Dynamics of exploitation on the American alligator


Material Information

Dynamics of exploitation on the American alligator environmental contaminants and harvest
Physical Description:
xii, 165 leaves : ill. ; 29 cm.
Rice, Kenneth G., 1963-
Publication Date:


Subjects / Keywords:
American alligator -- Florida   ( lcsh )
Wildlife Ecology and Conservation thesis, Ph. D
Dissertations, Academic -- Wildlife Ecology and Conservation -- UF
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1996.
Includes bibliographical references (leaves 154-163).
Statement of Responsibility:
by Kenneth G. Rice.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 023826523
oclc - 35822274
System ID:

This item is only available as the following downloads:

Full Text








Copyright 1996


Kenneth G. Rice

For Mary Lou, Rodney, and Luke

Mi familiar, mi fuerza


Let me begin by expressing my deepest gratitude to H.

Franklin Percival, my major professor and mentor in this

endeavor. He never allowed my schooling to get in the way

of my education. My remaining committee members, Ab

Abercrombie, Jon Allen, Susan Jacobson, and Jim Nichols,

offered their considerable assistance, insights, and


This study was fortunate to receive substantial funding

from many institutions including the American Alligator

Farmer's Association, the Florida Alligator Farmer's

Association, the Florida Wildlife Federation, the Florida

Game and Fresh Water Fish Commission (GFC), the St. Johns

River Water Management District, the U.S. Environmental

Protection Agency, the U.S. Fish and Wildlife Service, and

the Florida Cooperative Fish and Wildlife Research Unit

(GFC, University of Florida, U.S. National Biological

Service, and Wildlife Management Institute cooperating).

J.D. Ashley provided the impetus for initiation of

research culminating in Chapter 2. C.L. Abercrombie and

T.C. Hines assisted in developing the framework for data

collection and interpretation. For my attempt at modeling

an alligator population, I thank Allan Woodward and Steve

Linda of the Florida Game and Fresh Water Fish Commission

for providing parameter estimates and a framework of

questions; Ab Abercrombie of Wofford College for patience in

attempting to estimate survival; Jon Allen of the Department

of Entomology, University of Florida for assistance in model

preparation; and Jim Nichols of the Patuxent Wildlife

Science Center for having done this before. Jay Harrison of

IFAS Statistical Consulting provided invaluable assistance

in the analyses reported in Chapters 4 and 5. Clint Moore

provided the program for assigning unknown sized animals to

known categories for the night-light survey analyses in

Chapters 2 and 4. Phil Wilkinson was the impetus for and

the brains behind the nesting female capture study reported

in Chapter 5. Mike Jennings led the data collection efforts

prior to 1988 utilized in Chapters 2 and 4. Greg Masson and

I shared those duties from 1988-1992. Tim Gross provided

valuable insights into environmental contaminant effects on

alligator reproduction.

Debra Hatfield and Barbara Fesler offered not only

secretarial assistance and financial organization, but acted

as confidants and the core of our Coop Unit family. Ray

Carthy and I spent many hours discussing our wildlife

philosophies in our shared office.


The project has had the luxury of incorporating

the expertise of many individuals from multiple agencies

over the years. At the risk of eliminating many important

names, these people always had pride in referring to

themselves as members of a loosely organized group, the

Florida Alligator Research Team. Requiring special

reference are the following: J. Anderson, A. Brunell, D.

Carbonneau, J. Connors, R. Conrow, D. David, G. Davidson, L.

Folmar, M. Fuller, C. Hope, L. Hord, M. Jennings, W.

Johnson, L. Rhodes, T. Schoeb, S. Shrestha, R. Spratt, H.

Suzuki, C. Tucker, J. White, J. Wiebe, C. Wieser, and A.


Many friends have encouraged and sustained me

throughout my tenure here: Holly, Tom and Michelle, the

Jims, Scott, Laura, Craig, Cindy, Geoff, Phil, Woody and

Susan, Rick, Dwayne, Dennis and Ilonka, Paul, John, and many

others. Several other friends, though no longer a part of

my life, certainly provided me with their love and support:

Joni, Janet, Kristi, Wanda, Wandy, Karen, Sydney, and Beth.

Finally, my family has been there for me every step of

the way and, as in all things, their unwavering faith in my

abilities has made this possible. My thanks are gratefully

given to Mom, Rodney, Luke, Raye, Sandy, and Chad.



ABSTRACT . . .. x



Exploitation . . 1
Harvest . . 2
Environmental Contamination . 5
Study Areas .. . 8


Introduction . . .14
Methods . . 16
Study Areas . . 16
Removal . . 17
Survey Procedures . .. .22
Analysis . . 24
Size Distributions . 25
Results .......... .. 26
Nest Production . . 26
Night-light Surveys . 27
Cumulative Distributions . 29
Discussion . . 29


Introduction . . .43
Methods . .. .45
Alligator Population Model . .. .45
Survival . . 46
Transition rates . 49
Fertility . . 50


Other density dependent factors
EnvTi rnnmntal f r-c+rs

The model .
Sensitivity Analysis .
Elasticity Analysis .
Simulation of Various Harvest
Population health .
Hunter satisfaction .
Rancher satisfaction .
Aesthetic value .
Results . .
Sensitivity Analysis .
Proportion nesting .
Survival .
Transition rates .
Elasticity . .
Recommendations .
Simulation of Various Harvest
Density .
Economic value .
Numbers of animals in har
Population .
General Recommendations and Discuss

. 51



vest .

ion .


Introduction . . .
Methods . . .
Egg Collections . .
Clutch and Egg Characteristics .
Incubation . .
Hatchling Care . .
Viability and Clutch Size . .
Survey Procedures . .
Survey Analysis . .
Clutch Viability Analysis . .
Results . .
Clutch Viability and Clutch Characteristics
Population Trends . .
Discussion . . .


Introduction . . .
Methods . .
Nesting Female Capture . .
Nest and Clutch Characteristics .
Demographic Comparisons . .


. 81
. 84
S. 84
S. 86
S. 87
S 88
. 90
. 90
. 92
. 93
S. 94
S 94
S. 95
S. 97


. 114
. 120
. 120
. 121

Nesting and Nest Composition Trends .
Results . . .
Nesting Female Capture .
Nesting Female Demographics .
Nesting and Nest Composition Trends .
Discussion . .





. 124
. 125
. 125
. 127
. 131
. 131

. 143

. .. 153



Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Kenneth G. Rice

August 1996

Chairperson: H. Franklin Percival
Major Department: Wildlife Ecology and Conservation

Exploitation of American alligator (Alligator

mississippiensis) populations by humans was investigated by

viewing the dynamics of population level effects due to

controlled harvest and uncontrolled environmental

contamination. I found that alligator populations were

resilient to exploitative stresses.

Alligator populations in 2 Florida lakes, Griffin and

Jesup, were harvested at a level of 50% of annual production

from 1987 to 1991. No changes were found in numbers of

juveniles (P > 0.11). Numbers of subadult (P < 0.011) and

adult (P < 0.001) size-classes increased.

A simulation model was developed to investigate changes

in alligator population dynamics under various harvest

strategies. The simulated population was very sensitive to

errors in parameter values for fecundity and adult female

survival and growth. Simulations predicted that a 50%

harvest of eggs did not negatively affect population growth

nor did a 10% harvest of adults. The population decreased

when exposed to simultaneous egg and adult harvests.

On another area, Lake Apopka (a Super Fund Contaminant

Site), the alligator was exploited through environmental

contamination. Beginning in 1981, a major decline was

observed in juvenile alligators on the lake. The total

population of alligators decreased until 1989 then increased

through 1995 (P < 0.01). Over the same period, juveniles

decreased by 95% through 1989 then increased by 480% (P <

0.01). Clutch viability decreased by approximately 80%

through 1988 and then increased over 400% (0.12 to 0.53)

through 1995. However, viability remained depressed when

compared to a reference area, Lake Woodruff (P < 0.001).

Size distributions of adult nesting females were

estimated for lakes Apopka and Woodruff. Apopka's

distribution (x = 259.91 cm) consisted of larger animals

than Woodruff's (x = 246.61 cm; P < 0.01). Clutch viability

was higher for the medium-sized animals than for the very

large (-Ra,, = 0.659) indicating that a portion of the

continuing depressed viability on Lake Apopka is

attributable to adult female demographics. The loss of

several cohorts to the acute effects of environmental

contamination was implicated.

Although the alligator appears resilient to heavy

exploitative pressures, ongoing monitoring programs should

continue. Successful management programs will incorporate

ecological constraints with all stakeholder groups involved

in exploitation of the alligator.




Humans have exploited wildlife species and their

habitats since before the beginning of civilization (Ehrlich

1988). Herein exploitation is defined as the utilization of

a species either directly through mortality or indirectly

through adverse effects on biotic potential and habitat

quality. In crocodilians, the sources of this exploitation

have been diverse (see Thorbjarnarson 1992): harvest

(Glasgow 1991; Gibson-Carpenter 1992; Hines and Woodward

1980), habitat degradation (Ogden 1978; Maskey and Percival

1994), and environmental contamination (Rice and Percival

1996). Management alternatives must first recognize that

exploitation is inherent before sound guidelines can be

established. Certainly, successful crocodilian management

programs exist that utilize controlled exploitation for

conservation (Woodward 1987; Rose 1984). However, examples

of over-exploitation are both easily recognized (Rao and

Singh 1994; Thorbjarnarson 1992) and obscure (Rice and

Percival 1996).

This study examines the dynamics of alligator

populations exposed to exploitation (Fig. 1-1). Harvest and

environmental contamination are listed as primary

exploitative forces. The effects of these forces on

alligator populations in Florida are examined in terms of

demographic relationships, population trends, population

simulations, and reproductive potential.


Historically, the commercial exploitation of

crocodilians for their skins, meat, and other products has

been uncontrolled and destructive, often resulting in

depletion of species and populations. However, recent

management plans for crocodilians have incorporated

conservation principles that allow for sustained utilization

and preservation of the species (Thorbjarnarson 1992).

Successful management programs for crocodilians exist in

several areas of the world and with several species. Most

notably, these include Alligator mississippiensis in the

United States (Woodward 1987), Crocodylus niloticus in

Zimbabwe (Anon 1982), Crocodylus porosus in Australia (Letts

1987), and Crocodylus novaeguineae in Papua New Guinea (Rose

1984). The majority of these programs operate on the

principle of ranching. In ranching, wild collected eggs

and/or hatchlings are reared under artificial conditions and

harvested for meat, hides, and other products. An

alternative, farming, involves captive propagation of

animals and generally includes wild adult captures for at

least initial stocking purposes. Farming is generally

considered to be less valuable in terms of conservation of

the wild resource. Biologically, the reproductively active

portion of the population is most valuable and least

abundant. Conversely, early age class animals, utilized in

ranching, are considered most expendable to the population

(Hines et al. 1986)

In addition, management has begun to utilize the

monetary value of crocodilian species to construct economic

links between landowners and the habitat that crocodilians

inhabit. This link has been defined as value-added

conservation (Hines and Percival 1987).

Value-added conservation is the concept that a portion

of the return from the commercial harvest of wildlife can be

returned to the conservation and management of wildlife and

its habitat (Hines and Percival 1987). Several key

mechanisms are necessary for this type of management

alternative to succeed (Hines et al. 1986; Hines and

Percival 1987):

1. landowners who participate in the exploitation must
have vested interests in the conservation of the
species, thereby becoming political advocates for
habitat preservation;

2. revenues from licenses, hide tags, taxes, etc...
provide funding for habitat enhancement, species
management, conservation, and research; and,

3. these economic incentives gained from exploitation of
the species lead to modified land-use decisions by

In an unpublished survey by the Florida Game and Fresh

Water Fish Commission in 1989 and 1993, alligator hunters

listed monetary gain as a primary motivation for

participating in the harvest (M. Jennings and D. Carbonneau

unpub. data). Similarly, entire communities depend

economically upon the commercial harvest of wildlife species

(Gibson-Carpenter 1992). Alligator ranchers also

participate in harvests. Other communities utilize the

alligator resource through the commercial tourism industry

(D. David pers. comm.). In this study, each of these

motivations is incorporated into recommendations for trial

sustainable harvest programs based upon a simulation model.

Ludwig et al. (1993) contend that sustained use of

natural resources is an impossible goal because people are

by nature too greedy and shortsighted. They argue that if

sustained use is to ever occur, one must adhere to the

following five principles of management:

1. Human motivations need to be considered in the
management strategy.

2. Modifications to management strategies should be
incorporated before scientific consensus is achieved.

3. Scientists should be employed to recognize problems in
management, but not to remedy them.

4. Claims of sustainability should be distrusted, in light
of past resource use.

5. Uncertainty should be confronted and incorporated into
the decision-making process.

In this study, several of these principles are discussed in

terms of sustainable management programs involving harvests

of the American alligator.

Environmental Contamination

Due to high nutrient and pesticide levels, Lake Apopka

is considered Florida's most polluted large lake (U.S. EPA

1979; Schelske and Brezonik 1992). In 1981, a major decline

was observed in the number of the lake's juvenile

alligators (Woodward et al. 1993). Meanwhile, other lakes

studied in Florida showed numerically stable or increasing

alligator populations (Jennings et al. 1988; Woodward and

Moore 1990).

Since 1984, controlled incubation of alligator

eggs has demonstrated that Lake Apopka has the lowest clutch

viability rates (proportion that hatch from total eggs

deposited) of 7 Florida alligator populations examined

(Percival et al. 1992; Masson 1995). Most embryological

mortality occurred prior to or soon after oviposition

(Masson 1995). Possible causes of Apopka's low clutch

viability included detrimental effects of environmental

contaminants on reproductive success.

The lake is chronically exposed to agricultural

chemicals. Lake water is frequently pumped into and from

agricultural areas created over the last 5 decades by the

draining of about 8000 hectares of sawgrass (Cladium

jamaicense) marsh originally associated with Lake Apopka

(Conrow et al. 1993). Also, for years the lake was largely

encircled by extensive citrus groves (many of these were

severely damaged by freezes in 1983 and 1985). In addition

to agricultural chemicals, Lake Apopka receives tertiarily

treated sewage effluents from the municipality of Winter


In 1980, a spill from the Tower Chemical Company

contaminated the drainage into Gourdneck, the southwest

portion of the lake. The spill involved large quantities of

dicofol (among other chemicals), a compound known to have

substantial chemical contamination by DDT, DDE, and DDD

(compounds to which dicofol is structurally related; Clark

1990), and sulfuric acid. The area was designated a Super

Fund Contaminant Site by the Environmental Protection

Agency, a designation it continues to hold today. Following

the spill, an acute fish die-off and vegetational changes

were reported in the impacted drainage. Elevated

concentrations of dicofol subsequently were detected in the

lake proper (Fla. Dept. Envir. Reg. 1980). Monitoring has

been conducted since 1981 by various agencies.

An analysis of alligator eggs collected in 1985 from 3

Florida lakes showed that eggs from Lake Apopka had

significantly elevated levels of DDD, DDT, and DDE. DDE was

measured at 3.5 ppm (Heinz et al. 1991), a concentration

known to preclude normal egg production and embryonic

development in birds. The frequency of malformed embryos

and hatchlings did not appear different on Lake Apopka than

on other lakes studied during 1988-1992 (Percival et al.

1992). However, only gross external mutations were noted,

and hatchlings' internal organs were not examined.

Dicofol, DDT, DDE, and many other pesticides can act as

synthetic hormones (chemical signals) or as inhibitors to

natural hormones in the developing embryo (Guillette et al.

1996). Normal development of certain regions of the brain,

the reproductive tracts, and the gonads are dependent upon

specific hormonal signals (Guillette et al. 1995). This is

particularly important since several pesticides (including

those found in alligator eggs from Lake Apopka) are known to

act as synthetic estrogens (Colborn and Clement 1992).

This study examines long term trends in alligator

populations and clutch viability on Lake Apopka. In

addition, possible effects of environmental contamination on

the demographic structure of the population are


Study Areas

Several areas in central Florida were utilized as

experimental or control groups in the studies reported in

this manuscript (Fig. 1-2). In Chapter 2, lakes Griffin

(5,675 ha) and Jesup (4,805 ha) were chosen as harvest areas

and Lake Woodruff (6,477 ha) was selected as a control. In

Chapter 3, the Orange Lake (5,254 ha) alligator population

was utilized as the basis for a simulation model. Lake

Apopka (12,809 ha) was the experimental and Lake Woodruff

was the reference area in Chapters 4 and 5.

Lakes Griffin and Apopka are eutrophic, hardwater

natural lakes within the central physiographic region of

Florida (Table 1-1; Canfield 1981). Lake Griffin's southern

shoreline is highly developed, while much of the east marsh

has been drained and converted to agriculture. Most of the

525 ha wooded marsh occurs in a narrow band proximal to open

water in the northern half of the lake. True wet marsh

(1,221 ha) communities exist beyond the wooded shoreline.


About 96% of Lake Apopka's northern marsh was diked and

drained for agricultural purposes. Only a small remnant

(100 ha) of the original emergent marsh remains. Much (700

ha) of the remaining shoreline is characterized by a narrow

band of wooded swamp. The St. Johns River Water Management

District (SJRWMD) recently planted emergent vegetation to

expand the marsh fringe on the northern and western

shoreline. In 1990, the SJRWMD installed a filtration

marsh (The Marsh Flow-Way Project, 2,000 ha) on the northern

end of the lake near its outflow via the Beauclair Canal

(Conrow et al. 1993) to reduce nutrient inputs into the

lake. Inputs of water to the lake include a natural spring,

a stream in the southwest, rainwater runoff, backpumping

from the northern muck farms, and tertiarily treated sewage

effluent from the municipality of Winter Garden.

Lakes Jesup and Woodruff are eutrophic, alkaline,

natural lakes in the St. Johns River drainage basin in east-

central Florida (Table 1-1; Canfield 1981). Much of Lake

Jesup's western shore has been converted to improved

pasture. The undisturbed northeastern marsh (996 ha) is

dominated by sand cordgrass (Spartina baker) and giant reed

(Phragmites australis), whereas much of the southern half is

comprised of wooded marsh.

Lake Woodruff's wetlands are dominated by Spartina

bakeri (3,900 ha). The remaining wetlands consist primarily

of wooded swamp and remnant canals. The Lake Woodruff study

area includes several impoundments managed for migratory

waterfowl. Most of the study area is regulated by the Lake

Woodruff National Wildlife Refuge. Inputs include numerous

feeder creeks from the surrounding basin and DeLeon Springs.

Orange Lake, located in north-central Florida (Alachua

County), consisted of 1188 ha of potential alligator nesting

habitat (Table 1-1). Characteristic of the marsh area was

an accumulation of peat resulting in extensive floating

islands of vegetation. Most of the original emergent marsh

remains intact. Development is concentrated on feeder

streams and associated uplands. Vegetation consisted

primarily of Sagittaria lancifolia, Cladium jamaicensis,

Hydrocotyle umbellata, Myrica cerifera, Cephalanthus

occidentalis, and Decodon verticillatus (Deitz and Hines

1980). Inflow is primarily from the River Styx located in

the northeastern portion of the lake.

Table 1-1. Areal cover (ha) of principal habitat types in
lakes Apopka, Griffin, Jesup, Orange, and Woodruff.














































Percent total area given in parentheses.

Percent total area given

in parentheses.


Contamination Harvest

Chemical spills Ranchers
Agriculture Hun
Waste disposal -legalille

ability 1

-population trends
S............. ........... -sim elation
-reproductive effects

-crash _

Figure 1-1. Representation of links between forms of
exploitation, the American alligator, and studied effects.

Orange Lake

Lake diffini

Woodr ff
:e Jesup\

Florida, U.S.A.
Florida", U.S.A.

Figure 1-2. Location of study areas in Florida.



Harvest of wild crocodilians as eggs or young for

captive-rearing (ranching) has become an important aspect of

management programs throughout the world. Commercial

harvest and ranching programs have been initiated for

saltwater crocodiles (Crocodylus porosus) in Papua New

Guinea, Indonesia, and Australia (Onions 1982; Webb et al.

1987), New Guinea crocodiles (C. novaeguineae) in Papua New

Guinea and Indonesia (Solmu 1994), Nile crocodiles (C.

niloticus) in several African countries (Mkanda 1992; Hutton

and Child 1989; Kelly 1994), American alligators (Alligator

mississippiensis) in Louisiana and Florida (Joanen and

McNease 1987; Hines and Abercrombie 1987), and caimans

(Caiman sp.)in Latin America (Larriera 1994; Gorzula 1987)

among others(see Webb et al. 1987 for reviews). Currently,

ranching, under strict guidelines and regulation through

international trade treaties (CITES), is recommended as a

management practice with potentially high conservation

benefits and relatively little risk to population status

(David 1994).

In the late 1970s in Florida there was considerable

interest from the Florida Game and Fresh Water Fish

Commission (GFC) and the alligator industry in commercially

harvesting both adult and early age class alligators. By

1977, the GFC had instituted a nuisance alligator program

and had embarked on research to investigate the feasibility

of sustainable adult harvest on public waters (Hines and

Woodward 1980). Florida's alligator farmers were interested

in harvesting eggs and hatchlings for captive rearing to

supplement closed-system farming. In 1981, the GFC began an

experimental harvest of early age-class alligators to

determine feasibility of ranching in Florida (Jennings et

al. 1988).

This harvest was designed to evaluate the effects of a

50% removal of total annual production on several lakes in

Florida. The harvest level was based primarily on a

simulation model by Abercrombie (unpub report) and defined

as the presumed level at which population level effects

could be detected (such as changes in population size and

structure; Jennings et al. 1988).

After 6 years of this experimental harvest, no negative

effects were seen in the populations studied (Jennings et

al. 1988). The GFC initiated a statewide alligator egg and

hatchling harvest program in 1987 subsequent to this

positive evaluation of juvenile alligator harvest. However,

due to concerns over long-term sustainability of the

harvest, the GFC desired to extend the evaluation for an

additional 5 years.

Therefore, the objective of this chapter was to

evaluate the effects on alligator populations of the removal

of 50% of total estimated annual production from 2 study

areas over an 11-year period.


Study Areas

From Jennings et al. (1988), criteria for selecting

study areas included a relatively dense alligator population

(large sample sizes needed to detect relatively small

changes in population size and structure), >50 nests/year

(from an estimation of sample size requirements and

collection of sufficient animals to evaluate the viability

of a ranching program), and location within close proximity

to other study sites (to reduce environmental variation).

Lake Griffin in the Oklawaha River drainage and lakes Jesup

and Woodruff, our reference area, in the St. Johns River

drainage were chosen (see Chapter 1 for study area


Another area, Lake Apopka, was initially chosen as a

harvest site during the earlier study. However, due to

reductions in the alligator population for reasons unrelated

to this study (Woodward et al. 1993), the area was not

included (see Chapter 4).


I analyzed data on the removal of 50% of estimated

annual production on the 2 treatment lakes (Griffin and

Jesup) and survey data collected on the control area both

from the literature (1981-1986; Jennings et al. 1988) and

personal studies (1987-1991). Removals were accomplished by

collecting hatchlings during the fall and spring of the 1981

and 1982 nesting seasons and a combination of hatchlings and

eggs in 1983 (Jennings et al. 1988). Since egg collection

was more effective per unit effort and more economical

(Hines et al. 1986), only eggs were collected from 1984 to

1991. Techniques for early collections (1981-1986) were

discussed in more detail by Percival and Jennings (1986).

Egg collection techniques utilized from 1987-1991 are

discussed in Chapter 4 of this manuscript. After

incubation, the surviving hatchlings were placed on ranches

for captive-rearing.

Aerial nest surveys were conducted annually during

early incubation (late June to early July), mid-incubation

(late July), and late incubation (late August to early

September), to determine nest production and survival

(Jennings et al. 1988; this study).

Nest searches were conducted using helicopters flown at

30-50 m altitude and 30-50 kph except 1982 when a fixed-wing

aircraft (approx. 96 kph) was used (Jennings et al. 1988).

Marshes were initially searched by flying strips (50-100m)

parallel to the shoreline repeatedly until all marsh was

surveyed. Other nests were occasionally noted during egg

collection. To facilitate collection of eggs, collection

crews in ~4m airboats were led by the helicopter crew (the

pilot and myself) to nests obstructed from observation by

marsh vegetation. This repeated circling over nesting areas

allowed ground crews sufficient time to reach the nests.

During this circling period, many nests which were partially

hidden by vegetation or missed during the strip surveys were


During surveys, nest locations were recorded on aerial

photographs (1":800'; Florida Department of Transportation,

Tallahassee, Florida) and/or by loran or GPS. Nest status

was recorded as active or successfully hatched, depredated,

false, flooded, or of unknown fate (Jennings et al. 1988).

Active nests were defined as mounds of vegetation

approximately 1.5m in diameter and 0.5 to 1.0m in height

with no obvious scattering of vegetation, holes, or other

evidence of tampering. Other evidence of an active nest

site included the presence of an alligator of the

approximate size of a nesting female (1.8-3.0 m), well-used

trails leading to the nest, or fresh (green) vegetation

incorporated into the nest. Nests were deemed hatched

during late-incubation surveys if a semi-circular excavation

from the top center down to the bottom of the egg cavity

(Deitz and Hines 1980) was observed. Nests were considered

flooded if submerged by two-thirds or more (Jennings et al.

1988). False nests were small, contained vegetation that

was generally scattered and not neatly incorporated, and

sometimes located near large active nests. Experience with

the particular study area, knowledge of nest size when

comprised of a particular nesting material, and ground-

truthing were essential for determination. Depredated nests

were distinguished as being flattened with nest material

scattered and alligator eggs or shell fragments littering

the immediate area (Jennings et al. 1988).

Total production was estimated through a combination of

aerial counts and estimates of hidden nests and predation.

The latter estimates were necessary to effect the 50%

removal of total surviving production.

The following notation and equations were originally

discussed in Jennings et al. (1988). Overall proportional

predation rate (D,) was estimated by:

D, D,
D=1-[ (1- ) (1- ]
A1 A

where D, = number of depredated nests observed during aerial

survey and before egg removal; A, = total nests observed

during the first aerial survey (conducted prior to egg

removal); D2 = number of nests sighted during aerial survey

that were depredated after egg removal; and, A = number of

nests remaining after egg removal. A2 was calculated by:

A= [ (A -D) -R]+A

where R = number of clutches removed and A, = number of

nests found during subsequent surveys that were not found on

the first survey. Nests which lost identifiable visual

characteristics or located in dense vegetative cover which

obscured visibility in later surveys were considered to have

incurred predation rates in proportion to those nests with

known final status. On Lake Woodruff, annual predation

rates were calculated by dividing the total number of

depredated nests observed by the number of nests found

during aerial survey.

Due to obstructive vegetation (trees and other tall

vegetation), time of day (early morning shadows), and

experience of observers (Rice 1992) some unknown proportion

of nests were not observed during aerial survey. Therefore,

an attempt was made to estimate the number of hidden or

missed nests. In September and October of each year,

hatchling pod surveys were conducted on each study area

utilizing a 4-5m airboat (observer's seat approximately 2 m

in height) operated at approximately 5 km/hour. A 200,000

candlepower hand-held light was used to search for alligator

eye reflections. When a hatchling was found, the survey was

suspended until an adequate count of any associated animals

(pod) could be made and the location recorded on aerial

photos and loran or GPS position noted. Later, utilizing

known nest and hatchling pod locations, pods were associated

with known or hidden nests. Consequently, the minimum

number of nests that were not observed from the air (H) was

estimated by:


where P = number of hatchling pods found during post-

hatching night-light surveys that were not associated with a

nest observed from the air.

Finally, total nesting (N) was estimated by:

N=A +Ai+H

An implicit assumption was that nest survival for unobserved

nests equaled that of observed nests.

Survey Procedures

In accessible habitats, night-light counts can be used

to provide an index of crocodilian population changes (Wood

et al. 1985, Brandt 1989, Webb et al. 1990, Woodward and

Moore 1990). In this investigation, night-light counts were

used to measure response of alligator populations to an

annual removal of 50% of total estimated production.

The following methods generally follow those of

Jennings et al. (1988). Night-light surveys (2/year) were

conducted in late May or early June on lakes Griffin, Jesup,

and Woodruff. Survey routes generally followed the open

water-shoreline interface (Murphy 1977, Woodward and Marion

1978). Dense marsh, wooded swamp, and other inaccessible

alligator habitat were not surveyed. On Griffin, surveys

were confined to the open lake; peripheral canals were not

surveyed. On Jesup, all associated creeks were surveyed

when water conditions allowed access. On Woodruff, only

selected canals and creeks were surveyed. However, in each

case, surveys were standardized across years.

Searches for alligator eye reflections were conducted

with an airboat at a planing speed of 20-25 km/hr, depending

on water conditions. When dense groups of alligators were

encountered, the airboat was sufficiently slowed to allow a

thorough count. A 200,000 c.p. spotlight was used and size

of detected alligators judged as approaching them at normal

survey speed.

To estimate alligator size, both the snout length:total

length (TL) index described by Chabreck (1966) and a general

impression of size, periodically calibrated by catching and

measuring size-judged alligators, were used. Alligators

were classified in 30cm (1-ft) size classes when possible or

placed into broader TL categories (0-60 cm, 61-121 cm, 122-

182 cm, >122 cm, and >183 cm) when specific size class could

not be determined but other indications of size were evident

(e.a., habitat, eye reflection, bubble trails, size of

splash or wake). Size was classified as "unknown" when no

indication of size was apparent.

Water levels were recorded from U.S. Geological Survey

water level gauge stations in permanent locations on the

main water bodies. The "water level" analyzed for each

survey date was departure from mean water levels (LD1WL) for

all surveys (see Woodward and Moore 1990). Hydrilla

(Hydrilla verticillata) data represented proportion coverage

of the lake by hydrilla based on annual aquatic plant

surveys conducted by the Florida Department of Natural



For trend analysis, the size distribution of unknown-

size alligators was assumed consistent with the distribution

of known-size alligators. Unknown-sized animals were

apportioned accordingly in 4 TL classes [30-121 cm

("juvenile"), 122-182 cm subadultlt"), >30 cm ("total

population"), and >183 cm ("adult")] (see Woodward and Moore

1990). During the May-June surveys, hatchlings from the

previous year (approx. 9 months old) were considered to be

sufficiently dispersed to render their sighting

probabilities as independent, and were included in the


Within areas, tests for trends in count densities were

conducted by regressing log-transformed counts of alligators

in each general size class on elapsed time, DMWL, and

proportion coverage by hydrilla (Griffin only). I IIWL was

previously found to explain a significant proportion of

variation in night-light surveys (Woodward and Moore 1990).

Only Griffin had sufficient hydrilla coverage to allow the

additional variable to be considered. For comparing the 2-

covariate model for Griffin with the 1-covariate model for


other areas, adjusted R2 (Rawlings 1988) which allows valid

goodness-of-fit comparisons among regression models having

different numbers of parameters was reported (Woodward and

Moore 1990).

Size Distributions

I measured changes in the population size structure of

the study areas both before and after harvest and between

the experimental areas and the control. I compared both the

shape and means of the size structure distributions from

each area. For comparisons of distribution shape between

the areas, I employed the Kolmogorov-Smnirnov test for

differences in distributions (SAS Institute, Inc. 1988).

This test calculates a cumulative distribution function

(CDF) for each distribution, the proportion of values equal

to or less than some point, r, on the distribution. The

test then defines the maximum vertical distance between the

two CDF curves and performs a significance test based on

that distance. I tested the difference in the mean of each

distribution utilizing a t-test.


Nest Production

From 1987-1991, 726 clutches representing 33,895 eggs

and resulting in 15,028 hatchlings were collected from lakes

Griffin and Jesup (Table 2-1).

Removal rates of pods and clutches were variable,

although the target of 50% mean removal of the estimated

annual production was achieved in most years (Table 2-2).

On Lake Griffin, dense canopy cover obscured a higher

proportion of the nests and minimum nest estimates were more

variable. Nests were more visible on Lake Jesup occurring

principally among Phragmites australis in open marsh. Lake

Woodruff nests also were very visible with little canopy

cover and primarily occurred on levees and in open marsh

comprised of Spartina baker.

The importance of unobserved nesting (via aerial

survey) was evidenced by its large contribution to the total

nesting estimate on Lake Griffin (Jennings et al. 1988;

Table 2-2). However, consistency in nest surveys provided a

reliable indicator of true nesting effort on all areas. Few

pods not associated with known nests were found on lakes

Jesup and Woodruff (Table 2-2).


Nesting on Lake Woodruff was very stable throughout the

study, ranging from 52 in 1989 and 1991 to 56 in 1987 (Table

2-2). Nesting on Lake Griffin ranged from 160 in 1988 to

191 in 1989 (Table 2-2). Nesting estimates on Lake Jesup

ranged from 113 in 1989 to 135 in 1990. Mean observed

nesting was higher on lakes Griffin and Jesup in this study

than the study reported in Jennings et al. (1988) due to the

use of fixed-wing aircraft by Jennings in 1982 and, perhaps,

an increase in adult populations (Woodward and Moore 1990).

Removal rates in most years were reasonably close to the

target of 50% of estimated production (Table 2-2).

Depredation rates varied both within and among study areas

(Table 2-2). Lake Woodruff had a higher mean nest

depredation rate (0.384)than the other lakes (P < 0.04;

Griffin = 0.146; Jesup = 0.230). All lakes exhibited high

annual variation in depredation rates (Table 2-2).

Night-light Surveys

The regression model with independent variables year,

water level, and hydrilla (for Griffin) explained a

substantial proportion of the variation in counts of

juvenile, subadult, and the total population of alligators

(R2 dj > 0.30) for all areas (Table 2-3). However, for

adults, the model explained a meaningful (R',, > 0.63)


proportion of the variation only on harvest areas (Table 2-


Count densities of juvenile alligators increased (P =

0.006) on the control area, Woodruff, but showed no obvious

trends (P > 0.11) on harvest areas (Table 2-3; Figs. 2-1, 2-

2, 2-3). Count densities of subadult alligators increased

(P < 0.02) on Griffin and Jesup but no trend (P > 0.08) was

observed on Woodruff (Table 2-3; Figs. 2-1, 2-2, 2-3).

Densities of the total population increased on all areas (P

< 0.04; Table 2-3; Figs. 2-1, 2-2, 2-3). Count densities of

adult alligators increased on the harvest areas (P < 0.01)

but no trend (P > 0.80) was detected on the control area

(Table 2-3; Figs. 2-1, 2-2, 2-3).

Count densities of all size classes of alligators

increased (P < 0.01) with decreasing water levels on all

areas with the exception of adult animals on Lake Woodruff

(P > 0.70) and subadult animals on Lake Griffin (P > 0.40;

Table 2-3). On Griffin, count densities of juveniles and

the total population of alligators increased (P < 0.04) with

increasing hydrilla coverage, count densities of adult

alligators showed some evidence (P = 0.08) of a similar

response, and count densities of subadult alligators

decreased (P = 0.05) with increasing hydrilla coverage

(Table 2-3).

Over the 11-year study period, estimated population

changes of juvenile alligators ranged from an increase of

190% on Woodruff to an increase of 24% on Jesup (Table 2-3).

Changes in subadult populations ranged from an increase of

166% on Jesup to an increase of 83% on Woodruff (Table 2-3).

For the adult population, changes ranged from a decrease of

7% on Lake Woodruff to an increase of 137% on Lake Griffin.

Finally, changes in the total population ranged from an

increase of 50% on Lake Griffin to an increase of 96% on

Lake Woodruff.

Cumulative Distributions

Utilizing the Kolmogorov-Smirnov test, I could not

detect a significant difference in the shape of the size

distributions from pre- and post-harvest counts on either

the harvest or control areas (P > 0.90; Figs. 2-4, 2-5, 2-

6). Similarly, there was no difference in the means of the

distributions on either harvest or control areas (P > 0.50).


The adjustment to nesting estimates for unobserved

nests was conservative because:

1) pods from hidden nests in close proximity to other
known nests may have been indistinguishable from those
from observed nests;

2) some pods from unobserved nests were not detected
during night surveys; and,

3) some clutches from unobserved nests may have failed to
hatch due to reasons other than depredation.

Therefore, the estimates of removal rates may have been

biased high and the actual proportion of annual production

removed slightly less than estimated.

Count densities on treatment areas indicated that a 50%

harvest had no detrimental effect on the subadult population

of alligators. Based on mean growth rates of alligators,

all cohorts within the subadult size class were exposed to

harvest during their hatchling year. Trends in counts of

subadult populations indicate that recruitment to the adult

size class was stable or increasing. Because few cohorts

would have attained adult size, the consequences of

hatchling removals should have been minimal on the adult

size class. However, a trend was detected in increasing

adult size classes on both harvest areas. The increasing

juvenile population on Woodruff suggests that harvests on

treatment areas may have been compensatory at the point

where recruitment into the subadult population occurred.

Population trends during this 11-year study differed

somewhat from the those found by Jennings et al. (1988)

after evaluating the same harvests over the first 6 years.

However, I did not use pre-harvest (1980) data from lakes

Woodruff and Jesup as did they, and I evaluated different

size classes. The juvenile and total population size

classes included hatchlings 9 months of age. In general,

these data concurred with Jennings et al. (1988), indicating

that 50% removal of hatchlings had no adverse effects on


Although water level had a significant effect on the

number of subadult alligators counted on lakes Woodruff and

Jesup, no effect was found on Lake Griffin. Similarly, on

Lake Griffin, hydrilla coverage significantly affected

counts of all size classes except the subadult animals.

Subadult animals begin to change food habits to that of

adult alligators and, therefore, begin to compete for the

same resources. However, the subadult cohort is still

susceptible to predation by the larger size classes.

Perhaps, on Lake Griffin, the lack of cover in hydrilla beds

and offshore during low water levels might overcome the

advantage of concentrated food items in those areas. Also,

the cost to fitness of surviving in sub-optimum habitat

might be offset by the decrease in predation by larger


No trend in the adult size class on Lake Woodruff was

detected over the study period. However, a significant

increasing trend was detected in juvenile animal counts.

Emigration into adjacent marshes by adults and immigration

by juveniles was certainly possible. Other explanations

would include a decrease in survival in the subadult size

class due to overcrowding and an exodus from relatively

sheltered, but suboptimum, habitats to areas with

populations of large animals. Increases in the observed

populations of subadult animals on harvest areas provide an

excellent platform to study the dynamics of cannibalism in

alligator populations.

The minimum population change detected in this study

over 11 years was an increase of 50% (3.4% annual rate of

increase) for the total population on Lake Griffin. A 24%

and a 31% increase in the estimated counts were not

statistically significant. Accounting for major sources of

variation and reducing random experimental error lowers

overall variation and reduces both minimum detection size of

trends and number of years of data needed to discern trends

(Harris 1986). Increased power in trend detection could be

achieved by adding an additional independent survey each

year (Harris 1986; Woodward and Moore 1990). Additional

counts would most certainly be needed to detect similar

trends over shorter periods.

The increases in populations of alligators on both

harvest areas strongly suggest compensatory mechanisms at

work. No increase in juvenile animals was detected on

harvest areas similar to the increase on the control.

Conversely, an increase in subadult counts and the total

population was detected. Some immigration could have

occurred since both harvest areas are connected to much

larger riverine systems. However, the magnitude of change

in the populations points to other compensatory agents such

as increased survival and growth rates. This study did not

address those issues, but did provide a basis for further

research into the dynamics of exploited populations.

Other areas of future research should concentrate on

newly instituted management alternatives of the GFC

including the effects of simultaneous juvenile and adult

harvest, harvest on small scattered wetlands, and harvests

of the subadult size-classes.

Table 2-1. Number of alligator hatchlings and nests removed
from lakes Griffin and Jesup from 1987 to 1991.

Lake Griffin


Lake Jesup





















Totals 9170







rJumrber of eggs collected presented









in parentheses.

Table 2-2. Alligator nesting effort, nest success, and
estimated removal rates for 3 lakes in central Florida from

Nests Hidden Total Predation Removal
Year Observed Nests Nests Rate (%) Rate (%)

Lake Woodruff

1987 56 0 56 52

1988 50 5 55 31

1989 52 0 52 26

1990 54 0 54 43

1991 52 0 52 40

Lake Griffin

1987 142 21 163 8 50

1988 141 19 160 7 51

1989 154 37 191 24 47

1990 151 28 179 16 51

1991 155 22 177 18 52

Lake Jesup

1987 123 2 125 14 45

1988 123 4 127 32 40

1989 112 1 113 28 46

1990 135 0 135 30 42

1991 130 0 130 11 47

Table 2-3. Mean densities, trends, and regression results for control area, Lake Woodruff
and harvest areas, lakes Jesup and Griffin from 1981-1991 (1979-1991 for Griffin).

Year Water level Hydrilla
Study area x density Est. pop.
Size class (cm) (#/km) Change (%) trend/yr P r P r P R2ad

Woodruff (n=20)

30-121 5.32 +190 0.107 0.006 -0.720 0.001 0.657

122-182 1.35 +83 0.060 0.088 -0.655 0.001 0.465

231a 8.67 +96 0.067 0.035 -0.677 0.001 0.537

183" 1.89 -7 -0.007 0.830 -0.066 0.782 -0.110

Jesup (n=20)

30-121 9.73 +24 0.021 0.451 -0.858 0.001 0.716

122-182 2.18 +166 0.098 0.001 -0.760 0.001 0.782

231 14.46 +51 0.041 0.037 -0.789 0.001 0.818

183 2.46 +89 0.064 0.003 -0.692 0.001 0.663

Griffin (n=21)

30-121 18.24 +31 0.022 0.117 -0.573 0.001 0.497 0.030 0.487

122-182 1.34 +155 0.078 0.011 -0.106 0.498 -0.347 0.049 0.309

L31 23.52 +50 0.034 0.008 -0.485 0.001 0.452 0.036 0.534

183 3.74 +137 0.072 0.001 0.418 0.046 0.342 0.088 0.631

'Represents the total population.
"Represents total adults.

Total Population

1.2-1.8 m Size Class


,---------------i --.-

0.3-1.2 m Size Class

1980 1982 1984 1986 1988 1990 1992

Figure 2-1. Population trends of 3 size classes of
alligators based on night-light surveys conducted on Lake
Woodruff from 1981-1991. Counts adjusted for water level.

[ Total Population


1.2-1.8 m Size Class

0 *

0.3-1.2 m Size Class

* 0

1980 1982 1984 1986

1988 1990 1992

Figure 2-2. Population trends of 3 size classes of
alligators based on night-light surveys conducted on Lake
Jesup from 1981-1991. Counts adjusted for water level.

Total Population

0 0 v

1.2-1.8 m Size Class

V 0

(-^ ^0
^^** 0

0.3-1.2 m Size Class

@* .00-

0 0 00

~~ *

1098 1 0 1 2 1 4
1978 1980 1982 1984 1986

1988 1990 1992

Figure 2-3. Population trends of 3 size classes of
alligators based on night-light surveys conducted on Lake
Griffin from 1979-1991. Counts adjusted for water level
and hydrilla coverage.

lIill[11 Inill


.... ..................

0.8 -





4j 0.4



1981-1983 1989-1991

0 1 2 3 4
Size (m)

Figure 2-4. Cumulative distribution of alligator size
during night-light survey on Lake Jesup.




0.6 -


4J 0.4

0.2 -

1979-1981 1989-1991

0 1 2 3 4
Size (m)

Figure 2-5. Cumulative distribution of alligator size

during night-light survey on Lake Griffin.




- 0.6

4* 0.4



1981-1983 1989-1991

0 1 2 3 4
Size (m)

Figure 2-6. Cumulative distribution of alligator size
during night-light survey on Lake Woodruff.



In general, there are two types of population models:

theoretical and simulation models (Logan 1994). The

difference between the two lies with the principle of

parsimony. Theoretical models generally begin with the

simplest possible structure needed to address the questions

posed. Such models add parameters and variables only when

necessary. Simulation models on the other hand are

typically mechanistically detailed and may contain many

hundreds of parameters and variables.

"There is nothing inherently wrong with complex models,
just as there is nothing inherently correct with simple
models; it is more a question of appropriateness"
(Logan 1994, page 203).

However, regulatory agencies desire straightforward, simple

models that can be easily used in the regulatory process

(Heinz et al. 1994).

The Florida Game and Fresh Water Fish Commission (GFC),

as the management authority for the alligator in Florida,

required a systematic manner in which to evaluate possible

harvest alternatives. This evaluation entailed

consideration of population effects and the satisfaction of

several GFC constituencies. Therefore, at the request of

the GFC, I developed an alligator population model designed

to address certain questions and management concerns:

1. Utilizing data collected previously on Orange Lake
(the best studied alligator population in Florida),
what parameters are most important to population growth
potential under assumptions made in the model? What
information is most lacking in understanding alligator
population dynamics?

2. How does the Orange Lake population model react to
various politically and economically reasonable harvest

3. Can the model be written in sufficiently general form
such that it can be used by the GFC to examine
management questions about other alligator management
units in Florida?

One question in this study dealt with comparisons of

various alligator populations across Florida, many of which

had not been studied in any great detail. Necessarily, this

entailed ignoring some important processes (area-specific

effects such as water level, weather, etc.), combining

others (growth and survival rates across size classes), and

utilizing several parameters estimated only for the Orange

Lake population as representative of all areas. Therefore,

this model was of 'intermediate' complexity, somewhere

between the 'big ugly' models and the theoretical ones

discussed by Logan (1994). The degree of uncertainty in

this model was identified and acknowledged. Under this

assumption, the simple model can be effective (Heinz et al.



Alligator Population Model

Most crocodilian population growth and, therefore,

crocodilian models are typically deterministic with rather

low noise levels because factors affecting recruitment into

the adult population are fairly stable (Hall 1990).

Stochastic processes and rare events (e.g., loss of a year's

production from weather events) have little long-term effect

since the reproductive population is dispersed over many

annual cohorts (Abercrombie 1989; Hall 1990).

Age-specific models (i.e., each annual cohort is

modeled) for crocodilian populations have been described in

the literature (Nichols et al. 1976). However, stage-based

models (i.e., several annual cohorts present in a single

stage) can be more appropriate for species which are

difficult to age and place into distinct yearly cohorts and

which have distinct stages of exploitative or behavioral

strategies (Heppell et al. 1994; Heppell et al. 1996). Over

the lifespan of the alligator, an animal enters several

distinct life stages relating to size (Nichols 1987).

Survival and growth rates can be quite variable between

these stages but are assumed to be constant within each


This model considered 7 life stages: egg, hatchling,

juvenile (0.3-1.2 m size class), subadult (1.2-1.8 m size

class), adult female (>1.8 m female size class), adult male

(1.8-2.7 m male size-class), and mature male (>2.7 m size

class). Survival and transition (the model surrogate for

growth rates) were assumed to be equal within stages.


Survival rates for eggs were estimated directly from

area-specific data. Survival was a combination of nest

survival (P,) and clutch viability (proportion that hatch

out of total eggs deposited; V) or P, V (for the Orange

Lake population, P,, = 0.46). Nest survival was estimated

via direct observation by GFC and Florida Cooperative Fish

and Wildlife Research Unit (FCFWRU) personnel over the last

decade (for procedures see Chapter 2). Clutch viability was

similarly estimated by clutch inspection and artificial

incubation (for procedures see Chapter 4). Hatchling

survival was estimated by Woodward et. al (1987) and assumed

equal across areas (Ph = 0.41).

For other stages, a stable size distribution was

defined as the pre-harvest levels of the population in

question (measured by night-light counts see Chapter 2).

While in most areas, this assumption was not explicitly met,

all areas had been under governmental protection for over 15

years prior to recent harvests (Woodward et al. 1992).

Static life-table analysis was used to develop survival

rates under the assumptions of a stable size distribution

and constant environmental effects (Caughley 1977; Webb and

Smith 1987; Nichols et al. 1976). Alligators are a long-

lived species and adult alligators have an effectively

benign environment (low mortality) in Florida. Therefore,

extreme environmental effects on the population should be

most apparent in the early life stages (estimated separately

- see above; Hall 1990). Also, the 215 year period without

exploitation should have allowed incorporation of the unique

effects of environment and habitat of each area. In

addition, the question of relative effects of harvest across

several areas necessitated a computationally simple method

to assign survival rates with area-specific data and a

method in which the differences in population structure

across areas could be highlighted. Certainly, direct

estimation of survival would have been more desirable, but

was not possible in this study. Again, the objectives of

this effort were to compare relative effects of various

management strategies, not predict absolute population


Stage specific survival rates were assigned utilizing

(Caughley 1977):

P -L

where P,-j was the probability that an animal would survive

to stage j given it was alive at stage i, L, was the number

alive in stage i, and Lj the number alive in stage j. I

assumed that the survival across stage i was constant and

that the two stages, i and j, had independent survival

probabilities. Annual survival rates were then assigned

(not estimated) by:

6=( L/t; lit
L /t

where t, was the mean transition time from stage i to j, t,

was the mean transition time from stage j to j+1 and L,/t

represents the mean animals alive per time step within stage

i. Using this process, annual survival rates for stages of

the Orange Lake population were assigned as: juvenile

0.77, subadult = 0.74, adult female = 0.89, adult male =

0.88, and mature male = 0.86. Population response was

simulated monthly to allow for nesting and harvesting during

appropriate seasons; survival rates were adjusted


Transition rates

Growth rates and growth functions have been estimated

for several crocodilian species (Nichols et al. 1976;

Brisbin 1988; Abercrombie 1992). Transition rates between

life stages were of primary interest in this study.

Transition rates were defined, under the assumptions of a

geometric distribution, as:


where d, was the proportion of animals that developed into

the next life stage over a time step, 1-d, remained in the

same stage, and t was the mean time in stage x (Caswell

1989). The model was simulated monthly. Development was

assumed to peak in summer months and lull in winter as a

cosine wave, thus, mimicking the growth cycle in an

alligator population:

d=d,(1 +0.95cos((-)(m-6)))

where d, was the proportion of animals that developed into

the next life stage in a given month (m) and June (m = 6)

was assigned as maximum growth.


The only density dependent factor included (this

particular factor adjusted densities, but did not control

maximum population growth; a more comprehensive density

dependent control will be discussed later.) was the

proportion of adult females nesting in a given year (FJ).

This function was defined by:

F =f[- +1]
S t

where f was the minimum proportion of adult females that

nested in a given year, x, was the number of females in the

'stable' field population, and x, was the number of adult

females in the population that year. The proportion of

females nesting could vary between f and 2*f (for the Orange

Lake population this resulted in a proportion between 0.17

and 0.34). Since a monthly time step was used, all nesting

effort occurred during July and F, was set to 0 for the

remainder of a given year. Since reproduction was

calculated based on the number of surviving females from the

previous time step, F, was them multiplied by adult female

survival over a time step.


In this model, fertility was then defined as the number

of eggs produced per year or:


where R was the clutch size.

A function also was incorporated to maintain at least a

1 to 5 proportion of adult males to females. I felt that at

least this proportion of adult males should be maintained to

allow full reproductive potential (male to female

proportions in this general range are utilized in captive

breeding operations). If the modeled population falls below

this threshold, reproduction is reduced by a similar

proportion as:

x /xf
Eggs=Eggs ( )

where x, was the total number of adult males and x, was the

number of adult females. This equation was only used when

the proportion of males to females was below 0.20.

Other density dependent factors

Other density dependent factors probably are important

to alligator populations (Abercrombie 1989). Survival,

growth, cannibalism, egg viability, etc. all may be affected

to some degree by the density of the particular population.

A unique system of these relationships may have existed for

each alligator population in Florida. Few data were

available concerning these factors for even the most studied

alligator population, Orange Lake. Therefore, for the

purposes of the primary questions of this model, those

processes were ignored. I assumed (since the alligator is

long-lived and the age at first reproduction is >10 years)

any density dependent factor that exists will not affect the

population's reproductive potential for some length of time

(Abercrombie 1989). Therefore, by using an appropriately

short simulation time (20 years), those factors were either

ignored or assumed equal across areas. In addition, as

stated in the principle of parsimony, this model used only

those factors necessary to address the questions posed.

Certainly, cannibalism rates are important to alligator

populations. However, any effects of cannibalism to the

reproductive potential of the simulated population would

necessitate incorporating a time-lag function. Predation on

the smaller size classes by adults would decrease the number

of animals available to enter the larger size classes.

However, due to slow growth of alligators, this cannibalism

effect would not appear for some long (~15 years) period of

time. By using a short simulation time, the cannibalism

effect on reproduction should have been minimal. This

concession was important to this model because the

cannibalism functions were probably area-specific.

The density-dependent function for the proportion of

adult females nesting did not operate to fully control the

growth of the population. At the minimum assigned value for

the proportion nesting (0.17) and without harvest, the

population could increase to infinity over a long simulation

period (an exponential growth model). However, either with

harvest or over short simulation times the population

remained within reasonable bounds of density. I did

incorporate a function to assign a carrying capacity (or

maximum) for nests into the model. Obviously, the habitat

could only support a finite number of nesting sites

regardless of the number of females in the population. Even

so, the population did not approach this carrying capacity

during the simulation period used for reasons outlined above

(Fig. 3-1).

Environmental factors

Certainly the environmental and habitat constraints of

a given area were important in structuring the associated

alligator population. Water level in particular has been

shown to affect alligator nest production (Joanen et. al

1977; Percival et. al 1992). Nesting and hatching success

also are affected by water levels and flooding events

(Abercrombie 1989). However, these factors are extremely

variable among areas. For example, in a given year on 2

central Florida lakes, Lake Jesup might lose all production

to flooding of nests whereas Lake Griffin, with its water

control structure, might lose only a small percentage (pers.


I hypothesized that over time the stable size

distributions used reflected these area-specific and random

events, especially how they apply to the survival of the

reproductive stages. The manager should keep in mind

factors such as these and adjust the management scheme

chosen accordingly.

The model

The model was prepared using Matlab software (The

Mathworks, Inc.). I used a modified Leslie transition

matrix (Leslie 1945) to simulate the population over time

(Appendix A). The modifications consisted of diagonal terms

in the transition matrix to account for animals which did

not develop into the next higher size class in a given time-

step. Within a life stage, the transition matrix contained

terms for survival, development, and harvest. The fertility

term, summed over the reproductive female age groups, was

the input for the first life stage (eggs). Inputs to the

model included the stable size distribution (input as the

numbers of animals present in a given stage), harvest

levels, and mean stage occupancy time (to calculate

transition rates) by life stage, minimum proportion of

females nesting, clutch size, egg survival, and nest

survival. Outputs included graphs of the population

density, animals harvested, and gross value of harvested

animals over time. Gross value per animal by stage was

assigned as: eggs $12; hatchlings $20; 0.3-1.2 m stage -

$0 (assigned as $0 because this size class is not harvested

in Florida); 1.2-1.8 m stage $235; adult female stage -

$449; 1.8-2.7 m male stage $449; and, >2.7 m stage -


Sensitivity Analysis

Sensitivity analysis measures the relative importance

of errors in parameter values. These errors were inherently

dependent on the model assumptions (Conroy 1993).

Sensitivity analyses were conducted for proportions of

females nesting, survival across each size class, and

transition into the 1.2-1.8 m and adult size classes. Since

these parameters were proportions, they were varied between

0 and 1, by 0.01 increments. The Orange Lake population was

simulated (20 year simulations) at each increment. Each

parameter was then judged as to its inherent effect on the

simulated population's growth.

Elasticity Analysis

To further examine the relative importance of each

parameter in the model, I performed a population elasticity

analysis. This allowed comparisons of the proportional

effects of changes in the parameters on the population.

Elasticity analysis was normally performed by examining the

proportional change in the dominant eigenvalue (A = e') with

changes in the parameter values (Heppell et al. 1996;

Caswell 1989). However, in this model, I have allowed the

proportion of adult females nesting to vary with a density-

dependent relationship. Therefore, comparisons within a

matrix described by a stable deterministic growth rate were

not possible. Consequently, I modified the elasticity

analysis described by Heppell et al. (1996) to compare the

proportional change in a simulated stable population by (the

actual population size was substituted for the dominant

eigenvalue (A)):

N -N
Elas(x)=- x~' x'
N (2p)

where ., is the end of simulation population size at the

initial value of parameter x and p is the proportional

change in the value of x.

Simulation of Various Harvest Strateaies

A spectrum of harvest options was considered for

simulations on the modeled Orange Lake population. These

included 0, 25, and 50% egg harvests, 0, 5, 10, and 15%

adult harvest (>1.2 m), and each combination of both egg and

adult harvest. These simulations were ranked by: modeled

density (at year 20), annual monetary value of harvest,

annual number of animals in harvest, and the population

growth rate.

The measures utilized to compare harvest strategies

were chosen to optimize several factors both intrinsic and

extrinsic to the population and are outlined below.

Population health

Certainly, a politically and ecologically sound

alligator harvest program must not cause any dramatic

declines in the population. The alligator is a CITES

Appendix II species, and its harvest is carefully monitored

locally, nationally, and internationally (Thorbjarnarson

1992). This factor was measured by examining modeled

population density and growth.

Hunter satisfaction

A successful management program must have the support

of its constituents. The program will be ultimately

supported politically and economically (through user fees)

by these groups. An unpublished survey of hunters by the

GFC in 1989 and 1993 listed sport hunting and economic gain

as primary motivations for hunt participation (M. Jennings

and D. Carbonneau unpub. data). Increased season length and

number of alligators harvested were listed as improvements

needed in the program. Therefore, the number of animals

harvested and monetary value of the harvest were chosen to

evaluate hunter satisfaction.

Rancher satisfaction

Another key constituency of the alligator management

program in Florida was ranch owners. Ranchers utilize the

wild resource to stock their captive-rearing operations.

They are primarily businessmen and, as such, can provide

political support for management programs. Numbers of eggs

and the monetary value of the egg harvest were used as an

index to rancher satisfaction.

Aesthetic value.

Viewing of alligators, especially large animals, is

vital to the maintenance of certain tourism operations. The

GFC has received complaints concerning the loss of large

alligators for viewing in areas shared by the harvest

program and tourism concessions (D. David pers. comm.).

Whether this is due to actual harvest or increased wariness

of a hunted population was debatable. However, the number

of adult animals present in the final population was

utilized as an index to this factor.


Sensitivity Analysis

Proportion nesting

I allowed f to vary between 0 and 1, by 0.01

increments, and simulated the Orange Lake population (20

year simulations) at each increment (Fig. 3-2). As

expected, this variable was the most important to overall

growth of the population as evidenced by the final density

of the modeled population. While caution has been suggested

in utilizing this type of analysis to validate the values of

parameters (Conroy 1993), I saw that the value of f chosen

by field observation and utilized in the model fell within

reasonable bounds of density projection (actual observed

density of Orange Lake population was approximately 1.3

animals/ha.; eggs are not included in the plotted



I allowed the annual survival rates of each size-class

to vary independently between 0 and 1, by 0.01 increments,

and observed the impacts on the modeled Orange Lake

populations (20 yr. simulation at each increment). As

expected, a threshold for survival of eggs and hatchlings

existed above which the population remained fairly stable

and below which the population declined (note that egg

survival included both egg viability and nest survival;

Figs. 3-3 & 3-4 respectively). Estimates utilized in this

model were well above this threshold value. Also, note that

any error, even a fairly large one (due to density-dependent

effects or estimation error) would have made very little

difference in the simulated outcomes.

Adult male size-classes operated in much the same

fashion (Figs. 3-5 & 3-6). However, no threshold existed

due to the assumption that adult males need only be present

in a 1 to 5 proportion to nesting females for full

reproduction to occur. I felt comfortable with this

assumption due to the harvest levels considered and the

wariness of adult males after harvest initiation (Woodward

et. al 1992).

Survival of adult females, 1.2-1.8 m, and 0.3-1.2 m

animals was very important to the density of the population

(Figs. 3-7, 3-8, & 3-9 respectively). In some cases, an

error of 5-10% would double (or halve) the modeled

population. Interestingly, of the 3 size-classes listed

above, changes in survival of the 0.3-1.2 m size class

caused the greatest change in the magnitude of population

density. Presumably, this was due to the large number of

animals present in the size class and the relatively high

annual survival in the larger size classes; although there

was a 9 year lag between animals entering the 0.3-1.2 m and

adult female size class. The values estimated from the

assumed stable size distribution fell within reasonable

bounds of density (again, comparisons were to the 1.3

animal/ha, density found with the Orange Lake population) in

all cases.

Transition rates

Transition into the 1.2-1.8 m and adult female size

classes also was very important to the modeled outcomes

(Figs. 3-10 & 3-11). The plots represented the mean monthly

proportion of those animals eligible for promotion into the

next larger size class that were promoted. Note, there was

a portion of the curve where a small error (even as low as

0.01) could mean a large change in the population.

Unfortunately, the estimated parameters in this model fell

within that range. The flat portion of the curve above this

region was unreasonable in a wild population. In addition,

note that in this model the development into the adult

female size class was synonymous with the age at first



The results of this analysis generally followed that

outlined above where the parameters were examined across the

spectrum of possible values (sensitivity analysis). In this

instance, the parameter values were compared at 5% above and

below the values chosen with actual field data from Orange

Lake (Table 3-1).

Table 3-1. Population elasticity analysis with a 5% change
in model parameter, x, values.

x Elas(x)

Proportion adult females

Survival of adult females

Survival of 0.3-1.2 m
size class

Transition rate into
adult female size class

Survival of 1.2-1.8 m
size class

Transition rate into 1.2-
1.8 m size class

Survival of eggs

Survival of hatchlings

Survival of 1.8-2.7 m
male size class

Survival of >2.7 m size











To review, elasticity of a given parameter was defined as:
N -N
Elas(x)= X'p x-p
NX (2p)

where N, is the population size at the initial value of

parameter x and p is the proportional change in the value of


In general, the proportion of females nesting and the

density-dependent relationships under which it operates were

most important for understanding and, perhaps, future

research. Survival of the 0.3-1.2 m, 1.2-1.8 m, and adult

female size classes also was important. Density-dependent

relationships that might exist for survival and development

should certainly be evaluated for future constructions of an

alligator model. For comparisons of independent populations

of alligators, the transition rates into the subadult and

adult female size classes were important. Finally, the

environmental effects on survival, development, and

reproduction were ignored, but would be necessarily

important for any model that might approach meaningful


Simulation of Various Harvest Strategies

Each combination of adult and egg harvests were

simulated using parameter values representing the Orange

Lake population (Figs. 3-2 to 3-5).


Final densities (animals/km) of the simulated

population over the various proportional harvest strategies

(20 yr simulation) were provided (Table 3-2).

Table 3-2. Densities(#/km) of the simulated Orange Lake
alligator population at year 20 under various management

%Adult Harvest
%Egg Harvest Size class 0 5 10 15

0 all 2.75 1.90 1.35 0.92

>1.2 m 0.65 0.45 0.30 0.20
25 all 1.90 1.35 0.90 0.60

>1.2 m 0.55 0.35 0.20 0.15
50 all 1.20 0.80 0.55 0.35

>1.2 m 0.40 0.25 0.15 0.10

The density of the alligator population on Orange Lake was

approximately 1.3 animals/km (post egg-hatching). For

comparative purposes, to maintain this population under the

assumptions of this model, a 5-10% adult, up to 50% egg, and

a 25%:5% egg:adult combination harvest would all be

appropriate for examination. Both 10% adult and 50% egg

harvests have been utilized in Florida successfully

(Woodward et. al 1992; Percival et. al 1992; Chapter 2).

The number of adult (>1.2 m) animals was important for

consideration not only for their reproductive role but for

aesthetics in general (D. David pers. comm.). In this

model, the same harvest levels were important for

maintaining a visible adult population.

Economic value

Total monetary value of harvested animals per year

(adults + eggs) was calculated by assuming a mean dollar

price for meat yield, hide value, and egg value (Table 3-3)

The total value of harvest was maximized for a 10% adult

harvest. However, a 25% egg harvest could be incorporated

for ranching purposes. When combined with a 5% adult

harvest, the total value would decrease, but ranching fees

and an increase in the number of parties with an interest in

the harvest might increase overall return to the GFC.

Table 3-3. Total annual value (thousands of $) of simulated
Orange Lake harvests at year 20 under various management

A.u.ili t Hir'.'est
%Egg Harvest 0 5 10 15

0 0 50 60 52

25 11 50 50 45

50 19 45 40 30

Numbers of animals in harvest

Total numbers of harvested animals for the final annual

time-step were summed by adult and egg stages (Table 3-4).


Table 3-4. Total annual harvest of the simulated Orange Lake
population at year 20 under various management strategies.

%Adult Harvest
%Egg Harvest Size class 0 5 10 15

0 eggs 0 0 0 0

21.2 m 0 150 180 180

25 eggs 1000 700 450 300

>1.2 m 0 105 150 120

50 eggs 1600 1100 700 450

>1.2 m 0 100 90 80

Total number of adults harvested (which could be assumed to

increase hunter satisfaction) was maximized in the 10-15%

adult harvest range. Total egg harvest was maximized by a

50% egg harvest (assumed to increase rancher satisfaction).

For a combination harvest, the 50%:5% and 25%:5% egg:adult

harvests maximized hunter/rancher take.


The intrinsic rate of increase, r, was estimated for

the last time-step for each simulation as (Caughley 1977):

Nt tI

where N,+1 was the population size at the end of the

simulation and N, was the population size 1 time step

previous (in this instance, I used time steps 246 and 234

for N,,, and N, respectively which represent the last 2

reproductive time steps in the 250 month simulation).

The simulated population could maintain a 25%:5%

egg:adult combination harvest (Table 3-5). Interestingly,

the currently utilized 50% egg and 10% adult harvests also

maintained a positive rate of increase.

Table 3-5. Final estimated r for the simulated Orange Lake
population at year 20 under various management strategies.

"A:l]itl Harvest
Harvest 0 5 10 15

0 0.0169 0.0231 0.0053 -0.0147

25 0.0324 0.0095 -0.0088 -0.0356

50 0.0123 -0.0083 -0.0332 -0.0674


If each harvest strategy is judged as having met the

above criteria or not, the 5% adult combined with a 25% egg

harvest performed best in the modeled population. However,

this recommendation should be taken with caution; it should

be utilized only as a starting point for management. The

harvest levels suggested should be examined, populations

monitored, adjustments made, and the results incorporated

into future constructs of this model.

General Recommendations and Discussion

Following are general points, cautions, and

considerations for viewing model output:

1. Survival rates were assigned based on the assumption of
a stable size distribution. If the wild population was
actually increasing, there would have been fewer
animals in the larger size classes; thus, the survival
rates would have been biased downward. If the wild
population was actually decreasing there would have
been a greater proportion of alligators in the larger
size classes and the survival rates would have been
upwardly biased.

2. Density dependent survival factors (including
cannibalism) were not included in the model. Under
natural conditions some density dependent factors may
occur and under conditions of population decline,
compensatory survival could stabilize population

3. The model is most appropriate for short time periods.
Some decline or reduction in the population may not be
cause for concern. Rather, the magnitude of the
decline is important to note for the decision-making
process. Did the population decline abruptly over the
short term or did it appear to stabilize at some
reduced density?

4. Any management alternative suggested by this model
should be combined with other management tools such as
the night-light survey and nest counts to monitor
population response to harvest.

5. Any declines or increases in the population modeled may
be dampened or amplified in the wild population by
factors not present in the model such as density-
dependence, compensatory mortality, and environmental

Cautions and suggestions for the use of models for

wildlife management and conservation (Conroy 1993)should


1. Model predictions are not equal to knowledge and
careful scientific inquiry. Model output (numbers) and
parameter values utilized in the model should not be
considered as true and real.

2. "Tweaking" of model parameters to give results that
adhere to a managers concept of reality should be

3. Models should be viewed as a tool for prediction under
various management conditions. These predictions
should then be evaluated under field conditions and the
results incorporated into the model. Prediction rather
than prescription should be emphasized.

While not a wholly new concept, adaptive management has

become a "buzzword" in the ecological literature. The

concept entails that management of wildlife populations

should be conducted as a scientific experiment: a hypothesis

of management is conceived, tested, evaluated, and

adaptations made as needed (Lee 1993). However, unlike a

scientific experiment, adaptive management has the goal of

the achievement of management objectives. The essential

point concerns management for change; change as an integral

element of the management plan. Uncertainty in management

of wildlife species is inherent and has been recognized for

many years. I cannot model a population, especially a

spectrum of populations such as the alligator in Florida, to

such a degree that this uncertainty is accounted for in any

real way. Therefore, the real benefit of a model is to

identify a starting point for management. Managers should

utilize all the tools at their disposal to make decisions.

Again, if uncertainty is recognized, the manager can readily

adapt to the situation. Recently, Ludwig et. al (1993)

suggested that modifications to management strategies should

be incorporated before scientific consensus is achieved.

While this model contains many shortfalls and has ignored or

combined many important factors, I believe it provides the

manager with a starting point for scientific, adaptive

management of the alligator in Florida.

0.5' L i I
0 50 100 150 200 250
Time in months





2 2-


0 100 200 300 400 500
Time in months

Figure 3-1. Population density with no harvest for a
simulated alligator population. Density-dependent control
effects population growth only after long simulation



Estimated Field Population

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Proportion adult female s nest ing/yr

Figure 3-2. Simulated Orange Lake alligator population after
20 years with differing values for proportion of adult
females which nested.

? / Estimated Field Population
SII)........ ..........................................................

0 o. 5

0 0.1 0.2 0.4 0.5 0.6 0.7 0.8 0. .9 1
Survival f eigJs:

Figure 3-3. Simulated Orange Lake alligator population after
20 years with differing values for survival of eggs.


0 0. 2 0.3 0.4 0.5 0.6 ). 0.9 1
Survival of hatchlings

Figure 3-4. Simulated Orange Lake alligator population after
20 years with differing values for survival of hatchlings.



-H 1.35


1.3 Estimated Field Population


0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Survival of 1.8-2.7 m male stage

Figure 3-5. Simulated Orange Lake alligator population after
20 years with differing values for survival of the 1.8-2.7 m
size class.

Estimated Field Po ulation


).3 0.4 0.
Survival ot

0.6 0.7
m st age

.8 n.9

Figure 3-6. Simulated Orange Lake alligator population after
20 years with differing values for survival of the >2.7 m
size class.


e 3



r 1
, 0.5


Estimated Field Population
......................................... ...............

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Survival of adult females

Figure 3-7. Simulated Orange Lake alligator population after
20 years with differing values for survival of the adult
female size class.






a 2.5

S1.5 Estimated Field opulai


0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Survival of 1.2-1.8 m stage

Figure 3-8. Simulated Orange Lake alligator population after
20 years with differing values for survival of the 1.2-1.8 m
size class.


Estimated Field Population

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Survival of 0.3-1.2 m stage

Figure 3-9. Simulated Orange Lake alligator population after
20 years with differing values for survival of the 0.3-1.2 m
size class.


2 .


Estimated Field Population

S 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Development into 1.2-1.8 m stage

Figure 3-10. Simulated Orange Lake alligator population
after 20 years with differing values for transition rates
into the 1.2-1.8 m size class.


.H 2

.14 Estimated Field Population


0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Development into adult female stage

Figure 3-11. Simulated Orange Lake alligator population
after 20 years with differing values for transition rates
into the adult female size class.





- 1.3
c 1.2






I 100
E 50

0 50 100 150 200 250
Time in months



0 -

n -

x 104




100 150
Time in months

100 150
Time in months





Figure 3-12. Simulated population density, number of eggs
in harvest, and dollar value of harvest for a 50% harvest of
production in a simulated alligator population.


1 -

n L

1.7 -


E 1.5-
S1.4 -

a 1.3

1.2 -

01.1 -




x 104



1- 11.iii

100 150
Time in months

100 150
Time in months

Figure 3-13. Simulated population density, number of adults
in harvest, and dollar value of harvest for a 10% harvest of
adult animals in a simulated alligator population.

50 100 150 200
Time in months


, 200
( 100








. 1000



Time in months

x 104

100 150
Time in months

100 150
Time in months





Figure 3-14. Simulated population density, number of total
animals in harvest, and dollar value of harvest for a 10%
harvest of adults and a 50% harvest of production in a
simulated alligator population.


a. 1.5

0 1.3

ca 1.2
0 ~

Time in months


100 150
Time in months



I iJnUi ~ l

100 150
Time in months




Figure 3-15. Simulated population density, number of total
animals in harvest, and dollar value of harvest for a 5%
harvest of adults and a 25% harvest of production in a
simulated alligator population.


E 200







In the late 1970's there was considerable interest from

the Florida Game and Fresh Water Fish Commission (GFC) and

the alligator industry to commercially harvest both adult

and early age class alligators. The GFC had instituted a

nuisance alligator program and had embarked on research to

investigate the feasibility of sustainable adult harvest on

public waters (Hines and Woodward 1980). Florida's

alligator farmers were interested in harvesting eggs and

hatchlings for captive rearing (ranching) to supplement

closed-system farming. Alligator farmers and the GFC

developed a cooperative venture in 1979 to monitor alligator

populations by night light survey on 3 lakes (lakes Apopka,

Griffin, and Jesup) from which hatchlings were being

harvested. These lakes were considered as demographically

similar lakes in the St. John's River drainage having dense

populations of both adult and early age class animals.

Night-light surveys commenced in September 1979 and

hatchling harvests began in 1981.

In 1981, the Florida Cooperative Fish and Wildlife

Research Unit (FCFWRU; University of Florida, GFC, and

National Biological Service cooperators; the U.S. Fish and

Wildlife Service was a cooperator in 1981) entered into the

research funded by the Florida Alligator Farmers Association

(FAFA) in cooperation with the GFC. A population crash on

Apopka was observed between 1981-1984 in <182 cm animals

(Woodward et al. 1993). Relatively small pod (group of

hatchlings) size and an unusual number of unhatched eggs on

Apopka were noted during 1982. A low clutch viability rate

(number hatch/total eggs in a clutch) of alligator eggs on

Lake Apopka was observed in 1983 after studies began

examining the feasibility of collecting eggs rather than

hatchlings for ranching purposes. Although Apopka was less

important as a harvest site because of the alligator

population decrease, the lake continued to be a focus of

attention because of extreme differences in hatch rates from

other lakes. With the support of GFC, FAFA, and the

American Alligator Farmers Association (AAFA), subsequent

investigations expanded research to include variability in

clutch viability on a number of lakes, including Lake


Through 1987, long-term trends still indicated a

decline in juvenile alligators on Lake Apopka. Meanwhile,

other lakes in the St. Johns River drainage showed

numerically stable or increasing alligator populations

(Jennings et al. 1988; Woodward et al. 1993). Controlled

incubation of alligator eggs indicated that Lake Apopka had

the lowest clutch viability rates of any population

examined (Percival et al. 1992, Masson 1995). In 1988, the

Lake Apopka viability rate fell to a low of 0.04. These

data, derived from laboratory incubation, appear to

correlate closely with events occurring in the wild.

Investigations of neonatal wild pods demonstrated that Lake

Apopka had the lowest density of pods and the smallest pod

size of all study areas in Florida (Percival et al. 1992).

In 1988, the FCFWRU began to investigate causes of

Apopka's low clutch viability and population declines.

Coupled with a study investigating the reproductive

physiology of adult female alligators, the possibility of

environmental contaminant effects was examined. Lake Apopka

had substantial contaminant inputs from agricultural run-

off, sewage effluent, and a major chemical spill (U.S. EPA

1994). An analysis of alligator eggs collected in 1985 from

3 Florida lakes showed that eggs from Lake Apopka had

significantly elevated levels of DDD, DDT, and DDE (Heinz et

al. 1991). Other studies have documented increased

estrogenic activity (a possible effect of exposure to

environmental contaminants) in juvenile alligators from Lake

Apopka. Other endocrine effects and gonadal abnormalities

also were found (see Rice and Percival 1996).

This chapter examines long-term trends in both

alligator clutch viability and population estimates on Lake

Apopka. In particular, the hypothesis of the recovery of a

population following catastrophic reproductive failure is

tested. Further, the dynamics of the Lake Apopka alligator

population are compared with those of a reference population

from Lake Woodruff National Wildlife Refuge.


Egg Collections

Alligator clutches were collected from nests on Lakes

Apopka and Woodruff during the summers of 1994 and 1995.

Prior collections from 1983 to 1993 also were included for

long term trend analysis. Alligator nests were located via

aerial surveys and ground searches. For each collected

clutch, nest height, diameter, ambient temperature

(thermometers were calibrated to a standardized

thermometer), clutch cavity temperature, clutch depth,

approximate percentage of daily shade, flooding status,

female presence and behavior, and nesting habitat were


recorded. Each nest was completely dismantled and searched

completely because multiple nesting (two or more females

ovipositing in a single nest) had been observed (Percival

unpubl. data). Disturbed or depredated clutches were not

collected. All eggs present in the nest, including damaged

and non-viable ones, were collected.

Eggs were placed into plastic bus pans (61 cm X 36 cm X

13 cm) on 5 cm of natural nest material with additional nest

material cushioning layers of eggs when required. Eggs

along the bus pan perimeter were cushioned with 2-3 cm of

nesting material. An identifying plastic plant tag was

affixed with copper wire to a hole in the corner of the bus

pan rim. The order (approximately uppermost to lowermost)

in which eggs had been removed from the nest was maintained.

To ensure that data sheets could be matched later with

clutches, nest number, date, time, and collectors' names

were recorded on the tag, as well as on the data sheets.

Parts of eggs or eggs which were found crushed were sealed

in a labeled plastic bag.

Transportation of eggs followed recommendations by

Woodward et al. (1989). During transportation of clutches,

care was taken to avoid excessive vibration or shock, which

could kill embryos by detaching their membranes from the

shell (Ferguson 1985). Clutches were protected by

inflatable cushions and foam rubber within transport boats,

which were carefully maneuvered within sheltered waters to

avoid additional vibration and shock. Eggs were transported

to an incubator facility in Gainesville, FL within 24 hrs of

collection in the covered bed of a pickup truck, which was

cushioned like the transport boat (Woodward et al. 1989).

Clutch and Egg Characteristics

At the incubation facility, care was taken to avoid

vibration and rotation of viable eggs. Bus pans were

weighed with and without eggs to calculate a clutch weight.

All intact (unbroken) eggs were transilluminated to check

for viability and egg band development (Woodward et al.

1989). Anomalous eggs (those containing two yolks,

yolkless, or with aberrant calcified regions) were recorded

and, when possible, were measured. Non-viable eggs were

identified by the absence of an opaque embryo attachment

spot or band. Unbanded eggs were opened, and a visual

status determination was made. If band development was

visibly retarded, or if vascular color was not similar to

that of other, apparently healthy, eggs in a clutch, the egg

in question was opened and the developmental age and

embryonic status (good, weak or dead and presence of

deformities) of the embryo was determined. One healthy

representative egg was removed from each clutch and

sacrificed for determination of clutch age and expected


hatch date. Crushed eggs and empty eggshells were examined

to determine whether an embryo was present and its

developmental age. Whenever possible, embryo age was

determined using a combination of back-dating (date of

collection minus embryo age), an embryological developmental

chart (Masson 1995), and egg band progression (Ferguson

1985). Eggs were transilluminated at approximately 14-21

day intervals thereafter to identify and age dead embryos.

Eggs remaining unhatched at the end of the incubation period

also were opened and embryo age determined. The disposition

of each egg within a clutch was recorded.


All eggs were incubated in a modified 7.3 m x 3.7 m

portable building (Lark Industries, FL). The building was

layered for insulation with 2.54 cm blown-in foam

insulation, 7.6 cm fiberglass batting, a visqueen vapor

barrier, and 2.54 cm styrofoam sheeting. The humidity and

temperature were controlled by a Hawkhead International

heat/cool humidifier unit providing controlled high humidity

warm air. This unit utilized mercury switches for

temperature and wet-bulb humidity monitoring.

Thermostatically controlled electric fans and louvers, as

well as ceiling fans circulated air throughout the building.

The relative wet bulb humidity in the building was 94-96%

and mean temperature within the nests and in the building

was 30.6 C 0.5. However, nesting material had to be

moistened every 7-10 days to maintain the appropriate

moisture level.

Clutch integrity and within-clutch collection order

were maintained. All clutches were incubated in damp

sphagnum moss (a minimum of 2.5 cm on top, bottom, and all

sides) in bus pans placed on shelves within the incubator.

Each bus pan was covered with 50% shade screening cloth to

allow air circulation and to contain hatchlings. To

minimize premature hatching of alligators by audible cuing

from adjacent clutches, clutches with similar hatch dates

were grouped together.

Hatchling Care

Hatchlings were maintained in the aforementioned

insulated incubation building for approximately 14 days

before release at their original nest site. The building

contained 6 galvanized aluminum tanks separated into 24

individual 0.6 m x 0.7 m compartments. Compartments were

divided with a galvanized aluminum wall sealed with latex

caulking. Each compartment was on a 3.8 cm slope to provide

both dry and wet surfaces. To increase the dry surface area

and minimize stress from overcrowding, an elevated mesh

platform was added to each compartment (2.5 cm square mesh

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EJAUT71E1_8TOCIC INGEST_TIME 2013-10-24T19:58:40Z PACKAGE AA00017674_00001