Keywords: allometric growth, harbour porpoise, paedomorphosis, skeletal proportions
 |  |
| ||||
Copyright © Anatomical Society of Great Britain and Ireland 2005 Sexually dimorphic proportions of the harbour porpoise (Phocoena phocoena) skeleton Correspondence Anders Galatius, c/o Åse Jespersen, Department of Cell Biology and Comparative Zoology, Institute of Biology, University of Copenhagen, Universitetsparken 15, DK-2100 Copenhagen Ø, Denmark. T: +45 28710372; E: Email: agjorgensen/at/zmuc.ku.dk Accepted January 3, 2005. | |||||||||||||||
Abstract Sexual differences in growth, allometric growth patterns and skeletal proportions were investigated by linear measurements of skeletal parts on 225 harbour porpoises (Phocoena phocoena) from the inner Danish and adjacent waters. Females show larger asymptotic sizes and extended period of growth compared with males. Measurements of the skull and flipper bones show negative allometry, whereas those of the bones of the body generally show positive allometry. There are no statistically significant intersexual differences in allometry except for the pelvic bones, where the males show stronger positive allometry. Throughout the range of individual sizes, females have significantly larger skulls and shorter vertebral columns than males for similarly sized individuals. In fully grown specimens, the condylobasal length of females makes up a smaller proportion of total length, and the vertebrae make up a larger proportion as compared with males. As these characters show negative and positive allometry, respectively, it is suggested that males finish their development at an earlier stage than females, retaining more paedomorphic proportions of the skeleton. Paedomorphosis in fully grown males relative to females is also found in the vertebral epiphyses that mature later in males than females, although the males finish growth at a younger age. Keywords: allometric growth, harbour porpoise, paedomorphosis, skeletal proportions | |||||||||||||||
Introduction The harbour porpoise (Phocoena phocoena) is the dominant cetacean species of the inner Danish waters and one of the smallest cetacean species. Its life history is characterized by a short lifespan, rapid postnatal growth to reproductive size, early attainment of sexual maturity and the ability to reproduce annually; the last is exceptional in cetaceans (Read & Hohn, 1995). It exhibits sexual dimorphisms otherwise unusual among odontocetes: females grow to larger sizes than males and have an extended period of growth (Gaskin & Blair, 1977; Van Utrecht, 1978; Stuart & Morejohn, 1980; Gaskin et al. 1984; Miyazaki et al. 1987; Lockyer, 1995a; Read & Tolley, 1997). Stuart & Morejohn (1980) hypothesized that the larger size of the female was an adaptation for carrying a fetus of up to 80 cm and Read & Tolley (1997) suggested that the evolutionary reduction in body size in the harbour porpoise might have approached a limit in females with respect to carrying viable offspring. On the basis of cranial morphology, it has been theorized that the phocoenids have developed their small size and early attainment of sexual maturity relative to the related delphinids through paedomorphosis (Barnes, 1985). In a recent study of the skeletal development of the post-cranial skeleton of the harbour porpoise, Galatius & Kinze (2003) found that full fusion of vertebral epiphyses was rare and delayed compared with delphinid species, which was interpreted as a post-cranial counterpart to the paedomorphic characters encountered in the skull. Although male porpoises terminate their growth at a younger age than females, there was a tendency for males to attain physical maturity of the vertebral column later if at all, which might suggest a higher degree of paedomorphosis. With the exception of papers dealing with the pelvic bone (e.g. Andersen et al. 1992), hitherto no studies seem to have focused on sexual dimorphism in the harbour porpoise in other areas than the period of growth and overall size. Detailed descriptions of sexual dimorphism can help to clarify the background for the known difference in overall size. If the theory that female porpoises are larger in order to carry large fetuses holds true, one might expect them to show body proportions that optimize their potential for carrying a large fetus. Examining the development of male and female porpoises may also yield clues as to evolutionary pathways leading to the observed sexual dimorphisms, such as different degrees of paedomorphosis. In this paper, using the extensive collections of harbour porpoise skeletons from inner Danish and adjacent waters held in Denmark, Sweden and Germany, I compare growth models, allometry of the skeleton and skeletal proportions of growing and fully grown male and female harbour porpoises. | |||||||||||||||
Materials and methods Observations were made on 225 harbour porpoise skeletons from the collections of the Zoological Museum of Copenhagen, the Museum of Natural History of Gothenburg, the National Museum of Natural History in Stockholm and the German Oceanographic Museum in Stralsund. The majority of the specimens (n = 218) originate from by-catches, and the remainder from strandings. All of these animals originate from the inner Danish and adjacent Swedish and German waters, including the the German north coast from Schleswig to Rügen and the Swedish west coasts from Skåne to Bohus. Thus, the sample should represent the inner Danish waters population as defined by Andersen (2003). Age determinations are available for most of the specimens from Copenhagen and Stralsund, provided by Christina Lockyer and Hartwig Kremer, respectively. Age determinations of the remaining specimens from these collections and the animals from Stockholm and Gothenburg were performed using the protocol of Hohn & Lockyer (1995). The specimens in the sample were selected to be evenly distributed over the range of total length in the sample. Data on total body lengths were obtained from the databases of the museums. Measurements All measurements were made with a calliper to the nearest millimetre except where otherwise noted. The measurements were as follows:
The measurements are depicted in Fig. 1
Growth models In order to estimate the average length of fully grown individuals and assess the average timing for termination of linear growth, growth models were applied to the total length at age data of females and males, respectively. Using a non-linear least-squares approach, the length at age data were fitted to the Gompertz growth model:
where L(t) is the length (cm) at age t, A is the asymptotic value, b is the constant of integration, k is the growth rate constant, and t is age in years. Allometry of the measured characters In order to identify characters with early and late development and determine if there were intersexual differences in the growth pattern of each measured character as related to total length, an allometric model taking the form
was used, where Y is the length of the measured character (mm), L is the total length (cm) of the specimen, a is a constant determined by the value of Y when X is unity, and b is the growth coefficient. For regressions on the cranial measurements, total length was replaced by condylobasal length (mm). This model was fitted to the data using a least-squares approach. A growth coefficient significantly less than 1.0 would indicate negative allometry and hence relatively early development of the concerned character, a coefficient significantly greater than 1.0 would indicate positive allometry and hence relatively late development of the concerned character, and a coefficient not significantly different from 1.0 would indicate isometry and that the growth of the concerned character was directly proportional to the growth of total body length throughout development. Statistical analysis involving the growth coefficient tested the null hypothesis H0: b = 1, where the test statistic t = (b − 1)/SEb. Comparisons between male and female growth coefficients were performed, testing the null hypothesis Sexual dimorphism in skeletal proportions Comparison of skeletal proportions between males and females of similar size The relative size of each measurement as related to total body length was compared between males and females using ancova with the measured characters as dependent variables and total length as the covariate (both log-transformed to obtain linear relationships). Significant differences between males and females would indicate innate differences in skeletal proportions that remained stable throughout development, given that there were no significant sexual differences in the allometry of the measured character.Comparison of skeletal proportions between males and females at the same developmental stage The only possibility to make direct comparisons between the body proportions of males and females at exactly the same developmental stage is when the animals are fully grown. The skeletal measurements as fractions of total body length were compared between males and females that were found fully grown. Based on the growth models, animals in age classes 6 and above were selected. This sample consisted of 41 females and 27 males. Comparisons between males and females were performed using anova. If one sex ceased growth at a paedomorphic stage, fully grown specimens would be expected to have relatively larger proportions in the early developing characters (negatively allometric) and relatively smaller proportions in the late developing characters (positively allometric), given that no innate differences in the concerned characters were detected (ancova tests).For allometric analysis as well as ancova and anova procedures, cranial measurements were related to condylobasal length instead of total length. Observations of epiphyseal ankylosis In order to assess intersexual differences in timing of full physical maturity, specimens that displayed full ankylosis of all vertebral epiphyses with no sign of sutures were registered. Logistic regressions were performed to assess the timing of the attainment of physical maturity in both sexes. Specimens having attained physical maturity were scored as 1 for the regression, whereas physically immature specimens were scored as 0. Maximum-likelihood regression was used to estimate the parameters of the logistic equation:
where t is the age (class), P(t) is the proportion of the population expected to be physically mature in a given age class and a and b are the parameters of the model. Later attainment of physical maturity would indicate paedomorphosis. | |||||||||||||||
Results Growth models The resulting equations for the growth models are presented in Table 1; the fitted curves are displayed in Fig. 2.
For males the predicted asymptotic length is 148 cm; 95% of predicted asymptotic length is reached at an age of 3.9 years (95% CI = 3.2–5.1 years). The model of the females predicts an asymptotic length of 161 cm; 95% of predicted asymptotic length is reached at an age of 4.9 years (95% CI = 4.2–6.0 years). Allometry The condylobasal length and the hyoid bones show negative allometry as related to total length in both sexes. The same applies to the bones of the flipper. The 1st rib displayed isometric growth in both sexes, whereas the 5th rib shows slightly positive allometry. The length of the first thoracic vertebra shows isometry; the lengths of the other vertebrae all show positive allometry, although less so in the 14th caudal vertebrae. The pelvic bones show very strong positive allometry. The other measurements from the body – the scapula and the measurements of the sternum – also show positive allometry except the sternum length in males, which did not differ significantly from isometry. In short, the measurements of the head and flippers show negative allometry related to overall length whereas the measurements of the body generally show positive allometry.Of the cranial measurements, the occipital width shows negative allometry related to condylobasal length, and the mandible, maxilla, rostrum and occipital ridge lengths all show positive allometry, the last very strongly so. The only significant intersexual difference of allometry among the characters is the length of the pelvic bone, where the males show more positive allometry than the females. The results are presented in Table 2, and scatter plots showing the regression lines can be seen in Fig. 3.
Sexual dimorphisms Analysis throughout growth –ancova Significant differences between males and females detected using this procedure are only found in the axial skeleton, i.e. condylobasal length and the lengths of the vertebrae. The females have the higher intercept for condylobasal length and the males have significantly higher intercepts for the 1st thoracic, the 8th lumbar and the 14th caudal; males also show higher intercepts for the 5th thoracic and 1st lumbar, although not significantly so. This indicates that the females throughout the range of sizes of porpoises in this sample have longer skulls when similarly sized individuals are compared and vice versa for male vertebral lengths. The results are presented in Table 3 and graphs comparing the regression lines of the characters can be seen in Fig. 3.
Comparison between fully grown males and females –anova Significant intersexual differences in the body composition of fully grown individuals are found in a number of characters. Females have a longer rostrum and occipital ridge relative to condylobasal length, and a longer 1st lumbar relative to overall length. Fully grown males have longer condylobasal length, humerus, ulna, hyoid bone and pelvic bone. The results are presented in Table 4.
The seeming paradox of the longer condylobasal length and shorter vertebral lengths in females than in males throughout development (ancova procedure) and the opposite results for fully grown individuals (anova procedure) is illustrated in Fig. 4.
Vertebral epiphyseal ankylosis Full ankylosis of all vertebral epiphyses is rare in the sample. Fourteen females and two males show full development. With 41 females and 27 males of 6 years or older available, this is a significant deviation from chance distribution (χ2 = 5.06; 1 d.f.; P < 0.05), males with mature vertebral columns being less frequent among the fully grown specimens despite the fact that males generally reach their adult size earlier. The few cases of physically mature males prevent the opportunity of making an assessment of the timing of male maturation of the vertebral column with any certainty by the logistic regressions. Estimates of the age at which 50% of the population have attained full physical maturity are 13.6 years for females (95% CI = 11.7–16.1 years) and 17.4 years for males (95% CI = 12.5–39.1 years). Even so, the regressions (Fig. 5) also suggest that the vertebral column matures later (if at all) in males than in females, the male curve mostly falling outside the female 95% confidence interval.
| |||||||||||||||
Discussion Sexual dimorphism in the harbour porpoise is well known; the growth models applied here imply that females grow to lengths of 161 cm, whereas males only grow to 148 cm. The asymptotic lengths are slightly larger than the assessments of mean adult lengths of Lockyer's (1995b) study on British harbour porpoises (160 and 145 cm, respectively). The females in this study show a prolonged period of growth compared with the males, reaching 95% of their predicted asymptotic length at an age of 4.9 years, whereas the males already do so at 3.9 years. A prolonged period of growth in female harbour porpoises has also been proposed by Gaskin & Blair (1977), Van Utrecht (1978), Stuart & Morejohn (1980), Gaskin et al. (1984), Miyazaki et al. (1987), Kull & Berggren (1995), Lockyer (1995b) and Read & Tolley (1997). As suggested by Stuart & Morejohn (1980), the larger female size may be an adaptation to carrying a large fetus (up to 80 cm; Lockyer & Kinze, 2003), as the newborn calf must be of a certain minimum size to withstand excessive energy loss due to loss of body heat to the surrounding water. It may be speculated that the larger the mother, the greater the calf's chances of survival. This selection pressure for larger female sizes is countered by the specialization of the porpoises for early reproduction (Read & Hohn, 1995). The results from the allometric analysis differ somewhat from the analyses on external measurements performed by Van Utrecht (1978) on North Sea harbour porpoises, and Read & Tolley (1997) on Bay of Fundy harbour porpoises. For the flipper measurements, Van Utrecht found what constitutes negative allometry in females and isometry in males. Like Read and Tolley's study, my data show negative allometry for the flippers. The discrepancy between Van Utrecht's results and those of the present study and Read and Tolley's concerning the flippers are probably due to different approaches to data analysis. The latter studies employ allometric equations, whereas the former analyses fractions of body length by age. The latter procedure could yield different results if the sexes ceased growth at different developmental stages, which, as we shall see, is probably the case. Both Read & Tolley and Van Utrecht report negative allometry in the growth of the anterior region of the body. In the present study, the scapula, the sternum and the vertebrae in the anterior region of the body all exhibit positive allometry, but the 1st rib shows isometry and the 5th rib moderate positive allometry. Read & Tolley and Van Utrecht included the head in their longitudinal measurements of the anterior body, and the condylobasal length shows strong negative allometry in the current study. The isometry or moderate positive allometry shown by the ribs may correspond to the isometry registered in the measurements of girth in the study of Read & Tolley. It is not surprising that the lengths of the thoracic and lumbar vertebrae show strong positive allometry; in odontocetes such as the striped dolphin (Stenella coeruleoalba; Ito & Miyazaki, 1990), the finless porpoise (Neophocoena phocoenoides; Yoshida et al. 1994) and also the harbour porpoise (Galatius & Kinze, 2003) the thoracic (except the first few) and the lumbar vertebrae have been found to be the last to exhibit epiphyseal ankylosis, and should accordingly proceed to grow longitudinally at a stage where longitudinal growth in other areas of the vertebral column has ceased. The less positive allometry of the 1st thoracic and 14th caudal vertebrae conforms to the pattern of epiphyseal ankylosis found in the harbour porpoise and other species, as all investigations conclude that epiphyseal ankylosis in the first few thoracic and in the caudal vertebrae occurs relatively early. It is likely that the growth of the sternum and the scapula, being bones of longitudinal orientation, to some extent follows the pattern of growth shown by the vertebrae in the area, and these should depict the general longitudinal growth. With this in mind, the notion brought forward by Read & Tolley (1997), that the anterior part of the body of the harbour porpoise should show negative allometry, and reach final size before the termination of skeletal growth, seems debatable, being based on longitudinal measurements including the head, and not taking into account the pattern of epiphyseal ankylosis of the vertebral column, which has been reported in other small odontocetes. The isometry registered by Van Utrecht and Read & Tolley for longitudinal measurements of the posterior region of the body does not seem incompatible with the moderate positive allometry of the 14th caudal vertebra, as this is the current study's only measurement in this area, besides the pelvic bones, and less positive (or even negative) allometry can be expected from more posterior caudal vertebrae, as epiphyseal ankylosis is known to occur earlier in these. The very strong positive allometry of the pelvic bone is not surprising, given the rapid growth of the pelvic bone at the onset of sexual maturity in both sexes (Andersen et al. 1992). The fact that the flipper bones show negative allometry, while the scapula shows positive allometry, indicates that the young harbour porpoise must drive a relatively large flipper with a smaller muscle mass, causing a potential deficit in manoeuvrability. There are no statistically significant intersexual differences in allometry except the pelvic bones, for which dimorphisms in size and morphology are already well known (Andersen et al. 1992). This suggests that the skeletons of male and female porpoises follow the same general pattern of development; the same parts of the body grow similarly across the sexes. In the ancova tests for sexual dimorphisms, females are found to have larger condylobasal lengths while males have longer vertebrae and pelvic bones; there are no statistically significant results for the other measurements. This means that for similarly sized males and females throughout the range of sizes, males will have a longer vertebral column, while females will have a longer cranium. This will certainly not constitute a female adaptation for carrying a large fetus. In the anova tests of fully grown individuals, however, the females show significantly shorter condylobasal lengths than the males, while they have longer vertebrae, the 1st lumbar being longer with high statistical significance. The explanation for this paradox seems to be that development of the males is stopped before it has run its full course (compared with the females). This means that fully grown males will generally be relatively larger than females in characters that show negative allometry or early development and vice versa for characters that show positive allometry or late development. This pattern holds true for all the statistically significant differences detected, condylobasal length and the flipper and hyoid bones, where the fully grown males show larger measurements, and the 1st lumbar vertebra, the rostrum and the occipital ridge, where females show larger measurements. The exception to this rule is the pelvic bones, which as mentioned differ in morphology as well as function between the sexes, being attached to the penis in males (Tinker, 1988), and as so are not directly comparable. Extra strength is given to these conclusions by the fact that the specimens of age class 6 and above are involved in this analysis. Regarding the fact that the timing of attainment of asymptotic size differs between the sexes, if anything the concerned males should be at a more advanced ontogenetic stage than the females. The observations of ankylosis of vertebral epiphyses also suggest a higher degree of paedomorphosis in males, as they attain full development of the vertebral column later than females, if at all. In other words, fully grown males retain juvenile proportions and characteristics of the skeleton compared with females. It has been theorized that the phocoenids have developed their small size through paedomorphosis (Barnes, 1985). This theory is based on several characters of the skull of adult porpoises, among others a short rostrum, closely packed teeth and a small developed occipital ridge. Galatius & Kinze (2003) suggested that the markedly late development of the thoracic and lumbar vertebral epiphyses of the harbour porpoises is a post-cranial counterpart to the paedomorphic characters of the porpoise skull. The present study shows significant differences between fully grown males and females in the proportions of characters that have been considered diagnostic for paedomorphosis in phocoenids, such as occipital ridge and rostrum length. It seems that the sexual dimorphisms of the harbour porpoise, the well-known differences of overall size and period of growth along with the differences reported here all stem from a higher degree of paedomorphosis in male harbour porpoises relative to females. It remains to be resolved if the females have evolved a larger size relative to males by development of a lesser degree of paedomorphosis than males, or if males have developed their smaller size through more marked paedomorphosis. Regardless, the less paedomorphic body proportions of the adult females with larger body measurements in the thoracic and lumbar regions are better suited to carrying a large fetus, the proposed explanation to the sexual dimorphism in overall size.
In conclusion, varying degrees of paedomorphosis may well be the evolutionary tool of choice for size adjustment in the phocoenids. Further research might focus on sexual differences in paedomorphosis in other phocoenid species with other patterns of sexual dimorphism in overall size, such as Dall's porpoise (Phocoenoides dalli) and the finless porpoise. | |||||||||||||||
Acknowledgments The Zoological Museum of Copenhagen, the Museum of Natural History in Gothenburg, the National Museum of Natural History in Stockholm and the German Oceanographic Museum in Stralsund are thanked for access to their collections. Mogens Andersen of the Zoological Museum in Copenhagen, Petra Rudd of the Museum of Natural History in Gothenburg, Bo Fernholm, Olavi Grönwall and Peter Mortensen of the National Museum of Natural History in Stockholm, and Harald Benke, Klaus Harder and Gerhard Schulze of the German Oceanographic Museum in Stralsund are acknowledged for support at their respective institutions. The age determinations of Christina Lockyer and Hartwig Kremer are acknowledged. Carl Chr. Kinze and Christina Lockyer are thanked for comments on the manuscript. | |||||||||||||||
References
| |||||||||||||||
The seemingly innate differences in condylobasal and vertebral proportions (