at the same site on different dates) and determining
proportions of manatees not seen during routine surveys.
Resultant models of detection probability based on the most
influential factors may be constructed by regression tech-
niques (Gasaway et al. 1985; Samuel et al. 1987). Environ-
mental (e.g., depth, turbidity, and temperature) and behav-
ioral (e.g., resting, feeding) factors are treated as covariates
that are tied to strata or sighted groups. A theoretical exam-
ple of the relation among survey effort, water clarity, and
percentage of present manatees that are seen (Fig. 2) illus-
trates how information about an environmental variable
(water clarity) allows correction of manatee counts if effort
was measured or standardized. Correction factors may also
be based on direct observation of tagged individuals from
an airship to determine the proportion of timed intervals that
individuals are visible, given different environmental and
behavioral factors. Other new technologies should be ex-
plored, such as time-depth recorders and infrared multi-
spectral photography, that may be useful for obtaining finer
measurements of behavioral and environmental factors.
Although Packard et al. (1986) did not specifically iden-
tify availability bias as a type of visibility bias, their sight-
resight index was intended for the correction of short-term


U-
a) Clear water
a 100%-
a) Intermediate

a)
Turbid water
cE 50%-
E

-0
E

Effort (e.g., number of passes or
total survey time per survey unit)
Fig. 2. Theoretical relation of aerial survey effort with counts of
Florida manatees (Trichechus manatus latirostris) in clear,
intermediately clear, and turbid water. Establishing such
relations at important aggregation sites or in different survey
strata may allow a determination of minimum effort for
obtaining accurate counts or of correction factors if
conditions do not permit accurate counts.


changes in visibility that result from the appearance and
disappearance of manatees because of obstructions to view-
ing (what we define as availability bias). Lefebvre and
Kochman (1991) rejected use of the sight-resight index
(based on capture-recapture methods) for this purpose be-
cause of problems with the accurate re-identification of
individual manatees in successive passes over a survey unit
and because manatees in aggregation sites tend to be cap-
tured (resighted) in groups (i.e., their capture probabilities
are not independent).

Absence Bias

To quantify absence bias, regions should be identified
that encompass aggregations in winter with little short-
term probability of exchange of individuals with other
such regions. On the Atlantic Coast, manatees aggregate
in three major areas or regions during cold periods in
winter: Blue Spring, the upper Indian River, and Riviera
Beach to Port Everglades. On the Gulf Coast of Florida,
four regions can be distinguished: Crystal River, Tampa
Bay, Fort Myers, and Naples to the Everglades. During
cold periods in winter, most of Florida's manatee popu-
lation may reasonably be expected to be in these regions
and absence bias is therefore minimized. (A relatively
small number of manatees use industrial effluents in
northeastern Florida and in southeastern Georgia. These
regions could be added to the statewide survey, depend-
ing on the survey objectives and funding in a given year.)
The survey regions should be stratified so that survey
effort is proportional to anticipated or known manatee
densities and distribution. Survey methods can differ






LYNN W. LEFEBVRE ET AL. 71


among strata, e.g., multiple circling in small units in
high-density strata and single passes or transects in larger
units of low-density strata. The sampling fraction in each
stratum should be determined.
Packard and Mulholland (*1983) were optimistic that
an index of manatee abundance in winter aggregation sites
could be statistically compared among years by an analysis
of covariance or by some other regression technique. How-
ever, they concluded that survey data available to them at
the time were not suitable for such a comparison. Also,
they were not able to identify a set of conditions that would
predict optimal conditions for obtaining maximum counts
in each survey site. More recent count data from surveys
by the Florida Power and Light Company were analyzed
with additional temperature covariates to determine popu-
lation-size trends after correction for differences in winter
severity across years (Garrott et al. 1995).
The potential of obtaining consistent, relatively unbi-
ased estimates of abundance from warm-season surveys
should be investigated (Ackerman 1995). Weather and


water conditions may be considerably better in some sur-
vey regions during the non-winter months, and problems
with counting large groups would be avoided. A protocol
should be developed for surveys in non-aggregation areas,
so that trends in both types of season and regions can be
compared over years. However, such surveys should not
be attempted unless statistical power analysis suggests that
population sizes and precision of the estimates are suffi-
cient to determine trends over a reasonable time period
(Gerrodette 1987; Taylor and Gerrodette 1993).



Recommendations

The interaction of numerous factors affects our ability
to accurately census or estimate manatee populations
with data from aerial surveys (Fig. 3). To determine
manatee trends in abundance and, ultimately, to estimate
total manatee abundance, several strategies should be
employed: development of a protocol for a statewide


Factors Affecting Manatee Population Estimation


Solutions:
1. Sampling design modifications standard protocol, stratification, indirect incorporation of environmental conditions, more replication.
2. Use correction factor covariate adjustment, direct incorporation of environmental conditions.
3. Combine multiple estimates double sampling, estimation by consensus.


Fig. 3. Interrelation of factors of estimation of population sizes of Florida manatees (Trichechus manatus latirostris) based on aerial
surveys.






72 INFORMATION AND TECHNOLOGY REPORT 1


survey of aggregation sites, development of a sampling
protocol for surveys during warm seasons, and develop-
ment of correction factors for biased counts.
Because part of a protocol for a statewide survey is
dependent on the results of experiments on bias, these
experiments should be initiated as soon as possible.
Depending on the results of investigations into the ef-
fects of covariates such as water temperature, cloud
cover, or glare on counts, some covariates may be re-
moved from the survey; a tolerable range of survey
conditions could be established. Availability bias related
to manatee behavior and environmental conditions must
be measured by direct observation from appropriate plat-
forms or in experiments with double-sampling, so that
correction factors can be developed. Perception bias can
be determined by using two independent observers. Ab-
sence bias can be controlled by identifying regions in
which manatees aggregate during cold periods in winter,
and stratifying these regions so that survey effort is
proportional to manatee density and distribution in the
survey areas. After experiments are conducted to deter-
mine the best census conditions and other, less-control-
lable factors are modeled to determine their influence on
visibility bias, the final protocol can be developed for a
statewide survey of aggregation sites. The final protocol
must be clearly defined to allow reasonable comparisons
across areas and years and must be easily repeatable from
year to year. Alternatively, detection probabilities can be
estimated at specific aggregation sites every year by
developing sightability models for each site.
A protocol should also be developed for surveys in warm
seasons, to provide additional information on the manatee
population-size trend. In the early stages of protocol devel-
opment, resulting estimates should be analyzed to determine
whether trends can be detected, and, if so, the number of
surveys per year and total number of years required to detect


trends with the degree of confidence required to permit
timely response by managers.
A multi-agency, aerial-survey working group should
be established to ensure that research needs of survey
design, survey technique, and correction for survey bias
are logically and efficiently addressed (Table 4). This
group should meet periodically to review the progress of
aerial-survey research and must keep abreast of progress
in other areas of manatee population research, particu-
larly mortality patterns (Ackerman et al. 1995) and the
use of capture-recapture data from recognizable mana-
tees to estimate survival rates (O'Shea and Langtimm
1995). The use of survival rates to establish trends and
the potential development of population-size estimation
procedures with capture-recapture data may ultimately
replace the aerial surveys for all but manatee distribution
information.



Acknowledgments

We are grateful to T. J. O'Shea for encouragement
and for providing this opportunity to review the devel-
opment and current status of manatee aerial-survey
methods. We thank P. M. Rose and the Florida Depart-
ment of Natural Resources for cooperation and support
of the 1987 Manatee Aerial Survey Technical Meeting.
Participants in the 1987 Manatee Aerial Survey Techni-
cal Meeting and the 1992 Technical Workshop on Mana-
tee Population Biology contributed many of the ideas
presented in this paper: D. P. DeMaster, J. D. Nichols,
M. A. O'Connell, J. E. Reynolds, III, M. D. Samuel, and
B. L. Weigle. I. E. Beeler and R. Muller also provided
useful advice. J. Vorhees skillfully coordinated the 1987
Manatee Aerial Survey Technical Meeting. H. Marsh
reviewed an early draft of the manuscript and made many


Table 4. Areas of survey design, technique, and bias correction that must be further researched to improve aerial surveys
of Florida manatees (Trichechus manatus latirostris). Establishment of a multi-agency aerial survey group is an
important first step of setting research priorities, assigning responsibilities for tasks to organizations, and evaluating
progress.
Survey design Survey technique Survey bias
Stratification by habitat type Tandem observers Test assumption that absence
(bays, lagoons, rivers, canals, Recorders bias can be controlled by timing
complex coastlines) GPS of surveys
Stratification by habitat use Photography Double sampling experiments
(primary and secondary units) Proportion of calves (radio tags, airships)
Replication (temporal and Environmental data Correction factors
spatial) Models
Total counts versus sampling Determine sampling fraction
(aggregation site versus
non-aggregation site)






LYNN W. LEFEBVRE ET AL. 73


helpful suggestions. J. P. Reid provided helpful informa-
tion on manatee distribution in winter.


Cited References3

Ackerman, B. B. 1995. Aerial surveys of manatees: a summary
and progress report. Pages 13-33 in T. J. O'Shea, B. B.
Ackerman, and H. F. Percival, editors. Population biology of
the Florida manatee. National Biological Service Information
and Technology Report 1.
Ackerman, B. B., S. D. Wright, R. K. Bonde, D. K. Odell, and
D. J. Banowetz. 1995. Trends and patterns in mortality of
manatees in Florida, 1974-1992. Pages 223-258 in T. J.
O'Shea, B. B. Ackerman, and H. F. Percival, editors. Popu-
lation biology of the Florida manatee. National Biological
Service Information and Technology Report 1.
*Brownell, R. L., Jr., K. Rails, and R. Reeves. 1981. Report of
the West Indian manatee workshop. Pages 3-16 in *R. L.
Brownell, Jr. and K. Rails, editors. The West Indian manatee
in Florida. Proceedings of a workshop held in Orlando, Fla.,
27-29 March 1978. Florida Department of Natural Re-
sources, Tallahassee. 154 pp.
*Eberhardt, L. L. 1982. Censusing manatees. Manatee Popula-
tion Research Report Number 1, Technical Report 81, Florida
Cooperative Fish and Wildlife Research Unit, University of
Florida, Gainesville. 18 pp.
Eberhardt, L. L., and M. A. Simmons. 1987. Calibrating popula-
tion indices by double sampling. Journal of Wildlife Manage-
ment 51:665-675.
Garrott, R. A., B. B. Ackerman, J. R. Cary, D. M. Heisey, J. E.
Reynolds, III, and J. R. Wilcox. 1995. Assessment of trends
in sizes of mantaee populations at several Florida aggrega-
tion sites. Pages 34-55 in T. J. O'Shea, B. B. Ackerman, and
H. F. Percival, editors. Population biology of the Florida
manatee. National Biological Service Information and Tech-
nology Report 1.
Gasaway, W. C., S. D. Dubois, and S. J. Harbo. 1985. Biases in
aerial transect surveys for moose during May and June.
Journal of Wildlife Management 49:777-784.
Gerrodette, T. 1987. A power analysis for detecting trends.
Ecology 68:1364-1372.
*Hartman, D. S. 1974. Distribution, status, and conservation of
the manatee in the United States. Document Number PB
81-140725, National Technical Information Service, Spring-
field, Va. 246 pp.
Irvine, A. B. 1982. West Indian manatee. Pages 241-242 in D. E.
Davis, editor. CRC Handbook of Census Methods for Terres-
trial Vertebrates. CRC Press, Inc., Boca Raton, Fla. 397 pp.
Irvine, A. B., and H. W. Campbell. 1978. Aerial census of the
West Indian manatee, Trichechus manatus, in the southeast-
ern United States. Journal of Mammalogy 59(3):613-617.
Kinnaird, M. F. 1985. Aerial census of manatees in northeastern
Florida. Biological Conservation 32:59-79.
Lancia, R. A., J. D. Nichols, and K. H. Pollock. 1994. Estimating
the number of animals in wildlife populations. Pages 215-253


3 An asterisk denotes unpublished material.


in T. A. Bookhout, editor. Research and management tech-
niques for wildlife and habitat. The Wildlife Society, Be-
thesda, Md. 740 pp.
Lefebvre, L. W., and H. I. Kochman. 1991. An evaluation of
aerial survey replicate count methodology to determine
trends in manatee abundance. Wildlife Society Bulletin
19:298-309.
Lefebvre, L. W., T. J. O'Shea, G. B. Rathbun, and R. C. Best.
1989. Distribution, status, and biogeography of the West
Indian manatee. Pages 567-610 in C. A, Woods, editor.
Biogeography of the West Indies past, present, and future.
Sandhill Crane Press, Inc., Gainesville, Fla. 878 pp.
Marsh, H. 1995. Fixed-width aerial transects for determining
dugong population sizes and distribution patterns. Pages 56-
62 in T. J. O'Shea, B. B. Ackerman, and H. F. Percival,
editors. Population biology of the Florida manatee. National
Biological Service Information and Technology Report 1.
Marsh, H., and D. F. Sinclair. 1989. Correcting for visibility bias
in strip transect aerial surveys of aquatic fauna. Journal of
Wildlife Management 53:1017-1024.
*McGehee, M. A. 1982. Manatees (Trichechus manatus): abun-
dance and distribution in and around several Florida power
plant effluents during the winter of 1981-1982. Florida Power
and Light Company, Contract Report 31534-86419. 67 pp.
O'Shea, T. J. 1988. The past, present, and future of manatees in
the southeastern United States: realities, misunderstandings,
and enigmas. Pages 184-204 in R. R. Odom, K. A. Riddle-
berger, and J. C. Ozier, editors. Proceedings of the Third
Southeastern Nongame and Endangered Wildlife Sympo-
sium. Georgia Department of Natural Resources, Game and
Fish Division, Social Circle. 253 pp.
O'Shea, T. J., and C. A. Langtimm. 1995. Estimation of survival
of adult Florida manatees in the Crystal River, at Blue Spring,
and on the Atlantic Coast. Pages 194-222 in T. J. O'Shea,
B. B. Ackerman, and H. F. Percival, editors. Population biol-
ogy of the Florida manatee. National Biological Service Infor-
mation and Technology Report 1.
*Packard, J. M. 1985. Development of manatee aerial survey
techniques. Manatee Population Research Report 7, Techni-
cal Report 8-7. Florida Cooperative Fish and Wildlife
Research Unit, University of Florida, Gainesville. 68 pp.
*Packard, J. M., and R. Mulholland. 1983. Analysis of manatee
aerial surveys: A compilation and preliminary analysis of
winter aerial surveys conducted in Florida between 1977 and
1982. Manatee Population Research Report 2, Technical
Report 8-2. Florida Cooperative Fish and Wildlife Research
Unit, University of Florida, Gainesville. 119 pp.
Packard, J. M., R. C. Summers, and L. B. Barnes. 1985. Vari-
ation of visibility bias during aerial surveys of manatees.
Journal of Wildlife Management 49:347-351.
Packard, J. M., D. B. Siniff, and J. A. Cornell. 1986. Use of
replicate counts to improve indices of trends in manatee
abundance. Wildlife Society Bulletin 14:265-275.
Packard, J. M., R. K. Frohlich, J. E. Reynolds, III, and J. R.
Wilcox. 1989. Manatee response to interruption of a thermal
effluent. Journal of Wildlife Management 53:692-700.
Pollock, K. H., and W. L. Kendall. 1987. Visibility bias in aerial
surveys: a review of estimation procedures. Journal of Wild-
life Management 51:502-509.






74 INFORMATION AND TECHNOLOGY REPORT 1


Powell, J. A., and G. B. Rathbun. 1984. Distribution and abun-
dance of manatees along the northern coast of the Gulf of
Mexico. Northeast Gulf Science 7:1-28.
*Raymond, P. W. 1981. Manatees (Trichechus manatus): abun-
dance and distribution in and around several Florida power
plant effluents. Florida Power and Light Company, Contract
Report 31534-82511. 62 pp.
Reid, J. P., G. B. Rathbun, and J. R. Wilcox. 1991. Distribution
patterns of individually identifiable West Indian manatees
(Trichechus manatus) in Florida. Marine Mammal Science
7:180-190.
*Reynolds, J. E., III. 1989. Distribution and abundance of the
West Indian manatee (Trichechus manatus) around selected
Florida power plants following winter cold fronts: 1988-
1989. Florida Power and Light Company, Contract Report
B87806-00039. 48 pp.
*Reynolds, J. E., III. 1990. Distribution and abundance of the
West Indian manatee (Trichechus manatus) around selected
Florida power plants following winter cold fronts: 1989-
1990. Florida Power and Light Company, Contract Report
B89806-00264. 41 pp.
*Reynolds, J. E., III. 1991. Distribution and abundance of the
West Indian manatee (Trichechus manatus) around selected
Florida power plants following winter cold fronts: 1990-
1991. Florida Power and Light Company, Contract Report
B89806-00264. 39 pp.


Reynolds, J. E., III, and J. R. Wilcox. 1985. Abundance of West
Indian manatees (Trichechus manatus) around selected Flor-
ida power plants following winter cold fronts, 1982-1983.
Bulletin of Marine Science 36:413-422.
Reynolds, J. E., III, and J. R. Wilcox. 1986. Distribution and
abundance of the West Indian manatee, Trichechus manatus,
around selected Florida power plants following winter cold
fronts: 1984-1985. Biological Conservation 38:103-113.
*Rose, P. M., and S. P. McCutcheon. 1980. Manatees (Trichechus
manatus): abundance and distribution in and around several
Florida power plant effluents. Florida Power and Light Com-
pany, Contract Report 31534-86626. 128 pp.
Samuel, M. D., E. O. Garton, M. W. Schlegel, and R. C. Carson.
1987. Visibility bias during aerial surveys of elk in north-cen-
tral Idaho. Journal of Wildlife Management 51:622-630.
Seber, G. A. F. 1982. The estimation of animal abundance. Char-
les Griffin, London. 624 pp.
Shane, S. H. 1983. Abundance, distribution, and movements of
manatees (Trichechus manatus) in Brevard County, Florida.
Bulletin of Marine Science 33:1-9.
Taylor, B. L., and T. Gerrodette. 1993. The uses of statistical
power in conservation biology: the vaquita and northern
spotted owl. Conservation Biology 7:489-500.
*U.S. Fish and Wildlife Service. 1989. Florida manatee
(Trichechus manatus latirostris) recovery plan. Prepared by
the Florida Manatee Recovery Team for the U.S. Fish and
Wildlife Service, Atlanta, Ga. 98 pp.






HELENE MARSH 75


The Life History, Pattern of Breeding, and Population Dynamics
of the Dugong



by


Helene Marsh

Department of Tropical Environment Studies and Geography
James Cook University
Townsville, Queensland, 4811, Australia


Abstract. The literature and recent unpublished data on the breeding cycle and life history of the
dugong (Dugong dugon) are summarized. The studies were based on the analysis of carcasses from
animals accidentally drowned in shark nets or killed by native hunters in northern Australia and in
Papua New Guinea. Age was estimated by counting the dentinal growth-layer groups in the tusks, the
deposition rate deduced from the seasonal pattern of growth-layer group deposition. The maximum
estimated age was 73 years, and the minimum pre-reproductive period was 9 or 10 years in both sexes.
The pre-reproductive period was variable and ranged to 15-17 years in some animals. Neither males
nor females were continuously in breeding condition and breeding was diffusely seasonal. A single
calf was usually born after an estimated gestation period of about 13 months. Calves suckled for at
least 18 months. Estimates of mean interbirth interval based on pregnancy rates and placental scar
counts ranged from 3 to 7 years in various samples. Variation between years was marked in the
proportions of breeding males and females. Reliable data on age-specific fecundity or mortality are
not available. Population simulations indicated that, with even the most optimistic combination of life
history parameters and a low rate of mortality, a dugong population probably does not increase at more
than about 5% or 6%/year.
Key words: Dugong, Dugong dugon, life history, reproduction, age determination, population
dynamics.


The life history of the dugong (Dugong dugon) has been
studied in northern Australia and in southern Papua New
Guinea as part of the development of conservation and
management of this species. However, data from known-age
animals are not available. No longitudinal information com-
parable to that about wild Florida manatees (Trichechus
manatus latirostris; O'Shea and Hartley 1995; O'Shea and
Langtimm 1995; Rathbun et al. 1995) is available about
dugongs. In addition, unlike manatees, dugongs have rarely
been maintained in captivity and have never bred in captiv-
ity. All information has been obtained from the analysis of
carcasses of animals that were either killed in native fisheries
(Bertram and Bertram 1973; Nietschmann 1984; Hudson
1986) or opportunistically collected when animals inciden-
tally drowned in shark nets set for the protection of bathers
(Heinsohn 1972).
In this paper, I summarize the current understanding of
dugong life history and breeding ecology to provide a com-
parative perspective of these features of manatee biology.


My assessment is based on previously published accounts
(Marsh 1980, 1986; Marsh et al. 1984a, 1984b, 1984c) and
my emphasis is on aspects of dugong life history that are
relevant to the parallel research on manatees in Florida.



Life-history Parameters


Sex Ratio

Dugongs from a native fishery at Numbulwar (Northern
Territory, Australia) sampled by Bertram and Bertram
(1973) had a preponderance of females (88:143). In contrast,
a sex ratio close to parity is indicated by other large samples
from native fisheries (267:237, Western Islands of the Torres
Strait, Australia, Nietschmann 1984; 218:235, Daru near the
Papua New Guinean coast in the central Torres Strait, Hud-
son 1986) and from incidentally drowned animals in shark






76 INFORMANT AND TECiHOLOGY REPORT 1


nets at Townsville, northern Queensland, Australia (44:45,
Marsh et al. 1984c).

Maximum Age
The absolute ages of 75 (Mitchell 1976) and 160 (Marsh
1980; H. Marsh, James Cook University, unpublished data)
dugongs from various locations in northern Australia and of
96 dugongs from Darm (H. Marsh and B. Hudson, James
Cook University, unpublished data) were estimated. The
ages were estimated from the number of dentinal growth-
layer groups in the tusks (Fig. 1); the seasonal layer deposi-
tion indicates that one growth-layer group is deposited per
year (Marsh 1980). Estimated minimum ages (!34.5 years)
are available only of adult male dugongs because their tusks
erupt and wear after puberty (Marsh 1980). Some more-
than-40-year-old females also have erupted and worn tusks
so that their ages are underestimated because of wear The
estimated age of the oldest examined female was 73 years
(H. Marsh, James Cook University, unpublished data).
However, less than 1.5% of females whose absolute ages
were estimated were older than 60 years.

Age at Attainment of Sexual Maturity

Females
Dentinal-layer counts and reproductive organs are avail-
able from 47 females from Townsville and Morington
Island (Queensland, Australia) and from 28 females from
Daru. These data are appropriate for estimating the range of
ages at which sexual maturity occurs but are insufficient for
estimating the age at which 50% of the females are mature
or for quantifying age-related changes in fecundity.


The ovaries and uteruses of all less-than-8-year-old
females were small and undeveloped. Thoseoftwo 8-year-
old and four 9-year-old animals from Townsville showed
some enlargement that suggested approaching maturity;
macroscopic follicles were present in the ovaries of two of
the animals (Marsh et al. 1984c).
The most accepted definition of sexual maturity in
females is that the animal has ovulated at least once. The
youngest female with corpora lutea or corpora albicantia
in the ovaries was a 9.5-year-old nulliparous female from
Townsville. All 10-year-old or older females from Towns-
ville were parous and had one or more placental scars in
the uterus (Marsh et al. 1984a, 1984c). However, data
from Mornington Island and Darm suggested that the age
at which females bear their first calves is variable; three
females from Mornington Island were still nulliparous at
15-17 years (Table 1). Two of the animals were sexually
mature because their ovaries contained small numbers of
corpora lutea or corpora albicantia (Marsh et al. 1984c).
One 18-year-old female from Daru had recently had her
first calf.

Males
What constitutes sexual maturity in males is complex.
Attempts to estimate the age of sexual maturity in male
dugongs have been hampered by the asynchronous, dis-
continuous pattern of male sexual activity (Marsh et al.
1984b); difficulties with distinguishing the testicular his-
tology of pubertal males (those approaching first spermio-
genesis) from that of mature males with recrudescent
testes (Marsh et al. 1984b); and small sample sizes in the
pubertal age range. The data suggested a pre-reproductive










Fig. L. A longitudinally bisected tusk
of a female dugong (Dugong du-
gon) prepared for age determina-
tion. One dark band and one light
band are laid down each year, in-
dicating that this animal was 21
years old when it died.






HELENE MARSH 77


Table 1. Pre-reproductive periods of female dugongs (Dugong dugon) from various areas.
Youngest female Oldest female
Location with placental scar without placental scar Source of data
Townsville 10 9.5 Marsh et al. (1984c)
Mornington Island 14.5 17.5 Marsh et al. (1984c)
Daru 13 12a Marsh (1986)
a One 18-year-old female had recently given birth to her first calf.


period similar to that of females. Most male dugongs from
Townsville were sexually mature when they were about 9
or 10 years old, and all 12-year-old or older males from
Townsville had erupting or erupted tusks. However, some
dugongs from Mornington Island and Daru that were as
old as 16 years were immature and their tusks had not
erupted (Table 2).

Size at Sexual Maturity

Marsh (1980) developed growth curves of male and
female dugongs. The range of body lengths of dugongs of
a particular age is considerable, and body length as an
indicator of age is unsatisfactory except for young animals.
Similarly, the size at which dugongs mature sexually is
variable of animals of both sexes. The data of Marsh et al.
(1984c) suggested that less-than-2.2-m-long male and fe-
male dugongs from northern Australia were immature,
whereas those longer than 2.5 m were probably mature.
Dugongs with body lengths between 2.2 and 2.5 m may be
either immature or mature. There is some evidence (Marsh
1980) that the asymptotic body length of females is slightly
greater than that of males.

Size at Birth

Nine fetal and nine postnatal dugongs had body lengths
between 100 and 130 cm. The data are insufficient for the
50% interpolation method of estimating size at birth rec-
ommended by Perrin and Reilly (1984). Accordingly, I
estimated the mean birth length as 115 cm from the lengths
of the postnatal animals. With this method, size at birth
tends to be overestimated (Perrin and Reilly 1984); how-
ever, three fetuses were longer than 115 cm.


Gestation Period

I estimated a gestation period of 13.9 months with the
method of Huggett and Widdas (1951) and Laws (1959) and
data on the body lengths and dates of deaths of 26 fetuses
from Daru (Fig. 2). As a result of the small sample size and
the diffusely seasonal breeding pattern of the dugong, the
95% confidence interval for this estimate was so imprecise,
it was meaningless. However, the estimated mean was in
good agreement with the estimated 12-14 month gestation
period of captive Florida manatees. (*lOdell et al. 1992; Qi
Jingfen 1984).

Length of Lactation

Data on length of lactation are sparse, especially be-
cause of the ban on taking females with attendant calves
in the Daru fishery. A calf of an estimated age of 1.5 years
(on the basis of dentinal-layer counts) and its presumed
mother who was still lactating were caught in a shark net.
This suggested that lactation can last at least 1.5 years
(Marsh et al. 1984c), even though dugongs start eating
seagrass soon after birth (Marsh et al. 1982). Pregnancy
can occur during lactation. I noted one lactating female
with a 41-cm-long fetus at necropsy (Marsh 1989).

Litter Size

One fetus was found in each of the 34 pregnant females
that scientists examined in recent years. Reports of the
occasional occurrence of twin fetuses (Norris 1960; Jar-
man 1966; Thomas 1966; Bertram and Bertram 1968) are

SAn asterisk denotes unpublished material.


Table 2. Pre-reproductive periods of male dugongs (Dugong dugon) from various areas.
Youngest male Oldest male with Oldest male with
with immature unerupted
Location mature testes testes tusks Source of data
Townsville 9 6 10.5 Marsh et al. (1984c)
Mornington Island 15 15.5 15.5 Marsh et al. (1984c)
Daru 11 16 18 Marsh (1986)






78 INFORMATION AND TECHNOLOGY REPORT 1


y = 0.2993x + 15.0408

r2= 0.7286


350 0 50 100 150 200 250 300 350 0 50
350 0 50 100 150 200 250 300 350 0 50


Jan-01


Day of the year


Jan-01


Fig. 2. The relation between fetal length and day of the year (January 1 is day 1) in dugongs (Dugong dugon) from Daru in the Torres
Strait. Data from H. Marsh and B. Hudson (James Cook University, unpublished data).


vague and unsubstantiated. The mean litter size must be
close to one as in Florida manatees whose twin calves
account for a low proportion of the total births (Marmontel
1995; O'Shea and Hartley 1995; Rathbun et al. 1995).

Pregnancy Rate and Interbirth Interval

The annual pregnancy rate is usually estimated as the
percentage of mature pregnant females (including those
pregnant and lactating) divided by the length of gestation
in years (Perrin and Reilly 1984). The interbirth interval is
the reciprocal of the annual pregnancy rate. Calculation of
the annual pregnancy rate requires accurate knowledge of
the length of the gestation period. Accordingly, I used the
apparent pregnancy rate (proportion of pregnant females
uncorrected for the length of gestation) in the following
discussion and three estimates of the gestation period (12,
13, and 14 months) in the estimates of interbirth intervals
based on pregnancy rates.
Apparent pregnancy rates of three series of dugong
carcasses are available: 86 mature-sized females from the
native fishery at Numbulwar (Bertram and Bertram 1973),
18 mature females from the Townsville shark nets (Marsh
et al. 1984c), and 168 mature females from the native
fishery in Daru (H. Marsh and B. Hudson, James Cook
University, unpublished data). The females from Towns-
ville and Daru were classified as mature by the appear-
ances of their ovaries.
The probability of detecting a pregnancy increases as
gestation progresses because of the size of the fetus, and


some small embryos may have been overlooked, espe-
cially in examinations in the field. Accordingly, I esti-
mated the pregnant proportion based on only confirmed
pregnancies (which tends to provide an underestimated
pregnancy rate) and on possible pregnancies from females
with corpora lutea in their ovaries. This tends to provide
an overestimated pregnancy rate because not all corpora
lutea are associated with pregnancy (Marsh et al. 1984a).
Irrespective of the gestation period used in the model or
whether the pregnancy rate was based on confirmed or
possible pregnancies, all estimated interbirth intervals are
long, ranging from about 3 to 7 years (Table 3). These
estimates are generally higher than those of manatees
(Marmontel 1995; O'Shea and Hartley 1995; Rathbun
et al. 1995; Reid et al. 1995). The biases in the dugong
samples on which these estimates are based are unquanti-
fied. However, the sample from Daru was probably biased
in favor of pregnant females. These were regarded as a
delicacy (Hudson 1986), and there was a ban against
taking females with young calves (*Hudson 1981). Thus,
the estimated pregnancy rates of the Daru sample are
probably high.
A maximum of nine placental scars has been counted
in the uterus of a dugong (Marsh et al. 1984a). The inter-
birth interval can also be estimated by regressing the
number of placental scars against age of parous dugongs
if persistence of the scar is assumed. These estimates also
suggested that the interbirth interval was long: 6.6 years in
Townsville dugongs and 4 years in Mornington Island






HELENE MARSH 79


Table 3. Estimates of the interbirth interval of various dugong (Dugong dugon) populations based on the annual
pregnancy rate and three possible gestation periods of 12, 13, and 14 months. Estimates are based on confirmed
pregnancies (a) and all possible pregnancies (b).
Number of Estimated interbirth interval for three
mature possible gestation periods ( S.E.)
Location females 12 months 13 months 14 months Source of data
Numbulwar 86a 3.07 0.47a 3.32 0.55a 3.58 0.62a Bertram and Bertram
2.69 0.38b 2.91 0.43b 3.14 0.49b (1973)
Townsville 18 4.50 + 1.98 4.88 2.26 5.25 2.55 Marsh et al. (1984c)
Dam 168 5.79 0.98a 6.28 1.11a 6.76 1.25a H. Marsh and B. Hudson
4.94 0.76b 5.35 0.86b 5.76 0.97b (James Cook University,
unpublished data)


dugongs (Marsh et al. 1984c). The sample sizes on which
these estimates are based are too small for meaningful
associated confidence intervals. The difference in the
mean ages of the dugongs from Townsville with one and
two placental scars is also 6.6 years. Placental scars do not
persist in manatees and cannot be used to estimate the
number of parities (Marmontel 1995).
Evidence from Daru suggested marked temporal fluctua-
tions in the apparent pregnancy rate. Hudson (1986) pre-
sented anecdotal evidence that none of the 35 females caught
between October 1976 and July 1977 was pregnant. Car-
casses were collected during the succeeding 4 years (1978-
82). The proportion of mature females that was pregnant
increased monotonically from 0.09 to 0.35 during this time
(Table 4). The difference between years was significant
(Table 4) and was paralleled by a significant increase in the
proportion of males with active testes between 1978 and
1981 (Table 5), suggesting that a common factor was affect-
ing female and male reproduction during this period. Anec-
dotal reports (Johannes and MacFarlane 1991) suggested a
major die-back of seagrasses in the Torres Strait in the
mid-1970's, and Nietschmann (1984) reported that sea-
grasses were overgrazed in the Torres Strait during 1976-
77. Nietschmann and Nietschmann (1981) observed that


wati dangal (lean dugongs with poor-tasting meat) were
quite common in the Torres Strait during this period.

Incidence of Breeding

The samples from Townsville (Marsh et al. 1984c) and
Dam (H. Marsh and B. Hudson, James Cook University,
unpublished data) indicated that neither male nor female
dugongs are continuously in breeding condition. The ovaries
of non-pregnant females probably contain follicles or cor-
pora lutea in the second half of the year. Sterile cycles seem
to be common (Marsh et al. 1984a) and may occur also in
Florida manatees (Marmontel 1995, Rathbun et al. 1995).
Mature male dugongs do not continuously produce
spermatozoa (Marsh et al. 1984b). Histological examina-
tion of the testes of 41 pubertal and mature males from
northern Queensland and 141 mature male dugongs from
Daru showed that more than half the males in each sample
were infertile at the time of sampling. Some of these males
had regressed testes (sensu Marsh et al. 1984b), suggesting
long-term or permanent sterility (Marsh et al. 1984b).
Animals with developed testes (fully spermatogenic or
recrudescent testes sensu Marsh et al. 1984c) were a sig-
nificantly higher proportion of the sample from Daru


Table 4. Differences between years (July 1978-June 1982) in the apparent pregnancy rate of dugongs (Dugong dugon)
sampled at the Daru fishery in Papua New Guinea. Data from H. Marsh and B. Hudson (James Cook University,
unpublished data).
Number of Apparent pregnancy rate S.E.
Year mature females Confirmed pregnancies All possible pregnancies
1978-79 75 0.093 0.034b 0.107 0.036c
1979-80 47 0.191 0.057b 0.255 0.064c
1980-81 29 0.241 0.079b 0.276 0.083c
1981-82 17 0.353 + 0.116b 0.353 0.116c
a 1 July through 30 June.
b Difference between years G = 8.0001, 3 df, P = 0.046.
c Difference between years G = 8.677, 3 df, P = 0.034.






80 INFORMATION AND TECHNOLOGY REPORT 1


Table 5. The numbers (%) of mature male dugongs (Dugong dugon) with active and inactive testes between February
and May and between June and January, sampled at the Daru fishery each year between July and June 1978-1981
(H. Marsh and B. Hudson, James Cook University, Townsville, Australia, unpublished data). Animals with regressed
testes were omitted from the table because they may have been permanently sterile (Marsh et al. 1984b).
February-May June-January
Active Inactive Active Inactive
Year testes testes testes testes
1978-79a 5 (20) 20(80) 12(71) 5 (29)
1979-80 7 (37) 12 (63) 25 (83) 5 (17)
1980-81 6 (46) 7 (54) 14(93) 1(7)
a The effects of year, season, and testicular activity were not independent (G7 = 42.92, P < 0.001); the effect of year was independent of season and
of testicular activity (G5 = 10.07, P = 0.074); testicular activity was dependent on season (GI = 32.86, P < 0.001).


between June and January than between February and May
(Table 5). This seasonal pattern of gonadal activity over-
laps that in females. Spermatogenesis also decreased in
male manatees in Florida in winter (Hernandez et al.
1995), and the behavior of mating herds also revealed
seasonal shifts in frequency (Rathbun et al. 1995).
Births also seem to be diffusely seasonal in dugongs in
northern Australia. In the Townsville-Cairns area, du-
gongs give birth from August-September through No-
vember (Marsh et al. 1984c). Dugongs with near-term
fetuses (>110 cm body length) were sampled at Daru
between July and February (Fig. 2). Seasonal parity has
been reported of Florida manatees (Marmontel 1995;
O'Shea and Hartley 1995; Rathbun et al. 1995).

Natural Mortality

Reliable data on natural mortality in dugongs are not
available. An age-frequency distribution can provide life-
table information and hence survivorship curves, but only
when it is drawn from a population with a stable age
distribution and a known rate of change (Caughley 1977).
The age-frequency data from dugongs at Townsville and
Daru are not suitable for this analysis. The rates of popu-
lation change were not known, and the assumption that the
populations were stable was not valid. The sample from
Townsville was obtained as a result of a massive increase
in mortality from the introduction of shark nets. The sam-
ple from the Daru fishery was obtained during a time of
rapidly changing harvest levels (Hudson 1986) and was
not representative because of the ban on taking females
with attendant calves. An additional problem was that
minimum-age estimates are available only of mature male
dugongs because of the loss of growth layers when the
tusks wear (Marsh 1980). Admitting these limitations, a
crude estimate of the mortality rate calculated from the
age-frequency distribution of females drowned in the
Townsville shark nets (Marsh 1980) was 0.08 with a 95%


confidence interval of 0.06-0.10 (H. Marsh, James Cook
University, unpublished data).
In the absence of reliable natural mortality schedules of
dugongs, Marsh (1986) constructed a simple population
model by using two mortality schedules based on those
developed for population models of another paenungulate,
the African elephant (Loxodonta africana), by Hanks and
McIntosh (1973). This model was extended to three mor-
tality schedules (Table 6). The typically U-shaped mam-
malian mortality curve (Caughley 1966) was modeled as
a step function with levels that corresponded to four dif-
ferent age groups.

Population Dynamics

For each mortality schedule, I constructed population
models to determine the annual rate of increase of stable
dugong populations for various combinations of pre-re-
productive periods and interbirth intervals, chosen to span
the range of estimates of pre-reproductive periods and
interbirth intervals obtained from carcass analysis (Ta-
bles 1 and 3 unpublished data).
The models were based on six simplifying assump-
tions: (1) the reproduction rate is independent of age
during the reproductive part of the female's life span, (2)
females cease to bear calves at age 50 or 60 (this age has
a minimal effect on the results; Fig. 3), (3) no females live


Table 6. Mortality schedules used for the population
models of dugongs (Dugong dugon). Mortality is
expressed as % of population in various age categories
dying per year.
Age in years
0-4 5-45 45-55 55-60
Low mortality 5 1 5 50
Medium mortality 10 2 5 50
High mortality 20 4 5 50






HELENE MARSH 81


---- Low (60)

--- Low (50)

---- Medium (60)

----- High (60)


9 10 11


12 13 14 15


Age at which first calf born (yr)
Fig. 3. The effects of the mortality schedules defined in Table 6 and age at which the first calf is born at the annual rate of increase
of a dugong (Dugon dugon) population with a mean interbirth interval of 3 years. The two graphs of the low mortality schedule
show the effect of changing the age at which females bear their last calves at ages 50 to 60. For the medium and high mortality
schedules, the graphs are based on the assumption that females bear their last calves at age 60.


beyond age 60 (only 1.5% of females of which absolute
age estimates are available were older than 60 years), (4)
the sex ratio at birth is 1:1, (5) immigration and emigration
do not occur, and (6) the anthropogenic mortality is zero.
Each asymptotic rate of increase was determined by
executing the appropriate Leslie matrix model over a
100-year period with an approximate initial stable age
distribution that was determined from an analogous con-
tinuous model (Pollard 1973). The resultant asymptotic
rates of increase were confirmed by using them to modify
the conditions of the simulation model.
The models indicated that the expected maximum
asymptotic annual rate of increase of an unharvested
dugong population is only 6.3% even with the most
optimistic combination of life-history parameters de-
rived from the carcass samples and a schedule of low
mortality. If the parameters suggested by the Daru sam-
ple apply (pre-reproductive period 12 years, interbirth
interval 6 years), the estimated annual rate of increase is
only 2.4%. The simulations of Marsh et al. (1984c) and
Marsh (James Cook University, unpublished data) also
indicate the sensitivity of the models to changes in sur-
vivorship, particularly adult survivorship, and interbirth
interval (Fig. 3; Table 7). The models are less sensitive
to changes in the age at first reproduction (Table 7).
These patterns are consistent with those predicted by
Eberhardt and Siniff (1977) in marine mammals and by
Packard (*1985) and Eberhardt and O'Shea (1995) in
manatees.


Implications for Conservation

Despite the limited sample sizes, the analyses of speci-
mens from dugong carcasses from various locations in
northern Australia and Papua New Guinea indicated that
the dugong is a long-lived mammal with a low reproduc-
tion rate. The data suggested plasticity in the age and size
at which dugongs mature sexually, but the causes of such
variation are not known. Neither mature males nor females
are continuously in breeding condition, and breeding is
diffusely seasonal (Marsh et al. 1984a, 1984b, 1984c). The
data from Daru suggested considerable differences be-
tween years in the proportions of breeding males and


Table 7. Annual rate of increase of a stable dugong
(Dugong dugon) population at various combinations of
pre-reproductive period and interbirth interval; the
assumed mortality schedule is age 0-4 years (5%),
5-45 years (1%), 46-55 years (5%), 56-60 years
(50%), and females are assumed to reproduce until they
die.
Mean pre-reproductive Mean interbirth interval
period (year) 3 5 7
9 6.3 3.7 2.2
11 5.6 3.3 2.0
13 5.1 3.0 1.7
15 4.7 2.7 1.5


6

5

4

3

2

1



-1






82 INFORMATION AND TECHNOLOGY REPORT 1


females. Anecdotal evidence links these fluctuations in
breeding with changes in food availability.
Goodman (1981) pointed out that large mammals have
a particular life table; survival is high, fecundity is low,
and sexual maturity is usually late. The dugong is an even
more extreme example of this life-history strategy than the
Florida manatee (Eberhardt and O'Shea 1995; Marmontel
1995). If dugongs are to be conserved, survivorship must
be high and anthropogenic mortality, low.
As explained by Marsh (1995), the range of the dugong
in Australia extends over a vast area in which all causes of
anthropogenic mortality cannot be prevented. Such pre-
vention is unacceptable because it would not allow tradi-
tional hunting. A more practical approach is to provide a
high level of protection in areas that support large numbers
of dugongs. This protection must extend to the seagrass
habitat and to the dugongs. Such a policy of zonal ecosys-
tem management exists in the Great Barrier Reef Marine
Park (*Great Barrier Reef Park Authority 1983, 1985) and
is being extended to other regions by the Australian Gov-
ernment's Oceans 2000 program to establish a national
system of protected marine areas.



Cited References2

Bertram, C. K. R., and G. C. L. Bertram. 1968. The Sirenia as
aquatic meat producing herbivores. Symposium of the Zoo-
logical Society of London 21:385-391.
Bertram, G. C. L., and C. K. R. Bertram. 1973. The modem
Sirenia: their distribution and status. Biological Journal of the
Linnean Society 5:297-338.
Caughley, G. 1966. Mortality patterns in mammals. Ecology
47:906-918.
Caughley, G. 1977. Analysis of vertebrate populations. John
Wiley and Sons, London. 234 pp.
Eberhardt, L. L., and T. J. O'Shea. 1995. Integration of manatee
life-history data and population modeling. Pages 269-279 in
T. J. O'Shea, B. B. Ackerman, and H. F. Percival, editors.
Population biology of the Florida manatee. National Biologi-
cal Service Information and Technology Report 1.
Eberhardt, L. L., and D. B. Siniff. 1977. Population dynamics
and marine mammal management. Journal of the Fisheries
Research Board of Canada 34:325-428.
Goodman, D. 1981. Life history analysis of large mammals.
Pages 415-436 in C. W. Fowler and T. D. Smith, editors.
Dynamics of large mammal populations. John Wiley and
Sons, New York. 477 pp.
*Great Barrier Reef Marine Park Authority. 1983. Cairns section
zoning plan and Cormorant Pass section zoning plan. Great
Barrier Reef Marine Park Authority, Townsville, Australia.
55 pp.


2 An asterisk denotes unpublished material.


*Great Barrier Reef Marine Park Authority. 1985. Far northern
section zoning plan. Great Barrier Reef Marine Park Author-
ity, Townsville, Australia. 42 pp.
Hanks, J., and J. A. E. McIntosh. 1973. Population dynamics of
the African elephant (Loxodonta africana). Journal of Zool-
ogy 169:29-38.
Heinsohn, G. E. 1972. A study of dugongs (Dugong dugon) in
northern Queensland, Australia. Biological Conservation
4:205-213.
Hernandez, P., J. E. Reynolds, III, H. Marsh, and M. Marmontel.
1995. Age and seasonality in spermatogenesis of Florida
manatees. Pages 84-97 in T. J. O'Shea, B. B. Ackerman, and
H. F. Percival, editors. Population biology of the Florida
manatee. National Biological Service Information and Tech-
nology Report 1.
*Hudson, B. E. T. 1981. The dugong conservation, management
and public education program in Papua New Guinea: working
with people to conserve their dugong resources. Pages 123-
141 in H. Marsh, editor. The dugong. Proceedings of a semi-
nar/workshop held at James Cook University 8-13 May
1979. 400 pp.
Hudson, B. E. T. 1986. The hunting of dugong at Daru, Papua
New Guinea, during 1978-1982. Community management
and education initiatives. Pages 77-94 in A. K. Haines, G. C.
Williams, and D. Coates, editors. Torres Strait Fisheries
Seminar, Port Moresby, February 1985. Australian Govern-
ment Publishing Service, Canberra.
Huggett, A. St. G., and W. F. Widdas. 1951. The relationship
between mammalian foetal weight and conception age. Jour-
nal of Physiology 114:306-317.
Jarman, P. J. 1966. The status of the dugong (Dugong dugon
Millerr: Kenya 1961. East African Wildlife Journal 4:82-88.
Johannes, R. E., and W. MacFarlane. 1991. Traditional fishing
in the Torres Strait Islands. Australian Government Publish-
ing Service, Canberra. 268 pp.
Laws, R. M. 1959. The foetal growth of whales with special
reference to the fin whale, Balaenoptera physalus Linn. Dis-
covery Reports 29:281-308.
Marmontel, M. 1995. Age and reproduction in female Florida
manatees. Pages 98-119 in T. J. O'Shea, B. B. Ackerman,
and H. F. Percival, editors. Population biology of the Florida
manatee. National Biological Service Information and Tech-
nology Report 1.
Marsh, H. 1980. Age determination of the dugong (Dugong
dugon Miller) in northern Australia and its biological impli-
cations. Reports of the International Whaling Commission,
Special Issue 3:181-201.
Marsh, H. 1986. The status of the dugong in Torres Strait.
Pages 53-76 in A. K. Haines, G. C. Williams, and D. Coates,
editors. Torres Strait Fisheries Seminar, Port Moresby. Feb-
ruary 1985. Australian Government Publishing Service, Can-
berra.
Marsh, H. 1989. Mass stranding of dugongs by a tropical cyclone.
Marine Mammal Science 5:75-84.
Marsh, H. 1995. Fixed-width aerial transects for determining
dugong population sizes and distribution patterns. Pages 56-
62 in T. J. O'Shea, B. B. Ackerman, and H. F. Percival,






HELENE MARSH 83


editors. Population biology of the Florida manatee. National
Biological Service Information and Technology Report 1.
Marsh, H., P. W. Channells, G. E. Heinsohn, and J. Morrissey.
1982. Analysis of stomach contents of dugongs from Queens-
land. Australian Wildlife Research 9:55-67.
Marsh, H., G. E. Heinsohn, and P. W. Channells. 1984a. Changes
in the ovaries and uterus of the dugong, Dugong dugon
(Sirenia: Dugongidae) with age and reproductive activity.
Australian Journal of Zoology 32:743-766.
Marsh, H., G. E. Heinsohn, and T. D. Glover. 1984b. Changes in
the male reproductive organs of the dugong, Dugong dugon
(Sirenia: Dugongidae) with age and reproductive activity.
Australian Journal of Zoology 32:721-742.
Marsh, H., G. E. Heinsohn, and L. M. Marsh. 1984c. Breeding
cycle, life history, and population dynamics of the dugong,
Dugong dugon (Sirenia: Dugongidae). Australian Journal of
Zoology 32:767-785.
Mitchell, J. 1976. Age determination in the dugong, Dugong
dugon (Miiller). Biological Conservation 9:25-28.
Nietschmann, B. 1984. Hunting and ecology of dugongs and
green turtles in Torres Strait. National Geographic Society
Research Reports 17:625-651.
Nietschmann, B., and J. Nietschmann. 1981. Good dugong, bad
dugong; good turtle, bad turtle. Natural History 90:54-62.
Norris, C. E. 1960. The distribution of the dugong in Ceylon.
Loris 296-300.
Odell, D. K., G. D. Bossart, M. T. Lowe, and T. D. Hopkins.
1992. Reproduction of the West Indian manatee in captivity.
Pages 192-193 in T. J. O'Shea, B. B. Ackerman, and H. F.
Percival, editors. Population biology of the Florida manatee.
National Biological Service Information and Technology
Report 1.
O'Shea, T. J., and W. C. Hartley. 1995. Reproduction and early-
age survival of manatees at Blue Spring, upper St. Johns
River, Florida. Pages 157-170 in T. J. O'Shea, B. B. Acker-
man, and H. F. Percival, editors. Population biology of the


Florida manatee. National Biological Service Information
and Technology Report 1.
O'Shea, T. J., and C. A. Langtimm. 1995. Estimation of survival
of adult Florida manatees in the Crystal River, at Blue Spring,
and on the Atlantic Coast. Pages 194-222 in T. J. O'Shea,
B. B. Ackerman, and H. F. Percival, editors. Population biol-
ogy of the Florida manatee. National Biological Service In-
formation and Technology Report 1.
*Packard, J. M. 1985. Preliminary assessment of uncertainty
involved in modeling manatee populations. Manatee Popula-
tion Research Report Number 9, Technical Report Number 8-
9, Florida Cooperative Fish and Wildlife Research Unit,
University of Florida, Gainesville. 19 pp.
Perrin, W. F., and S. B. Reilly. 1984. Reproductive parameters
of dolphins and small whales of the family Delphinidae.
Reports of the International Whaling Commission (Special
Issue) 6:97-125.
Pollard, J. H. 1973. Mathematical models for the growth of
human populations. Cambridge University Press. 186 pp.
Qi Jingfen. 1984. Breeding of the West Indian manatee
(Trichechus manatus Linn) in captivity. Acta Theriologica
Sinica 4:27-33.
Rathbun, G. B., J. P. Reid, R. K. Bonde, and J. A. Powell. 1995.
Reproduction in free-ranging Florida manatees. Pages 135-
156 in T. J. O'Shea, B. B. Ackerman, and H. F. Percival,
editors. Population biology of the Florida manatee. National
Biological Service Information and Technology Report 1.
Reid, J. P., R. K. Bonde, and T. J. O'Shea. 1995. Reproduction
and mortality of radio-tagged and recognizable manatees on
the Atlantic Coast of Florida. Pages 171-191 in T. J. O'Shea,
B. B. Ackerman, and H. F. Percival, editors. Population biol-
ogy of the Florida manatee. National Biological Service In-
formation and Technology Report 1.
Thomas, D. 1966. Natural history of dugongs in Rameswaram
waters. Madras Journal of Fisheries 2:80-82.






84 INFORMATION AND TECHNOLOGY REPORT 1


Age and Seasonality in Spermatogenesis of Florida Manatees



by


Patricia Hernandezl

University of South Florida
Department of Biology
Tampa, Florida 33620


John E. Reynolds, III

Eckerd College
Department of Biology and Marine Science
4200 54th Avenue South
St. Petersburg, Florida 33711


Helene Marsh

James Cook University
Department of Tropical Environment Studies and Geography
Townsville, Queensland 4811
Australia


and


Miriam Marmontel2

University of Florida
Department of Wildlife and Range Sciences
Gainesville, Florida 32611


Abstract. A histological study of the testes and epididymides of the Florida manatee (Trichechus
manatus latirostris) was made to assess reproductive status. Testes or epididymides of 67 animals
were examined for presence and abundance of sperm. In addition, qualitative and quantitative
examinations of the seminiferous tubules from 31 manatees, ranging in length from 98 to 320 cm (total
length), revealed 10 stages in the spermatogenic cycle. The shortest manatee in which spermiogenesis
occurred was a 237-cm-long, 2-year-old. Males as short as 252 cm had fully spermatogenic testes, and
one 255-cm-long individual was 2-3 years old. Spermatogenesis was not continuous and was
significantly affected by season. In winter (December through February), sperm were found in either
the testes or epididymides of only 13% of the recovered manatees that measured 241 to 280 cm, and


'Present address: Museum of Comparative Zoology, Harvard
University, 26 Oxford Street, Cambridge, Mass. 02132.
2 Present address: Projeto Mamirua, Caixa Postal 0001, 69470-000 Tef6
AM, Brazil.






PATRICIA HERNANDEZ ET AL. 85


in 25% of those longer than 280 cm. In contrast, 75% of the 241- to 280-cm-long manatees and 93%
of the greater-than-280-cm-long animals recovered in non-winter months (March through November)
had mature sperm. Of the manatees recovered in winter, none was fully spermatogenic and most had
regressed seminiferous tubules (characterized by immature cell stages and reduced diameters),
suggesting reduction, if not cessation, in the ability to impregnate females.
Key words: Florida manatees, reproduction, spermatogenesis, testes, Trichechus manatus latirostris.


In contrast to the dugong (Dugong dugon) for which
excellent data on reproduction in males (at the gross and
microscopic levels) have been available for some time
(Marsh et al. 1984), relatively little effort has been fo-
cused on the reproductive tracts from male Florida mana-
tees (Trichechus manatus latirostris). Observations of
behavior revealed that mating may occur year-round but
may be seasonally depressed during winter (Hartman
1979; Rathbun et al. 1995), but the reproductive status
of males, in terms of spermatogenic activity, has not been
examined by season or by age and size of the animals.
A preliminary description of reproductive tracts in
male manatees (*3Odell et al. 1981) revealed that tes-
ticular weight increased exponentially with body length
in 275 to 290-cm-long animals; this observation led the
authors to the conjecture that sexual maturity in male
manatees occurred at a body length of 275 cm or greater,
corresponding to estimated ages of 9-10 years. The
figure of 9 to 10 years was based on the growth of some
captive-born manatees of known age. In contrast, Hart-
man (1979) suggested, based on behavior of wild mana-
tees, that sexual maturity occurs at about 3-5 years of
age.
Odell et al. (*1981) noted that even large manatee
carcasses recovered in winter had relatively small testes
that were brown and from which no evidence of sperm
could be found in smears. Conversely, testes from car-
casses with comparable body lengths that were recov-
ered in seasons other than winter were larger and white;
however, smears were not available from these car-
casses. Odell et al. (*1981) suggested a seasonal cycle in
testicular activity and significantly less activity in win-
ter.
With routine histological techniques and a recent tech-
nique for aging manatees (*Marmontel et al. 1990), we
describe the reproductive status of male manatees, including
some of known age that were recovered by the manatee-car-
cass-salvage network (*Bonde et al. 1983). The data provide
information about age at onset of sexual maturity, variability
in that age, and seasonality of spermatogenic activity. These
data are useful for an understanding of the life history and
population dynamics of manatees.


3An asterisk denotes unpublished material.


Materials and Methods

Tissues from reproductive tracts of 68 male manatees
were obtained between April 1975 and February 1985
through a carcass-salvage network (*Bonde et al. 1983).
At necropsy, collected data included body length and
condition of the animal prior to death (reflected by quan-
titative or qualitative assessment of blubber or body-fat
thickness). Animals recovered throughout the year were
included in this study. Gonads or epididymides from all 67
manatees were examined for the presence of sperm. Tis-
sues from 31 of the animals were of sufficient quality (i.e.,
had minimal autolysis) to permit analysis of seminiferous
tubules for precise stages in the spermatogenic cycle.
Manatees were arbitrarily separated into three size catego-
ries: 98- to 240-cm length, 241- to 280-cm length, and 281-
to 320-cm length. The lower limit of the final size category
corresponds approximately to the length at which Odell
et al. (*1981) found that testicular weight increased expo-
nentially, and the upper limit corresponds to the length of
the largest animals in this sample.
Although testicular weights of some animals were pro-
vided, records and notes at necropsies were unclear about
whether the weight was for one or both testes or whether
the epididymis had been completely trimmed. Therefore,
we did not use testicular weights but focused on the
histological determination of sexual maturity.
Ages were estimated from counts of growth-layer
groups in the dome region of the tympano-periotic com-
plex (*Marmontel et al. 1990). An age of 1 was used to
designate an animal 1-2 years of age, 2 indicated an
animal 2-3 years old, and so on. Bone resorption or other
changes in the layers precluded accurate age determina-
tion of some old manatees.
During necropsy, samples from the epididymides or
testes were excised for histological examination and
fixed in 10% buffered formalin. Although whole testes
were occasionally preserved, often only one piece was
excised. The precise locations from which these samples
were taken was generally not noted and was presumably
not consistent. Fixed tissues were dehydrated in an etha-
nol series, cleared in xylene, embedded in Paraplast, and
sectioned at 6 lim. Sections of testis were stained with
hematoxylin and eosin (H and E). Epididymides and
testes were also stained with Putt's stain (Appendix) to






86 INFOR ATONANDTECh LOnov REPORT I


detect the presence of sperm. No testes were serially
sectioned.
Many of the 67 manatees assessed for the presence of
sperm had seminiferous tubules that were severely
autolyzed. In specimens with good quality of the pre-
served tissues, seminiferous tubules were assessed in
two ways. Diameters of 50 seminiferous tubules were
measured at 400x with an ocular micrometer. The same
tubules were then analyzed with phase analysis, in which
a numerical value was assigned to each tubule based on
certain cellular associations in the seminiferous epithe-
lium. Each section from the testes typically contained at
least 50 round seminiferous tubules. In rare cases when
only small sections were available, two or more separate
pieces from the testes were examined to locate 50 round
tubules. If consecutive sections through the same piece
of tissue had been used, biased data may have been
collected because consecutive sections probably in-
cluded the same plane through the seminiferous tubule.
A series of 10 phases was recorded after the same
criteria as Marsh et al. (1984). Phases -4 to 0 (Fig. 1)
represent seminiferous tubules in which spermatogene-
sis was either arrested or incomplete and Phase -4 sug-
gest total inactivity. Phases 0 through +5 represent active
spermiogenesis as round spermatids mature into sperma-
tozoa (Fig. 2). The mean seminiferous tubule phase of
any given animal (termed the testicularstage) was used
for comparisons among animals. A manatee was consid-
ered fully spermatogenic if the testicular stage was a
positive number (i.e., the mean tubule phase was be-


tween 0 and +5). Testes at this stage were undergoing
active spermiogenesis, characterized by a distinct cellu-
lar organization of the seminiferous epithelium (Marsh
et al. 1984). With similar categories of Marsh et al.
(1984), 31 manatees were placed in three reproduction
categories: 29% immature; 29% mature in resting or
intermediate state, recrudescent (undergoing spermio-
genesis) or prepubescent; and 42% fully spermatogenic
(undergoing spermatogenesis). The term fully spermato-
genic is used here to imply a stage at which most seminif-
erous tubules were undergoing spermiogenesis, during
which spermatozoa were produced (Marsh et al. 1984).
Data were examined for normality and found not to
conform to necessary assumptions to justify the use of
parametric statistics. Nonparametric statistics were used
for comparisons of reproduction among seasons and size
classes and for correlation analyses. The Kruskal-Wallis
test was used to determine significant differences in
testicular histology and tubule diameter among the three
size categories of the carcasses, and between carcasses
that were recovered in winter and carcasses that were
recovered in seasons other than winter. Spearman's rank
correlation test was used to determine significant corre-
lations between testicular stage and animal length and
between testicular stage and diameter of seminiferous
tubules. The latter was tested to determine a fast and
simple method by which reproductive status can be de-
termined from histological samples. A probability value
of 0.01 was chosen to represent significant results from
all analyses.


D


44.t8 lE


Fig Early and regressed phases of
spermatogenesis inFlorida mana-
tees (Trichechs manaus latiros-
trs). (A) Phase -4, the most im-
mature stage, is characterized by
Sertoli cell nuclei (S) and sperma-
togonia (SP) lining the periphery
of the seminiferous tubule.
(B) Phase -3, similar to Phase -4
with a few primary spermatocytes
(P). (C) Phase -2, like the pre-
vious phases but with numerous
primary spermatocytes. Seminif-
erous tubules at this phase may be
open or closed. (D) Regressed tu-
bule in Phase-3. This condition is
characterized by a few primary
spermatocytes, increased diame-
ter of the lamina propia (LA), de-
creased tubule diameter and in-
creased cross-sectional area of
interstitial tissue (note Leydig
cells,L) All samples were stained
with hematoxylin and eosin and
photographed at 100X.






PATnuA HERNANDEZ EAL 87




ig. 2. Several seminiferous tubules
from Florida mnartees (Trchechus
manatus latirostris) undergoing
spermiogenesis. Diameters of
seminiferoustubules arelargerthan
in Fig. 1. (A) Phase 0. Primary and
secondary (SE) spernatocytes and
round spermatids (RT) are present
(B) Phase +2. During this phase,
spermatogonia and Sertoli cell nu-
clei line the basement membrane.
Primary and secondary spermalo-
cytes reorganized intolayerswith
elongating speratids (ET) along
the lumen. (C)Phase4. Thisphase
includes round(RT) andelongating
(ET) spermatids. (D) Phase +5.
Round spenmatids and spematooa
line the tubule lumen All samples
were stained with henatoxyhn and
osin and photographedat 100X


Results


Staged Testes

Different cellular associations of the seminiferous epi-
thelium characterized the different phases. Phase -4, the
least active phase, was represented by essentially solid
tubules lined with spermatogonia and occasional Sertoli
cells. The underlying lamina propria was relatively thick
(5-10 tm) during this phase. The next phase, -3, included
tubules with small numbers of primary spermatocytes in
addition to already mentioned cell types. Phase -2 tubules
had numerous primary spermatocytes. The lumen of the
tubule became evident during Phase -1, at which time the
different cells of the seminiferous epithelium became or-
ganized into layers that included some secondary sperma-
tocytes. Phase 0 tubules contained a few rounded sperma-
tids in addition to the already named cell types.
Tubules assigned to Phases 1 through 5 were found in
animals undergoing spermiogenesis, during which the
haploid spermatids transformed into spermatozoa (Fig.
2). Tubules in Phase I had Sertoli cells with basal nuclei
and spermatogonia lining the periphery of the tubules
and primary and secondary spermatocytes and round
spermatids organized into layers. Phase 2 tubules con-
tained spermatids beginning to elongate and becoming
associated with Sertoli cells. Phase 3 tubules were char-
acterized by well-organized spermatid bundles associ-
ated with Sertoli cells and maturation divisions among
primary and secondary spermatocytes. In contrast to
Phase 3, two generations of spermatids (round and elon-
gating) were present in Phase 4. Sperm (with tails) lined


the entire lumen of tubules in Phase 5. Phase 6, identified
by Marsh et al. (1984), represented by spent tubules at
the end of a period of sexual activity, was not identified
in this study.
Testicular stage and diameter of seminiferous tubules
strongly correlated (Fig. 3; n = 31; Spearman's rank cor-
relation test; r = 0.815). A higher correlation existed be-
tween diameter of seminiferous tubules and age (Fig. 4;
n = 26; Spearman's rank correlation test; r = 0.770) than
between testicular stage and age (Fig. 5; n = 18; Spear-
man's rankcorrelation test; r = 0.487). Increases in diame-
ter of seminiferous tubules and testicular stage were sharp
in animals that were longer than approximately 250 cm
(Figs. 6 and 7).
The tubule diameter (Fig. 6; n = 31; P = 0.0001), tes-
ticular stage (Fig. 7; n = 31; P = 0.0008), and presence of
sperm in the testis or epididymis (Fig. 8; n=29,
P = 0.0012) of manatees differed by size class. Testicular
stage (P= 0.05, n = 1) or diameters of tubules (P= 0.016,
n = 17) did not differ by size category during winter
months because regressed testes were histologically indis-
tinguishable from immature testes.
Lack of testicular weights made it impossible to differ-
entiate between mature resting (i.e., reproductively quies-
cent) and prepubescent animals.

Immature Manatees

Nine 98- to 235-cm-long manatees fit characteristics
ascribed to immature animals (Table 1). Where age was
known, these animals ranged from 0-2 years old. Testicu-
lar stage was always negative and approached or was equal







88 INFORMATION AND TECHNOLOGY REPORT 1


0)
CD
z 1




(D
-s
.o -1

I--


-b i I I I I
0 50 100 150 200
Tubule diameter (rim)
Fig. 3. Correlation between mean diameter of seminiferous
tubules and testicular stage in the Florida manatee
(Trichechus manatus latirostris; n = 31).


200-



150-


100-



50-


-2 2 6 10 14 18 22
Age
Fig. 4. Relation between age and diameter of seminiferous
tubules in the Florida manatee (Trichechus manatus
latirostris; n = 26).


0
0.
0




0 0

~I


180

160

140

120

100

80

60


40U 1 I I I I I I
50 100 150 200 250 300 350

Body length (cm)
Fig. 6. Relation between body length and mean diameter of
seminiferous tubules in the Florida manatee (Trichechus
manatus latirostris; n = 31).


S S
*' 0
0
0 5


0
O Non-winter
* Winter o
00
0
0 0
0
0
0
o0 *
0


c Po

0 0 O


4-
*

2-



S0-


S -2 -

1"--a *
**
-4- *


-6 I i I
-2 2 6 10 14 18 22
Age

Fig. 5. Relation between age and testicular stage in Florida
manatees (Trichechus manatus latirostris) sampled during all
seasons (n = 18).






PATRICIA HERNANDEZ ET AL. 89


0
o Non-winter o 0
Winter
08
0
o


0

0



o o
0 0 00 0

50 100 150 200 250 300


Body length (cm)
Fig. 7. Relation between body length and testicular stage in the
Florida manatee (Trichechus manatus latirostris; n = 31).


98 -240 241 280 281 350
Size classes (cm)
Fig. 8. Percentage of Florida manatees (Trichechus manatus
latirostris) with sperm in the testes or epididymides. Note
that, for the smallest age class, 10% represents only a single
animal (n = 29).


Table 1. Life history data of male Florida manatees (Trichechus manatus latirostris) with a total body length of
98-240 cm (n = 21). Tubule diameter (in lim) represents a mean of 50 measurements. Sperm density represents the
observed highest density in tissues of a particular manatee. 0 = no sperm; 1 = sparse sperm; 2 = moderate density; 3
= very dense sperm. Putt's (1951) stain, which is specific for spermatids and spermatozoa, was used to identify sperm
in the epididymides. Months 12, 1, and 2 represent winter; 3-5 represent spring; 6-8 represent summer; and 9-11
represent fall. The testicular stage is the mean phase of 50 evaluated seminiferous tubules/animal; the phases range
from -4 to 5.
Total body Specimen Tubule Sperm Age Month of Testicular
length (cm) field number diameter density (year) death stage
98 M-293 47.60 0 0 05 -4.00
117 M-213 0 0 12
119 M-178 0 0 12
130 M-403 44.84 0 0 09 -3.98
133 M-135 58.68 0 0 02 -3.58
146 M-75-5 0 0 04
172 M-79-1 52.75 0 1 01 -2.86
175 M-144 0 0 04
177 M-76-3 0 01
190 M-244 0 1 06
205 M-79-26 0 0 12
209 M-388 0 0 05
218 M-154 63.62 0 1 02 -3.46
219 M-314 0 11
219 M-82-8 0 02
222 M-81-66 50.30 0 12 -4.00
225 M-82-16 44.44 0 0 02 -4.00
225 M-79-9 51.62 0 2 02 -4.00
228 M-137 0 2 02
235 M-79-24 44.40 0 1 11 -3.00
237 M-308 107.20 1 2 08 -0.56






90 INFORMATION AND TECHNOLOGY REPORT 1


to -4. Seminiferous tubules were narrow (44-64 glm) con-
taining only spermatogonia, Sertoli cells, and primary
spermatocytes resting on a thick lamina propria (i.e., no
spermatids or spermatozoa were present). Intertubular tis-
sue was extensive among these largely solid tubules.

Prepubescent or Recrudescent Manatees

Nine 237- to 295-cm-long and 2 to 11-year-old mana-
tees (when age was determined) were placed in this
category (Tables 1-3). Mean diameter of the seminifer-
ous tubules was between 50 and 107 jim, and the lamina
propria seemed to be thinner than in immature manatees.
In addition to the cell types in immature animals, secon-
dary spermatocytes and a few round spermatids were
included in the seminiferous epithelium. Testicular stage
was always negative, although the value approached
zero in at least one individual (M-3084). This 237-cm-
long, 2-year-old individual was the most precocious of
the examined manatees; he already had some sperm in
his testes.

4 Identification numbers of manatees are provided when known for
completeness of information.


Fully Spermatogenic Manatees

Animals were considered fully spermatogenic when the
testicular stage was between 0 and +5. Active spermiogene-
sis occurred in thirteen 252 to 320-cm-long, 2 to 18-year-old
manatees (Tables 2 and 3). Average tubular diameters were
71-195 pm. Cross-sectional area of interstitial tissue (which
includes Leydig cells) was reduced in fully spermatogenic
animals (Fig. 2). Although testicular tissue of the oldest male
in this study (M-235; 25 years old) was not available, a
moderate amount of sperm was in the epididymis (Table 3).

Seasonality of Changes in Testicular
Histology

Several important factors of potential reproduction dif-
fered by season. Mean testicular stage (Kruskal-Wallis test;
n = 31; P = 0.009) and the presence of sperm in the
epididymis or testes (Kruskal-Wallis test; n = 67; P = 0.007)
differed by season in which manatees were recovered. Most
of the differences seemed to reflect changes in testicular
histology in winter, when regressed seminiferous tubules,
containing mostly immature cell stages, were typical, even
in testes obtained from large animals.


Table 2. Life history data of male Florida manatees (Trichechus manatus latirostris) with a total body length of 241-280
cm (n = 24). Tubule diameter (in gm) represents a mean of 50 measurements. Months 12, 1, and 2 represent winter;
3-5 represent spring; 6-8 represent summer; and 9-11 represent fall. The testicular stage is the mean phase of 50
evaluated seminiferous tubules/animal; the phases range from -4 to 5. Eight manatees died in December, January,
or February (winter). One of them (13%) produced sperm. Twelve of 16 (75%) produced sperm in non-winter seasons.
Total body Specimen Tubule Sperm Age Month of Testicular
length (cm) field number diameter density (year) death stage
241 M-219 0 2 01
244 M-220 0 1 01
247 M-55 1 02
252 M-428 50.00 0 9 02 -3.96
252 M-82-34 93.40 1 04 1.90
252 M-217 0 4 01
255 M-78-42 109.50 2 2 11 3.10
256 M-153 3 4 05
257 M-78-35 176.60 2 09 3.96
258 M-82-20 0 02
260 M-275 86.90 0 3 03 -1.80
263 M-79-23 151.70 1 4 10 3.34
263 M-239 0 08
267 M-280 88.60 0 7 04 -1.94
267 M-77-16 0 2 05
270 M-258 61.60 0 01 -2.50
271 M-223 0 7 02
271 M-79-7 1 03
271 M-79-18 3 20 08
273 M-174 3 5 10
274 M-108 1 5 03
277 M-166 3 3 08
278 M-76-28 1 7 11
279 M-156 3 6 06






PATRICIA HERNANDEZ ET AL. 91


Table 3. Life history data of male Florida manatees (Trichechus manatus latirostris) with total body length of more
than 281 cm (n = 22). Tubule diameter (in gtm) represents a mean of 50 measurements. Months 12, 1, and 2 represent
winter; 3-5 represent spring; 6-8 represent summer; and 9-11 represent fall. The testicular stage is the mean phase
of 50 evaluated seminiferous tubules/animal; the phases range from -4 to 5. Eight manatees died in December,
January, or February (winter). Two of them (25%) produced sperm. Thirteen of 14 (93%) produced sperm in
nonwinter seasons.
Total body Specimen Tubule Sperm Age Month of Testicular
length (cm) field number diameter density (year) death stage
284 M-83-2 79.00 0 4 01 -2.38
285 M-78-8 52.85 0 11 02 -2.62
288 M-82-29 1 11 03
290 M-74-2 124.80 2 06 2.26
291 M-78-15 0 10 02
292 M-78-29 1 10 06
292 M-353 2 14 12
292 MSW-4 105.40 2 01 -1.34
295 M-210 71.00 0 10 -3.00
295 G-75-76 138.00 1 10 3.00
297 M-208 160.20 2 7 10 1.60
297 M-80-27 0 12
299 M-209 71.25 3 3 10 1.68
301 M-381 153.60 1 9 03 3.24
303 M-391 167.60 1 7 09 3.40
312 M-78-10 0 20 02
314 M-255 195.30 2 18 11 0.26
318 M-235 2 25 05
318 M-299 168.90 2 15 06 2.48
320 M-146 2 >12 04
320 M-82-30 115.50 2 03 2.78
320 M-81-11 0 01


The presence of sperm in the reproductive tracts of 67
manatees was examined. In the smallest size class (98-
240 cm), sperm was present in 11% (1 of 9) of the
carcasses recovered in a season other than winter, and in
none of the 12 carcasses recovered in winter (Table 1).
In the second size class (241-280 cm), sperm was pre-
sent in 13% (1 of 8) of the carcasses recovered in winter
and in 75% (12 of 16) of the carcasses recovered in
seasons other than winter (Table 2). In the largest size
class (281-320 cm), sperm was present in 93% (13 of
14) of the carcasses recovered in seasons other than
winter but in only 25% (2 of 8) of the carcasses recovered
in winter (Fig. 8; Table 3). Manatees with sperm in the
testes or epididymides were 237-320 cm long and 2-25
years old.
Of the 31 manatees whose testes were quantitatively
assessed, none (0 of 9) in the first size class (98- to 240-cm
length) was fully spermatogenic; however, 44% (4 of 9) of
the 241- to 280-cm-long carcasses and 69% (9 of 13) of the
281 to 320-cm-long carcasses were fully spermatogenic
(Fig. 9; Tables 1-3). Among the carcasses that were longer
than 241 cm, none of 5 recovered between December and
February, 60% (3 of 5) recovered between March and May,
67% (2 of 3) recovered between June and August, and 89%


(8 of 9) recovered between September and November were
fully spermatogenic (Tables 2 and 3).


98- 240 241 280 281 350
Size classes (cm)
Fig. 9. Percentage of fully spermatogenic Florida manatees
(Trichechus manatus latirostris) in each size class.






92 INFORMATION AND TECHNOLOGY REPORT 1


Discussion


Odell et al. (*1981) predicted that male manatees be-
came reproductively active at a length of about 275 cm (at
an estimated age of 9-10 years based on growth of some
captive-born animals), when testicular weight increased
exponentially. Our results showed that males may be
capable of impregnating females at a much smaller size
and younger age. Increases in diameter of seminiferous
tubules and testicular stage at a body length as short as
237 cm and at an age as young as 2 years were notable
(Figs. 5 and 6). Whether the sperm from such young males
can impregnate females is not known.
The possibility of early successful reproduction, as
supported by histological evidence, corroborates observa-
tions of behavior by Hartman (1979), who suggested that
reproduction began in male manatees between 3 and 5
years of age, and more recent observations of Rathbun
et al. (1995), who noted the presence of free-ranging males
(whose ages were known) of even younger ages in herds
of males pursuing presumably estrous females in the Crys-
tal River. The dynamics of these herds suggest no clear
pattern of participation by males on the basis of size or age
(Rathbun et al. 1995). In controlled breeding of captive
animals, keeping small male manatees, considered imma-
ture (based on their size or age), with females may not
prevent pregnancy. Odell (Sea World of Florida, Orlando,
Florida, personal communication) noted that captive,
about 2-year-old male California sea lions (Zalophus cali-
fornianus) successfully mated and produced offspring and
that similar matings occur in the wild with unknown
results.
In manatees, total length seemed to be a better, but still
imprecise, predictor of reproduction than age. Onset of
spermatogenic activity in mammals in general may be
more closely related to somatic development of the indi-
vidual than to age (Johnson et al. 1970). The dependence
of sexual maturity on attainment of critical body mass in
some mammals has been suggested (Laws 1956). In our
study, testicular activity, as gleaned from phase in the
spermatogenic cycle and diameter of seminiferous tu-
bules, increased profoundly when the manatee reached a
length of approximately 250 cm, although this size did not
correspond to a particular age (Fig. 10; n = 26). Several
examples illustrated the point that size is not a good
predictor of age: one 9-year-old manatee (M-428) was
only 252 cm long, but a 2-year-old manatee (M-78-42)
was already 255 cm long; two 3-year-old manatees (M-
209 and M-275) were 299 cm and 260 cm long; a 20-year-
old manatee (M-79-18) was only 271 cm long, whereas
two 3-year-olds (M-166 and M-209) were longer than
270 cm.
Seasonal changes in testicular activity in our study
concur with findings of Odell et al. (*1981), who assessed


Ah hI -


300

E 250-
.c
S200-
1-

-o 150-
o -


100-

KO -


2 2 6 10 1 18
-2 2 6 10 14 18


Age
Fig. 10. Relation between age of male Florida manatees
(Trichechus manatus latirostris) and total body length
(n = 26).


testicular weight and appearance and suggested that activ-
ity was high in summer and fall and low in winter. Our
observations (e.g., presence of regressed seminiferous tu-
bules, characterized by immature cell stages and reduced
diameters) suggested diminution, if not cessation, of sper-
miogenesis in winter. Although some larger manatees
(13% from 241-280 cm long, and 25% >281 cm; Tables 2
and 3) recovered in winter had some sperm in the testes or
epididymides, none was fully spermatogenic.
Seasonal effects seem to be an important factor of
potential reproduction in male manatees. This is evi-
denced by the fact that no fully spermatogenic manatees
were observed during winter. Only one (M-210) of 21
longer-than-270-cm carcasses recovered in seasons
other than winter lacked sperm in the gonads or
epididymis. This animal was an unusual case because it
died in an emaciated condition in the Chesapeake Bay
(outside the range of Florida manatees). In other-than-
winter months, healthy male manatees longer than
250 cm may well undergo spermatogenesis, and in any
healthy individual longer than 270 cm, spermatogenesis
is almost certain.
Cessation of spermatogenic activity in winter was
identified in other taxonomic orders of mammals. These
changes in the capability of males to reproduce may be
due to varying lengths of photoperiod (Millar and Glover
1973; Grocock and Clarke 1975) that affect circulating
levels of gonadotrophins and testicular hormones. Sea-
sonal reproduction in ringed seals (Phoca hispida) was
due to pulses in secretion of hypothalamic gonadotro-
phin that released hormones that were influenced by
several cues including photoperiod, nutrition, and social


0 3 g 0
0
S

gO


w,






PATRIA HERNANDEZ ET AL. 93


stimuli (Ryg et al. 1991). These environmental factors,
stress from the onset of cold weather, and the resulting
need for hurried, often-lengthy migrations may play a
similarly prominent role in the control of spermatogene-
sis in the manatee. Although insufficient nutrition cannot
be ruled out as a cause of cessation of spermatogenesis
in some manatees, at least one large manatee that died in
winter (M-78-15; 291 cm) had no sperm but extensive
body fat. This animal, presumed well nourished and
IaeshLry; d~ed fiumr a co'ision witd a boat. This example
may suggest that seasonal effect on spermatogenesis is
not entirely due to nutritional stress during winter.
Only limited seasonality of spermatogenic activity
occurred in male dugongs near Australia (Marsh et al.
1984; Marsh 1995). Although some individuals had not
continuously produced sperm throughout the year, the
influence of season on the capability to reproduce was
less clear in these animals than in manatees in Florida
where seasonal effects strongly correlated with sperma-
togenic activity. If one assumes that seasonal effects on
reproduction are due, at least in part, to changes in
temperature, the difference between dugongs and mana-
tees may reflect that the sampled dugongs were from the
northern regions of Australia and southern Papua New
Guinea where seasonal changes in water temperature are
less marked than in waters occupied by manatees in
Florida. The examined dugongs may not have undergone
the types of environmental, nutritional, or energetic
stresses associated with changes in temperature experi-
enced by Florida manatees in winter. To isolate and
evaluate the importance of seasonal changes in tempera-
ture on spermatogenesis, an examination of West Indian
manatees would be helpful outside Florida (i.e., in coun-
tries bordering the Caribbean Sea) where water tempera-
tures are consistently warm. Similarly, an examination
of testes of dugongs in cooler parts of the species' range
(e.g., Moreton Bay) may be helpful to determine whether
spermatogenesis in these animals is influenced by sea-
sonal change in water temperature.
De Jong and Zweers (1980) confirmed monophyly of
the paenungulate Orders Hyracoidea, Sirenia, and
Proboscidea. Results from studies of seasonal effects on
spermatogenic activity among the paenungulates failed
to provide information that may unite the group by the
physiology of reproduction. Whereas Laws (1969) found
that the diameter of the seminiferous tubules of African
elephants (Loxodonta africana) is affected by seasons,
Short et al. (1967) did not identify marked seasonal
changes in androgen concentration in these elephants.
Seasonal effects on spermatogenesis were not shown in
all hyraxes (Order Hyracoidea) (Glover and Sale 1968;
Millar and Glover 1973). Animals far from the equa-
tor showed marked seasonality in their capabilities to


reproduce, whereas in the tropics, as seen in the dugongs,
hyraxes had no well defined breeding season. Animals
at lower latitudes may be incapable of detecting strong-
enough changes in photoperiod to affect seasonal
changes in spermatogenesis. Alternatively, either tem-
perature or nutritional value of foods may play a vital
role in seasonal reproduction (Millar and Glover 1973).
Moreover, as described by Marsh et al. (1984) in du-
gongs, male hyraxes in the tropics do not undergo coor-
dinated spermatogenic activity (i.e., not afl members of
the population undergo spermatogenesis simultane-
ously). As in most eutherian mammals, manatees share
a similar spermatogenic cycle as that seen in elephants,
hyraxes, and dugongs, but unlike in most other examined
paenungulates, the capability to reproduce (as defined by
presence of sperm and testicular stage) correlates with
season in manatees (at least in Florida).
Whereas dugongs and manatees differ in the degree
to which their spermatogenic capability is affected by
seasons, the species are comparable in body length at
which sexual maturity may occur. Dugongs as short as
219 cm may exhibit spermatogenesis (Marsh et al.
1984), whereas spermatogenesis occurs in manatees as
short as 237 cm. However, manatees reach a larger
maximum size than dugongs. Age at sexual maturity
differs considerably between the two species; the young-
est mature male dugong was 9 years old (Marsh et al.
1984), and the youngest mature male manatees were only
2 years old. The difference may reflect species-specific
differences in life history. Alternatively, age at sexual
maturity in Florida manatees may have dropped, possi-
bly because of faster growth from changing population
parameters and resource availability, whereas the exam-
ined dugong populations may be stable and occupy more
stable habitats. Age, but not body size at sexual maturity,
dropped in Antarctic fin whales (Balaenoptera physa-
lus), presumably because of exploitation of mysticetes
and subsequent greater availability of food (Gambell
1985); in other words, body size, not age, is the cue for
sexual maturity.
As noted earlier, spermatogenic activity diminishes in
winter in many mammalian groups; however, some spe-
cies in the Order Chiroptera are able to store viable sperm
throughout winter (Asdell 1946). Five manatees in this
study (M-108, M-209, M-255, M-353, and MSW-4) had
little if any sperm in the testes but had abundant sperm
in the epididymides. This suggested that manatees may
be able to store sperm in the epididymis after completing
a spermatogenic cycle. Because the viability of stored
sperm is not known, the importance of this phenomenon
remains unknown. Because we did not serially section
and examine whole testes, spermiogenesis may have
occurred in tissues we did not examine. In dugongs,






94 INFORMATION AND TECHNOLOGY REPORT 1


abundant spermatozoa were found only in epididymides
of animals at the height of spermiogenesis (Marsh et al.
1984).
Testicond mammals (mammals without scrota) in-
clude species in the Orders Hyracoidea, Proboscidea,
Edentata, Cetacea, Sirenia, and Insectivora. Spermato-
genesis proceeds as usual in testicond mammals without
apparent detriment from abdominal temperature (Car-
rick and Setchell 1977). Although lower temperature
seems to be more important for sperm storage than
production in most cases, a mechanism for maintaining
the epididymides at a cooler temperature in most testi-
cond mammals is not known (*Bedford 1977). Recent
anatomical evidence suggested that cetaceans have a
vascular counter-current exchange that may serve to cool
the testes and epididymides (Rommel et al. 1992). A
similar vascular arrangement, if present, may explain
viable sperm storage in the epididymides of manatees.
In conclusion, observations of manatee testes and
epididymides suggested near cessation of spermatogene-
sis in winter, an apparent peak in spermatogenic activity
in late summer and fall, and initiation of spermiogenesis
in manatees as short as 237 cm long and as young as 2
years. This type of information provides important life-
history data, and could influence the captive mainte-
nance of manatees. Future tissue analysis should be
made a priority, particularly to determine whether age at
sexual maturity or timing of spermatogenesis change
over time. In addition to examining seminiferous tu-
bules, future researchers should include measuring hor-
mone levels in blood, feces, or urine. Testosterone plays
a central role in controlling the spermatogenic cycle
(Sharpe et al. 1990). Although testosterone levels are
affected by stress, a direct measure of testosterone level
and an assessment of the relation between testicular
stage and testosterone level may be informative. Finally,
entire testes, especially of recovered carcasses in winter,
should be serially sectioned and examined to determine
whether spermatogenic activity occurs at a consistent
level throughout the gonad.


Acknowledgments

Reproductive tract tissues of male manatees were
collected under U.S. Fish and Wildlife Service Permit
Number PRT 2-8430. The authors are grateful to D. K.
Odell of the University of Miami and currently at Sea
World of Florida and his many graduate students and to
C. A. Beck, R. K. Bonde, A. B. Irvine, T. J. O'Shea,
G. B. Rathbun, and other U.S. Fish and Wildlife Service,
Sirenia Project, scientists who collected and provided the
tissues we examined. We are also grateful to B. L. Wei-
gle of the Florida Department of Natural Resources


Marine Research Institute for covering costs of slide
preparation; to I. Quintera and her many lab technicians
also at the Marine Research Institute for processing the
tissues and preparing slides; and to P. S. Botts at Penn-
sylvania State University, Erie, for statistical assistance.
The comments by the editors of the proceedings and by
two anonymous reviewers improved the manuscript.


Cited References5

Asdell, S. A. 1946. Patterns of mammalian reproduction. Com-
stock Publishing Company, New York. 437 pp.
*Bedford, J. M. 1977. Evolution of the scrotum: The epididymis
as the prime mover. Pages 171-182 in J. H. Calaby and C. H.
Tyndale-Biscoe, editors. Reproduction and evolution: Pro-
ceedings of the Fourth International Symposium on Com-
parative Biology of Reproduction. Australian Academy of
Science, Canberra.
*Bonde, R. K., T. J. O'Shea, and C. A. Beck. 1983. A manual of
procedures for the salvage and necropsy of carcasses of the
West Indian manatee (Trichechus manatus). National Techni-
cal Information Service. Springfield, Va. PB83-255273.
175 pp.
Carrick, F. N., and B. P. Setchell. 1977. The evolution of the
scrotum. Pages 165-170 in J. H. Calaby and C. H. Tyndale-
Biscoe, editors. Reproduction and evolution: Proceedings of
the Fourth International Symposium on Comparative Biology
of Reproduction. Australian Academy of Science, Canberra.
De Jong, W. W., and A. Zweers. 1980. Study of eye lens protein
confirms relationship of manatee with hyraxes and elephants.
Acta Amazonica 10:897-902.
Gambell, R. 1985. Fin whale Balaenoptera physalus (Linnaeus,
1758). Pages 171-192 in S. H. Ridgway and R. J. Harrison,
editors. Handbook of marine mammals, volume 3: the sireni-
ans and baleen whales. Academic Press, New York.
Glover, T. D., and J. B. Sale. 1968. The reproductive system of
male rock hyrax (Procavia and Heterohyrax). Journal of
Zoology 156:351-362.
Grocock, C. A., and J. R. Clarke. 1975. Spermatogenesis in
mature and regressed testes of the vole (Microtus agrestis).
Journal of Reproduction and Fertility 43:461-470.
Hartman, D. S. 1979. Ecology and behavior of the manatee
(Trichechus manatus) in Florida. Special Publication 5,
American Society of Mammalogists, Lawrence, Kans.
153 pp.
Johnson, A. D., W. R. Gomes, and N. L. Vandemark. 1970. The
testis. Volume 1: development, anatomy, and physiology.
Academic Press, New York. 684 pp.
Laws, R. M. 1956. Growth and sexual maturity in aquatic mam-
mals. Nature 178:193-194.
Laws, R. M. 1969. Aspects of reproduction in the African ele-
phant (Loxodonta africana). Journal of Reproduction and
Fertility Supplement 6:193-217.
*Marmontel, M., T. J. O'Shea, and S. R. Humphrey. 1990. An
evaluation of bone growth-layer counts as an age-determina-


5 An asterisk denotes unpublished material.




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs