[FONT="]E Powell still ... (yellow underlining is mine)
This utilization of non-metric traits as population indicators rested on the assumption[/FONT]
[FONT="]that this variation has a genetic, rather than environmental basis. Evidence for the[/FONT]
[FONT="]genetic nature of these traits in man came from a number of fan-dly studies (e. g. Montagu[/FONT]
[FONT="]1937,Torgersen 1951 ab, Selby, Garn and Kanareff 1955, Suzuki and Sakai 1960) where one particular type of trait was investigated. These studies generally concluded that the traits[/FONT]
[FONT="]were inherited, usually by a don-dnant gene with incomplete penetrance'. The observation[/FONT]
[FONT="](Brothwell 1959) that the frequency of any particular variant was constant in a given race,[/FONT]
[FONT="]and similar in related races was also suggestive of a genetic basis. However, the major[/FONT]
[FONT="]breakthrough in elucidating the mechanisms controlling trait expression occurred in the[/FONT]
[FONT="]1950s. This came from a series of investigations by Griineberg and his co-workers, at[/FONT]
[FONT="]University College, London, into the inheritance of skeletal traits in the mouse[/FONT]
[FONT="](sunumrised in Groneberg1, 963).[/FONT]
[FONT="]Groneberg discovered that a range of n-dnor variants in the skeletons of inbred strains[/FONT]
[FONT="]of laboratory nice, although manifesting as "all or none" characters, were inherited as[/FONT]
[FONT="]continuous variables rather than as Mendelian traits. These traits exhibit a wide range of[/FONT]
[FONT="]rnorphological expression but they are all distinguished by having a discontinuous[/FONT]
[FONT="]distribution based on an inherited underlying continuous variable - Gruneberg (1952) coined[/FONT]
[FONT="]the term 'quasi-continuous' for this type of discontinuous variation.[/FONT]
[FONT="]Gruneberg (1963) demonstrated that the actual[/FONT][FONT="] inherited entity is the size or rate of[/FONT]
[FONT="]formation of an embryonic rudiment and not the presence or absence of a variant in'the[/FONT]
[FONT="]mature skeleton. In the CBA strain of mice, for example, 18% of adults lack one or both of[/FONT]
[FONT="]their lower third molars. There are apparently genes which control tooth size as such,[/FONT]
[FONT="]shown by the increased variance of third molar size in the hybrid offspring of two inbred
[/FONT]
[FONT="]strains. Strains of mice in which the incidence of missing molars was high also had on[/FONT]
[FONT="]average, smaller molars than other strains. Grilneberg showed that in embryonic mice,[/FONT]
[FONT="]there was a range in size of the tooth germ at any particular stage, but that if the tooth[/FONT]
[FONT="]germ failed to reach a certain size by the sixth day after birth, the tooth germ regressed[/FONT]
[FONT="]and the tooth failed to develop.[/FONT]
[FONT="]Gruneberg also found that the factors determining whether a tooth will develop or[/FONT]
[FONT="]not are the environmental factors connected with maternal physiology. For example, tooth[/FONT]
[FONT="]loss was commonest in large litters and in first litters where the size of the young was small[/FONT]
[FONT="]at birth. Third molar size could be increased by fostering the young onto mothers whose[/FONT]
[FONT="]lactational performance was better than that of the natural mother, or decreased by[/FONT]
[FONT="]feeding the mother on a deficient diet which interfered with lactation.[/FONT]
[FONT="]...[/FONT]
[FONT="]For any inbred strain, the frequency of a trait was found to be constant in each[/FONT]
[FONT="]generation, and that the trait frequency was a characteristic of the gene pool in question.[/FONT]
[FONT="]Moreover this frequency was found to be largely independent of age, and usually sex, so that[/FONT]
[FONT="]it could be used as a genetical marker in population studies in approximately the same way[/FONT]
[FONT="]as the frequency of a blood group. Furthermore, there are in the mouse very few correlations[/FONT]
[FONT="]of the joint occurrence of pairs of variants.[/FONT]
[FONT="]...[/FONT]
[FONT="] [/FONT]
[FONT="] [/FONT]
[FONT="] [/FONT]
[FONT="] [/FONT]
[FONT="] [/FONT]
[FONT="] [/FONT]
[FONT="] [/FONT]
[FONT="]...[/FONT]
[FONT="]In summary, this study has shown that the characteristics of distances based on metric[/FONT]
[FONT="]traits are consistentw ith those expectedo f a genetic or taxonon-dcd istance. Th, ep attern of[/FONT]
[FONT="]relationship between the groups, as revealed by the metric plots, is also reasonable in the[/FONT]
[FONT="]light of present archaeological knowledge. Non-metric traits, however, produce distances[/FONT]
[FONT="]which behave in an erratic manner, and the population relationships suggested by nonmetric[/FONT]
[FONT="]plots is problematical. Metric variation, therefore, must be regarded as the more[/FONT]
[FONT="]-227-[/FONT]
[FONT="]Rq=RETATION OF RESULTS.[/FONT]
[FONT="]reliable indicator of population affinity. This study is open to the criticism that the metric[/FONT]
[FONT="]variablest having been taken by different workers, are subject to inter-worker error.[/FONT]
[FONT="]However, the consistency of the metric results, and the discovery that non-metric traits are[/FONT]
[FONT="]subject to a considerable degree of intra-worker error, yield the conclusion that the results[/FONT]
[FONT="]are nonetheless valid.[/FONT]
[FONT="]62. The genetic basis of non-metric traits - revisited.[/FONT]
[FONT="]In chapter 2, the evidence was reviewed which showed that non-metric traits have a strong[/FONT]
[FONT="]genetic basis. The disagreement between male and female non-metric distances does not[/FONT]
[FONT="]concur with this supposition. Recently, some workers have begun to question this[/FONT]
[FONT="]assumption of a strong genetic component. Richtsmeier and McGrath (1986), studying mice,[/FONT]
[FONT="]found that in only 4 of 35 traits could they arrive at a significant heritability value.[/FONT]
[FONT="]Noting that these results were consistent with those of Self and Leamy (1978) and Searle[/FONT]
[FONT="](1954), they suggest that "historically accepted assumptions about heritability of nonmetric[/FONT]
[FONT="]traits require continued close scrutiny".[/FONT]
[FONT="]The contention that non-metric traits are not genetic is not easily accepted in the face[/FONT]
[FONT="]of counter-evidencefr om numerouss tudiesi n mice (e.g . [/FONT][FONT="]Grilneberg1 963)m, acaques(e .g .[/FONT]
[FONT="]Cheverud and Buikstra 1981) and man (e. g. Saunders and Popovich 1978). Heritability[/FONT]
[FONT="]estimates were originally formulated for the study of continuous variation, and the .'[/FONT]
[FONT="]statistical difficulties accompanying their modification for binary data may decrease the[/FONT]
[FONT="]validity of these family siudies (Falconer 1981). Similarly, many so-called discrete traits[/FONT]
[FONT="]do, in fact, show continuous variation (Corruccini 1974); subjectively constructed thresholds[/FONT]
[FONT="]may not adequately express an underlying genetic basis.[/FONT]
[FONT="]Categorical scoring of traits has been adopted by many workers to increase the[/FONT]
[FONT="]consistencyo f scoring,t hough the scoresa re usually dichoton-dsedp rior to analysis. The[/FONT]
[FONT="]loss of information which accompaniesd ichoton-dsationis expresseda s measuremente, rror[/FONT]
[FONT="]in the analysis. In heritability studies, this type of measurement error appears as[/FONT]
[FONT="]environmental variance (Falconer 1981, p. 124). This may account for the low heritability[/FONT]
[FONT="]values in some studies, though why all studies are not affected is unclear. If this type of[/FONT]
[FONT="]1 1, -11ý[/FONT]
[FONT="]-228-[/FONT]
[FONT="]RqTERPRETATION OF RESULTS.[/FONT]
[FONT="]error appears to reduce the genetic component of the trait data, it is possible that this[/FONT]
[FONT="]accounts for anomalous results which are a frequent feature of non-metric population studies.[/FONT]
[FONT="]As stated by Richtsmeier and McGrath (1986):[/FONT]
[FONT="]When dichoton-tizing, discrete categorization is imposed on a process which is, in[/FONT]
[FONT="]fact, continuous. Sincew e cannotd emonstratet he biological correctnesso f the chosen[/FONT]
[FONT="]categories,lu mping of trait expressionm ay not reflect the biological baseso f trait[/FONT]
[FONT="]variation.[/FONT]
[FONT="]Since the methods of trait development are poorly comprehended, arbitrarily defined[/FONT]
[FONT="]scoring thresholds can have no biological meaning. Richtsmeier and McGrath conclude that[/FONT]
[FONT="]"until a fuller understanding of nonmetric trait etiology is developed, the pron-dse of[/FONT]
[FONT="]nonmetric traits will remain unfulfilled".[/FONT]
[FONT="]The differences between left and right sides also has genetic implications. If[/FONT]
[FONT="]asymmetry is the result of random environmental noise, which at the moment seems to be[/FONT]
[FONT="]the most commonly accepted theory, the difference in information in the sides is diffýcult to[/FONT]
[FONT="]explain. It is quite possible that this difference is an artefact caused by inadequate sample[/FONT]
[FONT="]size. If this is not the case, it suggests that'sampling by individuar, as advocated by[/FONT]
[FONT="]Korey (1980) and Buikstra (1972) may be unwise, and that the 'additive genetic effects'[/FONT]
[FONT="]I[/FONT]
[FONT="]theory of Berry and Berry (1967, R. J. Berry 1968, A. C Berry 1975) and the'genetic basis of[/FONT]
[FONT="]asyrnmetry' theory of Ossenberg (1981) should be re-exan-dned. Until this matter is[/FONT]
[FONT="]resolved, it may be expedient to derive bilateral trait frequencies using one side only.[/FONT]
[FONT="] [/FONT]
[FONT="] [/FONT]