SEASONAL ASPECTS OF THE REPRODUCTIVE BIOLOGY OF THE
GOPHER TORTOISE, GOPHERUS POLYPHEMUS
ROBERT WAYNE TAYLOR, JR.
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
I wish to express my appreciation to the members of my supervisory
committee, Dr. Walter Auffenberg. Dr. John H. Kaufmann, and Dr. Donald
A. Dewsbury for their assistance. I am especially grateful to my
chairman, Dr. Auffenberg, for his advice, understanding, and encourage-
ment during the period of my graduate study. I would also like to thank
Dr. Archie Carr, Dr. Martha L. Crump, Dr. Carmine A. Lanciani, Dr.
Elliott R. Jacobson, and Dr. Frank G. Nordlie for their guidance and
criticism. Dr. Lewis Berner and Dr. James H. Gregg graciously allowed
me to use their photographic equipment. The Department of Zoology and
the Florida State Museum, University of Florida, provided me with office
and laboratory space and funds for needed supplies.
Support was also provided by Dr. David Crews, who offered valuable
guidance and criticism. I thank Dr. Harvey H. Feder and Dr. Chris
Reboulleau who are responsible for the many steroid hormone determi-
nations. Without their generosity that portion of the study could not
have been completed. I am also indebted to Dr. Jacobson for making the
resources of his histology laboratory available to me. Dr. Gerald E.
Gause provided advice and optical instruments for the stereological
portion of the study.
Mr. Leon Norris provided valuable field assistance during the
collection of tortoises.
I am deeply indebted to all the members of my family, and
especially my wife, Gloria, for their help, encouragement, and under-
standing throughout the preparation of this dissertation.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS. . . ii
LIST OF TABLES. . . .. vi
LIST OF FIGURES . . .... vii
ABSTRACT. . . . ix
INTRODUCTION. . ... . 1
MATERIALS AND METHODS . . ... .. 5
Preliminary Hormonal Studies . 5
Collection and Housing of Animals . 5
Sampling Procedures ................ 5
Detailed Study of Annual Cycle . .. 7
Collection and Housing of Animals . 7
Processing of Animals .............. 8
Histological Procedures .............. 9
Analysis of Tissues. ................... 9
Female Reproductive Tract . .. 9
Male Reproductive Tract . 10
Chin Glands . . 11
Courtship Behavior. .. . 12
RESULTS .. . . . 14
Female Reproductive Cycle.. . ... 14
Development of Ovaries and Oviducts . 14
Vitellogenic Cycle .. . 16
Corpora Lutea ... . .. 20
Hormonal Fluctuations . .... 20
Chin Gland Structure. ... . 26
Female Chin Glands. .. .. .. 29
Maturity, Fecundity, and Young. . .. 32
Male Reproductive Cycle. . 34
Testicular Changes. ............... 34
Spermatogenic Cycle ................ 41
Testosterone Fluctuations . 49
Male Chin Glands. . .. 52
DISCUSSION. . .... . 55
SUMMARY AND CONCLUSIONS ................... 82
LITERATURE CITED. . ......... .. 84
BIOGRAPHICAL SKETCH ...................... 90
LIST OF TABLES
1 Summary of data regarding vitellogenesis and egg production
in female Gopherus polyphemus .. .... ..... 17
2 Monthly summary of numbers and percentages of female
Gopherus polyphemus possessing oviducal eggs, corpora
lutea, and/or ovarian follicles in each size category 18
5 Mass and total length measurements of 4 hatchling gopher
tortoises from a clutch of 4 eggs removed from a female
13 June 1979. . ... . .... 35
4 Characteristics of the four stages of spermatogenesis found
in male Gopherus polyphemus .. . ... 42
5 Monthly number of male Gopherus polyphemus in each
spermatogenic stage . . 46
6 Plasma testosterone levels (ng/ml) in male gopher tortoises 51
7 The role of estrogen and progesterone in the control of
ovarian and oviducal events in female Gopherus polyphemus .. .. 62
LIST OF FIGURES
1 Geographical distribution of Gopherus polyphemus 2
2 Changes in female Gopherus polyphemus reproductive
tissue weights during the annual cycle . 15
3 Representation of the progress of vitellogenesis and
egg production throughout the year for female Gopherus
polyphemus . . .. 19
4 Regression in size of corpora lutea following ovulation. .. 21
5 Mean monthly estrogen concentration in female Gopherus
polyphemus . . 22
6 Mean monthly progesterone concentrations in female
Gopherus polyphemus .... ........ ........ 23
7 Diagram of section through one of the chin glands of a
gopher tortoise. .. . . 27
8 Section through one of the chin glands of a gopher
tortoise . . 28
9 Variation in chin gland volume of female Gopherus
polyphemus . .. .. 30
10 Section through a gopher tortoise chin gland in the
"active" condition .. . 31
11 Section through a gopher tortoise chin gland in the
"inactive" condition . .. .. 31
12 Changes in male Gopherus polyphemus reproductive
tissue weights during the annual cycle 36
13 Relationship between relative testis size and total length
of male Gopherus polyphemus. .............. 37
14 Change in relative proportion of testicular tissue
composed of interstitial tissue and seminiferous
tubules in male Gopherus polyphemus during the annual
cycle. ... . ...... 39
15 Change in diameter of seminiferous tubules and their
lumens in male Gopherus polyphemus during the annual
cycle. . . ... 40
16 Section through a gopher tortoise testis in the inactive
condition (Stage 1). ........... 43
17 Section through a maturing gopher tortoise testis
(Stage 2). . .. . 43
18 Section through a gopher tortoise testis undergoing
maximum spermiogenesis (Stage 3) . 45
19 Section through a gopher tortoise testis with abundant
sperm in the tubule lumen (Stage 4). . 45
20 Change in diameter of the ductus epididymis and the
mass of sperm contained within it for male Gopherus
polyphemus during the annual cycle . .. 47
21 Section through the epididymis of a gopher tortoise
in late fall . . .. 48
22 Section through the epididymis of a gopher tortoise
in early fall. . . 48
23 Mean plasma testosterone concentrations in male
Gopherus polyphemus. ............. ... ... 50
24 Variation in chin gland volume of male Gopherus
polyphemus .. . . ...53
25 Summary of plasma testosterone fluctuation data from
important studies of male turtle reproductive
endocrinology. .. . 69
Abstract of Dissertation Presented to the Graduate Council of the
University of Florida in Partial Fulfillment of the Requirements for the
Degree of Doctor of Philosophy
SEASONAL ASPECTS OF THE REPRODUCTIVE BIOLOGY OF THE
GOPHER TORTOISE, GOPHERUS POLYPHEMUS
Robert W. Taylor, Jr.
Chairman: Walter Auffenberg
Major Department: Zoology
The reproductive cycles of male and female gopher tortoises
(Gopherus polyphemus) were investigated in detail between March 1979
and May 1980 following preliminary studies in 1977 and 1978.. Gross and
histological changes in gonadal and chin gland morphology were examined.
The role played by the endocrine system in the control of the reproduc-
tive cycles was investigated.
Vitellogenesis began in September and ovarian follicles were fully
enlarged by December. Most ovulation occurred between early April and
late May, and nesting was completed by mid June. Following a summer
period of minimal ovarian size, recrudescence began in early fall. The
period of courtship and mating has been reported to include the months
of March, April, and May, but evidence indicates that lowered levels
of reproductive behavior may be exhibited until fall. Plasma estrogen
and progesterone levels in females showed peaks associated with ovula-
tion and vitellogenesis.
Male reproductive tissue weights reflected the stage of spermato-
genesis. Testicular weights were lowest when the seminiferous tubules
were inactive, and greatest during spermatogenesis. After sperm were
produced, they moved to the enlarging epididymides, and testes weights
returned to minimal levels. Male testosterone levels peaked prior to,
and declined during, the spring mating season. A second peak was
coincident with fall spermatogenesis.
Changes in chin gland volume were similar in timing and magnitude
for males and females. The size of this secondary sexual characteristic
was greatest for both sexes during the spring mating season, but pre-
viously unreported secondary enlargements were seen in September.
Increases in chin gland size were correlated with peaks in plasma
estrogen (female) and testosterone (male) concentrations.
It is concluded that the endocrine system plays an important role
in the control of gonadal events and the development of secondary sexual
characteristics in both sexes of the gopher tortoise. The results of
this study generally confirm and expand upon what is known about
endocrine control of reproduction in other turtles. Other data indicate
that the gopher tortoise is characterized by a very low reproductive
potential. Taken together, these data provide a basis for making more
informed decisions about the management and possible protection of
Until recently, detailed studies of the reproductive biology of
reptiles have been uncommon. Although some knowledge had been gathered
for a large number of species, most of it was fairly cursory and was
usually limited to observations concerning mating, nesting, and
offspring. During the past few decades a few reptile species have been
examined more closely. In general, however, the reproductive cycles of
reptiles are still poorly understood when compared to those of mammals.
The gopher tortoise (Gopherus polyphemus) is a locally abundant
inhabitant of suitable habitats throughout the southeastern Coastal
Plain from extreme southern South Carolina to extreme eastern Louisiana
and peninsular Florida (Auffenberg and Franz 1978,. 1982; Fig. 1). The
species is restricted almost exclusively to areas with well-drained
sandy soils and vegetative associations that permit abundant sunlight to
reach the ground. Here the animals construct extensive underground
burrows and feed on low-growing grasses and forbs (Auffenberg 1969).
Gopherus polyphemus has experienced dramatic population size declines
throughout its range in recent years (Auffenberg and Franz 1982), with
most of the remaining animals concentrated in southern Georgia and
Information regarding the gopher tortoise's reproductive rate
(Douglass 1976; Iverson 1980; Landers et al. 1980), growth and maturity
Figure 1. Geographical distribution of Gopherus polyphemus.
Open circles represent isolated populations.
Adapted from Auffenberg and Franz 1978.
(Auffenberg and Iverson 1979; Landers et al. 1982), and population
composition (Alford 1980) has been compiled. The composite drawn from
the known data is that of a species whose individuals grow slowly,
mature at a relatively late age, and once mature have a very low
reproductive potential. Therefore, G. polyphemus must be categorized as
one of the most K-selected vertebrates studied (Pianka 1970).
Data bearing on the reproductive biology of the gopher tortoise are
included in the studies of mating and clutch development by Iverson
(1980) and Landers et al. (1980). Auffenberg (1966) and Weaver (1970)
described the courtship sequence, Douglass (1976) studied the mating
system, and Rose (1970) and Rose et al. (1969) analyzed the subdentary
(also called chin or mental) gland exudate thought to play a role in sex
and species recognition. However, no data regarding the physiological
factors thought to be important in the reproductive cycle were
available, nor had any information about the male cycle, other than
courtship behavior, been gathered.
The present study of the reproductive biology of Gopherus
polyphemus was initiated for two reasons. First, there was a general
lack of knowledge concerning reptilian reproductive cycles in spite of
their possible evolutionary significance relative to those of mammals.
Second, the continuing decline in gopher tortoise population size
suggested that such data should be gathered as soon as possible in order
that appropriate management and conservation policies regarding the
species can be formulated.
The specific objective of the present study was to examine the
reproductive cycles of male and female gopher tortoises throughout an
annual period and to compare the information gathered to similar data
for other vertebrates in general and other turtles in particular. Data
were to be gathered from studies of gonadal morphology (including
vitellogenic and spermatogenic cycles), changes in the paired glands
located in the underside of the lower jaw of all individuals (a
secondary sexual characteristic), aspects of endocrine system
physiology, and behavior associated with mating. These results would
then be combined with related previously existing information to
describe in detail the reproductive biology of G. polyphemus.
MATERIALS AND METHODS
Preliminary Hormonal Studies
Collection and Housing of Animals
Preliminary hormonal studies were conducted from May to October
1977, and March to August 1978. Male and female tortoises were
collected in Alachua and Lake Counties, Florida, between April 1977 and
April 1978. Most of the animals were captured in bucket traps placed at
the mouths of burrows (Taylor 1982). A few were picked up after being
discovered away from their burrows, usually when they were crossing
Tortoises were housed in a 275 m2 outdoor enclosure 8 km NW of
Gainesville and allowed to construct and utilize burrows. The number of
animals in the enclosure varied between 20 and 30. Food consisted of
naturally occurring grasses and forbs supplemented with a wide variety
of fruits and vegetables obtained from the refuse of local markets.
During each study period 3 to 7 individuals were periodically
selected from the penned population for blood sampling. During most of
the study some individuals were bled each week. After the tortoises
were weighed and their chin glands measured, 1.0 ml of blood was removed
via cardiac puncture. A sterile 1 or 1 1/2 inch 21 or 23 gauge needle
attached to a 3 ml disposable plastic syringe was inserted through a
1/16 inch diameter hole drilled through the plastron along the midline
at the junction of the pectoral and abdominal scutes. Insertion of the
needle at an angle perpendicular to the surface of the plastron
facilitated blood removal in the vast majority of cases. Prior to
drilling the hole in the shell, the area surrounding the site was
cleaned by scrubbing it with a toothbrush and Betadine scrub solution.
The area was then wiped with a sterile prep pad saturated with 70%
isopropyl alcohol. Immediately prior to sampling each individual, a
small amount of ammonium heparin (aqueous solution--1000 USP units per
ml) was introduced into each syringe as an anticoagulant. The needle
was inserted into the vial and approximately 1 ml of heparin was drawn.
Most of the solution was injected back into the vial, leaving only the
amount of anticoagulant that remained at the tip of the plunger and in
the lumen of the needle. The needle was inserted into the heart only as
far as necessary to allow blood to flow into the syringe.
The drawn blood was placed into heparinized Vacutainer (Sherwood
Medical Industries) blood collection tubes (7 ml; 13 x 100 mm; green
stopper). The tubes containing the blood were placed in a Junior Angle
Centrifuge (Model 1600) and spun at 3000 RPM for 10 minutes. During
centrifugation of the blood the tortoise was prepared for reintroduction
into the enclosure. A stick of dental compound was gently heated with
the flame of a cigarette lighter until the end became soft and began to
"droop". This softened end was touched to the 1/16th inch diameter hole
in the plastron of the tortoise so that a small amount of the compound
partially filled and completely covered the opening. In only a few
seconds the material hardened, providing a solid closure for the wound.
A piece of waterproof adhesive tape was placed over the dental compound
as an additional protective measure. The method worked well, for the
dental material never dislodged during subsequent activities of the
tortoises. It could, however, be easily removed later by prying it
loose with the blade of a pocket knife for additional blood sampling.
After centrifugation, the plasma was pipetted from atop the compacted
cellular-material and placed in a second heparinized Vacutainer tube for
storage at -20 C. Frozen plasma samples were shipped periodically to
the Institute of Animal Behavior, Rutgers University, Newark, New Jersey
for determination of steroid hormone levels by radioimmunoassay.
Detailed Study of Annual Cycle
Collection and Housing of Animals
Tortoises were collected in Alachua, Putnam, and Marion Counties,
Florida, between March 1979 and May 1980. All collecting sites were
within 40 km of each other. Animals were taken primarily by "pulling"
them from their burrows with a collecting hook (Taylor 1982). A few
individuals were bucket-trapped or found away from their burrows. No
animals were utilized that had been injured during the pulling process.
Tortoises were housed in an enclosure inside the Florida State Museum
until they were processed. All animals were examined within 48 hours
after capture. Water, but not food, was provided. Temperature was
maintained at 280C and the photoperiod was 12:12 L/D.
Processing of Animals
Weight and shell measurements were determined for each tortoise, as
well as the length, width, and height above the level of the skin of its
subdentary glands. Blood was then drawn for hormonal analysis following
methods similar to those used in the preliminary hormone study, with the
following exceptions: 1) blood sample size was 7 ml in order to
increase the plasma sample for radioimmunoassay. 2) the hole drilled in
the plastron was not plugged.
Each tortoise was killed by injecting 5 ml of 100% isopropyl
alcohol directly into its brain. This procedure rendered the animals
dead and completely relaxed within 10 to 20 seconds. Immediately after the
death of the tortoise its plastron was removed by cutting through the
bridge with a bone saw (Stryker Corp. autopsy saw, Model No. 9004-210).
The skin, muscle, and connective tissue were then separated from the
plastron with a knife, allowing the peritoneal cavity to be opened and
the viscera exposed.
The reproductive tissues were removed, excess liquid was drained,
and they were weighed on either an Ainsworth Type 21N Electronic Balance
or an Ohaus Dial-0-Gram Triple Beam Balance, depending upon their size.
The tissues were then fixed in either 10% neutral buffered formalin or
Bouin's Fluid. The intact chin glands were also dissected out and fixed
along with the other tissues. Tissues from 23 males and 31 females were
collected in this way. Six additional tortoises included in the study
were females obtained from a hunter who had butchered them. There was
no opportunity to obtain measurements of the live animals, blood
samples, or chin glands. Also, oviducal eggs may have been present and
removed from 3 of them.
The right testis and epididymis of each male and the right chin
gland of all individuals were prepared for microscopic examination.
Tissues were fixed for several days, rinsed in running water, and
transferred to 70% ethyl alcohol for storage. As time permitted, the
tissues were prepared for examination under light microscopy using
standard histological techniques (Humason 1979; Luna 1968). Paraffin-
embedded tissues were sectioned at 5 to 7 pm on an American Optical
rotary microtome with a steel knife. Sections were stained with
hematoxylin and eosin.
Analysis of Tissues
Female Reproductive Tract
In addition to the weights of ovaries and oviducts obtained upon
dissection, detailed analyses of ovarian follicles, corpora lutea, and
oviducal eggs were made from preserved tissues. The diameter of each
ovarian follicle greater than or equal to 5.0 mm was determined to the
nearest millimeter with dial calipers. Follicles were grouped into 5 mm
size categories for analysis of the data. Regardless of season, all
females had numerous undeveloped follicles less than 5 mm in diameter.
These were not counted. The number of corpora lutea and/or oviducal
eggs present was determined and the diameter of each one was measured
using dial calipers.
Male Reproductive Tissues
In addition to testes weights obtained upon dissection, several
groups of data were obtained from prepared microscopic sections using
stereological techniques (Chalkley 1943; Underwood 1970; Weibel and
Gomez 1962; Weibel and Elias 1967; Weibel et al. 1966). These pro-
cedures allowed conclusions to be drawn about the three-dimensional
testes based upon analysis of a two-dimensional section taken from those
tissues. A square grid reticle was placed in the ocular of the micro-
scope so that the grid was superimposed over the tissue section. The
grid used was composed of 11 horizontal and vertical lines, giving 121
points of intersection between those lines. Ten randomly selected
fields on a section from the center of each testis were examined using
the grid. The tissue types (seminiferous tubule wall or lumen, inter-
stitial cell of Leydig, vascular component, or connective tissue) under-
lying each of the 121 points was identified and recorded for each of the
ten random fields. From these data the relative proportion of each
testis made up of the component tissue types was determined and seasonal
Measurements of the diameter of the seminiferous tubules and
epididymides were made. A calibrated ocular micrometer was used to
measure 15 seminiferous tubules that were most nearly cut in perfect
circular cross section within the section from each testis. As many
circular cross sections through the ductus epididymis as possible were
measured (usually 5 to 8). Changes in the size of these structures were
determined for each of the stages of the spermatogenic cycle.
Approximately 15 measurements were also made of the distance
between adjacent seminiferous tubules within each testis. The number
varied slightly between individuals because the number and arrangement
of cuts through the tubules were not constant. Three representative
fields were selected for each testis. Within each field the distance
from a central tubule to its surrounding neighbors was recorded.
Fluctuations in these intertubular distances were related to changes in
the tubules themselves and the overall condition of the testis.
The process of spermatogenesis was divided into four stages based
upon the diameter of seminiferous tubules, thickness of and number of
cell layers in the wall of the seminiferous tubules, relative
proportions of cell types in the walls, and relative abundance of
spermatozoa within the tubules and/or epididymis of each testis. Each
male was categorized as having testes in one of these four stages.
The volume of the gland on the right jaw of each tortoise was
calculated from the length, width, and height above the skin
measurements taken prior to dissection. The shape of the gland was
taken to be that of a "sphere" in which the "radii" of the 3 dimensions
are not necessarily equal. The formula used was V = 4/3 i abc; where a
= 1/2 the length measurement, b = 1/2 the width measurement, and c = the
height measurement. The calculated volumes were used as a measure of
gross morphological change in the glands during the annual cycle.
Histologically prepared sections of the glands were examined under a
light microscope to detect changes in structure or cellular development.
Courtship and copulatory behaviors among captive (Auffenberg 1966)
and wild (Douglass 1976) gopher tortoises have been described. Previous
studies, however, have not attempted to quantify these behaviors or
relate them to known physiological conditions in specific animals.
Three attempts (1978, 1979, 1980) were made to gather detailed
information about behavioral sequencing of courtship and copulation, and
to relate these results to data concerning other reproductive parameters
in the same individuals. Unfortunately, these efforts did not result in
the accumulation of any useful information. I feel that the main cause
of this failure was the use of freshly caught tortoises in these
experiments. Animals that have been maintained in captivity for a
number of years are known to exhibit courtship periodically during the
breeding season (pers. obs.). It was my belief that the introduction of
animals recently obtained from the field into a restricted area would
cause increased levels of social interaction and facilitate the study of
mating behavior. Such was not the case, however. The tortoises'
primary interest was to hide themselves from view, and they spent
essentially no time interacting socially with each other. A lengthy
period of acclimation to captivity seems to be required before
individuals of this species can be expected to behave naturally. This
applies whether they are maintained indoors or out. Although no
behavioral information could be gathered, the other data regarding
gonadal development, chin gland morphology, hormonal fluctuations, and
the timing of the production of young could be correlated with the
results of previous studies of gopher tortoise reproductive behavior.
Female Reproductive Cycle
Development of Ovaries and Oviducts
Monthly mean weights of female reproductive tissue (Fig. 2) varied
during the year in a manner that reflected the progress of vitello-
genesis within the ovaries and the periodic presence of oviducal eggs.
Spring (March-April) and winter (December-February) weights were both
significantly greater than those of summer (May-September) (t = 5.42 and
5.43; p < 0.001 for both). Combined ovary and oviduct weights were
minimal in May. The single May female examined in detail had ovi-
posited. Three additional May females were obtained from a hunter, so
fresh tissue weights were not available. Their ovarian and oviducal
development was, however, similar to that of the more completely studied
individual. One of the 2 individuals in June contained 4 oviducal eggs,
causing the mean reproductive tissue weight for that month to be some-
what higher, but values in August and early September were consistently
low again. Dramatic increases in size occurred from late September
through November, but weights levelled off in December and remained
fairly constant through April. The combined weights of ovaries and
oviducts of females examined after vitellogenesis was complete but
before oviposition were 3 to 5 times as great as for those that had
oviposited but not yet initiated growth of the next group of follicles.
AJ 4-) 4.J 4.J
O M u r*
(D 0 0
SJ0 O U4
~0 u 41
d U *v
.co c C
H 1 a0 0 c:
z 0 0 Cw
S 01 (0 ,C
o mO LnA 1
-- I0 O t0
^ ." e A *r js
D J- 4 ) 4J C
\ 0 01 0 E
\r 0. Q
- t3- e CO 00 C
) 4 tr0 0
S ( 41 l 4-
o 0 I 0 0- 0
The month with the greatest mean reproductive tissue weight was April
1979, when 4 of 9 individuals contained oviducal eggs, including 2 that
had clutches of 10 eggs each. In general, the weight of ovaries
depended upon the stage of vitellogenesis shown by their follicles (see
below). Oviducts exhibited their greatest diameter and wall thickness
during the nesting season (late spring-early summer). Oviducal develop-
ment was somewhat reduced during late summer and reached its minimum in
Ovarian follicles were grouped into 5 mm size categories for each
female throughout the year (Table 1). A summary of these data by month
shows the trend toward increasing ovarian follicle size during the late
summer and fall (Table 2). I devised a measure of the stage of vitello-
genesis based on the degree of follicular development within each indi-
vidual. This measure (Vitellogenic Index) provides a means of assigning
a numerical value to the stage of development of each ovary. Based on
the size of her largest follicle, each female was given a number: less
than 5 mm = 1, 5-9 mm = 2, 10-14 mm = 3, 15-19 mm = 4, 20-24 mm = 5, 25-
29 mm = 6, 30 mm or greater = 7. The highest relative value (= 8) was
assigned to those individuals possessing oviducal eggs or corpora lutea.
The values for all females in each month were averaged to give the
overall measure of Vitellogenic Index (Fig. 3). The shape of this curve
graphically depicits the progress of the vitellogenic cycle in my sample
of female tortoises.
Table 1. Summary of data regarding vitellogenesis and egg
production obtained from female Gopherus polyphemus.
The lists of numbers in each column represent the quantity
of oviducal eggs, corpora lutea, and/or ovarian follicles
in each size category found in each female.
TORTOISE OVIDUCAL CORPORA 25- 20- 15- 10- 5-
ID DATE EGGS LUTEA e30mm 29mm 24mm 19mm 14mm 9mm <5mm
3 MAR 79
5 APR 79
7 APR 79
10 APR 79
14 APR 79
25 APR 79
25 APR 79
25 APR 79
27 APR 79
28 APR 79
24 MAY 79
13 JUN 79
14 JUN 79
2 SEP 79
2 SEP 79
LI SEP 79
24 SEP 79
25 SEP 79
30 SEP 79
4 OCT 79
23 OCT 79
24 NOV 79
15 DEC 79
I FEB 80
26 MAR 80
27 MAR 80
8 APR 80
15 APR 80
18 APR 80
31 MAY 80
31 MAY 80
31 MAY 80
6 6 I
* ALL INDIVIDUALS HAD NUMEROUS MINUTE FOLLCLES IN EACH OVARY.
LESS THAN 2mm IN DIAMETER.
MOST OF THESE WERE
CO C3 (
1- c t
oo o o
0 0 -
0 0 0 0 0
0 0 0 0 -
W I -
Z _j -
C L, -
SN -- I -
0 U < w
I I I I I I I
o 0 0ri
60 O4 fn
Corpora lutea were found in the ovaries of 12 females, during
April, May, June, and September. The mean corpus luteum diameter for
each individual was plotted against the date on which the female was
examined (Fig. 4). The equation for the resulting least squares
regression was calculated. The identification of small corpora lutea
became very difficult because of their tendency to become embedded in
the surrounding ovarian tissue and the lack of differentiation by color.
The smallest corpora lutea that could be positively identified as such
were about 5 mm in diameter.
Plasma estrogen and progesterone levels were determined for 28
females examined between March 1979 and May 1980. The mean value for
all individuals in each month was calculated (Figs. 5 and 6). Estrogen
levels were elevated in early September and from December through April
(the period of maximum follicular development). Both the spring (March
and April) and fall (September) peaks were significantly greater than
the combined summer (May-August) values (t = 2.32 and 2.94; p < 0.05 for
both). A more brief depression in estrogen levels occurred between the
late September and December peaks (Fig. 5).
Plasma progesterone also exhibited 2 peaks in concentration (Fig.
6). Dramatic increases occurred in April, when all individuals have
either recently ovulated or possessed enlarged follicles that were ready
to be released, and October. Following the April peak, progesterone
levels declined steadily but relatively slowly throughout the summer to
S I I I
/ / NIUa
/ >- ("
/ < c -
/ 2 ^
/ C en
/ 0 V
/ en (
/ 0) 0
/ ^ '-
/ oo &.
/ a) <
/ < rl
.r* / (
(uw) u..IviLVla mn31Lnl SndUOo
-CO 0 CO
0) C0 pr
4 H "
0 H $4
0 4) 00
o a u
OC J3 ,O
0 x a 0
p-i 0 1
-i oU 0
0 > 0 0
= w 4 0)
(U 0 >-
U .0 0)
a) c <
-s (34J (B,
(lUW/Bu) N3901S3 VY4SVld
I I I I I I I
o n o
(IU/Bu) 3NOi31SDOIId VlIVld
z e e
C > e
0 -4 .U
-I 0 3C
Q) nl )
S J 1 -
r- 4 *
minimal values in September (t = 3.11; p < 0.01). The trend following
the October peak was much different. Values declined rapidly to very
low levels in November, and remained there through February (t = 3.20, p
< 0.01, April vs. November-February).
The results from the preliminary hormonal study of 1977 and 1978
provide additional data that do not always agree with those of 1979-
1980. I do not doubt that the 1977 and 1978 determinations accurately
reflected the physiological condition of the tortoises that were
sampled, but I feel that these data may not be representative of wild
populations because 1) the samples were from individuals being kept in
captivity, 2) the reproductive condition of the tortoises was not known,
3) samples were not obtained from throughout the entire year 4) some
individuals were being sampled repeatedly and what effect this had on
the animals was unknown. In spite of these potential problems, the
amount of information on turtle reproductive endocrinology is so limited
that the results of the preliminary studies are of value.
Progesterone determinations were made for females between mid May
and late October 1977, and late February to late August 1978. In 1977,
peaks in concentration occurred in early June and mid July, which does
not conflict with the results of 1979-1980. The few determinations from
August, September, and October 1977 are all consistently low. The sharp
peak shown in October 1979, is conspicuously absent from the earlier
data, although a rapid increase and decrease could have been present but
not observed. The progesterone data for 1978 show an entirely different
picture. Whereas in 1977, concentrations in the 1 to 4 ng/ml range were
common (11 of 35 determinations [31.4%] were greater than 1.0 ng/ml),
the values from 1978 were much lower overall. Of 53 measurements in
1978, only 2 (3.8%) were greater than 1.0 ng/ml. Most samples were
taken in June and July of both years so timing should not be the cause
of this difference. As an extreme example of the low levels of 1978,
one individual was sampled 11 times between 24 February and 27 July, and
not once was her plasma progesterone level above the minimum detectable
concentration of 0.050 mg/ml.
Fewer determinations of plasma estrogen concentration were made
than for progesterone because of the limited amount of plasma in each
sample. Thirty-eight were made for 1978 individuals (vs. 53 for proges-
terone) and none in 1977. The data from the early part of the sampling
period in 1978 agreed with those obtained in 1979-1980, but determina-
tions from the later sampled months differed considerably. Levels in
February were high and approximately equal to those seen in February
1979. Concentrations were declining or low throughout March and April,
similar to the trend a year later. However, by late May and through
July, values were generally rising and high, although some temporary
declines were evident. This trend was quite different from that in
1979, when the lowest levels of the year were seen between May and
August. The possible causes, and the physiological significance, of the
progesterone and estrogen concentrations observed will be discussed
Chin Gland Structure
Each externally visible "gland" is actually a composite of 2 to 4
simple tubular holocrine glands, each of which has a duct leading to a
pore that can be seen at the surface (Figs. 7 and 8). The stratified
squamous epithelium of the epidermis is keratinized but scaleless. The
highly vascular dermis is composed of dense irregular connective tissue
and contains the glands proper. -Beneath the dermis is the somewhat less
dense hypodermis, and the underlying skeletal musculature.
Each of the individual secretary glands is surrounded by a dense
connective tissue capsule that is differentiated from, but grades into,
the connective tissue of the dermis. Septa that are continuous with
this surrounding capsule extend into and partially subdivide each gland.
There are no associated muscle fibers that might aid in secretion. The
cells making up each gland appear similar to those of mammalian seba-
ceous glands. Three structurally and functionally separate layers can
be identified within each gland. The basal or germinative zone is the
peripheral region of the gland adjacent to the surrounding capsule. It
is composed of only a few layers of small cells that have small nuclei
and little cytoplasm. In keeping with the holocrine nature of these
glands, it is assumed that the basal zone is a region of mitosis and the
source of the secretary cells. Inside the basal zone is the much
thicker proliferative zone. The cells in this layer possess nuclei that
are 2 to 3 times as large as those in the basal zone, and amounts of
cytoplasm that are relatively even more abundant. Presumably as these
cells are enlarging they are manufacturing the secretary product of the
. 0) 4
P 4-4 4-4
cd -H 44
1C g *ri
U 0 to1
0H C Ow
O >- <
R 60 0
Q 0: 4-
Section through one of the chin glands
of a gopher tortoise. Portions of the
basal, proliferative, and defoliative
zones and the duct can be seen (63X).
gland. The thickness of this zone varies with the activity of the
gland. If the gland is secreting, it may be up to 50 cell layers thick,
whereas in an inactive gland it may be almost unidentifiable. Inside
the proliferative zone is the relatively narrow defoliative zone, where
the secretary cells degenerate and the glandular product is released
into the duct. Before the cells reach the surface of the defoliative
zone they become reduced in size again, presumably as the metabolic
activity of the cell slows down. When the gland is active, the surface
of this layer has a ragged appearance as the cellular components are
cast off with the secretary product.
Female Chin Glands
Female chin gland development was maximal during April and May,
with decreased volumes from June to August (Fig. 9). A secondary
increase in chin gland size occurred in early September, after which the
trend toward reduced size was resumed. Minimal sizes were seen in
December, at which time the calculated volumes were only about 13% of
the May maximum. A trend toward increasing sizes began after December
and led into high values again the following April.
Microscopic examination of chin gland tissue revealed no cellular
differences between enlarged and shrunken glands, other than the
relative thickness of the proliferative zone (Fig. 10 and 11). The
inner and outer zones appear almost the same in all months, except for
the degree of "raggedness" of the defoliative zone. Even within the
proliferative zone, the cells themselves do not vary substantially in
size or shape, only in their number. When stained with the standard
U I__-----------------*------ f0
- 4 .
() 3 0 0 0
5 0 (A N
(eww) 3y4INOA aNVID NIHO
Section through a gopher tortoise chin gland in
the "active" condition. Note especially the
great width of the proliferative zone and the
uneven edge of the defoliative zone bordering
the duct (125X).
Section through a gopher tortoise chin gland in
the "inactive" condition. Compared to Figure 10
the proliferative zone has almost disappeared
and cells bordering the duct form a smooth intact
hematoxylin and eosin technique, no staining differences between active
and inactive glands could be detected.
Maturity, Fecundity, and Young
Values in the literature for size at sexual maturity for north
Florida gopher tortoises are females--220 to 230 mm plastron length
(Iverson 1980); males--230 mm carapace length (Auffenberg and Iverson
1979), which also equals 230 mm plastron length following the conversion
factors of Taylor (1982). It was not the purpose of the present study
to determine precisely size at sexual maturity, but some revealing data
regarding the subject were obtained. Using the presence of sperm in
either the seminiferous tubules or epididymis as the criterion for
maturity, all 23 males examined were mature. The smallest of these had
a plastron length of 182 mm, substantially smaller than the previously
established lower limit. Additionally, 6 other males had plastron
lengths less than 230 mm and were sexually mature. The data for females
do not entirely support the limits set by Iverson (1980). Females were
judged to be mature if they possessed oviducal eggs, corpora lutea, or
ovarian follicles greater than 5 mm in diameter. Of 31 females
examined, 4 were immature. The largest of these had a plastron length of
228 mm. The smallest mature female was 210 mm in plastron length,
although all other mature females were at least 220 mm.
Clutch size varied from 2 to 10 in 24 females that had either
oviducal eggs, corpora lutea, or enlarged follicles. Clutch size when
calculated for all criteria combined was 6.71 + 2.22 (mean + std. dev.).
When determined separately, mean clutch size based on oviducal eggs
(6.86 + 3.07, n = 7) is most similar to the overall mean. The cal-
culated mean based on number of corpora lutea (6.08 + 2.39, n = 12) is
probably lower because of the difficulty in identifying some of the
smaller ones within the ovaries of individuals that ovulated some time
ago. Conversely, the mean of enlarged follicles (7.25 + 1.36, n = 12)
is larger because not all follicles of "mature" size are ovulated. The
mean clutch size of my sample population (6.71) is much closer to the
figure derived by Landers et al. (1980) for G. polyphemus in south
Georgia (7.0) than to Iverson's (1980) estimate for north Florida
Five females were examined in which both oviducal eggs and corpora
lutea were present. In 3 of these the number of oviducal eggs and
corpora lutea in the right and left halves of the reproductive tract
were not equal, indicating transuterine migration. In 2 of these 3, net
migration was away from the side with the greater number of corpora
lutea. Iverson (1980) found 4 females with evidence of transuterine
migration. Net migration was away from the more productive ovary in 2
of those individuals, but Iverson concluded that transuterine migration
probably served to equalize reproductive tract volume. The few data
available on the subject do not constitute a basis for drawing any
conclusions about the adaptive significance of transuterine migration.
I examined the reproductive tracts of 7 females whose oviducts
contained shelled eggs. In 4 of these cases (5, 7, and 28 April 1979,
and 18 April 1980) calcification of the eggs was not very far
progressed. After removal from the females, these eggs were preserved
along with the remainder of the reproductive tracts. The ovaries and
oviducts of one female were obtained from a hunter (25 April 1979). The
6 well-calcified eggs from that individual were removed and eaten by the
hunter. The other 2 clutches of oviducal eggs were removed from indi-
viduals I had dissected (13 June 1979 and 26 March 1980) and looked as
if they were ready for oviposition. I incubated both of these clutches
but only one (13 June 1979) developed. All 4 of the eggs hatched, after
108, 109, 110, 111 days incubation at 28 C. Measurements of the
hatchlings were taken 2, 19, and 282 days after the last one emerged
from its shell (Table 3).
Male Reproductive Cycle
Mean monthly weights of male reproductive tissue (testes plus
epididymides) showed significant seasonal variation (Fig. 12). Testes
were lightest during the winter, spring, and early summer. Means of the
values from September and October (n = 7) were significantly greater
(0.278 vs. 0.134 g/100 g body weight) than values from the rest of the
year taken together (n = 16). In Figure 13, which depicts this
relationship, testes weights that have been adjusted for body size are
plotted against total body length. At all body sizes the values from
September and October are consistently greater than those from all other
months. The fall values (late September-October) were significantly
Table 3. Mass and total length measurements of 4 hatchling gopher
tortoises from a clutch of 4 eggs removed from a female
13 June 1979. Hatchlings are listed in order of emergence
from their eggshells. Mass in grams; Total shell length
in mm. Tortoise 122 died in December 1979.
4 October 1979 21 October 1979 10 July 1980
Tortoise ID Mass Total Length Mass Total Length Mass Total Length
L1 28.1 49.2 30.4 51.6 51.7 62.6
R1 28.8 47.7 29.4 49.2 50.5 60.1
L2 26.8 47.5 29.9 50.1 45.6 59.6
R2 24.9 42.8 28.1 57.5 -- --
I i 9-
> -L cc
r-f 0 j1
0 > >
/ 00 .S
I 4) CU C
'-e ..-e u
UJ / S 4 >o
L- U T4 u
I me .o
Z ~ z
2 \ a <
e \ 2 2
a. o 1
a, 1 -V
I 1' -Wa go
E E C4
(600 1/6) HDII3M S1 31
O o *HN
0 0 O 4
0L ba r-I 0
Y* w 0 0
41 r n 14
0 s *0
0 z S 0
a) a 0
I0 r 0-4
\ rn )
OOQ X H1TN3"1 IV10I 1H913M SlSJ31.
greater than those of July (t = 5.94; p < 0.01) and December (t = 4.19;
p < 0.01).
Changes in the tissue types within each male's testes were shown in
two ways. The relative proportion of each testis that was composed of
the various tissue types showed dramatic variation between months (Fig.
14). In the spring (the period of courtship and mating and when overall
testis size is minimal) the proportion of the testis that is composed of
seminiferous tubules was at its lowest level (only about 10% in April).
By September and October, however, the tubules had become predominant,
representing almost 60% of the testicular tissue. The combined mean for
the males examined in September and October (58.8 + 4.5%, n = 7) was
49.8% greater than that of individuals from all other months (39.2 +
4.2%, n = 16). The difference between September-October and April was
highly significant (t = 15.8; p < 0.001).
Changes in the diameter of the seminiferous tubules and their
lumens exhibited seasonal variation that corresponded to the propor-
tional changes. Tubules increased slowly in size during the summer, but
enlarged rapidly and reached their greatest diameter in the fall (t =
5.76; p < 0.001, September-October vs. April) (Fig. 15). Mean tubule
and lumen diameters for the September/October group (0.277 + 0.038 and
0.183 + 0.027 mm, n = 7) are 180 and 130% greater than the mean of those
individuals from the rest of the year (0.154 + 0.038 and 0.079 + 0.027
mm, n = 16). During the fall period of tubular enlargement, the lumens
showed a relatively greater increase in size than did the entire tubule
t = 6.28, p < 0.001, (September-October vs. April). Although the
0 1 O0 =
nM 5 tB
0 ca w
/ ^ 0 9 Cu
\ / C *
1d \ w3
4t cA- 4Cu
(H) r-4 r )
S\ U *r0
'~) 0. r 2H
/ \l- a<
o o Tr1 o3
O -I '
` N-- 3 0 -- Cu
I I I
0 0 -
y o 0
to C C
2 eC CT
absolute difference between the tubule and lumen diameters remained
about the same for all months (only one was outside the range of 0.06 to
0.11 mm), the dramatic increase in size by both resulted in a greater
proportional change in the lumen (maximum 3.7 times minimum vs. 2.7
times for tubule).
A series of spermatogenic stages was described to evaluate the
cellular changes taking place within the seminiferous tubules of each
male. The number of cell layers making up the wall of the tubule, the
relative numbers of the meiotic cell types contained within the wall,
and the presence of sperm in the tubule lumen were considered prior to
categorizing each male into one of the 4 stages (Table 4). Stage 1
(inactive) represents the period of minimal spermatogenic activity. The
tubule wall is only 1 or 2 cell layers thick and is comprised only of
spermatogonia and Sertoli cells (Fig. 16). Individuals classified as
Stage 2 (maturing) exhibit proliferation of spermatogonia, differenti-
ation of spermatogonia into primary spermatocytes, and production of
many secondary spermatocytes and some spermatids via meiotic divisions.
No sperm are present yet (Fig. 17). The period of maximal sperm produc-
tion (spermiogenesis) characterizes Stage 3. Large numbers of both
spermatids and sperm may be found within the tubules of males in this
stage, but only one or the other may be abundant, depending upon the
exact time of the cycle (Fig. 18). During Stage 4 sperm leave the
seminiferous tubules and pass out to the epididymis (spermiation).
Table 4. Characteristics of the four stages of spermatogenesis
found in male Gopherus polyphemus.
Stage Condition of Seminiferous Tubules
1 Inactive; only spermatogonia; tubules reduced
2 Maturing; mitosis of spermatogonia; many
spermatocytes, some spermatids, no sperm
present; tubule diameter increasing.
3 Spermiogenesis; few to many spermatids and
sperm; tubule diameter maximum.
4 Spermiation; sperm abundant in lumen or passing
into epididymis; few spermatocytes or spermatids
Section through a gopher tortoise testis in
the inactive condition (Stage 1). The tubule
wall is thin and contains only spermatogonia
and Sertoli cells (500X).
Section through a maturing gopher tortoise testis
(Stage 2). The enlarging tubule has a wall
composed of several layers of spermatogonia,
spermatocytes, and spermatids (500X).
Therefore, sperm may or may not be abundant in the tubule lumen. Unlike
Stage 3, few new spermatocytes or spermatids are forming (Fig. 19).
The number of males exhibiting each spermatogenic stage was
recorded by month (Table 5). The resulting trend in the timing of each
stage during the annual cycle correlates closely with changes in overall
testis size and detailed changes in seminiferous tubule diameter and
relative tissue composition of the gonad. Individuals exhibiting Stage
1 are found from December to April, when testis size, tubule diameter,
and relative tubule proportion are minimal. During Stage 2 (June and
July) tubules are enlarging and becoming a greater proportion of the
testicular tissue. Stage 3 extends from July to September and corre-
sponds to the period when gonadal weight, tubule diameter, and relative
tubule abundance all increase to their maxima. Throughout most of Stage
4 (September to December) these three variables are declining to their
Changes in the epididymides reflected what was occurring within the
testes (Fig. 20). During the spring and summer (the breeding season and
about 2 months afterward) the mean diameters of both the ductus
epididymis and the included sperm mass decreased slowly. In September,
however, they both increased dramatically in size, reaching levels that
were maintained throughout the winter and until the following breeding
season (t = 5.58 [ductus] and 11.9 [sperm mass]; p < 0.01 and 0.001,
September-October vs. July). The fall increase was a result of the
movement of sperm from the testes during and shortly after the period of
maximum spermatogenic activity (Figs. 21 and 22).
Figure 18. Section through a gopher tortoise testis undergoing
maximum spermiogenesis (Stage 3). The large diameter
tubule contains many spermatids and sperm (500X).
Section through a gopher tortoise testis with
abundant sperm in the tubule lumen (Stage 4).
Although sperm are numerous, few spermatocytes
or spermatids remain in the tubule wall. Tubule
diameter remains large (500X).
Table 5. Monthly number of male Gopherus
polyphemus in each spermatogenic
Month 1 2 3 4
Jul 2 1
Sep 5 1
Dec 2 1
r- o- =
I i I i
I I I I I I I I
Sod 0 d
- I I
40 u 0
S Co 0
0 C I
(d >' >
4J 0 (
Section through the epididymis of a gopher tortoise
in late fall. Spermiation is complete, resulting in
the large mass of sperm present within the ductus
epididymis. Only a small portion of the duct may be
seen in the field (125X).
Section through the epididymis of a gopher tortoise
in early fall. Maximum spermatogenesis has not yet
occurred. The number of sperm within the duct
represents the minimum amount seen during the annual
cycle. Compare the diameter of the duct to that in
Figure 21 (125X).
Plasma testosterone exhibited decreasing concentrations just prior
to and throughout the breeding season (February to May), and reached its
lowest level of the year in June (Fig. 23). Despite the large differ-
ence between the February and June means, the difference is not statis-
tically significant, however, due to the large amount of variability in
the data and small sample sizes. After June, the hormone's concentra-
tion rose again, reaching relatively high levels again in September and
October. This secondary peak in the fall preceded another decline
during the early part of the winter period of inactivity. The major
(spring) peak in plasma testosterone occurred at a time when the
tortoises were emerging from their burrows and becoming more active, but
declining values were well underway by the time courtship and copulation
would be expected. The late summer/fall increase corresponded to the
period of active spermatogenesis, with the highest values occurring
during maximum spermatogenesis (September). Chin gland volumes also
showed a secondary increase at the same time.
The preliminary results from 1977 and 1978 generally agreed with
those obtained in 1979-1980. Since data were collected between May and
October 1977, and between February and August 1978, it is not possible
to compare values for all months (Table 6). However, the two greatest
monthly means were both observed in February (1978 and 1979-1980).
Values declined rapidly in March of both years. The three months in
which the mean concentrations were the lowest were May 1977, May 1978,
and June 1979, indicating that late spring and early summer were
0 CO 0
4 0 w4
0 l 0(
0 41 <
0 1 >C
z) C U)
Table 6. Plasma testosterone levels (ng/ml) in male
gopher tortoises. Means were calculated
from the values of all males sampled each
month. The numbers in parentheses are
1977 1978 1979-1980
consistently the seasons when testosterone levels were minimal. The
secondary fall peak present in 1979-1980 was seen to a lesser degree in
1978, but was not observed in 1977. An increasing trend was indicated
between July and August 1978, so that values in September could have
approached the level of those from September 1979, but since data col-
lection in 1978 did not continue past August, this possibility could not
be confirmed. Since the animals sampled in 1977 had all been recently
captured and relocated, the hormone levels seen during the summer and
fall could have been affected by the stress of those activities.
Male Chin Glands
Maximum male chin gland volumes were observed in March and April
(Fig. 24). Following this portion of the breeding period volumes
declined until mid-summer (t = 5.15; p < 0.01, April vs. July), then
exhibited secondary increases in September and October (t = 4.46 and
6.31; p < 0.02 and p < 0.01, vs. July and December). Subsequently, they
declined again during late fall and reached their smallest size of the
year in mid-winter (December-January). Glands enlarged again in late
winter/early spring, prior to the next breeding season. The values from
the spring (April) and fall (late September-October) peaks were not
significantly different (t = 2.15; p < 0.1).
Microscopic changes in male chin gland structure were similar to
those found for females. There was no cellular or structural difference
between enlarged and reduced glands, except for a change in the thick-
ness of the proliferative zone. Once again, the factor causing the
overall change in chin gland volume was the number of cell layers
0 0 m
0 00 0
0 0 ;
^ -^ *s'0
o o o (U EL
O.- _D cuc
(glllU ^Ym r-l~ 0IH
present in that region, rather than a difference in cell type or size
anywhere within the gland.
Female gopher tortoises exhibit a distinct temporal separation of
the vitellogenic and ovulatory portions of the reproductive cycle.
There has been no evidence of the production of more than one clutch
annually by females of this species. All follicles ovulated each
nesting season (April through June) have undergone enlargement during
the previous fall and reached ovulatory size by December. A comparison
of maximum ovarian follicle sizes and the diameters of ova removed from
eggs by Iverson (1980) also indicates that no vitellogenic enlargement
of follicles occurs during the mating or nesting seasons. This scheme
differs from those of Chrysemys picta (Gibbons 1968), Trionyx muticus
(Plummer 1977), and Kinosternon subrubrum (Iverson 1979), all of which
can produce more than one clutch per nesting season. These species have
a period of vitellogenesis in the fall, but they also can form at least
one more clutch that undergoes rapid vitellogenesis shortly after the
first nesting of the year. The most extreme example of multiple nest-
ings among turtles involves the sea turtles. Female Chelonia mydas may
nest several times in a season and lay 100 eggs per nest, although they
do not nest every year (Carr 1967). A female of any species that
produces several clutches of eggs during one nesting season is precluded
by size constraints from possessing within her body more than a rela-
tively small percentage of the enlarged follicles at any one time.
By necessity, vitellogenesis must be progressing in a series of stages
during the nesting season. Domantay (1968) has reported that female
Chelonia mydas contain a graded series of follicles, with each set
successively moving up in size category as each successive clutch is
laid. The giant Aldabran tortoise (Geochelone gigantea) exhibits a
condition similar to that of the gopher tortoise with respect to vitel-
logenesis (Bourn 1977). This species undergoes vitellogenesis between
November/December and June/July, with ovulation of a single clutch
following soon after. Callard et al. (1978) reported that in Chrysemys
picta 3 distinct sets of ovarian follicles could be identified based on
their grouping into size categories. Callard concluded that only one set
of follicles was ovulated each year in females from the population that
was studied, and that that set had been undergoing vitellogenic growth
for 3 years.
The selective advantage of increasing clutch size and/or number of
clutches per season is clear, and many turtle species produce large
numbers of eggs per year by doing one or the other, or both (Moll 1979).
The factors placing limitations on these variables are not as easily
evaluated. Climate, diet, energetic efficiency of egg production, and
metabolic capabilities could all affect the ability of individuals of a
species to produce multiple clutches. The gopher tortoise possesses one
of the lowest annual reproductive potentials of any turtle, producing
only 1 clutch of 5-7 eggs per year (Moll 1979; Iverson 1980; Landers et
al. 1980). Additionally, the proportion of animals that survives to
hatch, grow, and become sexually mature is extremely small (Auffenberg
and Iverson 1979; Alford 1980). The ultimate causes of this low repro-
ductive potential and the question of whether the species can survive
under the conditions it faces today need immediate study (Auffenberg
1969; Auffenberg and Franz 1982).
Female reproductive tissue weights show the expected trends during
the annual cycle based upon the observed periods of vitellogenesis and
ovulation. Ovarian recrudescence follows oviposition by about 3 months,
after a summer-long period of minimal development. The occurrence of
mating and oviposition in the spring followed by vitellogenesis in the
fall is typical of other temperate turtle species (Gibbons 1968; Lewis
et al. 1979; Plummer 1977; Iverson 1979). Ovarian follicles grow at a
rate of about 5 mm per month between August and December. In August,
follicles grow out of the "less than 5 mm" category, and by December
have reached the maximum ovulatoryy) size of slightly over 30 mm.
The role played by plasma steroids in the control of the reproduc-
tive cycles of female turtles has not always been clear. The results of
the few studies that have been completed have not consistently indicated
the same relationships between ovarian and behavioral events and fluctu-
ations in measured hormone levels. Possible causes of these differences
include 1) the taxonomic diversity of the species examined, 2) the
range in geographic sources of animals, 3) various ecological positions
filled by the species examined, 4) different treatments and experimental
Klicka and Mahmoud (1977) gave exogenous progesterone injections to
mature preovulatory painted turtles (Chrysemys picta) to test for the
effect of this steroid on ovarian weight, follicle size, and oviduct
weight. While ovarian weight was not significantly changed in response
to this hormonal treatment, both follicle diameter and oviduct weight
were reduced. In experiments testing for the effects of progesterone on
oviposition in females bearing oviducal eggs, Klicka and Mahmoud found
that the hormone prevented oviposition to a great extent. Progesterone
was also shown to be very effective in preventing ovulation. No preovu-
latory turtle that received exogenous progesterone subsequently ovu-
lated. The authors concluded that progesterone acted as an antigonadal
agent in C. picta by affecting the pituitary and reducing production of
gonadotropin from that organ, rather than disrupting estrogen levels as
had been proposed for lizards (Callard et al. 1972). In addition to
affecting the reproductive tissues, the hormone caused a significant
reduction in pituitary weight compared to controls, but its antigonadal
effects were overcome when it was administered in conjunction with
pregnant mare serum gonadotropin. The physiological mechanism of pro-
gesterone control over gonadal events was not clearly elucidated by
Klicka and Mahmoud, but some insight into its action was gained. High
progesterone levels might inhibit ovulation as long as a previous clutch
of eggs remains in the oviducts but upon oviposition the level of the
hormone falls and ovulation of mature follicles could again occur. A
high progesterone concentration could also prevent oviposition by
blocking contraction of the smooth muscles of the oviduct, in a manner
similar to that in which progesterone from the placenta maintains preg-
nancy in mammals (Turner and Bagnara 1976).
The first workers to measure changes in circulating hormone levels
coincident with the reproductive cycle of a female turtle were Callard
et al. (1978). They examined the relationship between the plasma
steroids estrogen, progesterone, and testosterone and the reproductive
cycle of Chrysemys picta. Their study revealed dramatic increases in
the levels of all 3 hormones just prior to ovulation (July) and equally
rapid returns to low levels shortly thereafter. Levels of estrogen and
testosterone also rose significantly again during fall ovarian recru-
descence. Callard et al. concluded that estrogen facilitated the
production of vitellogenin in the spring (final maturation of preovula-
tory follicles) and fall (ovarian recrudescence), but questioned the
fact that the period of greatest follicular growth (fall OR) was charac-
terized by the lesser of the two peaks of estrogen. They also
questioned the absence of a significant progesterone peak in the fall,
when it might be expected that progesterone would exhibit a rise con-
comitant with that of estrogen, as was the case in the periovulatory
phase. Callard et al. hypothesized that the role(s) of progesterone in
the spring is/are to 1) play a role in follicular maturation by slowing
yolk uptake into the follicle just prior to ovulation, 2) facilitate
ovulation once follicles are mature, then cause a reduction in gonado-
tropic hormone level, 3) decrease oviduct motility after ovulation until
eggs can be shelled and readied for oviposition.
Licht et al. (1979) made repeated measurements of plasma steroid
levels in female green sea turtles (Chelonia mydas) throughout the
nesting season. Their results were similar to those of Callard et al.
(1978) in that for each ovulation there was a distinct peak in proges-
terone. Since these females nested repeatedly approximately every two
weeks, a series of periovulatory peaks resulted, separated from each
other by periods characterized by much lower levels. These data
supported the hypothesis of Callard et al. that progesterone is involved
in the final stages of follicular maturation. Licht et al. found,
however, that estrogen did not also show periovulatory peaks in the
green sea turtle. Since it was believed that estrogen played a role in
the accumulation of vitellogenin for each clutch of eggs, the observed
levels could not be explained. After having shown a moderate rise
during the mating (pre-nesting) period, the levels of the hormone
remained low throughout the nesting season. The cause for the observed
differences in the hormonal fluctuations between C. picta and C. mydas
is unknown, but Licht et al. hypothesized that it may be a result of the
different environments to which the two species are adapted (temperate
lakes vs. tropical oceans) or the striking differences between their
periods of ovarian growth and reproductive seasons (relatively short,
producing one or two clutches vs. prolonged, with several clutches).
Concentrations of progesterone and estrogen were determined in the
female snapping turtle (Chelydra serpentina) by Lewis et al. (1979).
Measurements were made in December (hibernation), May (prior to ovula-
tion), June (following ovulation), and September (ovarian recrudescence
and vitellogenesis). The possession of data from only 4 times during
the annual cycle makes evaluation of the results of this study somewhat
difficult, because of the possibility of missing significant
fluctuations during other periods. The conclusions to be drawn, how-
ever, are not substantially different from the general trends seen in
other studies. Estrogen concentrations were highest in May and
September, the periods of follicular maturation and ovarian recrudes-
cence. Progesterone was high in June, when oviducal eggs and corpora
lutea were present, but consistently low at the other 3 samplings.
In light of previous studies, the roles played by the steroid
hormones estrogen and progesterone in the control of ovarian and ovi-
ducal events in female Gopherus polyphemus may be evaluated (Table 7).
Near the time of ovulation, estrogen level is high, facilitating final
growth of follicles prior to their release. Rising and high levels of
progesterone prevent ovulation and end follicular growth. After
estrogen and progesterone concentrations fall, vitellogenesis ends (E)
and ovulation can occur (P), but oviposition is still prevented until
the levels of progesterone become low enough, allowing sufficient time
for albumin and eggshell deposition within the oviducts. During the
summer, estrogen is basal and no ovarian growth occurs. Progesterone
continues to decline, removing any inhibition to ovulation or oviposi-
tion that might not yet have occurred. In August and September the
estrogen level rises dramatically, resulting in substantial follicular
yolk accumulation and ovarian weight increases. Progesterone continues
to decline during this period. In October, progesterone rises and
estrogen falls sharply. The increased antigonadal effect of
progesterone and the removal of the stimulus for vitellogenin accumula-
tion become evident. The result is that ovarian growth is halted and
Table 7. The role of estrogen and progesterone in the
control of ovarian and oviducal events in
female Gopherus polyphemus.
Months) Fluctuation Events
P falls slowly
--follicles in final maturation
- ends follicular maturation
reduces excitability of oviduct
- vitellogenesis ends
- ovulation facilitated by sharp decline
oviposition facilitated when moderate
levels are reached
- vitellogenesis prevented
+ ovulation and oviposition may occur
antigonadal effect removed
ovaries prepared for fall vitellogenesis
-ovarian growth and vitellogenesis begins
+antigonadal effect not expressed
E drops sharply *vitellogenesis not
P rises sharply--antigonadal effect
November-December E rises again -+ follicular growth
P falls again -antigonadal effect
no ovarian growth
1) follicles reach
2) reproductive tissue
maximum and levels
December-February E high
+ follicles maintained in near-ovulatory
+ ovarian growth and maintenance not
follicles enlarge only slightly. A reversal of the hormonal condition
takes place in November, with estrogen returning toward maximal levels
and progesterone nearing zero. In response, yolk deposition resumes
again shortly thereafter. From December until nearly the next nesting
season estrogen remains high, progesterone remains low, and ovarian
follicle diameters remain near nesting season levels. As the period of
ovulation approaches in the spring, the concentrations of both hormones
begin their cyclic fluctuations once again.
These hypotheses regarding the relationship between the plasma
steroids and the female reproductive cycle generally agree with those
proposed by Klicka and Mahmoud (1977), Callard et al. (1978), and Licht
et al. (1979). My data differed substantially only from those of Licht
et al. who found no periovulatory or vitellogenic estrogen peaks in
Chelonia mydas. The facilitation of yolk accumulation by estrogen in
Gopherus polyphemus seems clear, as it does in Chrysemys picta (Callard
et al. 1978). The hormone or other factor that influences vitello-
genesis in the absence of elevated estrogen levels in the green sea
turtle is unknown.
The dramatic change in the relative concentrations of G. polyphemus
estrogen and progesterone during October was quite interesting.
Although Callard et al. (1978) did report a minor increase in progeste-
rone during September for C. picta, a sharp rise and decline during the
fall had not previously been reported for any turtle. This temporary
increase seemed to be the cause of striking effects on the concentration
of plasma estrogen and the rate of follicular growth. The estrogen
level plunged while progesterone was elevated, and the previously steady
increase in ovarian weight was reduced to zero at the same time.
Whether progesterone exerted this effect by disrupting the estrogen
concentration directly (Callard et al. 1972) or by reducing gonadotropin
secretion via some influence on the hypothalamo-hypophyseal pathway
(Klicka and Mahmoud 1977) is not known. Whatever the mechanism, the
evidence for a strong antigonadal effect by progesterone seems clear.
The significance of the pause in fall ovarian growth may also be
questioned. Why would a long-term physiological process be temporarily
suspended, only to be resumed later at about the same rate? The most
likely explanation is that it in some way saves energy. The production
and maintenance of a clutch of large follicles presumably involves a
great energetic cost. Perhaps under favorable conditions (food, water,
temperature), when follicular growth is proceeding rapidly, a mechanism
to slow ovarian enlargement would be advantageous by reducing the period
of time that the clutch of mature (large) follicles would have to be
maintained. An alternative hypothesis is that harsh environmental con-
ditions that occurred between September and November necessitated sus-
pension of ovarian development, and that this was mediated through the
The earliest that Iverson (1980) found oviducal eggs in a north
Florida gopher tortoise was 26 April. Based on the literature, the
females he examined, and personal observations of others, Iverson con-
cluded that most oviposition occurs between mid May and mid June. I
believe that the main portion of the nesting season begins earlier than
this in north Florida. Of the 12 females I examined in April, 5
possessed oviducal eggs. If eggs are retained in the oviducts about 2
weeks (Powell 1967; Gibbons 1968; Moll 1973; Tinkle and Gibbons 1977),
42% of my April sample would have oviposited no later than 12 May. An
additional 4 females with follicles 30 mm or greater in diameter were
examined before 15 April. If only 2 of these individuals had ovulated
within the next 2 weeks, 50% of the April females would have nested
before 15 May.
The discovery during this study of a female with well-calcified
oviducal eggs on 26 March was quite surprising, and would seem to add
even more support to the idea that the nesting season is well underway
by mid May. The exact reproductive condition of this female, however,
must be questioned. Primarily because of the presence of an additional
5 ovarian follicles greater than or equal to 30 mm in diameter, I wonder
whether the two well-calcified oviducal eggs were produced from the same
group of enlarged follicles. Also, these eggs failed to develop when
incubated, suggesting that they had either not been fertilized or were
adversely affected by oviducal retention. Although none of the other 12
individuals examined between August 1979 and March 1980 contained ovi-
ducal eggs, the possibility exists that the 2 eggs found on 26 March
represent a clutch (or part of a clutch) that was produced but not laid
the previous year. Potentially, these eggs could have been laid with
the 5 for 1980, assuming all 5 of those were to be ovulated. Although
the production of viable eggs by 26 March in north Florida cannot be
ruled out, all of the evidence from this and other studies indicates
that the nesting season of the gopher tortoise is not yet underway by
that date. I conclude that in north Florida the major period of ovi-
position begins between mid and late April rather than mid May.
It seems that most females have nested by mid June. Of four indi-
viduals examined in May, one (on the 24th) possessed no oviducal eggs,
corpora lutea, or follicles greater than 14 mm in diameter, indicating
ovulation and oviposition had occurred some time ago. The other three
(all on the 31st) all had corpora lutea, but had been obtained from a
hunter and the presence of oviducal eggs could not be determined. Of
two tortoises from 13 and 14 June, one had corpora lutea but no oviducal
eggs, so had nested. The other had oviducal eggs that later hatched,
indicating that they were ready to be laid. I have no evidence that any
of the other females examined would have nested after mid June.
In male gopher tortoises the period of maximum testicular size is
coincident with the period of spermatogenesis but temporally distinct
from the mating period. This condition is typical of several other
temperate turtle species including Chrysemys picta (Gibbons 1968;
Callard et al. 1976), Sternotherus odoratus (McPherson and Marion 1981;
Mahmoud and Klicka 1972), Trionyx muticus (Plummer 1977), Trionyx
spiniferus (Robinson and Murphy 1978), Trionyx sinensis (Lofts and Tsui
1977), and Testudo hermanni (Kuchling et al. 1981). Singh (1977)
reported, however, that largest testis size and greatest spermatogenic
activity were concurrent with the breeding period in the Indian pond
turtle Lissemys punctata granosa.
Environmental factors that are the ultimate causes of changes in
testicular size and spermatogenic stages are less well understood for
turtles than for lizards (Licht 1967, 1969, 1971, 1972; Licht et al.
1969; Noeske and Meier 1977). McPherson and Marion (1981) hypothesized
that in Sternotherus odoratus temperature and photoperiod interacted to
initiate the production and maturation of sperm and to bring about the
end of spermatogenesis. Singh (1977) felt that in addition to those two
factors, changes in precipitation were important in controlling the
testicular cycle in populations of Lissemys living in areas with dis-
tinct wet and dry seasons. One or more of these environmental variables
are also suspected of influencing the timing of spermatogenesis in G.
Changes in the development of the epididymides of the gopher
tortoise seem to be in response to testicular events and mating
activity. Hormonal or other factors may also be of importance in
influencing epididymal diameter, but the major determinant seems to be
the number of sperm contained within the duct. The very close
association seen between the diameters of the ductus epididymis and the
mass of sperm contained within it indicates that the duct expands in the
fall to accommodate the sperm that are leaving the testis and contracts
in the spring following ejaculation.
Few studies examining the role of the endocrine system in the
control of the male reproductive cycle in turtles have been completed.
Other than G. polyphemus, only Chrysemys picta (Callard et al. 1976),
Chelonia mydas (Licht et al. 1979), and Testudo hermanni (Kuchling et
al. 1981) have been examined. The geographic ranges and ecological
positions of these species differ, and the timing of specific spermato-
genic stages varies somewhat between them. The relationship between
fluctuating levels of plasma testosterone and testicular events within
each, however, is surprisingly similar (Fig. 25).
In G. polyphemus the plasma testosterone level declines during the
prebreeding and breeding periods, rebounds during the summer and peaks
during fall spermatogenesis, and then declines again during the winter
before peaking again prior to next year's mating season. Of the other
three studies that examined plasma testosterone fluctuations, only that
of Kuchling et al. spanned an entire year. Their results from T.
hermanni were similar to those of the gopher tortoise. Testosterone
declined during the mating season, rebounded during fall
spermatogenesis, and returned to low levels for the winter. The levels
of circulating testosterone Kuchling et al. found at the "breeding peak"
were relatively low compared to those in the fall. These results seem
not to conform to mine, since values increased between January and
April, a period that includes part of the mating season. I propose,
however, that because no determinations were made in February or March,
the main part of the "spring peak" was not observed. Substantially
higher concentrations of plasma testosterone were probably present in
February and/or March, so that the value in April was part of a declin-
ing trend rather than a peak as it seemed due to a lack of data points
in the preceding two months. Licht et al. (1979) made testosterone
determinations between March and June, months that spanned the
z 0 w; u
SH 0 8 .
o o 02
4 H 0 i
( (0 MO
'0 0 0
S3 (U 0C6
0 4 q1
4-o 40 -H C0
0 0 1w
0 0 w.
2(0 1 04
v;t il 0
i- a ii l
&.e i -
3 4|1< (
0 0 0 0
prebreeding and mating periods of the C. mydas population they studied.
For this portion of the annual cycle the trend is the same as that of G.
polyphemus. Values peak during the prebreeding period and decline just
prior to and during the mating season. Callard et al. (1976) studied C.
picta from April until December. The mating period included April and
May, and was a time when plasma testosterone concentrations declined
precipitously. Levels were basal during the summer, but rose dramati-
cally in the fall during spermatogenesis. If endocrine control of the
spermatogenic cycle of C. picta is similar to that of G. polyphemus, I
suspect that the increasing testosterone concentration Callard et al.
found in December would not lead directly to the high levels present in
April. A mid winter period of lowered levels probably separates two
peaks in the annual cycle.
In male mammals, both spermatogenesis and the behaviors associated
with courtship and mating are dependent on the presence of androgens.
Testosterone (and its "active form," the metabolite dihydrotestosterone)
is one of the main circulating androgens of testicular origin (Lofts
1968; Turner and Bagnara 1976). In seasonally-breeding mammals elevated
plasma testosterone concentrations are correlated with concurrent semi-
niferous tubule development and reproductive behavior. Spermatogenesis
in most turtles that have been studied, however, occurs months after the
breeding season at a time when the testicular interstitium is atrophied,
but when there is a second peak in plasma testosterone.
When considering the role played by the endocrine system in the
control of reproductive events in turtles, some questions come to mind.
If testosterone levels are high in both spring and fall, why would
courtship occur only in the spring and spermatogenesis only in the fall?
Part of the answer hinges on the fact that the few data gathered so far
for turtles do not clearly indicate the relative magnitudes of the
spring and fall peaks. By relying on the information from the gopher
tortoise as a basis for hypothesizing, it could be concluded that the
levels of testosterone circulating in the fall are not high enough to
elicit reproductive behavior, or at least to the degree seen in the
spring. Although the spring and fall peaks in testosterone concentra-
tion are not statistically different, the higher levels seen in the
spring may nonetheless be of biological importance. If, in the gopher
tortoise, a behavioral response to hormone concentration is dose-
dependent, as has been proposed for the lizard Anolis carolinensis
(Tokarz and Crews 1980), then the overall greater values seen in the
spring should correspond to the period of more intense courtship.
Courtship in captive animals has been repeatedly observed as late in the
year as November (Douglass 1976), and it would not be surprising that
reproductive behavior and elevated plasma testosterone are positively
correlated in the fall as well as the spring. The significant increases
in chin gland volume occurring in September and October lend credence to
the idea that some level of fall mating might occur. When considering
possible explanations for a lack of androgen-dependent spermatogenesis
in the spring, it must be remembered that the data gathered so far in
the study of turtle reproductive endocrinology have only been for tes-
tosterone levels in the peripheral circulation. Levels of testosterone
within the seminiferous tubules have not been determined, even though it
is at that location where the concentration is important when consider-
ing spermatogenesis. Harris and Barthke (1974) showed that in the rat,
testosterone concentration could be 20 times greater within the tubules
than in the general circulation. In the fall, levels within the semi-
niferous tubules of the gopher tortoise may be much higher than those
measured, whereas the cirulating concentration in the spring may not be
great enough to stimulate tubular development. If the conclusion is
that spermatogenesis occurs in response to testosterone from a source
other than the testicular interstitium, then another possible origin
must be identified. A likely candidate is the Sertoli cell, present
within each seminiferous tubule and shown in at least one study (Lofts
and Bern 1972) to exhibit a secretary appearance during the production
of sperm. Lacy (1960, 1973) and Lofts (1972) have for some time
advanced the idea that the Sertoli cell might play a role in regulating
spermatogenesis through the production of hormones.
The genus Gopherus is unique among the testudinids in that its 4
species possess integumentary glands on each mandibular ramus
(Auffenberg 1969). Agassiz (1857) first reported these structures, and
pointed out that those of G. berlandieri were more developed than those
of G. polyphemus. The glands were subsequently reported in G. agassizii
(Grant 1936) and G. flavomarginatus (Legler and Webb 1966). Smith and
Brown (1948) incorrectly stated that the only species to possess these
glands was G. berlandieri, and that only in males of that species were
Despite the fact that their existence has been known for some time,
these subdentary glands have received little attention. In only a very
few studies have attempts been made to examine them in detail or to
investigate their function. In his initial description of the courtship
of G. polyphemus, Auffenberg (1966) reported that the glands (and pre-
sumably their exudate) were employed in the behavioral sequence. When
courting, male gophers vigorously bob their heads in a vertical arc
directly in front of females. Auffenberg concluded that this served to
waft the scent from the glands toward the female. Auffenberg (1969)
also reported that females rubbed their chins with their forelegs during
the courtship sequence, presumably causing the secretary product of the
glands to be expressed. The legs were then extended anteriorly and
"presented" to the male facing her. Auffenberg concluded that these
actions served to facilitate olfactory communication between the court-
ing tortoises and indicated a readiness to mate. The results of the
histological portion of the present study support Auffenberg's
hypothesis. Since there is no muscular tissue incorporated into the
structure of the gland, accumulated secretary product can be forced out
of the duct during courtship encounters only by vigorous contraction of
the deep underlying skeletal muscle or by pressing on the gland
Rose et al. (1969) studied the chin glands of both sexes of all
four Gopherus species and described electrophoretic differences between
the glandular products of each. Rose et al. stated that chin glands are
largest in male G. berlandieri and G. agassizii, and weakly developed in
all females. They found that the glandular exudate contained esterases,
phospholipids, triglycerides, free fatty acids and cholesterol. Rose et
al. concluded that chin glands probably served as olfactory as well as
visual cues during courtship and/or combat. They also speculated that
the lipid component of the chin gland exudate might be a factor that
elicits the behavioral response in gophers since that fraction has a
characteristic odor to humans.
Rose (1970) found nine fatty acids in the chin gland secretary
product of male G. berlandieri, and tested the effect of painting the
head of a model tortoise with a solution containing those compounds in
the correct concentrations. Rose found a highly significant increase in
the number of positive responses of both sexes to the "painted" model as
compared to those directed toward the model before the fatty acid
solution was added. The behavioral responses in these tests were
combat-like rather than courtship-like (for example, no mounting
attempts by males). Rose speculated that this might have been because
the test solution was patterned after the chin gland exudate of males.
The small size of G. berlandieri females' chin glands prevented the
collection of sufficient secretary product to allow completion of a
complimentary analysis and test.
Weaver (1970) tested the ability of G. berlandieri to differentiate
between male and female conspecifics. Weaver found that the tortoises
used a variety of cues for sex recognition, including chemical signals
originating from the chin glands. He also found that behavioral
responses following encounters occurring when the chin glands were
secreting differed from those seen when the glands were inactive.
Weaver reported that male G. polyphemus rubbed their forelimbs against
the jaw in a manner previously described for females.
Literature references to chin glands have usually included the
terms "active," "secreting," "enlarged," or evertedd" and their antonyms
to refer to the physiological state of the structures. It had become
accepted that the glands were "active" during the breeding season and
"inactive" throughout the rest of the year, although I am not aware of
any efforts to try to quantify the changes in size. I wanted to examine
this secondary sexual characteristic and to correlate changes in gland
size with other reproductive variables including hormonal fluctuations.
Several points can be drawn from the chin gland volume data, some
of which are unexpected. Males and females exhibited almost exactly the
same trends in chin gland size fluctuations, with glands becoming
enlarged and reduced almost simultaneously at most times of the year.
Male glands are only slightly larger than those of females. Female
glands are usually about 20 mm; smaller than the males', or only about
10% less when both are swollen. Most surprising is that the glands of
both sexes show significant enlargement in the fall as well as during
the spring breeding period. There are no reports of this in the litera-
ture and I am not aware of anyone who has observed this previously.
It is assumed that, because they are secondary sexual characteris-
tics, gopher tortoise chin glands would be under the influence of the
endocrine system. This hypothesis is supported by the increases in
estrogen and testosterone (for females and males respectively) that
occur in association with greater activity of the glands. Hormonal
levels are elevated in the spring and fall, and chin glands enlarge in
the same seasons. However, close examination reveals that the spring
size increase follows the spring hormone peak by about 2 months (April
vs. February for males; May vs. March for females), whereas the fall
glands peak up to one month after the hormones. Possible explanations
for these phenomena include 1) other hormones are involved--but the data
for the only other hormone for which there is sufficient information do
not support this, 2) the endocrine system is not involved in control of
chin gland size, 3) other factors are acting synergistically with
hormone levels to influence chin gland size--including climatic factors,
physiological condition, reproductive state, or refractoriness. At this
time the last explanation seems most likely.
Regardless of the exact relationship between hormonal fluctuations
and changes in chin gland size, the tissues certainly seem to play a
role in the social interactions of gopher tortoises. This is true of
the behaviors seen in the spring breeding season, as has been reported
in the literature previously, but also during the fall. Courtship has
been observed in the field in April (Kenefick 1954) and May (Layne in
Douglass 1976), and the breeding period has generally been considered to
be the spring and early summer (Carr 1952; Ernst and Barbour 1972). In
captivity, however, tortoises have exhibited reproductive behavior in
August, September, and November (Douglass 1976). The occurrence of
enlarged chin glands in the fall suggests that courtship observed in
that season may not be an artifact of captivity. The significance of
possible fall mating in the wild and the number of offspring resulting
from it (if it occurs) are unknown, but the evidence of a second or
extended breeding period is increasing. Since females do not possess
fully enlarged ovarian follicles until December, any sperm transferred
by the males during fall matings could not immediately fertilize ova,
and would therefore have to be stored by the female. There has been
speculation (Iverson 1980) but no evidence that female gopher tortoises
can store sperm.
Combat between tortoises is also commonly observed in the fall
months. Douglass (1976) noted many instances of agonistic behavior
during his study of the mating system of G. polyphemus, and also
reported related observations of other persons. Although these records
were not systematically gathered, Douglass stated that combat between
males was observed in March, August, September, and October. A compila-
tion of Douglass' reports of others' observations reveals that most of
them occurred in the same months. Weaver (1970) reported combat in the
spring and summer. Taken together, these data suggest that there may be
2 periods when combat is most common. In the spring such behavior is
presumed to be associated with dominance hierarchies, maintenance of
territories, and access to females by males (Douglass 1976). If mating
does occur in the fall, enlarged chin glands would presumably serve the
same purpose then that they do in the spring. If fall courtship only
occurs as a result of captivity, then an alternate explanation for the
significance of swollen glands and combat during August, September, and
October is needed. Such behavior could serve to reaffirm the social
system that was established in the spring and to reassert ownership of
burrows. It is possible that this "system" occurs in the fall because
of a general reduction in away-from-burrow activity and social contacts
during the midsummer months. Although tortoises may occasionally be
found away from burrows on hot summer days, even at midday (pers. obs.),
the animals suffer thermal and water stress if subjected to such
conditions for extended periods (Bogert and Cowles 1947; pers. obs.). A
reduction in the opportunity to communicate may occur on the longest
days of the year because of the necessity of spending part of the day
down in the burrow, whereas the more tolerable period in the fall may
allow dominant individuals to reinforce their position through increased
The question of whether or not the gopher tortoise warrants pro-
tection from hunting is a current topic of debate (Matthews 1979;
Anonymous 1979). Dramatic reductions in the number of tortoises have
been documented and further declines have been predicted (Auffenberg and
Franz 1982). There are two primary causes for the smaller population
sizes seen today--habitat destruction due to urbanization and agricul-
ture, and hunting of the animals for food by man. Disagreement occurs
among persons who attempt to determine the relative effects of these two
factors in contributing to the plight of G. polyphemus. In light of
predicted population size declines this point is moot, however, since it
is the cumulative effect of all the factors that determines whether or
not the species will remain viable. Steps need to be taken now to
reduce the loss both of the tortoises and their preferred habitats,
regardless of the exact relationship between the contributing factors.
The results of the present study do not support the contention that
the gopher tortoise can sustain continued human predation at the level
currently being experienced. A previous study (Taylor 1982) showed that
on the average, 20% of mature individuals were removed from a population
during each episode of human predation. Because of the large body size
(i.e., high age) required to attain sexual maturity and the small clutch
size of mature females, replacement of mature individuals through repro-
duction is extremely slow. The results of the attempted behavioral
analyses in the present study draw attention to another potential
problem in current gopher tortoise management policy. Groups of tor-
toises are sometimes relocated from an area that is scheduled to be
developed or disturbed, in an attempt to reestablish the population in
another suitable area. While these efforts are certainly more benefi-
cial to the animals than doing nothing to save them, the positive
effects of such actions may not be as dramatic as anticipated. The
disruption of an established social system (Douglass 1976) and the
relocation of a portion of a colony into an unfamiliar locality may have
severe enough effects that a period of years may be required before the
mating system of the group becomes reestablished to the extent that
significant numbers of young are produced.
The current game management regulations for Florida (Florida Game
and Fresh Water Fish Commission 1981) specify a possession limit of five
gopher tortoises. This rule, however, is based on no biological
evidence whatsoever, and tortoise populations have declined precipi-
tously while it has been in effect. The relationship between the
species' reproductive potential and its status as an exploited game
animal needs thorough investigation. Such studies have recently been
initiated by the Florida Game and Fresh Water Fish Commission. The
management policy to be followed while these important data are being
gathered is still being debated, however. Should the gopher tortoise be
protected immediately and throughout the period that data are being
collected, with the final decision concerning protection to be made at
some time in the future? Or should exploitation be allowed to continue
under current guidelines while studies are being conducted to determine
if the species can withstand such pressure? If the latter course is
followed, and if it is determined that the level of exploitation has
been excessive, then the ability of G. polyphemus to recover, even under
strict future controls, may be seriously affected. In light of
ecological problems and faunal extinctions that have resulted from the
ignorance and disregard of biological principles, the prudent course of
action is to conserve this resource until the determination is made that
it can withstand exploitation. Since human predation of the gopher
tortoise is for the most part recreationally and economically motivated
(Taylor 1982), the protection of the species and the enhanced survival
of this unique and biologically important animal would more than
outweigh the cost to hunters.
SUMMARY AND CONCLUSIONS
1. Ovarian follicles that are ovulated in April and May have undergone
enlargement between September and December of the previous year.
The six to seven eggs produced by a typical female are laid between
mid April and mid June.
2. In females, plasma estrogen and progesterone concentrations exhibit
periovulatory peaks, and secondary surges during vitellogenesis.
The antigonadal effect of progesterone is indicated by the
cessation of ovarian growth and declines in estrogen levels
following increases in its concentration.
3. Maximum spermatogenesis is coincident with greatest testicular
weight and seminiferous tubule diameter in September and October.
Testis and seminiferous tubule development are minimal during the
spring breeding period.
4. In males, plasma testosterone peaks just prior to the mating
season, and again during fall spermatogenesis. Both courtship and
sperm production seem to be under endocrine control, but
differences in peripheral and intratubular testosterone
concentrations may be the cause of the temporal separation between
5. Both males and females exhibit spring and fall increases in chin
gland size that are associated with peaks in steroid hormone
concentrations. Fall mating may be of greater significance than
6. Additional protection for the gopher tortoise is warranted.
Overexploitation of this slowly-reproducing resource must be
prevented in order to ensure its future survival.
Agassiz, L. 1857. Contributions to the natural history of the United
States of America. Little, Brown, and Co., Boston. 452 p.
Alford, R. A. 1980. Population structure of Gopherus polyphemus in
northern Florida. J. Herp. 14(2): 177-182.
Anonymous. 1979. Law enforcement protects the gopher tortoise. Fla.
Wildl. 32(6): 46.
Auffenberg, W. 1966. On the courtship of Gopherus polyphemus.
Herpetologica 22: 113-117.
Auffenberg, W. 1969. Tortoise behavior and survival. Biol. Sci.
Curric. Study, Patterns of Life Series, Rand McNally & Co.,
Chicago. 38 p.
Auffenberg, W., and R. Franz. 1978. Gopherus polyphemus. Catalogue of
American Amphibians and Reptiles: 215.1-215.2.
Auffenberg, W.,and R. Franz. 1982. The status and distribution of the
gopher tortoise (Gopherus polyphemus), p. 95-125. In: E. B. Bury
(ed.). U.S. Fish Wildl. Svc., Wildl. Res. Rept. No. 12.
Auffenberg, W., and J. B. Iverson. 1979. Demography of terrestrial
turtles, p. 541-570. In: M. Harless and H. Morlock (eds.).
Turtles: Research and perspectives. Wiley-Interscience, New
Bogert, C. M., and R. B. Cowles. 1947. Moisture loss in relation to
habitat selection in some Floridian reptiles. Am. Mus. Novitates
Bourn, D. 1977. Reproductive study of giant tortoises on Aldabra. J.
Zool., Lond. 182: 27-38.
Callard, I. P., G. V. Callard, V. Lance, and S. Eccles. 1976. Seasonal
changes in testicular structure and function and the effects of
gonadotropins in the freshwater turtle, Chrysemys picta. Gen.
Comp. Endocrinol. 30: 347-356.
Callard, I. P., S. W. C. Chan, and M. A. Potts. 1972. The control of
the reptilian gonad. Am. Zool. 12: 273-287.
Callard, I. P., V. Lance, A. R. Salhanick, and D. Barad. 1978. The
annual ovarian cycle of Chrysemys picta: Correlated changes in
plasma steroids and parameters of vitellogenesis. Gen. Comp.
Endocrinol. 35: 245-257.
Carr, A. 1952. Handbook of turtles: The turtles of the United States,
Canada, and Baja California. Cornell University Press, Ithaca.
Carr, A. 1967. So excellent a fishe. The Natural History Press,
Garden City, New York. 248 p.
Chalkley, H. W. 1943. Methods for the quantitative morphologic
analysis of tissues. J. Nat. Cancer Inst. 4(1): 47-53.
Domantay, J. S. 1968. Notes on the development and breeding habits of
Chelonia mydas. Acta Manilana Ser. A 4(2): 87-109.
Douglass, J. F. 1976. The mating system of the gopher tortoise,
Gopherus polyphemus, in southern Florida. M. A. Thesis, Univ. of
South Florida. August 1976.
Ernst, C. H., and R. W. Barbour. 1972. Turtles of the United States.
Univ. Kentucky Press, Lexington. 347 p.
Florida Game and Fresh Water Fish Commission. 1981. Florida hunting
handbook: Rules and regulations summary for 1981-1982.
Tallahassee. 32 p.
Gibbons, J. W. 1968. Reproductive potential, activity, and cycles in
the painted turtle, Chrysemys picta. Ecology 49(3): 399-409.
Grant, C. 1936. The southwestern desert tortoise, Gopherus agassizi.
Zoologica 21: 225-229.
Harris, M. E., and A. Barthke. 1974. Concentration of testosterone in
testis fluid of the rat. Endocrinology 95: 706-707.
Humason, G. L. 1979. Animal tissue techniques, 4th ed. W. H. Freeman
and Co., San Francisco.
Iverson, J. B. 1979. Reproduction and growth in the mud turtle,
Kinosternon subrubrum (Reptilia, Testudines, Kinosternidae), in
Arkansas. J. Herpetol. 13(1): 105-111.
Iverson, J. B. 1980. The reproductive biology of Gopherus polyphemus.
Am. Midl. Nat. 103(2): 353-359.
Kenefick, J. H. 1954. Observations on egg laying of the tortoise
Gopherus polyphemus. Copeia 1954: 228-229.
Klicka, J., and I. Y. Mahmoud. 1977. The effects of hormones on the
reproductive physiology of the painted turtle, Chrysemys picta.
Gen. Comp. Endocrinol. 31: 407-413.
Kuchling, G., R. Skolek-Winnisch, and E. Bamberg. 1981. Histochemical
and biochemical investigation on the annual cycle of testis,
epididymis, and plasma testosterone of the tortoise, Testudo
hermanni hermanni Gmelin. Gen. Comp. Endocrinol. 44: 194-201.
Lacy, D. 1960. Light and electron microscopy and its use in the study
of factors influencing spermatogenesis in the rat. J. Microsc.
Soc. 79: 209-225.
Lacy, D. 1973. Androgen dependency of spermatogenesis and the
physiological significance of steroid metabolism in vitro by the
seminiferous tubules, p. 493-532. In: V. James, M. Serio, and L.
Martini (eds.). The endocrine function of the human testis, vol.
1. Academic Press, New York.
Landers, J. L., J. A. Garner, and W. A. McRae. 1980. Reproduction of
gopher tortoises (Gopherus polyphemus) in southwestern Georgia.
Herpetologica 36(4): 353-361.
Landers, J. L., W. A. McRae, and J. A. Garner. 1982. Growth and
maturity of the gopher tortoise in southwestern Georgia. Bull.
Fla. State Mus. 27(2): 81-110.
Legler, J. M., and R. G. Webb. 1961. Remarks on a collection of bolson
tortoises, Gopherus flavomarginatus. Herpetologica 17: 26-37.
Lewis, J., I. Y. Mahmoud, and J. Klicka. 1979. Seasonal fluctuations
in the plasma concentrations of progesterone and oestradiol-17 beta
in the female snapping turtle (Chelydra serpentina). J. Endocr.
Licht, P. 1967. Environmental control of annual testicular cycles in
the lizard Anolis carolinensis. I. Interaction of light and
temperature in the initiation of testicular recrudescence. J. Exp.
Zool. 165: 505-516.
Licht, P. 1969. Environmental control of annual testicular cycles in
the lizard Anolis carolinensis. III. Temperature thresholds for
photoperiodism. J. Exp. Zool. 172: 311-322.
Licht, P. 1971. Regulation of the annual testis cycle by photoperiod
and temperature in the lizard Anolis carolinensis. Ecology 52:
Licht, P. 1972. Environmental physiology of reptilian breeding cycles:
Role of temperature. Gen. Comp. Endocrinol. Supply. 3: 477-488.
Licht, P., H. E. Hoyer, and P. G. W. J. Van Oordt. 1969. Influence of
photoperiod and temperature on testicular recrudescence and body
growth in the lizards Lacerta sicula and Lacerta muralis. J. Zool.
Licht, P., J. Wood, D. W. Owens, and F. Wood. 1979. Serum gonadotropins
and steroids associated with breeding activities in the green sea
turtle Chelonia mydas. 1. Captive Animals. Gen. Comp. Endocrinol.
Lofts, B. 1968. Patterns of testicular activity, p. 239-304. In: E.
J. W. Barrington and C. Barker-Jorgensen (eds.). Perspectives in
endocrinology. Academic Press, New York.
Lofts, B. 1972. The sertoli cell. Gen. Comp. Endocrinol. Suppl. 3:
Lofts, B., and H. A. Bern. 1972. The functional morphology of
steroidogenic tissues, p. 37-126. In: D. R. Idler (ed.). Steroids
in non-mammalian vertebrates. Academic Press, New York.
Lofts, B., and H. W. Tsui. 1977. Histological and histochemical changes
in the gonads and epididymides of the male soft-shelled turtle,
Trionyx sinensis. J. Zool., Lond. 181: 57-68.
Luna, L. G. (ed.) 1968. Manual of histologic staining methods of the
Armed Forces Institute of Pathology, 3rd edition. McGraw-Hill
Book Co., New York.
Mahmoud, I. Y., and J. Klicka. 1972. Seasonal gonadal changes in
kinosternid turtles. J. Herpetol. 6: 183-189.
Matthews, E. L. 1979. The gopher. Fla. Wildl. 32(5): 38-40.
McPherson, R. J., and K. R. Marion. 1981. Seasonal testicular cycle of
the stinkpot turtle (Sternotherus odoratus) in central Alabama.
Herpetologica 37(1): 33-40.
Moll, E. 0. 1973. Latitudinal and intersubspecific variation in
reproduction of the painted turtle Chrysemys picta. Herpetologica
Moll, E. 0. 1979. Reproductive cycles and adaptations, p. 305-332.
In: M. Harless and H. Morlock (eds.). Turtles: Research and
perspectives. Wiley-Interscience, New York.
Noeske, T., and A. H. Meier. 1977. Photoperiodic and thermoperiodic
interaction affecting fat stores and reproductive indexes in the
male green anole, Anolis carolinensis. J. Exp. Zool. 202: 97-102.
Pianka, E. R. 1970. On r- and K-selection. Am. Nat. 104: 592-597.
Plummer, M. V. 1977. Reproduction and growth in the turtle Trionyx
muticus. Copeia 1977(3): 440-447.
Powell, C. N. 1967. Sexual cycles of female Chrysemys picta and
Clemmys inscultpa in Nova Scotia. Canad. Field Nat. 81: 134-140.
Robinson, K. M., and G. G. Murphy. 1978. The reproductive cycle of the
eastern spiny softshell turtle (Trionyx spiniferus spiniferus).
Herpetologica 34(2): 137-140.
Rose, F. L. 1970. Tortoise chin gland fatty acid composition:
Behavioral significance. Comp. Biochem. Physiol. 32: 577-580.
Rose, F. L., R. Drotman, and W. G. Weaver. 1969. Electrophoresis of
chin gland extracts of Gopherus (tortoises). Comp. Biochem.
Physiol. 29: 847-851.
Singh, D. P. 1977. Annual sexual rhythm in relation to environmental
factors in a tropical pond turtle, Lissemys punctata granosa.
Herpetologica 33: 190-194.
Smith, H., and S. 0. Brown. 1948. A hitherto neglected integumentary
gland in the Texan tortoise. Proc. Trans. Tex. Acad. Sci. 30: 59.
Taylor, R. W. 1982. Human predation on the gopher tortoise (Gopherus
polyphemus) in north-central Florida. Bull. Fla. State Mus., Biol.
Sci. 28(4): in press.
Tinkle, D., and J. W. Gibbons. 1977. The distribution and evolution of
viviparity in reptiles. Misc. Publ. Mus. Zool. Univ. Mich. No.
Tokarz, R. R., and D. Crews. 1980. Induction of sexual receptivity in
the female lizard, Anolis carolinensis: Effects of estrogen and
the antiestrogen CI-628. Hormones and Behavior 14: 33-45.
Turner, C. D., and J. T. Bagnara. 1976. General Endocrinology (6th
ed.). W. B. Saunders Co., Philadelphia. 596 p.
Underwood, E. E. 1970. Quantitative Stereology. Addison-Wesley Publ.
Co., Inc., Reading, Mass.
Weaver, W. G. 1970. Courtship and combat behavior in Gopherus
berlandieri. Bull. Fla. State Mus. 15: 1-43.
Weibel, E. R., and H. Elias. 1967. Quantitative methods in morphology.
Springer-Verlag, New York.
Weibel, E. R., and D. M. Gomez. 1962. A principle for counting tissue
structures on random sections. J. Appl. Physiol. 17(2): 343-348.
Weibel, E. R., G. S. Kistler, and W. F. Scherle. 1966. Practical
stereological methods for morphometric cytology. J. Cell. Biol.
Robert Wayne Taylor, Jr. was born January 12, 1950, in Leesburg,
Florida. He attended public school in Leesburg and graduated from
Leesburg High School in 1967. He attended Auburn University from 1967
until 1971, graduating with the Bachelor of Science degree in Biological
Mr. Taylor served as a Supply Corps Officer in the United States
Navy from 1971 until 1976, and attained the rank of Lieutenant.
In 1976, he entered the University of Florida as a graduate student
in the Department of Zoology. He was allowed to by-pass the master's
program and to enter the Ph.D. program in June 1978. While at the
University of Florida, he was employed as a Graduate Teaching Assistant
in the Department of Zoology and as a Graduate Research Assistant in the
Florida State Museum. He is a member of the American Society of
Ichthyologists and Herpetologists, the Herpetologists' League, and the
Society for the Study of Amphibians and Reptiles.
Mr. Taylor is married to the former Gloria Gene Cays. They have a
daughter, Meredith Leigh, born February 22, 1980.