Regression equations between body and head measurements in the broad-snouted caiman (Caiman latirostris)

Equações de regressão entre medidas de corpo e cabeça em jacarés-de-papo-amarelo (Caiman latirostris)

REGRESSION EQUATIONS BETWEEN BODY AND HEAD MEASUREMENTS IN THE BROAD-SNOUTED CAIMAN (Caiman latirostris)

VERDADE, L. M.

Laboratório de Ecologia Animal, ESALQ, Universidade de São Paulo, C.P. 09, CEP 13418-900, Piracicaba, SP, Brazil

Correspondence to: Luciano M. Verdade, Laboratório de Ecologia Animal, ESALQ, Universidade de São Paulo, C.P. 09, CEP 13418-900, Piracicaba, SP, Brazil, e-mail: [e-mail]

Received March 4, 1999 ¾ Accepted December 22, 1999 ¾ Distributed August 31, 2000

(With 5 figures)

ABSTRACT

In the present study, regression equations between body and head length measurements for the broad-snouted caiman (Caiman latirostris) are presented. Age and sex are discussed as sources of variation for allometric models. Four body-length, fourteen head-length, and ten ratio variables were taken from wild and captive animals. With the exception of body mass, log-transformation did not improve the regression equations. Besides helping to estimate body-size from head dimensions, the regression equations stressed skull shape changes during the ontogenetic process. All age-dependent variables are also size-dependent (and consequently dependent on growth rate), which is possibly related to the difficulty in predicting age of crocodilians based on single variable growth curves. Sexual dimorphism was detected in the allometric growth of cranium but not in the mandible, which may be evolutionarily related to the visual recognition of gender when individuals exhibit only the top of their heads above the surface of the water, a usual crocodilian behavior.

Key words: relative growth, sexual dimorphism, size estimates, broad-snouted caiman, Caiman latirostris.

RESUMO

Equações de regressão entre medidas de corpo e cabeça em jacarés-de-papo-amarelo (Caiman latirostris)

No presente estudo, equações de regressão entre medidas de comprimento do corpo e cabeça de jacarés-de-papo-amarelo (Caiman latirostris) são apresentadas. Idade e sexo são discutidos como fontes de variação para modelos alométricos. Quatro medidas de comprimento corpóreo, 14 medidas de comprimento da cabeça e dez proporções relativas entre medidas foram tomadas de animais selvagens e cativos. Com excessão da massa corpórea, a transformação logarítmica não incrementou as equações de regressão. Além de auxiliar na estimativa do comprimento corpóreo a partir de dimensões da cabeça, as equações de regressão evidenciaram alterações na forma craniana durante processos ontogênicos. Todas as variáveis dependentes da idade mostraram-se também dependentes do tamanho (e conseqüentemente da taxa de crescimento), o que está possivelmente relacionado à dificuldade em prever a idade de crocodilianos com base apenas em curvas univariadas de crescimento. Dimorfismo sexual foi detectado no crescimento alométrico do crânio, mas não da mandíbula, o que pode estar evolutivamente relacionado ao reconhecimento visual do sexo quando os indivíduos exibem apenas o topo da cabeça acima da superfície da água, um comportamento normal em crocodilianos.

Palavras-chave: crescimento relativo, dimorfismo sexual, estimativas de tamanho corpóreo, jacarés-de-papo-amarelo, Caiman latirostris.

INTRODUCTION

Allometric relations can be useful for estimating body size from isolated measures of parts of the body (Schmidt-Nielsen, 1984). Population monitoring of crocodilians usually involve night counts when frequently only the heads of animals are visible. Thus, the relationship between length of head and total body length is usually employed to establish size-class distribution for the target populations. As an example, Chabreck (1966) suggests that the distance between the eye and the tip of the snout in inches is similar to the total length of Alligator mississippiensis in feet. Choquenot & Webb (1987) propose a photographic method to estimate total length of Crocodylus porosus from head dimensions. In order to improve these techniques, Magnusson (1983) suggests that a sample of animals should be captured and measured. Thus, relationships between estimates and actual animals' dimensions could be established and observers' bias could be corrected. The interesting point of this method is that it permits a quantification of the actual observers' bias.

In the present study, regression equations between body and head length measurements for both wild and captive broad-snouted caiman (Caiman latirostris) are presented. Age and sex are discussed as sources of variation for allometric models. Sexual dimorphism, ontogenetic variation and morphometric differences between wild and captive individuals are discussed in more detail by Verdade (1997).

MATERIAL AND METHODS

Body and head measurements were taken from 244 captive and 29 wild animals. The captive animals were located at Escola Superior de Agricultura "Luiz de Queiroz", University of São Paulo, Piracicaba, State of São Paulo, Brazil. Information about their age, sex, date of birth, and pedigree are available at the regional studbook of the species (Verdade & Santiago, 1991; Verdade & Molina, 1993; Verdade & Kassouf-Perina, 1993; Verdade & Sarkis, in press). The wild animals were captured on small wetlands associated with tributaries of Tietê River in East-Central São Paulo State from October 1995 to May 1996.

Capture techniques consisted of approaching the animals by boat at night with a spotlight. Juveniles (< 1.0 m total length) were captured by hand, similar to the method described by Walsh (1987). Noosing, as described by Chabreck (1963), was tried unsuccessfully for adults. The adult caimans were too wary and usually submerged before the noose was in place, similarly to what was experienced by Webb & Messel (1977) with Crocodylus porosus in Australia and Hutton et al. (1987) in Zimbabwe. Rope traps (adapted from Walsh, 1987) were also tried unsuccessfully for both adults and young. Captive individuals were taken either by hand or noose according to their size, on daytime in October 1996.

The captured animals were physically restrained during data collection. No chemical immobilizion was used. Body measurements (body-size variables) were taken with a tape measure (1 mm precision). Head measurements (head-size variables) were taken with a steel Summit Vernier caliper (.02 mm precision, second decimal unconsidered). Body mass was taken with Pesola hanging scales (300 x 1 g, 1,000 x 2 g, 5,000 x 5 g, 20 x 0.1 kg, 50 x 0.1 Kg, depending on individual body mass). Animals were sexed through manual probing of the cloaca (Chabreck, 1963) and/or visual examination of genital morphology (Allstead & Lang 1995) with a speculum of appropriate size.

Four body-size, fourteen head-size, and ten ratio variables were taken from wild and captive animals (Fig. 1, Table 1). Eight head-size variables are "length" measurements in the sense that they are longitudinal in relation to the body. The other six head-size variables are "width" measurements in the sense that they are transversal in relation to the body. Ten head-size variables are located on the upper jaw and cranium, whereas the other four head-size variables are located on the lower jaw. Four ratio variables represent relative length, whereas the other six represent relative width. Eight ratio variables are located on the upper jaw and cranium, whereas the other two are located on the lower jaw. One of these measurements, PXS, the length of the premaxillary symphysis, is not visible in live animals but is closely approximated by the distance from the snout tip to the anterior tip of the first tooth posterior to the prominent groove in the snout behind the nares (usually the 6th or 7th tooth).

"Size" and "shape" are difficult to define in biology (Bookstein, 1989). Unidimensional length measurements do not express the multidimensionality of size. However, since length and size are positively correlated in caimans, length measurements are called size-variables in this paper for the sake of simplicity. The morphometric variables used in this study were adapted from Iordansky (1973). They are based on linear distances between landmarks (body- and head-size variables) or ratios between measurements (ratio variables). The use of ratios present several disadvantages. Ratios tend to be relatively inaccurate, not-normally distributed, and discontinuous (Sokal & Rohlf, 1995). However, since ratios are still used by some authors (Hall & Portier, 1994) they have been included and discussed in the present study for comparative purposes.

Hall and Portier call these ratios relative growth indices. Relative growth represents change of proportions as body size increases. The study of relative growth has been characterized by Gould (1966) as the study of size and its implications in ontogeny and phylogeny. However, disregarding growth processes and size implications, these ratios express non-metric variables in the sense that they represent relative length and width instead of absolute values.

All statistical analyses were done in Minitab for Windows (Minitab, 1996) and their procedures are shown when adequate.

ALLOMETRIC RELATIONS

Table 2 and Fig. 2 show the regression equations and respective plots between body- and head-size variables and the snout-vent length (SVL) in wild individuals. Table 3 and Fig. 3 show the regression equations and respective plots between ratio variables and SVL in wild individuals. Due to the relatively small sample size, wild males and females are presented together. Table 4 and Fig. 4 show the regression equations and respective plots between body- and head-size variables and the snout-vent length (SVL) in captive animals. Table 5 and Fig. 5 show the regression equations and respective plots between ratio variables and SVL in captive animals.

With the exception of body mass (BM), log-transformation did not improve regression equations for either wild or captive animals. Logarithmic transformation is a simple device that may ease and improve diagrammatic and statistical descriptions of the effect of body size on other attributes (Peters, 1983). Regression equations for captive animals presented a higher coefficient of determination (r²) than the ones for wild animals. Body- and head-size variables presented a significantly higher r² than ratio variables for both wild and captive animals. They varied from 0.826 (OL) to 0.979 (CW) for body- and head-size variables (Table 2), and from 0.002 (RLSS) to 0.581 (RLST) for ratio variables (Table 3) for wild animals. For captive animals, in their turn, they varied from 0.916 (OW) to 0.993 (SW) for body- and head-size variables (Table 4), and from 0.003 (RLSS) to 0.934 (RLST) for ratio variables. The range of SVL relative to each equation can be found on the plots of Figs. 2 to 5.

The coefficients of determination of wild and captive animals concerning body- and head-size variables can be considered extremely high. Their main biological meaning is the apparent lack of morphological variation on the patterns studied, which could be expected for captive but not for wild animals. They also mean that most of the head-size variables studied can be useful for predicting body length. This can be particularly interesting for the study of museum collections, or even poaching wastes, in which only crania are usually preserved or found relatively intact. However, the present study lacks adult wild individuals.

Some precaution is advised when using ratio variables for predicting body length. Some of these regression equations are not statistically significant (P-value > 0.100). This is the case for the following variables: ROW, RLSS, and RWM for wild, and ROW and RLSS for captive animals). Plots in Figs. 3 and 3 help to visualize these patterns.

Besides helping to estimate body-size from head dimensions, the regression equations of the present study stress skull shape changes during the ontogenetic process. Non-linear equations express changes on the proportions of the skull, "accelerated" or "decelerated" on the inflexion points. For instance, the cranium of captive animals becomes relatively narrower as body size increases (see plot of CW in Fig. 4).

A similar and expected pattern can be seen on the mandible (see plot of WSR in the same figure). In both cases, regression equations are quadratic with the coefficient of the quadratic element being negative (see Table 4).

A somewhat sigmoid shape can be perceived on the relative growth curve of the eye-orbit length (OL) and width (OW) in captive animals. A positive quadratic and a negative cubic element in the allometric equations of both cases show a period of fast relative growth in young followed by a period of slow relative growth of these regions in adult animals. The smaller coefficient of the linear element of the OW equation than of the OL equation express the ontogenetic process of "elongation" suffered by the eye-orbits during initial development of the animals.

Age and Sex as Covariates of Body Size

Table 6 shows the analysis of covariance (ANCOVA) of sex and age of captive animals in relation to the regression equations between morphometric variables and snout-vent length (SVL).

ANCOVA may be used to compare males and females' equations. It may also be useful to separate age from body-size effect on the regressions analyzed.

All body- and head-size variables, and all but three ratio variables (RWI, RWN, and RPXS) are significantly affected by body size (P-value > 0.100), or in other words, they can be considered size-dependent. One body-size (BW), six head-size (CW, SL, OL, OW, PXS, and WSR), and one ratio variable (ROL) are significantly affected by age (P-value > 0.100), i.e., they can be considered age-dependent.

At last one body-size (BW), two head-size (OL and OW), and five ratio variables (RCW, RLST, ROL, ROW, and RWN) are significantly affected by gender (P-value > 0.100).

Webb & Messel (1978) report a perceptible sexual dimorphism in Crocodylus porosus involving interorbital width, which is not perceived in the present study. Hall & Portier (1994) found sexual dimorphism for 21 of 34 skull attributes, including DCL, ML, PXS, CW, OW, IOW, WCR, WN, and WSR. However, their results are possibly optmistic because they could not include age as a covariate of body size in their study of allometric growth of Crocodylus novaeguineae. Some variation actually caused by age (independent of size) may be erroneously accounted as a difference between sexes, or sexual dimorphism.

The fact that all age-dependent variables are also size-dependent explains why it is so difficult to predict age of crocodilians based on single variable growth curves (see Verdade, 1997, for discussion).

All of the sex-dependent variables are also size dependent, with the exception of RWN. However, its efficiency in predicting individual sex through discriminant analysis is low. Four sex-dependent variables (BW, OL, OW, and ROL) are also age-dependent, but the remaining four, all of them ratio variables (RCW, RLST, ROW, and RWN), are not. Age-dependent as well as sex-dependent variables are primarily located on the cranium. Only one age-dependent (WSR) and sex-independent variable is located on the mandible.

Sexual dimorphism was detected in the allometric growth of BW, OL, OW, RCW, RLST, ROL, ROW, and RWN. With the exception of BW, all of these morphometric variables are located in the cranium and none in the mandible. This may be evolutionarily related to the visual recognition of gender when individuals exhibit only the top of their heads above the surface of the water, a usual behavior of crocodilians. A multivariate approach for the study of sexual dimorphism is discussed by Verdade (1997).

Acknowledgments — This study is a part of the dissertation presented to the Graduate School of the University of Florida in partial fulfillment of the requirements for the degree of Doctor of Philosophy. This program was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico ¾ CNPq (Process N. 200153/93-5) and the University of São Paulo, Brazil. I am thankful to Prof. F. Wayne King, J. Perran Ross, Lou Guillette, Richard Bodmer, Mel Sunquist, George Tanner, Phil Hall, and Irineu U. Packer for their ideas and comments on the manuscript. Edson Davanzo, Fabianna Sarkis, and Rodrigo Zucolotto helped to measure the animals. Data set is available with the author.

• ALLSTEAD, J. & LANG, J. W., 1995, Sexual dimorphism in the genital morphology of young American alligators, Alligator mississippiensis Herpetologica, 51(3): 314-25.
• BOOKSTEIN, F. L., 1989, "Size and shape": a comment on semantics. Systematic Zoology, 38(2): 173-80.
• CHABRECK, R., 1963, Methods of capturing, marking and sexing alligators. Proc. Ann. Conf. Southeastern Assoc. Game Fish Comm, 17: 47-50.
• CHABRECK, R. H., 1966, Methods of determining the size and composition of alligator populations in Louisiana. Proc. Ann. Conf. SE Assoc. Game Fish Comm, 20: 105-12.
• CHOQUENOT, D. P. & WEBB, G. J. W., 1987, A photographic technique for estimating the size of crocodilians seen in spotlight surveys and quantifying observer bias in estimating sizes. pp. 217-224. In: G. J. W. Webb, S. C. Manolis & P. J. Whitehead (eds.), Wildlife Management: Crocodiles and Alligators Surrey Beatty & Sons Pty Lim., Chipping Norton, Australia.
• GOULD, S. J., 1966, Allometry and size in ontogeny and phylogeny. Biol. Rev, 41: 587-640.
• HALL, P. M. & PORTIER, K. M., 1994, Cranial morphometry of New Guinea crocodiles (Crocodylus novaeguineae): ontogenetic variation in relative growth of the skull and an assessment of its utility as a predictor of the sex and size of individuals. Herpetological Monographs, 8: 203-225.
• HUTTON, J. M., LOVERIDGE, J. P. & BLAKE, D. K., 1987, Capture methods for the Nile crocodile in Zimbabwe. pp. 243-247. In: G. J. W. Webb, S. C. Manolis & P. J. Whitehead (eds.), Wildlife Management: Crocodiles and Alligators Surrey Beatty & Sons Pty Lim., Chipping Norton, Australia.
• IORDANSKY, N. N., 1973, The skull of the Crocodilia. pp. 201-262. In: C. Gans & T. S. Parsons (eds.), Biology of the Reptilia Vol. 4. Morphology D. Academic Press, London.
• MAGNUSSON, W. E., 1983, Size estimates of crocodilians. Journal of Herpetology, 17(1): 86-88.
• MINITAB, 1996, Minitab for Windows Release 11 Minitab, Inc., State College, PA, USA.
• PETERS, R. H., 1983, The Ecological Implications of Body Size Cambridge University Press, Cambridge.
• SCHMIDT-NIELSEN, K., 1984, Scaling: Why is Animal Size so Important? Cambridge University Press, Cambridge.
• SOKAL, R. R. & ROHLF, F. J., 1995, Biometry: the Principles and Practice of Statistics in Biological Research W. H. Freeman and Company, New York.
• VERDADE, L. M., 1997, Morphometric Analysis of the Broad-snouted Caiman (Caiman latirostris): An Assessment of Individual's Clutch, Body Size, Sex, Age, and Area of Origin Ph.D Dissertation. University of Florida, Gainesville, Florida, USA, 174p.
• VERDADE, L. M. & KASSOUF-PERINA, S. (Eds.), 1993, Studbook Regional do Jacaré-de-papo-amarelo (Caiman latirostris): 1992 / 1993. Sociedade de Zoológicos do Brasil. Sorocaba, São Paulo, Brasil.
• VERDADE, L. M. & MOLINA, F. B. (Eds.), 1993, Studbook Regional do Jacaré-de-papo-amarelo (Caiman latirostris): Junho 1991-Junho 1992. ESALQ, Piracicaba, São Paulo, Brasil.
• VERDADE, L. M. & SANTIAGO, M. E. (Eds.), 1991, Studbook Regional do Jacaré-de-papo-amarelo (Caiman latirostris). CIZBAS/ESALQ/USP. Piracicaba, São Paulo, Brasil.
• VERDADE, L. M. & SARKIS, F., In press, Studbook Regional do Jacaré-de-papo-amarelo (Caiman latirostris): 1993/1997. ESALQ, Piracicaba, São Paulo, Brasil.
• WALSH, B., 1987, Crocodile capture methods used in the Northern Territory of Australia. pp. 249-252. In: G. J. W. Webb, S. C. Manolis & P. J. Whitehead (eds.), Wildlife Management: Crocodiles and Alligators Surrey Beatty & Sons Pty Lim., Chipping Norton, Australia.
• WEBB, G. J. W. & MESSEL, H., 1977, Crocodile capture techniques. Journal of Wildlife Management, 41: 572-575.
• WEBB, G. J. W. & MESSEL, H., 1978, Morphometric analysis of Crocodylus porosus from the north coast of Arnhem Land, northern Australia. Australian Journal of Zoology 26: 1-27.
• WERMUTH, H. & MERTENS, R., 1961, Schildkröten. Krokodile Brückeneschsen Veb Gustav Fisher Verlag, Jena, Germany.

Publication Dates

• Publication in this collection
  23 Feb 2001
• Date of issue
  Aug 2000

History

• Accepted
  22 Dec 1999
• Received
  04 Mar 1999