Site by Bennett Web & Design Co.
2013 SOUTHEASTERN NATURALIST 12(1):27–40
Growth in Body Length and Mass of the Florida Panther:
An Evaluation of Different Models and Sexual Size
Tad Bartareau1,*, Dave Onorato2, and Deborah Jansen3
Abstract - Knowledge of growth in body dimension and mass is important to understanding
fundamental elements of wildlife biology and ecology. We evaluated five classical
growth models (Gompertz, Logistic, Monomolecular, Richards, and von Bertalanffy) in
describing body length and mass growth curves as a function of age to determine which
best fit wild Puma concolor coryi (Florida Panther). When used for inferences on body
length and mass growth curves of both genders, the von Bertalanffy function proved to
be the best-fitting theoretical equation to our data set because it used the fewest parameters
derived directly from metabolic laws, had lowest residual standard deviation of data
points about fitted model, with lower Akaike Information Criterion value, and largest
Akaike weight. The von Bertalanffy model estimated that male asymptotic body length
was 9.4% longer and mass was 33.2% heavier than for females. Both genders grew in
body mass for a longer duration than length. Male-biased sexual size dimorphism develops
in part because males grew faster and for a more prolonged period. Our results should
prove useful in future studies of Panther energetics, reproduction, and in developing
conservation and management policies for this species.
Growth in body dimension and mass is a fundamental element in the biology
and ecology of all species that influences important life-history traits such
as minimum size or age at reproductive maturity (Stearns 1992). Growth trends
are heritable and affected by environmental factors, but one cannot continuously
measure changes in body size of an individual during growth. Therefore, a
practical method for estimating the growth curve of a species is the use of body
size-at-age measurements fitted to algebraic models that describe how individuals
grow as a function of age (Ratkowsky 1983). This approach gives one the
opportunity to interpolate non-observed measurement intervals and describe
consistent changes in the underlying growth processes. Growth models are also
useful in wildlife research because they summarize growth data with uniform
model parameters that provide quantitative indices for asymptotic size and relative
growth rate. These variables can be used to reveal variation caused by the
environment and facilitate comparison between genders and among populations,
or of the same population at dif ferent times (Ratkowsky 1983).
1Habitat and Species Conservation, Florida Fish and Wildlife Conservation Commission,
298 Sabal Palm Road, Naples, FL 34114. 2Fish and Wildlife Research Institute, Florida
Fish and Wildlife Conservation Commission, 298 Sabal Palm Road, Naples, FL 34114.
3Big Cypress National Preserve, 33100 Tamiami Trail East, Ochopee, FL 34141. *Corresponding
author - Tad.Bartareau@myfwc.com.
28 Southeastern Naturalist Vol. 12, No. 1
Puma concolor L. (Puma), also known as Mountain Lion, Cougar, or Panther,
is a sexually dimorphic species with males larger in body dimension and mass
than females (Currier 1983, Gay and Best 1995). The territory of dominant males
usually encompasses the home ranges of several females who are predominantly,
but not exclusively, mating partners of the resident male (Logan and Sweanor
2001, Maehr et al. 1991). Males are aggressive, and their reproductive success
is correlated with fighting ability and territory ownership (Logan and Sweanor
2001, Maehr et al. 1991). Females are generally less aggressive or territorial
(Logan and Sweanor 2001, Maehr et al. 1991), and their reproductive success
is dependent on energetic investment in rearing offspring (Ackerman et al.
1986, Laundré 2005, Robinette et al. 1961). Body mass tends to increase with
increasing latitude throughout the range of the Puma (Iriarte et al. 1990), but no
apparent geographic variation in relative sexual size dimorphism among adult
skulls was detected (Gay and Best 1995). The causes responsible for both why
and how sexual size dimorphism arises in this species are generally thought to be
associated with intraspecific and interspecific interactions (Gay and Best 1995),
although the root causes are still not well understood.
Assumptions about Puma growth characteristics (Maehr and Moore 1992,
Zullinger et al. 1984), reproductive productivity (Robinette et al. 1961), energetics
(Ackerman et al. 1986), and prey requirements (Laundré 2005) have been based
on body-mass growth curves. Historically, the generalized Richards function has
been the model of choice for describing Puma body length and mass growth curves
(Laundré and Hernández 2002, Maehr and Moore 1992). However, when a particular
model is chosen independently of the data and used to approximate the growth
curve as a basis for inference, uncertainty in the model’s selection is assumed to be
zero (Akaike 1981, Burnham and Anderson 2002). If model selection uncertainty
is overlooked, then precision can be overestimated and the accuracy of predictions
could suffer (Burnham and Anderson 2002). Despite the importance of testing concordance
among different growth models fitted to the same data set (Ratkowsky
1983, Zullinger et al. 1984), few studies have done so.
Puma concolor coryi Bangs (Florida Panther) is a rare and endangered Puma
subspecies that is native to the southeastern United States (Onorato et al. 2010).
We used an information-theoretic approach (Akaike 1981, Burnham and Anderson
2002) to evaluate five classical models (Gompertz, Logistic, Monomolecular,
Richards, and von Bertalanffy) in describing body length and mass growth curves
as a function of age to determine which best fit wild Florida Panthers (hereafter
Panther). Then, we compared gender differences in growth curves using the best
model from these alternatives and examined the ontogeny of sexual dimorphism
in this species.
The study area encompassed the current breeding range of the Panther, bordered
by the Caloosahatchee River to the north, coastal mangrove swamps of the
2013 T. Bartareau, D. Onorato, and D. Jansen 29
southern Everglades National Park to the south, and the urban centers of Miami–
Fort Lauderdale and Naples–Fort Myers to the east and west, respectively
(Onorato et al. 2010). Habitat types included several varieties of upland and lowland
forests, open-canopy marshes and prairies, and agricultural lands (Land et al.
2008, Onorato et al. 2011). The climate of south Florida is characterized by a wet
season during summer and autumn and a dry season during winter and spring.
Panthers were captured and monitored with radiotelemetry from 1982 to
2012 by biologists from the Florida Fish and Wildlife Conservation Commission
and National Park Service using methods described by Jansen et al.
(2010) and the Florida Fish and Wildlife Conservation Commission (2011).
Successive locations of radiocollared females in the same location were
noted as a potential indication of the commencement of denning behavior
and resulted in investigative trips to the area to search for neonatal den sites.
Kittens examined in neonatal dens were implanted with passive integrated
transponders. From this undertaking, standard growth data were recorded for
139 female and 155 male wild Panthers first examined as kittens in neonatal
dens. These Panthers included 27 females and 21 males eventually recaptured
for radiocollaring as larger dependent kittens, subadults, or adults. Growth
data included body length (cm) and/or mass (kg) and age (yr). Body length
was measured as the distance from the tip of the nose to the end of the last tail
vertebra along the contour of the spine while the Panther was aligned laterally.
Body mass was measured with calibrated spring scales. The age (yr) of kittens
examined in dens was estimated to within a few days based on a combination
of the movement patterns of radiocollared dams immediately prior to establishing
the neonatal den and developmental characteristics of kittens when
examined in dens (e.g., size, mobility, tooth development, eye opening).
Maehr and Moore (1992) excluded data from recaptures in growth-curve
construction to avoid potential autocorrelation. However, doing so violates a
basic assumption of random samples for each age class because kittens would be
excluded from future samples in older age classes (i.e., equal probability of being
sampled; Zar 1999). Therefore, growth data for individual Panthers in multiple
years were considered to be independent. As did Laundré and Hernández (2002),
we considered all Panthers in the population at a given age to be representative
of that age regardless of prior sampling.
We used a two-parameter body mass-length power function to compare body
shape allometry by gender: Mass = aLengthb. The coefficient evaluates the rate
of change, while the exponent is a measure of mass at unit length indicating an
isometric relationship when b = 3 or one that is positive or negative allometric
when b > 3 or b < 3, respectively. The correlation coefficient (r2) of data points
about a fitted curve was used to assess the general goodness of fit for body mass
and dimension data to power function.
30 Southeastern Naturalist Vol. 12, No. 1
The five growth models chosen in the study represent diminishing returns
behavior (monomolecular), sigmoidal with a fixed point of inflection (Gompertz,
logistic, von Bertalanffy), and sigmoidal with a variable point of inflection (Richards),
as described in Ratkowsky (1983). Each of the four 3-parameter models
can be derived from the generalized Richards model (Richards 1959) that encompasses
different growth functions for particular values of an additional shape
parameter (m) which determine the position of the inflection point. All models
used were from general forms of the equations and have similar properties (e.g.,
unweighted nonlinear regressions; behavior describing body size to an asymptotic
function of age as the response variable; asymptotic body size and maturing
index have the same biological meaning). Equations 1–5 are the monomolecular
(Brody 1945), Gompertz (Gompertz 1825), logistic (Verhulst 1838), von Bertalanffy
(von Bertalanffy 1957), and Richards (Richards 1959) growth models,
 A(t) = A∞ · [1 - be–Kt]
 A(t) = A∞ · e[–e–K(t–I)]
 A(t) = A∞ · [1 + e–K(t–I)] –1
 A(t) = A∞ · [1– e–K(t–T)]p
 A(t) = A∞ · [1 + be–Kt] – (1/m)
The parameter A(t) is an observed body length (cm) or mass (kg) at age t (yr),
and A∞ is the estimated asymptotic size of the sampled population. The parameter
K is the estimated relative growth rate or “maturing index” that describes the rate
at which asymptotic body length or mass is reached (yr–1). The parameter b is an
integration constant that represents a time-scale factor, and I is the age (yr) at the
inflection point. T is a fitting constant and is interpreted as the hypothetical age
(yr) of an individual at zero body length or mass, assuming the equation to be
valid at all ages. The exponent p determines the model for body length ( p = 1)
and mass ( p = 3).
The monomolecular model is the simplest equation used in this study and
it describes the progress of an irreversible first-order reaction (Brody 1945).
The Gompertz model arises from an equation that describes self-limited growth
where the rate decreases exponentially with time (Gompertz 1825). The logistic
model is more complex, and arises from an equation that describes three stages
of growth; early exponential growth where the rate is proportional to size, linear
growth where resources are devoted to maintenance, and diminishing growth
as a maintenance balance is approached (Verhulst 1838). The von Bertalanffy
model arises from an equation that describes metabolic laws and the balance
between anabolism and catabolism (von Bertalanffy 1957). The Richards model
does not have the underlying biological basis of the previous models and arises
on an empirical basis as a theoretical advancement that allows application to both
exponential and sigmoid growth curves (Richards 1959). The point of inflection
is defined as the theoretical time of maximum growth, and for monomolecular,
Gompertz, logistic, and von Bertalanffy models, is fixed at 33.3 % (A∞ · [1/3]),
36.8% (A∞/e), 50.0% (A∞/2), and 29.6% (A∞ · [8/27]p) of asymptotic body size,
2013 T. Bartareau, D. Onorato, and D. Jansen 31
respectively. The point of inflection for the Richards model is flexible and occurs
at any fraction of asymptotic body size.
Growth data were fitted to each model using the Levenberg-Marquardt-
Nash algorithm (Analytical Software 2008). The small-sample, bias-corrected
form of the Akaike information criterion (AICc) was used to evaluate and compare
models (Burnham and Anderson 2002). The Akaike weight (wi) of each
model was calculated to evaluate the weight of evidence that i is the best model
within the available set of models (Burnham and Anderson 2002). The best
model had the lowest AICc value and the highest wi, recognizing that models ±
2 AICc are considered equal (Burnham and Anderson 2002). Residual sums of
squares (RSS), pseudo R2, and residual standard deviation of data points about
fitted models (RSD) were used to evaluate general goodness of fit and accuracy
of each model to observed growth data (Sokal and Rohlf 1995). T-tests for
independent samples (Sokal and Rolf 1995) were used to evaluate differences
between model parameter estimates by gender. Statistical analyses were conducted
using Microsoft Excel® (Microsoft Corporation, Redmond, WA) and
Statistix® 9.0 (Analytical Software 2008), and we employed an alpha value of
0.05. All means are presented ± standard error.
The body length and mass of Panthers aged 0.01 to 12 yrs differed by gender.
Body length ranged from 39.0 to 192.5 cm for females and 44.0 to 219.0 cm for
males. Body mass ranged from 0.43 to 51.03 kg for females and 0.45 to 72.6 kg
for males. Body mass increased with length (Fig. 1.A), and the early growth rate
was rapid, reaching a maximum at the point of inflection in the curve and then
declining to zero at asymptotic body length (Fig. 1B) and mass (Fig. 1C).
The power functions for females (a = 2.64 x 10-5 ± 1.1 x 10-4, b = 2.7 ± 0.8,
n = 12, r2 = 0.916, P < 0.001) and males (a = 7.9 x 10-6 ± 2.1 x 10-5, b = 2.9 ±
0.5, n = 13, r2 = 0.975, P < 0.001) indicated that growth in body mass relative
to length was nearly isometric and negative allometric. The power functions
did not differ significantly by gender; females had a slightly larger rate of
change in mass per unit length than males (t23 = 0.173, P = 0.864) and were
slightly smaller per unit length (t23 = 0.274, P = 0.787).
The Gompertz, logistic, monomolecular, and von Bertalanffy models were
fitted without difficulty, and convergence criteria were met after 7 to 14 iterations
(Table 1). Difficulties were encountered when fitting the Richards model,
for which convergence criteria were met after 21 to 56 iterations (Table 1). Each
of the five models fitted the growth data well for both genders and pseudo R2
were ≥98.7%. For both genders and body size measurements, the von Bertalanffy
model performed the best with largest pseudo R2 and lowest RSS, RSD and AICc
values (Table 1). The von Bertalanffy model accounted for ≥45.8% of model
weight (wi). The Gompertz, monomolecular, Richards and logistic models were
least supported by the data with wi ≤ 35.4%, ≤ 31.0%, ≤ 18.5%, and ≤ 10.1%,
respectively (Table 1).
32 Southeastern Naturalist Vol. 12, No. 1
The growth data for both genders fit the von Bertalanffy model well in
younger Panthers, but became more variable in specimens >1 yr of age (Fig.
1B, C). Individual variation in body length and mass growth data were greater
than differences due to gender among panthers age ≤0.08 and 0.38 yr, respectively.
The difference between estimated and actual body length were 6.1 ± 0.9
Figure 1. Body
length and mass
growth curve of
wild female (□)
and male (■) Puma
as estimated from
function, and von
model for (B) body
length (C) and
mass by age. Fitted
lines reflect growth
curve for female
(dashed line) and
male (solid line).
2013 T. Bartareau, D. Onorato, and D. Jansen 33
cm (3.3% of asymptote) for females and 5.2 ± 1.1 cm (2.5%) for males. The
difference between estimated and actual body mass were 1.1 ± 0.2 kg (2.9%)
for females and 1.1 ± 0.2 kg (1.9%) for males. The estimated body length and
mass at birth was 37.95 cm and 0.58 kg for females and 36.0 cm and 0.55 kg
for males, respectively.
For each growth model (Table 2), the asymptotic body length and mass values
were greater in males than in females. The von Bertalanffy model estimated
asymptotic body length and mass values for males were 9.4% longer (t23 = 3.41,
P = 0.002) and 33.2% heavier (t359 = 18.32, P < 0.001) than that for females.
The corresponding maturing index values were smaller in males than in females.
The von Bertalanffy model estimated maturing index values for female
body length and mass were 6.1% and 21.7% larger than for males, respectively,
with significant gender differences in relative growth rate evident for mass
(t359 = 4.14, P < 0.001) and not body length (t23 = 0.05, P = 0.961).
Table 1. Assessment of five body length and mass growth models of wild Puma concolor coryi
(Florida Panther) based on number of model parameters (P), number of necessary iterations for
convergence (I), the relative model fit using residual sums of squares (RSS), pseudo R2 (R2), residual
standard deviation of data points about fitted growth curve (RSD), minimum small-sample
bias-corrected form of the Akaike’s information criteria value of model i in the candidate set (Δi),
and weight of evidence that model i is the best of the available set (wi).
Model P I RSS R2 RSD Δi wi
Body length (cm)
Female (n = 12)
Bertalanffy 3 17 564.91 0.999 7.92 0.000 0.470
Monomolecular 3 11 565.91 0.997 7.94 1.252 0.251
Gompertz 3 9 631.60 0.995 8.38 2.065 0.167
Logistic 3 9 672.07 0.993 8.64 3.085 0.101
Richards 4 56 631.61 0.995 8.89 7.625 0.010
Male (n = 13)
Bertalanffy 3 7 552.63 0.999 7.42 0.000 0.458
Monomolecular 3 14 553.83 0.998 7.43 0.785 0.310
Gompertz 3 11 657.88 0.997 8.11 2.265 0.148
Logistic 3 8 730.05 0.993 8.54 3.619 0.075
Richards 4 28 657.89 0.997 8.55 7.837 0.009
Body mass (kg)
Female (n = 171)
Bertalanffy 3 15 987.97 0.999 2.42 0.00 0.537
Gompertz 3 12 989.67 0.995 2.43 1.72 0.227
Richards 4 24 990.09 0.991 2.44 2.13 0.185
Logistic 3 12 1015.70 0.989 2.46 4.73 0.050
Monomolecular 3 9 1062.10 0.987 2.51 12.38 0.001
Male (n = 190)
Bertalanffy 3 13 1644.20 0.999 2.97 0.00 0.523
Gompertz 3 8 1651.00 0.993 2.98 0.78 0.354
Richards 4 21 1653.05 0.991 2.98 2.89 0.123
Logistic 3 16 2016.00 0.989 3.28 38.74 0.000
Monomolecular 3 12 2021.60 0.987 3.29 39.27 0.000
34 Southeastern Naturalist Vol. 12, No. 1
Table 2. Estimated parameters of five growth models of wild Puma concolor coryi (Florida Panther). Estimates (± SE) are shown of asymptotic body length
and mass (A∞), maturing index (K), integration constant (b), age at the inflection point ( I), fitting constant (T), and shape parameter (m).
Model A∞ K (yr-1) b I (yr) T (yr) m
Body length (cm)
Bertalanffy 186.15 (4.15) 2.28 (0.39) - - -0.10 (0.03) -
Monomolecular 186.15 (4.16) 2.28 (0.39) -0.10 (0.03) - - -
Gompertz 185.98 (4.55) 3.09 (0.47) - 0.37 (0.06) - -
Logistic 185.86 (4.35) 4.21 (0.57) - 1.11 (0.10) - -
Richards 185.98 (4.28) 3.09 (5.83) -9.45 (9.37) - - 5.43x10-5 (3.46)
Bertalanffy 205.47 (3.86) 2.14 (0.27) - - -0.09 (0.02) -
Monomolecular 205.47 (3.87) 2.14 (0.27) -0.09 (0.02) - - -
Gompertz 203.24 (3.66) 3.06 (0.34) - 0.46 (0.06) - -
Logistic 201.95 (3.59) 4.22 (0.42) - 1.28 (0.11) - -
Richards 203.24 (4.99) 3.06 (1.68) -9.35 (2.75) - - 5.44x10-5 (1.47)
Body mass (kg)
Bertalanffy 38.31 (0.47) 2.03 (0.09) - - -0.14 (0.02) -
Gompertz 37.94 (0.45) 2.67 (0.12) - 1.38 (0.06) - -
Richards 37.94 (0.48) 2.67 (0.47) -8.18 (3.03) - - 7.06x10-5 (0.21)
Logistic 37.25 (0.43) 5.44 (0.29) - 3.64 (0.18) - -
Monomolecular 39.11 (0.54) 1.18 (0.06) 0.02 (0.01) - - -
Bertalanffy 57.34 (0.89) 1.59 (0.06) - - -0.15 (0.02) -
Gompertz 56.08 (0.81) 2.11 (0.08) - 1.42 (0.05) - -
Richards 56.08 (0.88) 2.11 (0.24) -9.16 (4.48) - - 2.55x10-5 (0.11)
Logistic 53.62 (0.74) 4.19 (0.19) - 3.54 (0.14) -
Monomolecular 60.35 (1.30) 0.86 (0.05) 0.02 (0.01) - -
2013 T. Bartareau, D. Onorato, and D. Jansen 35
The asymptotic body length and mass estimates of the von Bertalanffy and
monomolecular models were higher and the maturing index lower than those estimated
by all other models. The asymptotic body length and mass estimates for the
logistic model were smaller and the maturing index larger than those estimated
by other models. The asymptotic body length or mass and maturing index estimates
for Gompertz and Richards models were nearly identical and fell between
those of other models. The von Bertalanffy model placed the inflection point for
body length and mass growth curves at age 0.05 yr and 0.4 yr in females and
0.07 yr and 0.5 yr in males, respectively. The estimated age to reach asymptotic
body length was 4.6 yr for females and 4.9 yr for males. Growth in body mass
continued for a longer duration than length, and females were estimated to reach
the asymptotic value at age 4.9 yr, about 1.6 yr earlier than males.
Historically, inference and estimation of Puma growth curves and their
precision relied solely on the Richards model (Laundré and Hernández 2002,
Maehr and Moore 1992), but other models may be more appropriate for some
populations (Zullinger et al. 1984). Growth curves for Pumas can differ between
genders and populations (Laundré and Hernández 2002, Maehr and Moore 1992),
and assessing multiple models through an information-theoretic approach should
delineate which model provides the best representation of observed growth data.
This application of an information-theoretic-approach to five classical growth
models for body length and mass at-age data from wild Panthers is, to our
knowledge, the first such study for Pumas. Also, the data set of the current study
is unique based on the large sample of Panther body-size measurements over a
wide range of known ages. Our results suggest that when used for inferences on
the Panther, the von Bertalanffy model provided the least-biased point estimates
of growth in body length and mass as a function of age.
The different models produced variable results when fitted to the growth data.
In examining ease of convergence to obtain parameters as initial criterion to
evaluate candidate models for goodness of fit, our analyses revealed problems of
convergence for the Richards model resulting in more iterations required to fit the
data compared to the Gompertz, logistic, monomolecular, and von Bertalanffy
models. The Richards model is flexible, but it presents certain disadvantages such
as a shape parameter that has no obvious biological interpretation (Ratkowsky
1983), and its parameters exhibit significant colinearity and are sometimes
numerically unstable (Zeide 1993). The Richards model was most difficult for
practical use in describing the growth curve of Panthers and, in this study, is an
example of over-parameterization.
Based on a pseudo R2 comparison alone, with a minimum and maximum value
of 0.987 and 0.999, no model was appreciably better than the other. However,
model selection according to RSS, RSD, and AICc revealed that the von Bertalanffy
model had the lowest values, indicating the best fit to growth data for
both genders relative to the other models. The monomolecular and Gompertz
models provided a similar fit to von Bertalanffy body length and mass models,
36 Southeastern Naturalist Vol. 12, No. 1
respectively (ΔAICc < 2). The logistic and Richards models provided the least
best fit relative to the other models (ΔAICc > 2). Akaike weights indicated that the
Gompertz, logistic, monomolecular, and Richards growth models have considerably
less support (wi ≤ 35.4%) than does the von Bertalanffy model (wi ≥ 45.8%).
While not specifically tested, this finding is in agreement with Maehr and Moore
(1992), who suggested that Gompertz and logistic forms of the Richards model
were unsuitable for modeling body mass-at-age growth curves of Panthers. Our
results were consistent with Laundré and Hernández (2002), who noted that the
body mass growth curves of Pumas in their study were similar to the von Bertalanffy
form of the Richards model.
The difficulty in comparing results of this study with other studies lies in the
different data sets that were examined. For instance, Maehr and Moore (1992)
found that gender differences in body mass growth rates to be inconsistent
because of the sparse data for 1- to 4-yr-old Panthers in their data set. Like Laundré
and Hernández (2002), our body-mass data set contained ample specimens
for this age span, and the model results indicated a gender difference in growth
rates. Maehr and Moore (1992) also estimated unreasonably high birth mass due
to a lack of growth data for Panthers aged ≤0.33 yr. In contrast, the presence of
growth data for Pumas aged ≤0.02 yr in the data set of Laundré and Hernández
(2002) gave reasonable model estimate of birth mass. We had data for Panthers
aged 0.01 yr and estimated realistic birth body length and mass sizes when compared
to those reported in the literature (Currier 1983, Robinette et al. 1961),
indicating that our data set and choice of model was suitable for describing the
early postnatal growth curve of Panthers.
Growth in body dimension and mass for both genders followed a sigmoid
curve, a form of progressive development that is prevalent among determinately
growing animals (Ratkowsky 1983). The estimated age at inflection point for
body length (≤0.07 yr) and mass (≤0.5 yr) models indicated that growth was more
rapid during the early postnatal period than at juvenile or adult ages, and most of
a Panther’s lifetime was spent near asymptotic size. The failure of estimates for
the generalized Richards model to be more strongly correlated to data than those
for the von Bertalanffy model is an indication of a rational effect of fixing the
point of inflection at about 29.6% of asymptotic size in a Panther growth curve.
We conclude from these results that the von Bertalanffy model is the best-fitting
theoretical equation to our data set for estimating the body length and mass
growth curve of Panthers from birth to age ≤12 yr .
Animal growth can exhibit complex seasonal patterns that depend on the
growth data selected. For instance, it can be difficult to separate continuing mass
growth from non-monotonic, oscillating changes in mass due to accumulation
and utilization of fat. Pumas can consume up to 10 kg of prey in a single feeding
(Ackerman et al. 1986), and variation in body mass independent of age is
expected (Jansen and Jenks 2010, Kohlmann and Green 1999). However, the
difference between estimated and actual body mass of 1.1 ± 0.2 kg in this study
is well within the range that might be expected from measurements of Panthers
2013 T. Bartareau, D. Onorato, and D. Jansen 37
made during a random capture. This difference could reflect variation in stomach
contents such as the time since last feeding and the amount of food consumed
(Ackerman et al. 1986) and fluctuation in mass of females when pregnant or
rearing offspring (Jansen and Jenks 2010). The large data set with inclusion of
size-increment growth data used to construct our growth models should reduce
bias associated with limited sampling. Thus, we believe that our von Bertalanffy
model parameter estimates reliably explain the observed growth data and accurately
depict the gender-specific body length and mass growth curves of wild
Panthers as a function of age.
The results of the von Bertalanffy growth models indicated that male Panthers
reached physical maturity at an older age than females, and significant gender
differences in growth rates were evident for body mass but not length. Both genders
continued growth in body mass for a longer duration than length. A nearly
isometric relationship of body length and mass is indicated by the similar value
in the exponent of the power curve for females (b = 2.7) and males (b = 2.9). The
von Bertalanffy mass function is based on a cubic relationship, and the maturing
index values for female and male body mass were 89.0% and 74.3% of those for
length, indicating that any fraction of mass growth was reached later than the
same fraction of length.
The estimated asymptotic body length and mass of Panthers indicated malebiased
sexual size dimorphism. Adult males were 19.32 cm (9.4%) longer and
19.0 kg (33.2%) heavier than females. Our results support findings from other
studies that report male Pumas are about 10% longer and 30% heavier than
females (Currier 1983). The asymptotic size estimates were not the largest observed
body length or mass attained by any individual in the population, but
instead they represent the mean maximum size attained in the population at
maturity, as argued from metabolic laws. Possible explanations for this sexual
size dimorphism may be that males grow faster, continue growing for a longer
duration, or a combination of these developmental processes.
The growth curves indicated that both genders grew at about the same rate
after birth, but there was well-defined body length and mass sexual size dimorphism
at age ≥0.09 and ≥0.39 yr, respectively. The fact that males ultimately
grew to a larger asymptotic body length and mass than females of the same age
revealed that a male must have a greater absolute rate of growth than female. Logan
and Sweanor (2001) found that sexual size dimorphism in Puma body mass
can develop as early as weaning by age 0.14 yr due to faster male growth. Panther
kittens are weaned by age 0.3 yr (Maehr 1997), and independence and dispersal
of offspring generally occurs at just over 1 year of age (Hostetler et al. 2012,
Maehr et al. 2002). Therefore, the proximate cause of gender differences in the
rate of body length and mass growth apparently begins to operate after offspring
are weaned and continues through to family dissolution and adulthood.
The gender differences in growth curves have biological and ecological
implications because of the correlation of Puma age at first reproduction with
growth (Robinette et al. 1961). Female Puma body mass is often attributed to
38 Southeastern Naturalist Vol. 12, No. 1
the accommodation of developing offspring, and variation in the age to reach an
estimated minimum size at sexual maturity of 36.3 kg to some extent is related to
differences in growth rates (Robinette et al. 1961). Our growth models estimated
that female Panthers reached 36.3 kg at 1.8 yr old, an age coinciding with 1.75
yr, when individuals are known to first reach sexual maturity (Benson et al. 2011;
Hostetler et al. 2012; Maehr 1997; Maehr et al. 1989, 1991). A female at age 1.8
yr is estimated to have reached 98.7% and 94.3% of asymptotic body length and
mass, respectively. The first sexual encounters for male Panthers typically occurs
at about 3 yrs old based on radio-collared panthers of known ages (Maehr 1997,
Maehr et al. 1991), but genetics revealed that males may become breeders as early
as 1.4 years old (Johnson et al. 2010; Florida Fish and Wildlife Conservation
Commission, Naples, FL, unpubl. data). A male at age 1.4 yr is estimated to have
reached 95.9% and 76.6% of asymptotic body length and mass, respectively.
These results indicated that growth in body length of both genders was largely
complete by age at sexual maturity, but males continue growth in mass for longer
duration than females after the onset of sexual maturation.
The modeling of Puma growth curves is a useful tool for the derivation of gender-
specific body mass estimates at different ages (e.g., Laundré 2005). The social
structuring of Puma populations makes growth information valuable in studying
population dynamics because male reproductive success depends on the ability to
grow large enough in size to acquire territory, thwart competitors, and mate successfully
with females in estrus (Logan and Sweanor. 2001, Maehr 1997, Maehr
et al. 1991). Growth curves are also important components of bioenergetic models
(Brody 1945) and could be combined with activity data to estimate food and space
requirements for the population (e.g., Ackerman et al. 1986, Laundré 2005). Describing
food and space requirements of populations is essential for examining
range quality and carrying capacity, and has importance in developing conservation
and management policies for Pumas (e.g., Laundré 2005). Future data collection,
especially with regard to body size and age at different developmental stages (e.g.,
birth, weaning, sexual maturity, first reproduction) would enhance our understanding
of the ontogeny of sexual size dimorphism in this species.
We thank O.L. Bass, R.C. Belden, J.E. Benson, M. Criffield, M. Cunningham, J. Kellam,
M. Lotz, A. Johnson, E.D. Land, D.S. Maehr, R.C. McBride, R.T. McBride, R.M.
McBride, C.E. McBride, J.W. McCown, J.C. Roof, S. Schulze, and D.B. Shindle for their
assistance in collecting the samples used in this study. The Florida Fish and Wildlife
Conservation Commission and National Park Service funded panther research, and this
paper is dedicated to that ongoing work. Additionally, we thank the citizens of Florida
for their contributions to the Florida Panther Research and Management Trust Fund via
the purchase of “Protect the Panther” license plates. Thanks to L. Pulliam for diligence in
sourcing literature, B. Crowder, J. Gore, J. Hostetler, E. Leone, and M. Oli for constructive
comments on the earlier manuscript.
2013 T. Bartareau, D. Onorato, and D. Jansen 39
Ackerman, B.B., F.G. Lindzey, and T.P. Hemker. 1986. Predictive energetics model for
Cougars. Pp. 333–352, In S.D. Miller and D.D. Everett (Eds.). Cats of the World:
Biology, Conservation, and Management. National Wildlife Federation, Washington,
Akaike, H. 1981. Likelihood of a model and information criteria. Journal of Econometrics
Analytical Software. 2008. Statistix 9.0. Tallahassee, FL.
Benson, J.F., J.A. Hostetler, D.P. Onorato, W.E. Johnson, M.E. Roelke, S.J. O’Brien, D.
Jansen, and M.K. Oli. 2011. Intentional genetic introgression influences survival of
adults and subadults in a small, inbred felid population. Journal of Animal Ecology
Brody, S. 1945. Bioenergetics and Growth. Reinhold Publishing, New York, NY.
Burnham, K.P., and D.R. Anderson. 2002. Model Selection and Multimodel Inference: A
Practical Information-Theoretic Approach. Springer-Verlag, New York, NY.
Currier, M.J. 1983. Felis concolor. Mammalian Species 200:1–7.
Florida Fish and Wildlife Conservation Commission. 2011. Annual report on the research
and management of Florida panthers: 2010–2011. Fish and Wildlife Research Institute
and Division of Habitat and Species Conservation, Naples, FL.
Gay, S.W., and T.L Best. 1995. Geographic variation in sexual dimorphism of the
Puma (Puma concolor) in North and South America. The Southwestern Naturalist
Gompertz, B. 1825. On the nature of the function expressive of the law of human mortality,
and a new mode of determining the value of live contingencies. Philosophical
Transmission of the Royal Society of London 1 15:513–585.
Iriarte, J.A., W.L. Franklin, W.E. Johnson, and K.H. Redford. 1990. Biogeographic variation
of foods habits and body size of the American Puma. Oecologia 85:185–190.
Jansen, B.D., and J.A. Jenks. 2010. Estimating body mass of Pumas (Puma concolor).
Wildlife Research 38:147–151.
Jansen, D., J. Kellam, and A. Johnson. 2010. Florida Panther (Puma concolor coryi)
research and monitoring in Big Cypress National Preserve 2009–2010 annual report.
National Park Service, Ochopee, FL.
Hostetler, J.A., D.P. Onorato, B.M. Bolker, W.E. Johnson, S.J. O’Brien, D. Jansen, and
M.K. Oli. 2012. Does genetic introgression improve female reproductive performance?
A test on the endangered Florida Panther. Oecologia 168:289–300.
Johnson, W.E., D.P. Onorato, M.E. Roelke, E.D. Land, M. Cunningham, C. Belden, R.
McBride, D. Jansen, M. Lotz, D. Shindle, J. Howard, D.E. Wildt, L.M. Penfold, J.A.
Hostetler, M.K. Oli, and S.J. O’Brien. 2010. Genetic restoration of the Florida Panther.
Kohlmann, S.G., and R.L. Green. 1999. Body size dynamics of Cougars (Felis concolor)
in Oregon. Great Basin Naturalist 59:193–194.
Land, E.D., D.B. Shindle, R.J. Kawula, J.F. Benson, M.A. Lotz, and D.P. Onorato. 2008.
Florida Panther habitat-selection analysis of concurrent GPS and VHF telemetry data.
Journal of Wildlife Management 72:633–639.
Laundré, J.W. 2005. Puma energetics: A recalculation. Journal of Wildlife Management
Laundré, J.W., and L. Hernández. 2002. Growth-curve models and age estimation
of young Cougars in the northern Great Basin. Journal of Wildlife Management
40 Southeastern Naturalist Vol. 12, No. 1
Logan, K.A., and L.L. Sweanor. 2001. Desert Puma: Evolutionary Ecology and Conservation
of an Enduring Carnivore. Island Press, Washington, DC.
Maehr, D.S. 1997. The comparative ecology of Bobcat, Black Bear, and Florida Panther
in south Florida. Bulletin of the Florida Museum of Natural His tory 40:1–176.
Maehr, D.S., and C.T. Moore. 1992. Models of mass growth for 3 North American Cougar
populations. Journal of Wildlife Management 56:700–707.
Maehr, D.S., J.C. Roof, E.D. Land, and J.W. McCown. 1989. First reproduction of a
Panther (Felis concolor coryi) in southwestern Florida, USA. Mammalia 53:129–131.
Maehr, D.S., E.D. Land, and J.C. Roof. 1991. Social ecology of Florida panthers. National
Geographic Research and Exploration 7:414–431.
Maehr, D.S., E.D. Land, D.B. Shindle, O.L. Bass, and T.S. Hoctor, 2002. Florida Panther
dispersal and conservation. Biological Conservation 106:187–197 .
Onorato, D., C. Belden, M. Cunningham, D. Land, R. McBride, and M. Roelke. 2010.
Long-term research on the Florida Panther (Puma concolor coryi): Historical findings
and future obstacles to population persistence. Pp. 452–469, In D. Macdonald and A.
Loveridge (Eds.). Biology and Conservation of Wild Felids. Oxford University Press,
Onorato, D.P., M. Criffield, M. Lotz, M. Cunningham, R. McBride, E.H. Leone, O.L.
Bass, Jr., and E.C. Hellgren. 2011. Habitat selection by critically endangered Florida
Panthers across the diel period: Implications for land management and conservation.
Animal Conservation 14:196–205.
Ratkowsky, D.A. 1983. Nonlinear Regression Modeling: A Unified Practical Approach.
Marcel Dekker, New York, NY.
Richards, F.J. 1959. A flexible growth curve for empirical use. Journal of Experimental
Robinette, W.L., J.S. Gashwiler, and O.W. Morris. 1961. Notes on Cougar productivity
and life history. Journal of Mammalogy 42:204–217.
Sokal, R.R., and F.J. Rohlf. 1995. Biometry. Third Edition. W.A. Freeman and Company.
New York, NY.
Stearns, S.C. 1992. The Evolution of Life Histories. Oxford University Press, Oxford, UK.
Verhulst, P.F. 1838. Notice sur la loi que la population suit dans son accroissement. Correspondance
Mathématique et Physique 10:113–121.
von Bertalanffy, L. 1957. Quantitative laws in metabolism and growth. Quarterly Review
of Biology 32:217–231.
Zar, J.H. 1999. Biostatistical Analysis. Fourth Edition. Prentice-Hall, Upper Saddle
Zeide, B. 1993. Analysis of growth equations. Forestry Science 39:594–616.
Zullinger, E.M., R.E. Rickleps, K.H. Reddord, and G.M. Mace. 1984. Fitting sigmoidal
equations to mammalian growth curves. Journal of Mammalogy 65:6 07–636.