Comparison of linear enamel hypoplasia frequencies in two populations of Thailand gibbons (Hylobates lar lar)
Cristina Elyse Alcorta ‘12
Department of Human Evolutionary Biology, Harvard University
This study investigates the effects of environmental stress on the prevalence of linear enamel hypoplasia (LEH) and M2 asymmetry in two Thailand gibbon (Hylobates lar lar) populations. Although a few studies have previously described the incidence of these two physiological stress markers across primate taxa, there has not been an intraspecies investigation of the gibbon dentition. A study conducted by Hannibal (2000) in West Africa found that lowland gorillas had a significantly higher frequency of LEH relative to a sample of mountain gorillas due to a higher prevalence of nutritional stress and disease-related stress due to closer human proximity. Their counterparts in the mountains are more removed from human populations and have diets that remain relatively stable, which helps reduce the prevalence of biological stress. This study takes a similar investigative approach, hypothesizing that gibbons of the mountainous region of Thailand (Inthanon Doi) would experience less stress than their counterparts in the lowlands (Chieng Dao) for the same reasons given in the gorilla study. Stress was measured based on two parameters: (1) the number of linear enamel hypoplasias present on the maxillary and mandibular canines and (2) the asymmetry of the second molar (M2). Within a sample group of 88, it was found that highland gibbons have a significantly reduced frequency of LEH as compared to the lowland gibbon population. However, there were no significant measures of asymmetry of maxillary and mandibular M2 found within either population. This study provides support that environmental stress is a cause for LEH and that differing degrees of stress affect LEH expression. The fact that the populations do not experience asymmetry in M2 calls for further investigation of how physiological stress is manifested in the dentition and whether the stress levels between mountain and lowland primates is essentially significant.
As the human presence on Earth grows and populations inhabit more remote areas of the world, it is important for societies to consider the effects such growth is having on our wildlife. Increased industrialization and habitation could have a variety of detrimental effects on ecosystem health. This, in turn, could lead to serious problems higher up the food chain. Diseases from humans may potentially be transmitted to other species, specifically other primates. Given that primates are social creatures, larger populations would allow transmitted diseases to spread wildly. It is therefore interesting to investigate whether specific populations of species indicate greater levels of physiological stress in areas densely populated by humans than in less-populated areas.
Enamel hypoplasias are deficiencies of enamel thickness that disrupt the contour of a tooth’s crown surface. Such an event is initiated during enamel secretion, which occurs in the secretory phase of amelogenesis in crown development (Hillson 2002). The majority of hypoplasias are arranged in a band around the circumference of the crown, following the lines of the perikymata. In the 18th century, it was proposed by Dr. Robert Bunon that hypoplasias found on the unerupted teeth of children with rickets, measles and small pox were a response to stress and that they formed during crown formation. Later in the 1800s, J. Berten found that the defects coincided with the layering of crown development and could be matched with teeth across the dentition. Defects come in a wide variety, including: pit-type defects which can be found on cusps or crown sides; plane-type defects in which large areas of brown stria planes are exposed; and furrow defects, often referred to as linear enamel hypoplasia or LEH (Goodman & Rose, 1990). This study is concerned solely with the third type of hypoplasia, LEH.
Linear enamel hypoplasias are only one type of developmental enamel defect (DED) widely used in dental literature. They are sensitive and non-specific dental indicators of physiological and/ or psychological stress that provide a permanent record of systemic stresses. Causes include nutritional deficiencies, fever-inducing diseases, and weaning (Guatelli-Steinberg, 2000). Occurring only in the imbricational zone of crown formation, LEHs represent an exaggeration of the perikymata (Hillson 2002). According to Goodman and Rose (1990), the category of linear enamel hypoplasias includes both faint and more pronounced linear defects. LEH has been used extensively as an indicator of stress in humans but is only recently gaining increasing attention in nonhuman primate studies. Investigations of tooth patterns in monkeys and great apes show that canines (particularly mandibular) and maxillary incisors are the most commonly affected teeth (Guatelli-Steinberg, 2000). It has also been found that monkeys exhibit a reduced frequency of LEH when compared to great apes (Hannibal, 2004).
There have been several investigations comparing the expression of LEH across primate taxa. It has been found that taxonomic influence on LEH prevalence is potentially related to a combination of factors, including: differences in enamel development between species; variable tooth morphology; life history features, which include duration of dental crown formation and exposure to sources of stress during the infant and juvenile periods; and experience of physiological stress (Guatelli-Steinberg, 2000). Despite these extensive interspecies studies, very few pursue in-depth intraspecies comparisons.
In 2005, Darcy Lee Hannibal published a study that found differing LEH frequencies between two West African gorilla populations: one from the lowlands and one from the mountainous region. It was found that highland gorillas had a significantly lower frequency of linear hypoplasia than those of the lowlands. Initially, one would think that seasonality in the lowlands would affect food availability, which could in turn cause an increased frequency of LEH due to nutritional deficiency during certain times of the year. Investigation of the gorillas’ diet, however, revealed that both populations relied on eating leaves, which are both abundant and available year-round in both regions. Hannibal therefore suggests that exposure to pathogens in lowland versus highland habitats or proximity to human populations (which would increase disease and encroachment stresses) plays a role in the differing prevalence of LEH in the gorilla subspecies of the sample.
Hannibal’s study inspired the investigation of whether a similar pattern of LEH frequencies would occur within primate populations of a different species under the same conditions. Hannibal’s explanation as to why the difference in LEH frequency occurred is plausible enough to suspect that other nonhuman primates have the same experience in other parts of the world. Guatelli-Steinberg showed that gibbons exhibit one of the highest frequencies of LEH within the primate taxa, nearly equivalent to that of gorillas (2000). Gibbons are a class of lesser apes, more closely resembling monkeys than the great apes by being smaller, exhibiting low sexual dimorphism (Gaulin et al., 1983), and having pair-bonding behavior (Brockelman et al., 1998). Gibbons are omnivores with a diet of approximately 75% fruit. This could make them more susceptible to nutritional stress during seasonal climate changes, where the availability of fruit could change. The investigation of LEH frequency between two populations of gibbons could provide insight on how lesser-ape primates are responding to variable environments.
This study investigates two populations of white-handed gibbons (Hylobates lar lar) from northwestern Thailand. One population was from Chieng Dao, a lowland region under 457.2 meters above sea level. This lowland region experiences seasonal weather, with climate changes that could potentially affect food source and prevalence of pathogens (Guatelli-Steinberg 2000, Department of National Parks, Wildlife and Plant Conservation, 2004). Doi Inthanon (formerly Mount Angka) is a mountainous region, within an elevation range of 609 to 2,565 meters above sea level. Historically, it is Thailand’s most elevated region and experiences constant cool climate year-round with little variation. The two areas have different human population densities. The more industrialized Chieng Dao district had a population of approximately 87,922 (46.7 per km2) in 2000 while Doi Inthanon’s district had a mere 57,214 in 2008 dispersed among remote villages around the National Park and not within (Internet World Stats 2011). The Karen, ethnic Thais, and Hmong tribes are the main inhabitants living within the 1,000 km2 that constitutes Doi Inthanon National Park, from where the gibbon specimens were collected. As of the early 2000s, approximately 4,500 tribal villagers live within Doi Inthanon National Park, the gibbon sanctuary (Zeppel, 2006). Another 8,000 live in 14 villages within a 5 km radius of the park’s boundaries (Hvenegard and Dearden, 1998). Such a difference in human populations potentially could greatly affect the level of exposure to disease that animals experience within their habitats, as humans are a source of fecal and industrial waste. It is suspected that Chieng Dao is a region of higher pathogen exposure because of its relatively high human population density compared to that of Doi Inthanon.
This study hypothesizes two conditions regarding LEH. (1) In the highland region of Doi Inthanon, where the climate is stable, the environment is stable with regards to food source, and pathogen exposure is low due to a lower human regional population, the number of environmental stresses as measured by LEH and molar asymmetry would be lower than that of the lowlands. Gibbons of this region would therefore experience a lower frequency of LEH on their canines. (2) In the lowland region of Chieng Dao, where the climate is variable and changes with season, the environment is relatively less stable in terms of food source, and pathogen exposure is higher due to a higher human regional population, the number of environmental stresses would be higher than that of the mountainous region. Gibbons would therefore experience a higher prevalence of LEH on their canines.
As a second parameter of comparison for the quantification of stress, the prevalence of asymmetry of the 2nd molar (M2) was investigated. In previous studies, asymmetry has been considered an indicator of Selyeian stress during the developmental period. Experimental studies with rats and mice have shown that homeostatic stressors, such as changes in temperature or nutrition, increase asymmetry (Sciulli et al., 1979). Other studies suggest that environmental factors have much more of an effect than inherited factors (Potter et al., 1976). High levels of asymmetry might therefore suggest similarly high levels of physiological stress (Hillson 2002). This study hypothesizes that individuals residing in the Chieng Dao region, which supposedly has the higher level of environmental stress, would display directional asymmetry of M2, while little or no asymmetry would be expected in the mountainous gibbon population.
Eighty-eight specimens from the Museum of Comparative Zoology of Harvard University were used in this study. Specimens were collected from two sites in Thailand: Chieng Dao (21: 10 Males, 11 Females) and Doi Inthanon (67: 34 Males, 33 Females) that were collected during the Asiatic Primate Expedition of the 1930s. Gibbons were obtained from various elevations on Doi Inthanon, with most samples coming from 1310 meters. All samples were either feral or free-ranging until death. Only individuals with fully erupted third molars (M3) were considered for this study. Infants and sub-adults were excluded entirely.
Tools used for measuring tooth dimensions and scoring included a Digital Caliper (Mitutoyo-Absolute Digimatic), a magnifying glass or hand-lens (5X), and an incandescent light source.
Number of linear enamel hypoplasias
Maxillary and mandibular canines were observed under consistent lighting, using only an overhead fluorescent light and a second incandescent desk lamp. The reason canines were chosen and not other teeth was because they displayed linear enamel hypoplasias that were more readily detectable with the naked eye than any other tooth type. The number of LEH was counted using a 5X hand lens. Canines were scored based on the appearance of any observably exaggerated perikymata, both faint and pronounced. To count as an LEH, the furrow needed to have an antimeric pair. Such a convention would help ensure that the defect arose from a systemic physiological disturbance rather than localized trauma that could have occurred ante- or post-mortem. If half or more of the tooth was not visible due to natural causes, such as wear, or if the tooth had experienced post-mortem breakage, the canine pair would not be scored. Teeth were considered LEH negative if they did not have a matching pair of furrows on the antimeric tooth. In the case where the antimeric tooth was not present at all, the individual would not be included in the sample. Therefore, the number of matched defects in antimeric pairs was defined as the LEH count for that pair (following Guatelli-Steinberg 2000). Quantitative measurements recorded included number of hypoplasias on antimeric pairs, which yielded one value for the pair. Qualitative measurements recorded included spacing of LEH and definition of furrow (pronounced versus faint).
Canine length was measured using a digital caliper (Mitutoyo-Absolute Digimatic). Measurements were taken from the cementum enamel junction (CEJ) to the cusp of the crown on the labial surface of the tooth. All measurements were taken in millimeters (mm). The length was then used to create a scaling factor that measured the number of hypoplasias on a given pair divided by the average length of the canine pair. Since gibbon teeth exhibit apical wear over time, samples with sufficient wear (i.e. dentine was visible, chipped) were excluded from the data. Given that all teeth analyzed were taken from the same age group of adults with erupted third molars, the wear of the canines was comparable across samples. The frequency was therefore defined as absolute # LEH/ average length in mm.
Knowing that buccal-lingual and mesiodistal diameters have both been shown to be sensitive to stress in regards to expressing asymmetry (Hillson 2002) justified the use of only the buccal-lingual diameter as the measurement to consider for asymmetry. The second molar was chosen because it is not as worn as M1, and it is known to develop roughly around the same time as the canines. This ensured that stress occurring during the development of one tooth also occurred in the development of the other. Given the available tools, buccal-lingual diameter was the dimension that could be most easily and accurately measured. It also guaranteed that diameters would not be affected by variances in attrition, whereas the mesiodistal diameter would not. The measurement was taken on the widest portion of the crown surface, which was usually midway on the CEJ-crown cusp axis and midway on the mesial-distal axis. The caliper was oriented perpendicular to the dental arcade and to the surface of the alveolar bone. Measurements were recorded in millimeters (mm).
To analyze differences among the sexes and populations in the incidence of hypoplasia and asymmetry, a Mann-Whitney U test was performed. The test, also known as the Wilcoxon rank-sum test, is a non-parametric statistical hypothesis test for assessing whether two samples of independent measurements are significantly different from one another. The significance of each analysis was determined by the exact p-value measurement. The following describes values and the groups in which they were compared using the Mann- Whitney U test analysis: (1) male canine lengths of each population (Chieng Dao, lowland and Doi Inthanon, highland) were compared to the lengths of females; (2) the frequency scale (Number of LEH/length of canine) by sex and by locality; (3) analysis of whether asymmetry of M2 occurs in either of the two populations, lowland versus highland.
The statistical program SPSS was used to make all Mann-Whitney U calculations. Bar graphs were constructed from calculation of means in Microsoft Excel.
Qualitative observation of LEH maxillary and mandibular canines
Overall, the hypoplasias of the Chieng Dao (lowland) gibbons were more pronounced with deeper grooves going completely around the canine. In the Doi Inthanon (highland) population, the grooves were generally shallower, with fainter grooves following a single perikymata.
An interesting feature in the Chieng Dao (lowland) population was that many of the upper canine LEH within the sample were spaced evenly, with approximately 0.5-2 mm between hypoplasia depending on the individual. Such an observation indicates that seasonal patterns could have potentially contributed to their appearance. The LEH of the highlands exhibited a more sporadic expression.
A number of gibbons from both populations appeared to have one significantly pronounced LEH near the cervical region on the lower canine. This could very well be a feature of the species. It was almost impossible to determine whether this line was due to a single event in all individuals. Perhaps this line corresponds to dispersal of the natal group or weaning.
There was no apparent difference observable between males and females. Teeth were often marked by tartar pigments, especially those from older individuals.
Canine length: male versus female
Before running stress-specific analyses, male versus female dichotomies were assessed to see whether the features were dimorphic enough to consider separating the sexes in comparing the two populations of gibbons. Canine length was assessed first. It was found that the lengths of canines differed between males and females, with males having significantly larger canines than females (Table 1). Comparing canine lengths of both sexes by locality showed that there was no significant difference between the average lengths of canines of the Chieng Dao and Inthanon Doi populations (Table 2). Figure 3 shows a visual comparison of canine lengths.
Analysis of LEH frequency by sex
The incidence of hypoplasias between the sexes within the specific groups was considered, using the absolute number of hypoplasias counted. In Chieng Dao, there was a significant difference between the sexes for incidence of hypoplasias in the upper canines but not for the lower. Doi Inthanon showed the opposite result: there was significant difference between the sexes in the incidence of hypoplasias in the lower canines but not in the upper canines (Table 3). In both cases of significant differences, males had a higher incidence of LEH, possibly due to the fact that they have longer teeth on average.
This information, however, does not control for the length of canines. For that reason, a frequency measure was devised in which the absolute number of hypoplasias for each group was divided by the average length of its respective antimeric canines: [# LEH/ Canine Length (mm)]. This approach controls for length of canines, since those of males are longer on average (Figure 3). The differences in maxillary and mandibular frequencies between the groups were analyzed to see whether they were significantly different, and they were not (Figures 4-6, Table 4). Given that these results would provide more accurate information about the frequency of LEH, the males and females of each separate population were pooled, which left locality as the primary variable of comparison.
These results show that gender is not determining LEH expression, further supporting the idea that environmental factors affect LEH frequencies in these two gibbon populations. Such findings validate the decision to pool male and female groups by locality.
Analysis of LEH frequency by locality
Males and females of Chieng Dao were pooled together to represent the lowland population. The same was done for the gibbon population of Doi Inthanon. The next step was to analyze whether or not these frequencies were statistically significantly different by locality (Figure 7). According to the hypothesis, it is expected that the population of Chieng Dao (lowlands) would exhibit a higher average frequency of LEH than that of Doi Inthanon (highlands). Data analysis showed that gibbons from Chieng Dao had significantly higher frequencies of LEH in both the maxillary and mandibular canines (maxillary: p < 0.001; mandibular: p = 0.001). Table 5 shows a summary of the mean frequency values for LEH by locality.
Analysis of asymmetry in M2
Asymmetry in M2 was used as a second parameter of comparison for measuring stress levels of each population. The buccal-lingual diameters of M2s of the mandible and maxilla were compared within each species. According to the hypothesis, it was expected that asymmetry would be found in the Chieng Dao population. Had both populations exhibited asymmetry, a Mann-Whitney U Test would have been conducted to see if the magnitude of one population’s asymmetry was larger than the others. However, asymmetry was found in neither the lowland nor the highland populations (Figures 8 and 9). The p-values reported by the Mann-Whitney U test rested high above the 0.05 cutoff (Table 6). Asymmetry was not found to differ between the sexes for either population.
This investigation provides evidence that gibbon samples from two different regions in Thailand had significantly different levels of LEH expression, which provides support that LEH are highly influenced by environmental factors. Although it is important to study the differences across taxa, it is essential to note that certain characteristics can differ within a species. Under the notion that each species has its own unique pattern, LEH frequencies were previously used to group primate dental remains into specific taxa (Guatelli- Steinberg, 2000). This study rules out the method of categorizing individual primates into species based on LEH frequencies, given that environment can cause significant variation in LEH incidence within a species. The fact that this study is specific to a single species makes it advantageous in that it reduces the probability that variation in LEH expression is due to genetic differences.
The LEH frequencies reported here are higher than those reported in the Guatelli-Steinberg article; over 36% of gibbons in this experiment were recorded to have experienced multiple stress events. This is most likely due to the fact that Guatelli-Steinberg compared various tooth types while this study only looked at canines (2000). This further raises the issue of inter-observer error in linear enamel hypoplasia studies. Methods for measuring hypoplasias have not been standardized. After using hand-lenses in her experiments, Guatelli-Steinberg suggested using one future method, in which casts of teeth are collected and then viewed under a scanning electron microscope (SEM). This would allow one to see the faintly accentuated lines not visible to the naked eye. Others believe it is sufficient to use a hand lens and that counting the minutia might give a false reading of LEH frequency. Guatelli- Steinberg supports limiting data recoding to the more prominent lines or grooves of varying width and depth with the support of a 10X hand lens (2000). Lighting is another issue in scoring LEH. It is difficult to keep lighting consistent when making observations in different museums. Even within the same museum, natural light can influence how hypoplasias can be seen, regardless of how one controls synthetic light sources. Curvature of the canine could have lead to inaccurate measurements. Spacing between perikymata was not measured.
Another issue is the difficulty of dividing the sample by sex. Initially, only the absolute number of hypoplasias was considered for comparison for males and females of both groups. However, this measurement does not account for the longer length of male canines on average. It was found that males had a higher average absolute number of LEH in both mandibular and maxillary canines (see Figure 3). Therefore, it is better to find the number of hypoplasias over a specific length. In this study, females and males were pooled on the basis that frequency controls for an important difference between the sexes, that is, the length of their canines. Although it might not be a perfect measure, it is more accurate than counting absolute values. This form of measurement revealed an interesting piece of evidence: there was no difference between LEH frequencies of males and females, whether or not the data was pooled by locality. This evidence suggested that expression of LEH is independent of sex. However, pooling males and females could potentially be a source of error, becoming an issue of over-generalization in a very small sample size, which could give false readings.
Pooling the sexes and comparing the populations by locality, this study found that gibbons of the lowlands had higher frequencies of LEH expression than their counterparts in the highlands, which is what the hypothesis had predicted. The key differences between the two environments were their elevation (and therefore climate) and their exposure to human population. Such data suggests that increased LEH may be due to physiological stressors such as increased pathogen exposure or nutritional deficiency, assuming that larger human exposure increases pathogen exposure. Seasonality would create periods of unstable food supply, particularly in fruits, which make up ~75% of a gibbon’s food supply. In Carpenter’s field notes, data gathered from both highland and lowland sites indicate reliance on the same type of primary food sources, at least during the months of March to June (1940). Evenly spaced LEH suggest seasonal stress, which has led some to conclude that variation in food availability would be the culprit. However, the structurally complex environment of the Thailand regions buffers seasonality, meaning that gibbons would be allowed to maintain a year-round intake of ripe fruit (Guatelli-Steinberg, 2000), ruling out nutritional deficiency as a physiologic stressor. Future studies may want to consider whether there is actually a larger genetic than environmental component. Perhaps these two populations experienced allelic mutations or changes in allele frequency due to regional separation over time.
This experiment failed to find an adequate second parameter by which to measure the level of stress in both gibbon populations. This may be due to the measurement of a single dimension of M2. It could also be that asymmetry is found within a region other than the largest area of the crown’s circumference, such as at the cervix. On a similar note, it is difficult to get a consistent measure across a sample of teeth, even if it is the same observer. There is also the question of whether the proper tooth was used—perhaps another molar is more prone to show uneven dimensional growth. The fact that the populations do not experience asymmetry in M2 calls for further investigation of how physiological stress is manifested in the dentition and whether the stress levels between mountain and lowland primates is essentially significant. Although asymmetry has been used in previous studies, there have been some scientists who encountered problems using this parameter, such as DiBennardo and Bailit, who investigated a large group of Japanese children (Hillson 2002). The relationship between asymmetry and stress therefore remains unclear. For future studies, it might be beneficial to incorporate established indicators of stress, such as reduction in long bone growth.
The findings in this study support the expected outcome that lowland gibbons of Chieng Dao have a higher frequency of LEH than highland gibbons living in Doi Inthanon. Increased frequencies of linear enamel hypoplasias may very well be a sign of maladaptation. Theory and research suggest that this maladaptation may be result from increased exposure to humans. Such exposure may cause an increase of fecal and industrial waste, which in turn increases the number of foreign pathogens that can be transmitted between species. The increased exposure to human populations may not only increase physiological stress in gibbons, but in other species as well. It is therefore valuable to understand how our wildlife is reacting to a rapidly changing human-inhabited environment and how such might affect the ecosystem as a whole.
Future studies should investigate different geographic regions for both primate and non-primate species that might be experiencing similar pattern. Such research may lead to pinpointing specific factors contributing to the highest level of stress in certain species. One of the greater challenges of such investigations will be obtaining a large enough sample size. However, any form of indication that a specific population may be undergoing unwarranted increases in physiological stress may help identify and reverse the effects of the source that is causing harm.
It is a pleasure to thank those who made this paper possible. Through her passion for dental anthropology and generous support, Dr. Tanya Smith taught me to appreciate the importance of the study of dentition, helping develop an understanding for the subject.
Thank you to Judy Chupasko and Mark Omura of the Harvard University Museum of Comparitive Zoology Mammalogy Department, who helped me navigate through the primate collections.
Lastly, I would like to thank Ron Serko for his endless support and advice throughout the writing process.
Brockelman, W. Y., Reichard, U., Treesucon, U., & Raemaekers, J. (1998) Dispersal, pair formation and social structure in gibbons (Hylobates lar). Behavioral Ecolog and Sociobiology, 42 (5): 329-339
Carpenter, C. R. (1940) A field study of Siam of the behavior and social relations of the gibbon (Hylobates lar). Comparative Psychology Monographs, 16: 1-212.
Department of National Parks, Wildlife and Plant Conservation. <http://www.dnp.go.th/parkreserve/nature.asp?lg=2>. Accessed 1 December 2011.
Gaulin, S.J.C., & Sailer, L.D. (1983) Sexual Dimorphism in Weight Among the Primates: The Relative Impact of Allometry and Sexual Selection. International Journal of Primatology, 5 (6): 515-535
GBGM Kingdom Kids Foundation. (2011). Angel’s Haven Missionaries. <http:// www.thailandumc.com/2011/05/hmong-tribe-cell-group-in-doi-thanon. html> Accessed 30 November 2011.
Goodman, A. H. & Rose, J. C. (1990) Assessment of systemic physiological perturbations from dental enamel hypoplasias and associated histological structures. Yearbook of Physical Anthropology, 33: 59-110.
Guatelli-Steinberg, Debbie (2000). Linear Enamel Hypoplasia in Gibbons (Hylobates lar carpenteri). American Journal of Physical Anthropology, 112: 395-410.
Hillson, S. (2002) Dental Anthropology. Cambridge University Press, New York.
Hannibal, D. L. (2005) Linear enamel hypoplasia in the great apes: Analysis by genus and locality. American Journal of Physical Anthropology, 127: 13-25
Hvenegaard, G and Dearden, P. (1998). Ecotourism versus tourism in a Thai National Park. Annals of Tourism Research. 25: 700–720
Internet World Stats: Usage and Population Statistics. Miniwatts Marketing Group. <http://www.internetworldstats.com/stats.htm>. Accessed 30 November 2011.
Potter, R. H., Nance, W. E., Yu, P. L. & Davis, W.B. (1976) A twin study of dental dimension. II. Independent genetic determinants. American Journal of Physical Anthropology, 44: 397-412.
Sciulli, P. W., Doyle, W. J., Kelley, C., Siegel, P. & Siegel, M. I. (1979) The interaction of stressors in the induction of increased levels of fluctuating asymmetry in the laboratory rat. American Journal of Physical Anthropology, 50: 279-284.
Zeppel, H. (2006). Indigenous Ecotourism: Sustainable Development and Management. Cromwell Press, Trowbridge.