Canadian Re View of Physical Anthropolog Y Revue Canadienne D

CANADIAN RE VIEW OF PHYSICAL ANTHROPOLOG Y

REVUE CANADIENNE D 'ANTHROPOLOGIE PHYSIQUE

VOLUME 2 NUMBERS 1-2

anadian Associahon for Physical Anhropolog~ lAssociahon pour Mnthropologie Phpique au Unada Managing Editor

William D. Wade Department of Anthropology University of Manitoba Winnipeg, Manitoba R3T 2N2

Editorial Board

Braxton Alfred Susan Pfeiffer University of British Columbia University of Guelph Linda M. Fedigan Dwight A. Rokala University of Alberta University of Manitoba Francis Forest Shelley R. Saunders Universite de Montreal University of Toronto Christopher Meiklejohn Mark F. Skinner University of Winnipeg Simon Fraser University

Editorial Assistant

Louis Allaire University of Manitoba

The Reviewl~evueis published by the Canadian Association for Physical Anthro- pology/l'Association pour lYAnthropologiePhysique au Canada. Articles, letters, book reviews and other materials relevant to physical anthropology and its allied disciplines are invited. These may be submitted in either French or English, but articles must include an abstract in both languages. Material submitted for publication must follow the Wistar Institute Guide for Authors, which appears in the first issue of each year of the American Journal of Physical Anthropology. Membership inquiries, advertising copy and address corrections should be sent to the Secretary-Treasurer, Dr. N. S. Ossenberg, Department of Anatom3, Queen's University, Kingston, Ontario K7L 3N6.

44.>.

ISSN 0225-9958 The Skeletal Remains from the Taber Child Site, Taber, Alberta

ROBERT I. SUNDICK Department of Anthropology, Western Michigan University, Kalamazoo, Michigan 49008

KEY WORDS Early Man North America

ABSTRACT An analysis of a four to nine month old human infant with a possible date of at least 40,000 years, from Taber, Alberta, is presented. The specimen was found by A. MacS. Stalker in 1961 and has been reconstructed by the present author. The skeletal material derives primarily from the left side of the skull and face and, based on the degree of mineralization of the bone and the density of the matrix in which the material was embedded, it appears likely that the specimen is as old as has been suggested. The fact that only one side of the skull and face is present would indicate that erosion of the missing right side occurred before the specimen was embedded in the matrix. Loss of the right side after mineralization would have been unlikely because of the density of the matrix.

RESUME L'Ctude d'un enfant humain 2gC entre quatre et neuf mois, associC A une date possible de 40,000 ans, provenant de Taber, Alberta, fait l'objet de cette communication. Le specimen fut dCcouvert par A. MacS. Stalker en 196 1 et a 6tC reconstituk par l'auteur. Les parties du squelette appartiennent principalement au c6te gauche du criine et de la face. Sur la base du degrC de minkralisation de 1'0s et de la densite de la matrice lors de sa dkcouverte, il semble possible que le sujet soit aussi ancien qu'on le sugghre. Le fait qu'un cbtC seulement du cr2ne et de la face ait 6tC prCservC semble indiquer que 1'6rosion de la partie droite manquante a eu lieu antkrieurement A l'ensevelissement du specimen dans la matrice de sa dkcouverte. La perte du c6tC droit apr6s minkralisation aurait CtC peu probable en raison de la densite de la matrice.

The bones to be discussed herein are those and hence we have available for study only a which were discovered by A. MacS. Stalker and very small sample of fragmentary infant skel- his field staff from the Geological Survey of etal remains. The bones were first sent to Mr. Canada on July 1 1, 196 1, near Taber, Alberta. H. L. Shearman, of the National Museum of Their find was made accidentally while Stalker Canada who removed some of the surrounding and his colleagues were mapping surficial matrix from the bone, but left all of the pieces deposits in the area. At the time of the dis- together within the matrix. This lump of matrix covery Stalker was not aware that the small, with bone fragments extending from it was fragmentary bones which were almost entirely studied by Langston and Oschinsky ('63). Late embedded in a dense matrix were from Homo in the 1960's the material was given to the sapiens (Stalker '69). If this were known present author by Dr. David Hughes, of the Stalker admits that a more diligent search of the University of Toronto, for further study. I, in area for additional bones would have been turn, brought the material to Dr. A. Gordon undertaken (Stalker '69)' but this was not done Edmund of the Royal Ontario Museum,

CAN. REV. PHYS. ANTHROP. 2:l-6 1 ROBERT I. SUNDICK

Toronto, Ontario, who very generously re- molars. The bones recovered are primarily from moved the entire skeleton from the matrix in the skull and are described below: which it was still embedded. Once the bone was removed from its matrix it was possible to Parietals reconstruct parts of the cranial vault and to The left parietal (fig. 1) is virtually com- identify additional parts of the skeleton which plete, missing only the lateral anterior corner were included in the matrix. which articulates with the greater wing of the This report is a review of the evidence for sphenoid. The bone measures 104 mm. along the antiquity of the material and a description the sagittal suture and 92 mm. along the lamb- of the skeletal material itself, based on the most doidal suture. It is 2 mm. thick at lambda and complete reconstruction which has been under- 1 mm. at bregma. There is a good deal of boss- taken to date. The material now resides in the ing at the parietal tuberosity, which would be National Museum of Man, National Museums expected in an individual of this age. The articu- of Canada, Ottawa, Ontario (Catalog No. lation between the two parietals and the occipi- xv-C5). tal indicate that the posterior fontanelle is The site which yielded the material is closed. This should occur in the first few located approximately 3 miles north of Taber, months after birth. The inner surface (fig. 2) Alberta on the east bank of the Oldman River. shows grooves for the frontal branch of the More specifically it came from a cliff referred to middle meningeal vessels, which are identical to as the Woodpecker Island Bluff (Stalker '69). those seen in modern day man. Stalker has recently presented a number of The right parietal (fig. 3), less complete than arguments detailing the reasons why he believes the left parietal, consists of an 80 mm. section in the great antiquity of this specimen (Stalker along the sagittal suture extending from lambda '69, '77), and concludes the following: anteriorly. It ex tends laterally to the parietal "The Taber child is the most direct evidence for tuberosity. the presence of man on the Canadian prairies before retreat of the last Wisconsin glacier. It Occipital consists of the skull and some other bones from The squamous part (fig. 4) of the occipital a four-month-old child, found in alluvial sand some 60 feet below prairie level at a fresh bone is fairly complete. The superior median exposure along the Oldman River just north of fissure is fused. The mendosal suture or the line Taber, Alberta. Because the sand beds lie of union between the upper and lower parts of beneath a classical Wisconsin till, the child is at the squama cannot be observed on either side. least 18,000 years old. Further, the bones com- Grooves on the inner surface for the superior pare in mineralization and preservation with sagittal venous sinus and the transverse sinuses mid, or perhaps classical, Wisconsin bones re- are deep and easily discernible. The superior covered from near Medicine Hat to the east. sagittal venous sinus turns into the right trans- Elsewhere the unit that contained them has yielded radiocarbon dates on wood of > verse venous sinus to drain eventually into the 32,000 and > 49,000 years. The best estimate right sigmoid venous sinus of the temporal. A of their age is about 40,000 years." (Stalker small portion of the right exoccipital including '77: 135). the posterior two-thirds of the occipital condyle is present. There is no posterior condy- THE SKELETAL MATERIAL lar canal. The skeletal remains are those of an im- Temporal mature individual aged somewhere between four and possibly nine months of age. The early The left temporal (fig. 5) is intact except for age is assigned on the basis of the fact that the a portion of the squamous part anterior to the two halves of the mandible at the symphysis squamomastoid suture and posterior to the menti are not yet fused together. This is gener- zygomatic process. Maximum length of the ally thought to take place by six months of age. bone in the sagittal plane is 56 mm. The pos- The older age is based on the degree of calcifica- terior limb of the tympanic ring is fused to the tion of the lower, first and second deciduous petrous part through its entire length, while TABER SITE SKELETAL REMAINS 3

only the most superior part of the anterior limb based on X-rays of contemporary Boston chil- is fused to the squamous part. There is a small dren and the applicability of the standard to growth of bone extending down from the most this specimen is questionable. superior part of the anterior limb toward the posterior limb, forming a foramen of Huschke. Vertebrae This stage of development corresponds to The left lateral mass of the atlas and two stage 4 in Anderson's ('62) report on the devel- neural arch halves from the cervical or thoracic opment of the tympanic plate. The incus and region are present. malleus are present on this side. Clavicle Frontal The left clavicle with a midshaft circumfer- The frontal bone consists of two fragments ence of 14 mm. is present. A small segment of from the left side. One piece measuring 67 mm. the medial end is broken off and missing. by 39 mm. articulates with the left parietal and includes part of the edge of the anterior fonta- Tibia nelle. The other piece, a fragment of the orbit, A very small segment of the proximal end extends from the sphenofrontal suture medi- (31 mm. maximum length) of the right tibia is ally to a point just past the depression for the present. lacrimal gland. The zygomatic process of the Other small fragments of bone are present frontal bone is present. which cannot be positively identified. One fragment, however, appears to be a nonhuman Maxilla vertebral body. A barely recognizable fragment of the left Table 1 is a tabulation of the bones which maxilla is present. The alveolar sockets for the are now identified and those which were identi- lateral incisor, permanent and deciduous fied by Oschinsky ('63). As can be observed canines, first and second deciduous molars are there is a discrepancy between this report and exposed as the external surface of the alveolar the earlier one and I believe that this can be border is missing. There is a small part of the attributed to the fact that when Oschinsky nasal process of the maxilla present. studied the material it was still embedded in its matrix. He has, in fact, mistakenly identified a Mandible tibia as being a femur and skull fragments as The left half of the mandible is present and being a scapula. This points out that we should complete (fig. 6). Fusion has not yet occurred make every attempt to clean and reconstruct at the symphysis menti. The height of the bone our skeletal material before analyzing it in at the symphysis menti is 13 mm. and its depth order to prevent errors from occurring and per- is 10 mm. A single mental foramen is present sisting through time in the literature. below the anterior root of the first deciduous Additionally it should be noted that the molar. The lateral and central deciduous in- bones of the face and skull came almost entirely cisors and the deciduous canine are missing. from the left side with only a fragment of the The crown of the lateral permanent incisor is right parietal present. This indicates to me that present but cannot be observed. The first and the skull was most likely intact at the time of its second deciduous molars are present in their deposition and that differential preservation sockets. The first molar can be seen to be at a has led to the destruction of the right side. This stage where one-fourth of its root is developed. could have occurred if the right side were ex- An X-ray of the second molar shows it to have a posed to the elements before the cementing of completely calcified crown. The respective ages the bones together into their dense matrix. I for these teeth according to 50th percentile think it is unlikely that the destruction took standards for development of deciduous teeth place recently during the exposure of the by Fanning ('61) would be between 0.57 and deposits in which they were found, because of 0.92 years for the first molar and 0.83 and 0.88 the extremely dense nature of the matrix which years for the second molar. These standards are almost completely surrounded the bone.

ER SITE SKELETAL RE

Fig. 5 Left temporal,

one

hagnaents present present

Heft neural &arch

right absent kft Be fz absent two pr125d417 t 6 ROBERT I. SUNDICK

CONCLUSION ACKNOWLEDGEMENTS This specimen found by Stalker in Taber, I am grateful to Dr. A. Gordon Edmund of Alberta is an extremely important one in the Royal Ontario Museum, Toronto, Ontario, delineating when man first arrived in the New Canada who most graciously took the time and World. Stalker, based on the geological evi- put the effort into removing this skeletal dence, is firmly convinced that the specimen is material from the matrix in which it was em- as old as 40,000 years, if not older. Jennings bedded. Without his help this material never ('78) indicates that there was a land bridge would have been as completely reconstructed connecting Asia to North America at this early as it now is. I also express my thanks to Dr. time period, over which man could have trav- Jerome S. Cybulski of the National Museum of elled and Dragoo ('75) in a review of the evi- Man, National Museums of Canada who made dence for man's arrival in the New World indi- some of the information contained in the cates that this arrival could have occurred more report available to me. The photographs were than 30,000 years ago. The specimen itself can- taken by Mr. John Glover of the University of not be dated by the C-14 technique because Toronto. there is so little of it and because it has been treated with an unknown preservative. The LITERATURE CITED matrix surrounding the bone is unlike any that I Anderson, J. E. 1962 The development of the tym- have observed in more recent specimens and is panic plate. Contributions to Anthropology 1960, probably consistent with an early date. Part I. National Museum of Canada, Bulletin No. The fact that only this small sample is pres- 180. ent should not be taken to indicate that man Dragoo, D. W. 1976 Some aspects of Eastern North did not arrive in North America this early. I American prehistory: A review 1975. Am. Antiq. would presume that if man arrived in North 41: 3-27. America as early as 40,000 years ago that he Fanning, E. A. 1961 A longitudinal study of tooth formation and root resorption. New Zeal. Dental J. arrived in small hunting bands, who would not 57:202-217. have left significant amounts of skeletal or arti- Jennings, J. D. 1978 Origins. In: Ancient Native Amer- factual remains in one place for later discovery. icans. J. D. Jennings, ed. W. H. Freeman, San Fran- It is probably not unreasonable to assume that cisco. in a group such as this, that if a four to nine Langston, W. and L. Oschinsky 1963 Notes on Taber month old infant were to die he would be left "Early Man" site. Anthropologica.5: 147-150. behind in some convenient area for nature to Stalker, A. M. 1969 Geology and age of the early man take its course. This, in fact, might lead to the site at Taber, Alberta. Am. Antiq. 34:425-428. type of preservation seen in the Taber speci- -1977 Indications of Wisconsin and earlier man from the Southwest Canadian Prairies. Ann. New men. It is hoped that further investigations in York Acad. Sci. 288: 119-136. this region will lead to more definite evidence of man's early arrival in the New World. Etude Preliminaire de Certains Iroquo'iens Prehistoriques du Quebec ROBERT LAROCQUE et NORMAN CLERMONT D6partement dYAnthropologie, Universitd de Montrdal, MontrBal, P.Q. H3C 357

MOTS CLES Iroquoiens Laurentiens VariabilitC Biologique

RESUME L'analyse de caracteres mktriques, morphologiques, discrets et odontologiques, de quatre Cchantillons de crhes attribuCs i des Iroquoiens prC- historiques de la vallCe du St-Laurent, rCvCle une grande variabilitk intergroupe. La prCsence possible de quelques individus non Iroquoiens, dans un ou deux 6chantillons, pourrait partiellement expliquer cette importante hCtCrogCn6itk. 11 est d'autre part propos6 que des facteurs environnemen taux auraien t pu favoriser la differen tiation de ces groupes humains.

ABSTRACT Observations of metrical, morphological, discrete and odontological characters, on four samples of crania identified as prehistoric laurentian Iroquoians, show that these samples exhibit an important intergroup variability. This he tero- geneity could partly be explained by the possible presence of a few non4roquoian individuals in one or two samples. On the other hand, it is proposed that environmental factors could have favoured the differentiation of these human groups.

INTRODUCTION d'Hochelaga. I1 est n6anmoins Cvident qu'il doit En 1860, des travaux d'excavation effect& &re plus ou moins contemporain des sCjours de au coin des rues Sherbrooke et Mansfield, i Cartier au Nouveau-Monde. MontrCal, ont mis au jour les restes d'un village Les Ctudes biologiques ont 6th jusqu'i iroquoyen. Un inter& majeur de ce site a depuis rCcemmen t handicapbes par l'apparen te dispa- toujours rCsidC dans le fait qu'il a 6tC considCrC rition d'une partie du materiel osseux recueilli par Dawson. Cette lacune vient d'gtre partielle- comme &ant celui d'Hochelaga, visit6 par ment comblke, et c'est presque d'une toute Cartier en 15 35. Cette possibilitC peut en effet nouvelle collection dont nous parlerons ici. Elle &re envisagke, Ctant donne la situation gCo- rev& d'autant plus d'intCrCt que, depuis 1'6tude graphique du site, et la presence d'objets de d'Anderson ('72), d'autres squelettes iro- manufacture europkenne, dans le matCriel quoiens ont 6tC mis au jour dans la vall6e du rCcupkrC i lY6poque,mais cette interprktation a St-Laurent . Ainsi, sommes-nous dCsormais plus aussi 6t6 contestCe (Trigger et Pendergast, '72). en mesure de situer les Iroquoiens du site Le materiel, maintenant 6parpillC dans plusieurs muskes, et les rCcits de cartier, ont-CtC ' Ce texte est celui d'une communication presentee 6tudiCs par quelques auteurs, sans que la ques- par Robert Larocque au 7e Congrks Annuel de 1'Asso- tion soit pour autant dkfinitivement rCsolue, et ciation pour 1'Anthropologie Physique au Canada, nous ne pouvons espCrer trancher le dCbat i Ste-Adkle, Quebec, 1979. 11 a kt6 revise par l'auteur, l'aide de 1'6tude des squelettes humains. ainsi que par le Dr Norman Clermont, qui avait anterieurement fait une analyse prkliminaire des I1 n'est peutCtre pas si important de savoir Bchantillons des sites de Mandeville, de la Place que le site Dawson ait 6tC ou non celui du village Royale et de Lanoraie.

CAN. REV. PHYS. ANTHROP. 2:7-18 8 ROBERT LAROCQUE et NORMAN CLERMONT

Dawson, sur la base de leurs caractbres bio- (pp. 3 1 1, 3 14), on ne saurait les inclure dans la logiques, par rapport aux autres groupes d7Iro- collection du site Dawson, auquel ils ne sont pas quoyens de la vall6e du St-Laurent. associes. De plus, il n'est mCme pas acquis qu'ils soient contemporains des habitants de ce site. HISTORIQUE DE LA COLLECTION Nous devons donc traiter ces 2 Cchantillons DU SITE DAWSON comme s'ils reprksentaient 2 groupes distincts. Le site a d'abord 6t6 rCvC16 par la dkcouverte de I1 y a quelques mois, l'un de nous a pu squelettes humains. Dans deux publications de retracer d'autres pieces squelettiques, dans des 17Cpoque, Dawson a d6crit quelques caractbres mus6es affili6s 1 17UniversitC McGill. La col- mCtriques et morphologiques des 8 cranes les lection comprend d6sormais au moins 19 boites plus complets, laissant entendre qu'il y en eut criiniennes. Un vingtibme individu, iigC plusieurs autres dans un plus mauvais 6tat de d'environ 6 ans, est reprCsent6 par un fragment conversation (voir Trigger et Pendergast, '72 d'arcade alveolaire. app. 2 et 3). Au moins 20 squelettes auraient Cependant, ces 20 individus ne sont pas tous CtC dkgagCs, dit-il. associ6s de fa~oncertaine au site Dawson. Seu- I1 rapporte dans ses notes que les 8 crhes lement 10 le sont, les autres 176tant implicite- avaient une tendance la dolichocrdnie, une ment. Mtme parmi les crdnes certainement a trouvCs au site Dawson, il demeure possible que capacitC criinienne importante, et qu'ils prC- sentaient une grande variabilitk dans les dimen- certains soient ceux d'individus 6trangers h la population locale. sions. Cette variabilitC aurait 6tC telle qu'il en vin t croire que 176chantillonreprksentait plus Les considerations prC&dentes, relatives a a l'identification de ces 20 individus a la popula- d7unetribu, ou que la population 6tait m6tissCe. tion du site Dawson, nous obligent a entrevoir I1 signale aussi qu'en gCnCra1, les dents Ctaient les 3 possibilitCs suivantes: fort usCes, et que plusieurs Ctaient cariCes ou 1. ils proviennent tous du site Dawson, et tomb6es ante mortem. Ctaient tous membres de la communautC Au debut des ann6es '70, J. E. Anderson a locale. repris l'analyse h partir des squelettes du musee 2. ils proviennent tous de ce site, mais certains McCord (Anderson, '72, pp. 31 1-323). En se 6taient des 6trangers qui furent ensevelis basant sur les maxillaires et les mandibules pour une raison ou une autre, sur la terre de presents, il estima ?I au moins 10 le nombre la communautC locale. minimum d'individus de la collection, alors que 3. ils ne proviennen t pas tous de ce site. 4 crdnes seulement, plus ou moins complets, Tant que nous ne saurons a laquelle des ont pu Ctre observ6s. Aucun de ces criines ne possibilit6s nous rallier, les 2e et 3e se rCduisent put Ctre identifi6 comme l'un des 8 dCcrits par en fait a une seule. Or, dans 1'6tat actuel de la Dawson (ibid, p. 3 14). Une important partie recherche, on ne peut dire laquelle de ces pos- des squelettes rCcupCr6s par Dawson Ctait donc sibilit6s est la plus probable, car parmi le groupe manquante. d'individus qui ont certainement CtC trouvCs au MalgrC la maigre collection dont Anderson site Dawson, la variabilitC est importante. On ne disposait, il arriva A la conclusion que les peut donc pas raisonnablement exclure de ce individus Ctudi6s pouvaient ttre consid6r6s groupe, aucun des autres individus, sur la base morphologiquement comme des Iroquoi'ens, et de ses caractkres biologiques. que le groupe d71roquoYenss'en rapprochant le plus Ctait celui de Roebuck, site ontarien de la En comparant les descriptions publikes par Dawson, celles des individus prksentement a vall6e du St-Laurent, datant de 1 390 f 100 ans h ap. J. C. I1 nota aussi que certains attributs de la l'Ctude, on rCalise qu'il manque toujours des denture traduisaient un regime alimentaire criines de la collection originale. On arrive tout bas6 h la fois sur l'agriculture et la chasse- au plus A reconnaytre 2 des 8 criines d6crits par cueillette (ibid; pp. 3 19-320). Dawson. Dans quelques autres cas, les observa- Quant aux 4 crdnes en parfait 6tat de con- tions morphologiques correspondent, mais pas servation, trouv6s sur les flancs du Mont-Royal les observations mktriques, ou le contraire. Une h Westmount, et dont Anderson fait mention grande partie du nouveau matCriel, s'il vient du site Dawson, serait vraisemblablement celui, IROQUOIENS PREHISTORIQUES DU QUEBEC 9 trop fragmentaire, que Dawson n'a pas cru bon Les miiles ont de plus un profil aux lignes plus de dCcrire. MCme si on tient compte des im- arrondies, celui des femelles ktant plus angu- ponderables relies aux faits qu'il a pris ses leux, vu l'aplatissement davantage marque de la mesures en pouces, et qu'il a fallu ensuite les region post-bregmatique et de la rkgion sus- convertir en mm, que les dkfinitions de ses lambdoi'de, ou un degrk de courbure plus mesures pouvaient Ctre diffkrentes des netres, prononck entre ces 2 rkgions (fig. 1). et qu'avec le temps, certains criines ont pu Ctre Parmi les quelques tendances ou constantes endommages, on ne peut raisonnablement qui ressortent, signalons un relief sus-orbitaire reconnaItre plus de 4 des 8 criines qu'il a modkrk chez les miiles, et attknuk chez les observes et decrits. femelles, I1 est en forme de V dans presque tous Quant a la collection ktudike par Anderson, les cas. Le front est bas, et meme chez les miiles, elle fut retrouvke presque intkgralement, sauf il est relativement bien bombk, quoique moins quelques fragments. anguleux compark A celui des femelles. Sur plus de la moitik des sujets ou l'obsemation ktait DESCRIPTIONS DES COLLECTIONS possible, on remarque une dkpression ou un aplatissement de la rkgion post-bregmatique. La Dawson rkgion sus-lambdoi'de est aplatie sur la majoritk Des 20 individus, aucun n'est reprksentk des criines, et l'occipital est trks gknkralement par un criine complet. Quelques-uns ne le sont bombk. que par la calotte ou un fragment de calotte. La La norma posterior montre une pente determination de l'iige et du sexe de ces derniers biparietale prononc6e. En cons6quence, les reste delicate et sujette i caution, mais ils ont bosses pariktales sont relativement basses, ainsi aussi permis peu d'observations, et influencent que les euryons, et surmontent des murs donc peu les rksultats. Les iiges et sexes se tombant plutbt droit. MCme si les types ovoi'des repartiraient comme suit: tendent a dominer, en norma superior, on en - 2 enfants (6 et 12 ans), 8 jeunes adultes, 5 relkve quelques-uns de type plus ovalaire, mais adultes d'iige moyen, et 5 adultes iiges. qui se distinguent entre eux par la largeur - 3 individus de sexe indetermink (dont les 2 maximum du criine, relativement A la longueur enfants), 11 milles (3 sont discutables), et 6 maximum ou au dkveloppement lateral du femelles. frontal. En premiere analyse, un attribut immCdi- Sur les caractbres de la face, il y a peu 6 dire, atement evident de ce t 6chantillon est sa grande deux sujets seulement ayant une face bien con- variabilite morphologique et me trique. On semke. Cette lacune est moins grande pour les peu t souvent dire d'un Cchan tillon de criines mandibules, plus nombreuses et mieux con- d'une mCme population, qu'il s'en degage un servkes, dont on parlera a l'occasion des com- "air de famille", tels ceux de Mandeville ou de paraisons entre les collections. Westmount, qui seront dkcrits plus loin. Mais Notons que, des 3 tendances qui avaient kt6 tel n'est pas le cas de 1'6chantillon de Dawson. notkes par Dawson, deux d'entre elles ont aussi I1 est en effet plus difficile d'en dresser une kt6 relevees sur l'kchantillon prksent, a savoir la diagnose. Aussi, quelques individus tombent grande capacitk criinienne et la grande vari- en dehors du type iroquoi'en d6crit par Ander- abilitk individuelle . Quant h l'indice criinien son, pour certains caractbres. Les individus sur moyen, il indique bien, a 74, 6, une tendance A lesquels sont baskes ces constatations sont la dolichocrinie comme le disait Dawson, mais pourtant certainement associks au site Dawson, contradictoirement, les crhes auxquels il se et parmi les plus complets. rCfCrait alors avaient un indice moyen de 77,6. Le dimorphisme sexuel parait aussi plus important que celui observe dans les autres collec- Westmount tions, et en conskquence il a 6tk relativement La collection de Westmount a aussi des plus facile de determiner le sexe des individus origines obscures. Seule la localisation gko- les plus complets. Le volume de la boi'te graphique est indiscutable, bien qu'elle soit criinienne notamment, k tait trbs discriminant, imprkcise. Les 4 criines qui se trouvent celui des miiles ktant en moyenne trks grand. aujourd'hui au Muske National de 17Homme,a 10 ROBERT LAROCQUE et NORMAN CLERMONT

Ottawa, et qui sont identifies au catalogue du indice moyen de 77,7, s'6loignant ainsi de ceux musee comme Ctant des "Hochelagans and de Dawson dont l'indice moyen Ctait de 74,6. Mohawks from an ancient Iroquois site at West- La largeur maximum est aussi relativement plus mount, Montreal" (Anderson, '72, p. 3 l4), reculCe, et le front parait plus Ctroit par rapport auraient 6tC trouvCs dans un cimetiere pres du A la largeur maximum. Les boites crgniennes reservoir situC le long de la Cbte-des-Neiges, sur sont donc plut6t de forme ovo'ide que de forme le flanc ouest du Mt-Royal (ibid, p. 314, ovalaire. note 1). I1 est permis de douter qu'ils viennent Les arcades zygomatiques sont proCmi- tous de cet endroit, d'autant plus que de nom- nentes, et cela explique en partie que la face soit breux autres crhes ont 6tC trouvks B West- large, l'indice facial supkrieur 6tant inf6rieur a mount (Lighthall, 1898, 1899, '24). Un cinq- 50. Les orbites sont relativement larges et dis- uikme crine, qui aurait aussi 6tC trouvC B West- posCes horizontalement, plut8t qu'inclin6es mount, est actuellement dans les collections du vers les cetds, tandis que le nez est Ctroit. Le musee McCord. palais est large et tend A Cpouser la forme Ces dCcouvertes sont presque totalement d'un U. dCpourvues de tCmoins culturels. Seuls les modes de sCpulture et les caracteres biologiques Mandeville permettraient d'identifier Cventuellement la Le site de Mandeville est situC sur la rive population d'origine de ces 5 crines. Or, les ouest du Richelieu, pres de son embouchure. Ce modes de sepulture ont 6tC qualifies d'algon- site iroquoi'en de la fin du Sylvicole supkrieur a quiens par Lighthall ('22, pp. 73-75), qui avait 6tC fouillC B diverses reprises au cours des cependant d'abord cru qu'ils 6taient iroquo'iens annkes '70, sous la direction de Girouard et (1899, p. 206). De son c8t6, Anderson les range BarrC. I1 a livrC 6 fosses funeraires, 2 simples et 3 dans son type iroquo'ien (p. 3 14). Les moyens doubles, dont le contenu Ctait dans un Ctat de de datation font aussi dCfaut, si ce n'est conservation mediocre (Clermont et Falardeau, l'absence d'objets de manufacture europbenne, '77). qui suggere que ces sCpultures Ctaient antCri- La collection de crines est composke de 9 eures B la p6riode du contact. individus: 5 miles adultes, 1 femelle adulte, 2 Sauf celui du musee McCord, auquel il femelles subadultes, et un jeune enfant. Ici manque la mandibule, tous ces crines sont com- encore, ils sont beaucoup plus homogenes que plets. 11s sont en gQnCral beaucoup plus sem- ceux de Dawson, et le dimorphisme sexuel est blables entre eux que ne le sont ceux de Daw- rkduit. Le relief sus-orbitaire, en V, est modCr6. son. On y peut identifier 4 miles adultes et une Le front est gCnCralement bas chez les 2 sexes, adolescente. et chez les miles il est plus fuyant qu'il ne l'est Les capacitks criniennes different tres peu dans les autres collections. Le profil prCsente, l'une de l'autre, mdme en incluant l'adoles- chez les 2 sexes, des lignes adoucies (fig. 3). La cente, et sont en moyenne bien infkrieures B calotte est, en aucun cas, pyramidale, contraire- celles des individus du site Dawson. Le relief ment aux les collections prbckdentes. Le toit sus-orbitaire, en V, et la fuite du front, sont du crine est aussi moins 6levC. L'aplatissement modCrCs. En norma lateralis, on distingue 2 sus-lambdo'ide et le bombement de l'occipital formes de calottes: celles oh le frontal est sont tous deux plus faibles. MalgrC ces observa- plut6t fuyant et le vertex nettement dCgagC, et tions, aucun individu n'a un profil arrondi celles oi~le frontal est plus droit et le vertex comme il en existe parmi les sujets du site difficile B localiser (fig. 2). Si dans la collection Dawson . de Dawson des types semblables existent, on ne Par rapport B la largeur maximum de crine, trouve cependant pas dans celle de Westmount le front est plus large. L'indice crinien de 76,4 le type de calotte regulihrement arrondie, qui les rend lkgbrement brachycrines, et le crine, caractkrise quelques individus du site Dawson. vu en norma superior, est ovoide. Les bosses pariktales sont A la fois plus Les observations sur la face sont trop peu dCveloppCes et relativement plus ClevCes que sur nombreuses pour qu'on s'y attarde. Seuls 2 les crines du site Dawson. La norma superior individus ont la face presque complete, et l'un rCvble qu'ils sont plut6t brachycrines, avec un d'eux est affect6 d'une dkformation. IROQUOIENS PREHISTORIQUES DU QUEBEC 11

Fig. 1 Site Dawson. A) Un des crgnes de males qui B) Un crane femelle, au profil plus anguleux et au ont, en norma lateralis, une forme plut6t arrondie. front plus fuyant.

Fig. 2 "Site" Westmount. A) Un crftne de male, au B) Un crane de male, dont le vertex est nettement toit de forme pyramidal. moins degage.

Place Royale remarquer un aplatissement sus-lambdoi'de, Ce site du Sylvicole supbrieur, situe dans la une proeminence de l'occipital, un relief sus- ville de Quebec, a kt6 fouillk en 1977 par orbitaire modere et en V, et un relief musculaire Girouard et Mandeville. I1 contenait, en deux marque, surtout au niveau des crstes temp- fosses, les squelettes de 2 hommes et de 2 orales et des mandibules, ces dernihres Ctant femmes adultes, et de 6 enfants de moins de 6 particulihrement robustes. Le seul individu ans (Clermont, '78). dont on pouvait estimer l'indice crbnien, serait Pour fins de comparaisons, nous n'avons nettement brachycrbne, l'indice se situant aux retenu que les adultes, dont 1'Ctat fragmentaire alentours de 80. permet peu d'observations. On y peut toutefois ROBERT LAROCQUE et NORMAN CLERMONT

meme diffkrence aura en effet d'autant plus de chances d'etre significative, que les effectifs seront klevks. Nos comparaisons relevent donc surtout d'une perception de type intuitif. Dans le tableau 1 sont portkes les moyennes de 12 variables qui traduisent le dkveloppement de la boite crinienne chez les miiles. Pour toutes ces variables, les moyennes sont plus elevkes chez les habitants du site Dawson que chez ceux des 2 autres kchantillons, et parfois de faqon nette. Ce fait est conforme a l'observation formulke plus haut, a savoir que les crines miles du site Dawson sont plus volumineux que les autres. Mdme si plusieurs de ces variables sont corrklkes positivement, cela ne diminue en rien Fig. 3 Site Mandeville. Chez ce male, la norma l'ampleur du phknombne. laterah montre des lignes adoucies, comme chez tous Une autre variable s'inscrit dans cette ten- les individus de ce site, ainsi qu'un front trCs fuyant. dance. Mdme s'il n'a kt6 possible de calculer le module criinien que d'un seul mile du site Enfin, signalons la prksence, au Mdske Dawson, il est supkrieur de 6 et 8 unitks aux National de lYHomme,a Ottawa, des restes de moyennes des miles de Westmount et Mande- deux adultes et d'un enfant trouvks prbs du ville respectivement. Et ce mile est, dans village de Lanoraie, sur la rive nord du St- l76chantillon de Dawson, celui dont la capacitk Laurent. 11s ont kt6 offerts par Wintemberg, il y crinienne parait 6tre la plus rkduite. a quelque 50 ans. Le peu d'observations qu'on Malgrk les contraintes poskes par les effectifs y peut faire permet nkanmoins de confirmer rkduits, le test de t a kt6 applique aux variables l'identitk iroquoienne de la population qui dont l'effectif est au moins de 4 dans les kchan- occupa ce site. De meme, les modes de skpul- tillons de Dawson et Westmount. Sur ces 10 ture s'apparentent ceux d'autres Iroquoi'ens variables, 4 ont une moyenne significativement de la plaine de Montrkal (Clermont, '79). plus klevke chez les sujets du premier kchantil- Que peut-on conclure de ces quelques obser- lon. S'il arrive que d'autres crdnes viennent vations faites sur les quatre premieres collec- s'ajouter aux collections actuelles, a la suite tions? D'abord, que les quatre s'apparentent a d'autres trouvailles dans des muskes, et que les des degrks sensiblement diffkrents, au type tendances se maintiennent, d'autres differences iroquoi'en d7Anderson. Mais ce qu'il faut sur- pourraient Ctre significatives. tout retenir, c'est que les groupes de Mandeville Les effectifs des variables de la face, ktant et de Westmount, spkcialement le premier, encore plus rkduits, ne permettent pas un traite- prksentent une variabilitk intragroupe trks ment statistique de ces donnkes. Nous obser- rkduite, alors que celui de Dawson est beaucoup vons nkanmoins que les dimensions externes du plus hktkrogbne, et diffbre globalement des 2 palais presentent certaines tendances. En effet, autres groupes. Nous comparerons maintenant les gens de Dawson ont en moyenne un palais les variables mktriques et les variables discretes, plus long, plus large, et relativement plus ktroit. et pour terminer, les observations sur la L'indice moyen de l'arcade alvkolaire supkri- dentition. eure est de 1 1 2, comparativement a 1 2 1 et 122 chez les gens de Westmount et de Mandeville ANALYSE COMPARATIVE respectivement. Ce fait ne serait vraisemblable- Caracteres metriques ment pas Lik Q un degrk different de prog- nathisme, puisque les moyennes de l'angle Des kchantillons d'effectif rkduit empd- form6 par le nasion, le cprosthion et le basion chent gknkralement un test de t de dkceler une sont toutes autour de 74 . Enfin, notons que les diffkrence significative entre 2 moyennes. Une cavitks orbitaires, dkja relativement larges sur IROQUOIENS PREHISTORIQUES DU QUEBEC 13 les individus de Mandeville et de Westmount, L'utilisation du test de la diffkrence entre 2 le sont encore plus sur ceux de Dawson. variances, le test de F, pose les m6mes con- Dans les comparaisons entre mandibules, les traintes que celles du test de t. sujets qui attirent notre attention sont ceux de La variabilitk a kt6 testke h 3 niveaux. Ont la Place Royale, dont les mesures de robustesse d'abord kt6 comparks, les groupes mlles, en- sont toutes plus klevkes, en moyenne, que celles suite les groupes complets, et finalement les calculkes aux autres sites. Elles sont prksentkes mlles aux femelles d'un m6me Cchantillon. dans le tableau 2. On verra plus loin que ce fait De tous les tests sur l'kgalitk de 2 variances, est vraisemblablement en relation avec les un seul a rkvklk une diffkrence significative observations sur la den tition. Pour l'kpaisseur a entre 2 groupes mlles. C'est celui sur l'indice de la M2g7 la moyenne est significativement plus hauteur-longueur des groupes de Mandeville et &leveedans l'kchantillon de la Place Royale que de Westmount. Le test ne peut 6tre appliquk a la celles des kchantillons de Dawson et de Mande- collection de Dawson, un seul indice ayant pu ville . &re calculk. D'autre part, ont ktk comparks les

TABLEAU 1

Mandeville Dawson Westmount - - N x (mm) N x (mm) N x(mm) longueur 3 185.0 4 194.5 4 183.3 largeur 3 139.3 4 147.3 * 4 140.8 larg. front. min. 3 94.3 6 97.0 4 94.3 corde Na Br 3 107.3 5 119.4 * 4 110.8 corde Br Ld 5 109.0 7 113.4 4 107.8 corde Ld Op 3 97.7 3 100.0 4 96.0 arc Na Br 3 117.0 5 132.4 * 4 123.5 arc Br Ld 5 120.4 7 126.0 4 118.3 arc Ld Op 3 117.0 3 122.3 4 116.0 arc Br Po 5 151.4 4 157.3 4 150.0 p6rim. horiz. max. 3 506.7 4 533.8 * 4 506.0 arc Po Po 5 306.8 5 315.6 4 303.5 * identifie une difference significative (p =G 0.05)

TABLEAU 2 Mesures de robustesse de la mandibule (males et femelles)

Mandeville Place Royale Dawson Westmount N x(mm) N x(mm) N X(mm) N x(mm) Cpais. a la M2g 6 18.0 * 3 21.0 * 6 19.0 4 19.0 Bpais. i la symp. 5 16.4 3 17.0 7 16.4 4 16.3 haut. a la M2g 2 24.0 3 30.3 2 25.5 3 27.3 haut. a la symp. 5 33.2 3 36.3 6 33.2 4 32.5 * identifie une diffirence significative (p < 0.05) 14 ROBERT LAROCQUE et NORMAN CLERMONT

coefficients de variation des variables don t prCsentCs au tableau 4: 11 coefficients sont plus l'effectif mile est au moins de 4 dans les Cchan- ClevCs au site Dawson, contre 3 au site de tillons de Dawson et de Westmount, sauf celles Mandeville. Si la variabilit6 intragroupe est sur- de la mandibule, dont l'effectif est de 3. Les tout expliquke par le dimorphisme sexuel, alors miles de Mandeville n'ont pas 6tC retenus, les ce dernier est plus important chez les gens de effectifs Ctant trop souvent infkrieurs B 4. Des Dawson, sauf nouveau B la mandibule. 15 variables ainsi sClectionn6es, 11 ont un coeffi- La rCsorption alvkolaire a cependant pu cient de variation plus Clev6 dans la collection fausser des mesures prises au niveau de la 2e de Dawson que dans la collection de West- molaire, mais certainement pas celles prises au mount, et dans 4 cas c'est le contraire (tab- niveau de la symphyse mandibulaire, 06 aucun leau 3). cas de rksorption n'a 6tC remarque. I1 a d'autre I1 semble donc que les mensurations tendent part Cte vCrifi6 que la prCsence de 1 ou 2 in- a etre plus variables chez les miles de Dawson, dividus subadultes dam le calcul du coefficient sauf celles de la mandibule, oh cette tendance de variation de certaines variables, ne suffit pas est, ou peu marquCe, ou inverse. 2t expliquer les tendances observCes, tant B la La comparaison des groupes complets a bo?te crinienne qu'a la mandibule. Cette pr6- donnC des rksultats semblables. Le test de F n'a cision vaut aussi pour toutes les comparaisons port6 que sur les 6chantillons de Mandeville et jusqu'ici faites, bien que dans plusieurs cas elle de Dawson, oh le sex-ratio est sensiblement ne s'applique pas Cgal, et diffkrent de celui de 1'6chantillon de Pour savoir s'il y a des differences significa- Westmount. On observe alors que pour la tives entre les moyennes des males et des fe- largeur maximum et l'arc Br-Po, la variance est melles d'un mCme groupe, le test de t a CtC significativement plus grande parmi les sujets appliqu6 B diverses variables des groupes de du site Dawson. Le calcul des coefficients de Mandeville et Dawson. I1 arrive qu'il n'y a pas variation des variables dont l'effectif est au plus de differences significatives chez l'un que moins de 5 dans les 2 groupes, variables pour chez l'autre. I1 semble donc que la variabilitk lesquelles les sex-ratios sont, sinon Cgaux, tout intragroupe des variables metriques, plus au moins trks semblables, a donn6 les rksultats grande dans lY6chantillonde Dawson, soit due B

TABLEAU 3 TABLEAU 4 Coefficients de variation (m8les) Coefficients de variation (mdles et femelles)

Dawson Westmount Mandeville Dawson N x(mm) N V N V longueur 4 2.49 longueur 5 4.31 8 6.24 largeur 4 1.68 largeur 6 2.48 * 8 6.75 larg. front. min. 4 5.64 larg. front. min. 5 2.74 9 6.34 corde Na Br 4 2.79 corde Br Ld 8 6.05 12 7.23 corde Br Lb 4 3.96 corde Ld Op 5 1.98 5 4.84 arc Na Br 4 2.14 arc Br Ld 8 6.45 12 8.72 arc Br Ld 4 4.80 arc Ld Op 5 5.4 1 5 7.26 arc Br Po 4 2.83 arc Br Po 8 3.04 * 7 7.07 pCrim. horiz. max. 4 1.46 arc Po Po 8 4.05 8 6.23 arc Po Po 4 1.46 indice cCphalique 5 2.17 8 5.35 indice cCphalique 4 3.94 indice fronto-parietal 5 2.76 7 6.06 indice fronto-pariCtal 4 4.18 Cpais. i la M2g 6 9.94 6 6.65 Cpais. 5 la M2g 3 17.22 Cpais. i la symp. 5 11.07 7 5.94 Cpais. i la symp. 3 6.25 haut. 5 la symp. 5 10.08 6 9.23 haut. i la symp. 3 6.93 * identifie une variance significativement plus grande dans la collection Dawson (p < 0.05) IROQUOIENS PREHISTORIQUES DU QUEBEC 15 une plus grande variabilitk h lYintCrieur de Les frkquences des caracteres CpigBnktiques chaque sexe - tout au moins chez les miles - doivent toutefois &re interprCtCes avec pre- plut6t qu'i un plus fort dimorphisme sexuel. La caution, quand on songe h la subjectivitk inter- distribution bimodale d'une variable donnee venant dans 17apprCciationdu degrC d'expres- presenterait un plus grand Ctalement pour la sion de ces particularitks. I1 a en effet 6tC collection de Dawson que pour celle de Mande- dCmontrC rCcemment que, dans une batterie ville, sans qu 'il y ait une plus grande diffCrence d'observations r6pCtCes par le meme observa- entre les modes. Mais, pratiquement, on ne peut teur, certaines manquaient de fidClitC au point parler ici de distribution bimodale, vu les effec- oii l'erreur intraobservateur pouvait fausser tifs rkduits. lYinterprCtation relative aux affinitCs biologiques (Molto, '7 9). Caracteres discre ts Des quelques 20 caracteres dont la frk- Caracteres odon tologiques quence a 6tC relevee, environ la moitiC permet h Tout comme pour les mesures de la mandi- un groupe de se dCgager des autres. Les frkquen- bule, 17intCr&des observations sur la dentition ces de ces 9 caractkres sont donn6es dans le est centrC sur les habitants du site de la Place tableau 5. Ce sont les groupes de Mandeville ou Royale. Les chiffres du tableau 6 rnontrent les de Westmount qui se ddtachent le plus souvent faibles taux de caries et de dents tombkes ante des autres, alors que celui de Dawson est inter- rnortem h ce site. I1 a d'autre part 6tC Ctabli par mediaire, ou prks de l'un des deux autres. Clermont et Falardeau ('77), que les habitants

TABLEAU 5 Frtquences de quelques caractkres discrets

Mandeville Dawson Westmount

trous sus-orbit. 611 1 54 11/25 44 trous sus-orbit. doubles 0111 0 4/23 17 carene sagittale 518 62 3/18 16 aplatis. sus-lamb. 117 14 7/15 46 occipital bomb6 418 50 11/11 100 dthiscence tymp. 4/15 26 1/17 6 arc mylohyoYdien 218 25 219 apoph. geni fortes 516 83 218 menton bilateral 015 0 117

TABLEAU 6 Donntes odontologiques

Mandeville Place Roy ale Dawson Westmount N % N % N % N % dents possibles 210 100 101 100 196 100 135 100 pertes ante mortem 33 16 2 2 49 25 12 9 pertes post mortem 28 13 33 33 45 23 98 73 dents en place 149 71 66 65 102 52 25 18 caries 24 16 6 8 19 17 4 16 incisives en pelle 18 72 3 75 10 48 - - 16 ROBERT LAROCQUE et NORMAN CLERMONT des sites Mandeville et Dawson partageaient les Une partie de cette usure diffkrentielle serait m6mes attributs odontologiques. La dkcou- due aux nombreuses pertes ante mortem de verte rkcente d'autres individus ne fait que con- molaires. Ainsi, les dents antkrieures auraient- firmer ces observations antkrieures, qui dkja elles davantage kt6 utiliskes, pendant que les faisaient croire que les Iroquoi'ens de la rkgion quelques molaires toujours en place ne pou- de Montrkal avaient un rkgime alimentaire plus vaient plus s'user sur les molaires de l'arcade dkpendant des cultigknes que celui des Iro- dentaire opposke affectke de pertes ante quoyens de la region de Qudbec, ces derniers mortern (fig. 5a). Mais il est des cas oh cette ayant des modes de subsistance plus semblables explication parait insuffisante. Chez un h ceux des nomades (Clermont, '78). Rappelons individu du site Mandeville par exemple, l'usure que les documents ethnohistoriques et archko- est trbs avancke jusqu'a l'extrkmitk postkrieure logiques abondent dans le meme sens. de la face occlusale des deuxiemes prkmolaires, Les individus de Westmount se rangent sans qu'il y ait de traces perceptibles d'usure la plut6t du c6td de ceux de Mandeville et face mksiale des premieres molaires (fig. 5b). Dawson, bien que le taux de chutes ante Une discontinuitk de ce type peut difficilement mortem soit intermediaire entre celui de la &re 'expliquke comme une conskquence exclu- Place Royale et celui des 2 autres groupes. sive de la mastication alimentaire. L'usure des dents est en general importante, et plus avancke sur les dents antkrieures que sur CONCLUSION les molaires, sauf a la Place Royale, ou l'usure Si prkliminaire soitelle, cette ktude com- est plus uniforme. Cette usure diffkrentielle est parative de ces habitants de la vallke du St- surtout marquee au site Mandeville, ou les inci- Laurent permet nkanmoins d'en dkgager une sives sont parfois uskes jusqu'au collet, alors nette hk tkrogknkitk. que les sillons des molaires de m6mes individus Les individus du site Dawson se distinguent sont toujours apparents (fig. 4). des autres, non seulement par des boites cr5- niennes aux dimensions moyennes supkrieures et a des formes qui leur sont propres, mais aussi par une plus grande variabilitk l'intkrieur de chacune de ces catkgories de caract6res. De plus, les gens de la Place Royale avaient a la fois l'appareil masticateur le plus robuste et la denture la plus saine. Ces deux faits sont vraisemblablement liks entre eux, ainsi qu'a un rkgime alimentaire plus semblable a celui des nomades que ne l'aurait kt6 celui des habitants de la rkgion de Montreal. Malgrk le faible effectif des kchantillons, qui nous interdit toute gknkralitk a l'kchelle des populations, il est tout de m6me permis d'entrevoir certaines possibilitks relatives a l'explication de l'hktkrogknkitk observke. Tout se passe comme si certaines populations dYIroquoi'ens de la vallke du St-Laurent avaient entretenu avec les autres groupes de la vallke ou d'ailleurs, des rkseaux de relations fort diffkrents. L'homogCnkitC de l'kchantillon de Mandeville fait penser B un groupe particulierement endogame, alors qu'au site Dawson, il semble que tout se soit pass6 comme s'il y avait eu un important brassage de genes, toujours Q Fig. 4 Site Mandeville. Une usure diffkrentielle trbs la condition qu'aucun ktranger la marquCe entre les I' et la C droite, et les molaires. communautk locale se soit glissk dans IROQUOIENS PREHISTORIQUES DU QUEBEC 17

Fig. 5 Site Mandeville. A) La chute ante mortem de B) Un gros plan des faces occlusales de la P2d et de molaires inferieures a contribuk i l'usure diffkrentielle la M1 d. observke A l'arcade dentaire supkrieure.

1'6chantillon. D'autre part, la situation du certainement moins le cas selon un axe est- site de la Place Royale, a l'extreme nord de ouest. Puis, cette differentiation initiale a pu l'IroquoiXe, a pu engendrer une adaptation donner lieu une autre source de diffkrentia- propre un environnement au climat plus tion, A savoir l'ktablissement et le maintien, par rigoureux, et peut-etre moins permissif pour les communautks des rkgions de Montrkal et de l'agriculture que la plaine de Montrkal. Qukbec tout au moins, de relations avec des I1 est de plus lkgitime d'envisager la possibili- groupes humains diff6rents. Cette dernibre tk que l'exploitation de ressources alimentaires source de diffkrentiation a pu se manifester distinctes, ou tout au moins dans des propor- notamment par le dkveloppement de patri- tions diffkrentes, ait engendrk certaines diffkr- moines gknktiques distincts, et aurait kt6 ences entre les habitants des regions de d'autant plus importante que la rkgion de Montrkal et de Qukbec, non seulement dans les Montrkal ktait un carrefour. schbmes d'k tablissemen t, l'organisation sociale, I1 est nkcessaire d'observer d'autre restes la culture matkrielle ou les phknotypes, mais humains pour donner du poids, s'il y a lieu, aux aussi dans les kchanges gkniques. La pauvretk tendances relevkes. D'autres recherches nous des documents ostCologiques de la Place Royale permettront peut-&re de localiser le reste du nous interdit d'Ctre plus catkgorique, mais les material trouvk h Westmount, et aux sites documents ethnohistoriques laissent entendre Dawson et de la Place Royale. L'analyse des que les groupes de Stadaconk et d'Hochelaga individus du site Roebuck, et de ceux des sites 6taien t rivaux, et qu'ils en tre tenaien t des rela- qui restent a dkcouvrir, contribuera aussi a tions diffkrentes avec les groupes de l'intkrieur. kprouver le modkle proposk, relatif aux rela- L'important brassage de ghes qui parait tions qu'entretenaient les IroquoYens de la s'6tt.e pass6 dans la rkgion de Montrkal, pourrait vallke du St-Laurent, entre eux et avec d'autres s'expliquer par le fait que cette rkgion aurait kt6 groupes humains. plus propice a une assimilation culturelle et biologique de groupes plus ou moins distincts. REMERCIEMENTS Le St-Laurent aurait donc pu favoriser la La dkcouverte d'autres crines du site differentiation des groupes humains qui occu- Dawson a Ctk possible grace aux demarches de paient ses rives, et cela de deux fa~ons.D'abord, M. Conrad Graham, du muske McCord, qui a de son orientation, grossibrement selon un axe plus consen ti a pr6 ter les crines de ce tte collec- nord-sud, aurait favorisk une diffkrentiation tion. Nous l'en remercions infiniment, de meme culturelle et phknotypique des groupes exploit- que Mrs. Allison, du muske Redpath, qui a aussi ant des environnements diffkrents, ce qui est 18 ROBERT LAROCQUE et NORMAN CLERMONT contribuk A la dkcouverte de quelques cranes. OUVRAGES CONSULTES Nous remercions aussi le Dr. Jerome S. Anderson, J. E. 1972 Dawson site physical anthropol- Cybulski, de Muske National de l'Homme, qui a ogy. In: Cartier's Hochelaga and the Dawson site, bien voulu preter les cranes de la collection de B. G. Trigger et J. F. Pendergast, Cd. McGill- Westmount, et M. Laurent Girouard, qui a Queen's University Press, MontrCal. permis l'utilisation de donnkes encore inkdites Clermont, N. 1978 Rapport sur les squelettes Syl- sur les collections de Mandeville et de la Place vicoles de la Place Royale. Manuscrit. Royale. -1979 Analyse des restes humains de Lanoraie. Manuscrit. Robert Laro cque Clermont, N. et L. Falardeau 1977 L'analyse des Norman Clermont sCpultures du village iroquobn de Mandeville (BiFg-1). Manuscrit. , La presentation de la version initiale de ce Lighthall, W. D. 1898 A new Hochelagan burying ground. The Canadian Antiquarian and Numis- texte, au 7e Congrbs Annuel de l'ACAP, a valu 1 matic J. 3e drie, 1: 149-159. un des auteurs de la prksente version, de par- , 1899 Hochelagans and Mohawks: a link in Iro- tager le prix McKern-Oschinsky, attribuk A l'kt- quois history. MCm. et comptes rendus de la Soc. udiant dont la communication a kt6 jugke la Roy. du Canada, 2e sdrie, vol. 5 sec. 2: 199-211. meilleure. Je tiens donc A remercier tous les -1922 The Westmount stone-lined grave race. membres de l'association qui, de prbs ou de MCm. et comptes rendus de la Soc. Roy. du loin, ont eu un r61e a jouer dans l'attribution de Canada, 3e drie, vol. 16, sec. 2:73-75. ce prix. I1 est a souhaiter que l'attribution d'un -1924 Hochelaga and "The hill of Hochelaga". tel prix, ainsi que l'invitation faite aux laurkats MCm. et comptes rendus de la Soc. Roy. du Canada, 3e drie,vol. 18, sec. 2:91-106. (-tes) de publier leur texte, deviennent une Molto, J. E. 1979 The assessment and meaning of tradition de notre association, tant il est stimu- intraobserver error in population studies based on lant pour un 6tudiant de recevoir ces honneurs. discontinuous cranial traits. Am. J. Phys. Anthrop. 51:333-344. Robert Larocque Trigger, B. G. et J. F. Pendergast 1972 Cartier's Hochelaga and the Dawson site. McGill-Queen's University Press, Montreal. Possible Pre-Columbian Treponematosis on Santa Rosa Island, California JEROME S. CYBULSKI Archaeological Survey of Canada, National Museurpt of Man, National Museums of Canada, Ottawa, Ontario KIA OM8

KEY WORDS Treponematosis Cranial Syphilis Radiocarbon Dates California Prehistory

ABSTRACT Two crania from the Skull Gulch site on Santa Rosa Island, Cali- fornia, suggest treponematosis. The bone changes are described in view of recently proposed diagnostic criteria in dry bones. One cranium suggests vault lesions of the caries sicca sequence while the other mainly exhibits naso-palatine destruction. The site context suggests pre-European contact for both items. Pre-Columbian antiquity is strongly indicated for at least one skull on the basis of radiocarbon dates which, when corrected for fluctuations in atomospheric C-14, yield calendar dates of A.D. 1340 (error range from 1260 to 1390) and A.D. 1100 (error range from 1030 to 1190). The latter date is based on collagen from the right tibia believed associated with the skull.

RESUME Deux crhnes provenant du gisement de Skull Gulch dans I'ile de Santa Rosa en Californie semblent indiquer un cas de trCponCmatose. I1 est possible de dCcrire lYaltCrationosseuse i partir de critbres diagnostiques rCcemment reconnus dans 1'0s sec. L'un des crhnes suggbre un cas de lCsions de la voute affectant le sequence caries sicca alors que l'autre cas montre surtout une destruction naso- palatine. La contexte du gisement suggbre un contact prC-EuropCen dans les deux cas. Les datations radiometriques sugghrent Cgalement un iige pre-colombien pour l'un des crhnes; en effet, une date de 1340 avant J.C. (variant entre 1260 et 1390) et une date de 1100 avant J.C. (variant entre 1030 et 1 190) on 6tC obtenues en tenant compte des fluctuations de C- 14 atmosphCrique. La dernibre date est bas6e sur le collagbne du tibia droit que l'on croit Ctre associC au crhne.

The question of pre-Columbian venereal difficult, if not impossible, to distinguish from syphilis in the New World has long been those due to venereal syphilis. Hence, when debated by many scholars largely on the basis referring to the bone lesions in archaeological of skeletal remains (cf. Williams, '32; Crosby, materials, it is better to speak of treponemato- '69). It has been theorized, however, that vene- sis or treponemal infection than to assign a real syphilis is but one of four syndromes in a specific etiology. common organismal disease, treponematosis, Major problems in the debate have centered the others being pinta, yaws, and endemic (non- on the frequency of lesions in New World skele- venereal) syphilis (or treponarid), and that en- tal remains, correctness of diagnosis in putative vironmental circumstances will dictate the specimens, and accuracy of dating. Stewart form the disease takes (cf. Steinbock, '76, ('73) and, more recently, Hackett ('76) have pp. 90-94). Moreover, the bone lesions that suggested that a general survey of the New result from yaws and endemic syphilis are often World material might resolve these problems.

CAN. REV. PHYS. ANTHROP. 2:19-25 19 20 JEROME S. CYBULSKI

Literature reviews of published specimens as The crania were first brought to my atten- well as proposed new material have been re- tion in 1973 while I was teaching at the Univer- ported by Steinbock ('76) and El-Najjar ('79). sity of California. At the time, eight long bones, This report places on record two potentially a sacrum, two innominate bones, and a mandi- pertinent additional specimens. Together , the ble were boxed with the Burial 11 skull. two crania are not without problems in terms of Though some analysis was then undertaken, it certainty of diagnosis and temporal context. was not until I was settled in a new position at However, the lesions strongly suggest the National Museum of Man in Ottawa that I treponemal infection. Two different manifesta- was able to complete the study. In 1976 I tions are present, the crania are from the same requested that the materials be sent to Ottawa, archaeological locality, and for at least one of and only the long bones and mandible were the cases pre-Columbian antiquity is strongly received with the skull. The remaining parts indicated. This paper details the context of the could not be found. Because the long bones had items, the efforts made to establish their dating, not originally been numbered, there is a mea- and the lesions manifested. sure of uncertainty regarding their field association with the skull. However, because of infor- MATERIAL AND CONTEXT mation in the field notes, similar preservation in The two crania are in the collections of the substance and in staining, and commonness of Santa Barbara Museum of Natural History, sex identification, the long bones and skull are catalogued NA-CA-13 1.2B-12-49 (formerly believed to be of the same person. The mandi- #4246) and NA-CA-13 1.2B-12-6 1 (formerly ble articulates well with the skull and there is #4181). They were recovered early in the little question of its anatomical association. 1950's by Phil C. Orr, then curator of the Both crania and the long bones were judged to museum, from the Skull Gulch site (SBMNH- be of adult females, the crania assessed at 131.2) located on the northwest coast of Santa middle-age on the basis of the degree of suture Rosa Island in the Santa Barbara Channel. Cali- closure. The long bones include the left fornia. The site, intermittently excavated over a humerus and radius, right and left ulnae, right 14-year period (Orr, '68), consisted of a num- femur and fibula, and right and left tibiae. ber of house-pits and three cemeteries desig- nated A, B, and C. The skulls were recovered DATING from two different parts of Cemetery B, desig- Overall, the Skull Gulch site appears to date nated eastern and western. from prior to European contact: The eastern part of Cemetery B mainly con- Radiocarbon dates indicate the Skull Gulch site sisted of individual interments, some disturbed, was inhabited from about 2000 BP to 400 BP. including 26 people. The western part was pri- This latter date would indicate that the village marily "an unbelievable charnel of concen- may have been abandoned about the time of trated human bones" in which there was "70 Cabrillo's visit to the Island in 1542 . . . ,but if skulls and hundreds of skeletal parts, jaws, and it was occupied later than that, no glass or fragments of skulls." Ten skeletons were also metal found its way into the portions we have present, "though most showed signs of having excavated (On,'68, p. 191). been disturbed, by a missing leg or arm." (Orr, Orr considered Cemetery B to be younger '68, pp. 199-201). than Cemetery A, the latter yielding radiocar- According to field notes on file at the bon dates, based on shell, of 900 f 100 C-14 museum, NA-CA-13 1.2B-12-49 was given the years (UCLA-178) and 1820 f 90 C-14 years field designation "Burial 5 0" while NA-CA- (UCLA-135).' One radiocarbon date was ob- 131.2B-12-6 1 was "Burial 1 1". The latter was a disturbed individual in the eastern part of the 'Walker ('78), in a dental study of materials from the cemetery and the former was one of the 70 site, incorrectly reported the dates as "1,050 + 80 B.P. and 130 * 90 B.P.". Both Orr ('68) and the UCLA skulls in the western part. Their field positions Laboratory (Fergusson and Libby, '63, p. 4) list the are schematically illustrated in a burial map of dates as reported in the present study, although Orr Cemetery B in Orr's published report ('68, gives an error term of + 80 for the first date. Orr ('68) p. 198, fig. 64). regarded the older date as the more accurate of the two. TREPONEMATOSIS ON SANTA ROSA ISLAND 2 1 tained from Cemetery B, specifically from unsuccessful. The laboratory returned the jaw black seeds of red maid (Calandrinia ciliata) noting that it contained too much preservative which filled the "interstices" of the bones of (glue and plaster). Burial 13 in the eastern part. The date was 600 + 70 C-14 years (UCLA-102) which when LESIONS corrected for fluctuations in atmospheric C-14 As stated in the introduction the two skulls yields a calendar date of A.D. 1340 with an are pathologically affected in different ways. error term from A.D. 1260 to A.D. 1390 In NA-CA- 13 1.2B-12-6 1 only the cranial (MASCA, see Ralph et al., '74). vault is affected (fig. 1) including the frontal Burial 13 lay to the west of Burial 1 1. Orr's bone, parietals, and the occipital squama. The published description ('68, p. 200) and his field temporal bones, facial skeleton (including the notes suggest that the seeds could have been mandible), and cranial base, all of which are deposited sometime after the interment of intact, appear to be free of changes. A general Burial 13, thereby making the time of death of area of involvement extends from the central that individual earlier. Although the single part of the superior half of the frontal bone, radiocarbon date might be sufficient ground to through bregma, and through the medial halves conclude pre-Columbian antiquity for the buri- of the parietal bones - the left more exten- als, it was deemed desirable to obtain a more sively than the right - almost to obelion. A direct date for Burial 1 1 and, not incidentally, a small area of involvement is at left stephanion, verification date for this portion of the cem- and the area around right stephanion is more etery. In 1977 I submitted the right tibia from extensively involved. Most of the planum the long bones believed associated with the occipitale is involved, the lesions infringing NA-CA-13 1.2B-12-6 1 skull to the Saskatche- upon but not extending below the nuchal ridge. wan Radiocarbon Laboratory. The date re- There is a relatively small, near isolated lesion in turned, based on collagen extraction, was the posterior half of the right parietal and a 865 f 65 C-14 years (S-1286, NMC-902) or, small isolated lesion in the left parietal pos- corrected, A.D. 1 100 with an error term from terior to the eminence. A.D. 1030 to A.D. 1 190 (MASCA). The areas of involvement at the top of the Burial 50 may or may not be contemporane- vault, about right stephanion, and in the occi- ous. Orr regarded the western part of Cemetery pital show irregular knurled heaping of healed B to be younger than the eastern part. The mass bone around foci of destruction. What has been of bones was said to lie "somewhat higher" described as the "stellate lesion characteristic than the latter, and their disposition suggested of cranial syphilis" (Steinbock, '7 6,p. 129) is to Orr that the bones represented "reburials" in especially clear in the area of right stephanion. the manner found among late sites on the Santa "Caries sicca" has long been considered Barbara coast. There is the possibility, however, pathognomic of cranial syphilis (cf. Williams, that the bones could represent disturbance of '32). In an attempt to clarify problems in the earlier intact burials by later inhabitants of the identification of the lesion in dried bones, site. This is suggested by the presence and con- Hacket ('76) has recently proposed the "caries dition of the ten skeletons, noted above, in the sicca sequence" as a diagnostic criterion of western part of the cemetery. syphilis based on a study of 424 pathological As no long bones were associated with the crania in European medical museums. Of these, Burial 50 (NA-CA-13 1.2B-12-49) skull, a radio- 28 1 were "labelled" syphilis (33 with authentic carbon date was not attempted. Although a diagnoses), while the remainder were labelled date might be obtained from the skull itself, I with other diseases. Sound diagnostic criteria was reluctant to have it processed because of its were additionally proposed for tuberculosis, potential comparative value for diagnosing the leprosy, some neoplasms, Paget's Disease, presence of pre-Columbian treponematosis in trauma, and hyperostosis frontalis interna, all the New World. For this reason as well, the of which might be confused with syphilis NA-CA- 13 1.2B-12-6 1 skull was not directly (treponematosis) at some stage in their develop- dated. An attempt to date the mandible proved ment.

24 JEROME S. CYBULSKI

The changes in NA-CA-13 1.2B-12-6 1 most state of "syphilis" than that deemed diagnostic closely fit Hackett's diagnostic criteria of of the disease. syphilis. He illustrates several changes in the caries sicca sequence (see esp. p. 31, fig. B) DISCUSSION which are represented in the skull: "focal None of the eight long bones believed associ- superficial cavitation" and "circumvallate cavi- ated with the NA-CA-13 1.2B-12-6 1 skull tation" at the top of the vault, and "cir- appear pathological, and x-rays show no clear cumvallate cavitation" and "radial scar" in the internal abnormalities. Though treponematosis right stephanion area of involvement . Overall, may affect most bones of the skeleton, confine- the most extensive pathology is on the exterior ment of lesions to the skull is not uncommon, surface of the cranium, as is expected in cranial particularly in known cases of venereal syphilis syphilis. However, radiographs (now on file at (Steinbock, '76,pp. 113-1 14,127). Indeed, the the Santa Barbara museum) suggest thickening lack of involvement in the long bones might, into the diploetic space and, as viewed through therefore, rule out yaws and endemic syphilis the foramen magnum, there is at least one focus since the long bones are most often involved in of porosity on the inner surface at the top of these two syndromes. Further, the cranium is the vault. These internal changes are also con- "rarely affected" in endemic syphilis while in sistent with the diagnosis. yaws it is "infrequently involved" (Steinbock, In NA-CA-13 1.2B-12-49 the pathology is '76,pp. 139,143). mainly in the facial skeleton (fig. 2). There is Applying the stringent criteria of Hackett irregular destruction of the hard palate, pri- ('76) to the present study, it does appear that marily through the midline and directed later- both Santa Rosa Island crania exhibit trepo- ally; destruction of the alveolar bone between nemal infection. He emphasizes that the .bone the first premolar sockets; destruction of the lesions of venereal syphilis, yaws, and endemic nasal borders of the maxillae with some infold- syphilis cannot at present be separated. ing and healing; partial destruction and remain- The overall context of the site suggests pre- ing malformation of the nasal bones with heal- European contact for both skulls, and pre- ing; destruction of internal parts of the nasal Columbian antiquity is strongly indicated for passage including the vomer and maxillary sinus NA-CA-13 1.2B-12-6 1 on the basis of radio- walls. There is a small patch of porous osteitis carbon dates. The use of correction factors as on the right maxilla near the nasal aperture. applied here is now widely considered neces- According to Hackett ('76, p. 65), naso- sary to provide more accurate and consistent palatine destruction is a diagnostic criterion of calendar information. The error terms for the syphilis "when extensive and healed". Specifi- dates further enhance the conclusion. cally he refers to "an empty nasal cavity with The Santa Rosa Island crania are not the first smooth lateral walls presenting a smooth, California specimens said to demonstrate possi- 'bored out' tunnel-like passage-way" as charac- ble pre-Columbian treponematosis. Roney teristic of his specimens labelled syphilis ('66) has suggested "syphilis" to account for (P. 63). bone changes in nine individuals from a north- NA-CA-13 1.2B-12-49 does not exhibit this ern California coastal site dated to the 6th to "advanced state" and its unhealed portions 3rd centuries B.C. More recently, Brothwell may give the impression of neoplasm (Hackett, and Burleigh ('75) reported "possible evidence '76, p. 65). However, the healing in the nasal of treponemal infection" in a Berkeley museum area and infolding of the maxillary borders is specimen which they radiocarbon dated at more characteristic of what Hackett has 857 If: 52 C-14 years with a corrected calendar described as syphilis. Further, there is a small date of A.D. 1 105. Details concerning the speci- area of healed bone change in the frontal to the men, including site of discovery, represented left of midline that appears to include a de- skeletal parts, and manifested lesions were not pressed "radial scar" of the caries sicca provided. Given these previous reports and, sequence. On the basis of all evidence discussed especially, the present data, a survey of existing by Hackett, it appears that the naso-palatine California archaeological skeletal material as changes in the skull represent a less advanced well as new excavations may well enhance our TREPONEMATOSIS ON SANTA ROSA ISLAND 25 knowledge of pre-Columbian treponematosis in Fergusson, G. J. and W. F. Libby 1963 UCLA radio- the New World. carbon dates 11. Radiocarbon 5: 1-22. Hackett, C. J. 1976 Diagnostic Criteria of Syphilis, ACKNOWLEDGEMENTS Yaws and Treponarid (Treponematoses) and of Some Other Diseases in Dry Bones (for Use I thank Murchie Corcoran and Robin in Osteo-Archaeology). Springer-Verlag, Berlin- Krogfoss for their invaluable assistance with the Heidleberg-New York. field records of the Skull Gulch site excavations Orr, P. C. 1968 Prehistory of Santa Rosa Island. Santa and Doctor D. Travis Hudson, Curator of Barbara Museum of Natural History, Santa Anthrop.ology at the Santa Barbara Museum of Barbara. Natural History, for entrusting me with the Ralph, E. K., H. N. Michael and M. C. Han 1974 Radiocarbon dates and reality. Archaeol. Eastern skeletal material and for his advice and encour- N. Amer., 2:l-20. Reprinted from M.A.S.C.A. agement during the analysis. Paul Heuston, Newsletter, Vol. 9, No. 1,August, 1973. with the Department of Anthropology at the Roney, J. G. Jr. 1966 Palaeoepidemiology: an exam- University of California at Santa Barbara, took ple from California. In: Human Palaeopathology. the photographs. Facilities and financial sup- S. Jarcho, ed. Yale University Press, New Haven port were provided by the National Museum of and London, pp. 99-107. Man, National Museums of Canada. Steinbock, R. T. 1976 Paleopathological Diagnosis and Interpretation. C. C. Thomas, Springfield. LITERATURE CITED Stewart, T. D. 1973 The People of America. Charles Scribner's Sons, New York. Brothwell, D., and R. Burleigh 1975 Radiocarbon Walker, P. L. 1978 A quantitative analysis of dental dates and the history of treponematoses in man. J. attrition rates in the Santa Barbara Channel area. Archaeol. Sci. 2: 393-396. Am. J. Phys. Anthrop. 48: 101-106. Crosby, A. W., Jr. 1969 The early history of syphilis: a Williams, H. U. 1932 The origin and antiquity of reappraisal. Amer. Anthrop. 71:2 18-227. syphilis: the evidence from diseased bones. Arch. El-Najjar, M. Y. 1979 Human treponematosis and Pathol. 13:779-814; 931-983. tuberculosis: evidence from the New World. Am. J. Phys. Anthrop.51:599-618. Sexually Dimorphic Variation in the Ossification of the Hand and Wrist in Japanese Children C. E. EYMAN Department of Archaeology, University of Calgary, Cizlgary, Alberta T2N 1N4

KEY WORDS Ossification Center Sexual Dimorphism Growth Japan

ABSTRACT Carpal and metacarpal ossification center maturity scores were recorded, and the skeletal/chronological age ratio as well as the ossification sequences were examined in a cross-sectional sample of 3,283 inbred and non-inbred ("control") Japanese children. All hand-wrist ossification centers were examined for the presence or absence of bony variations as well as for "missing" centers. Extensive statistical testing of the data indicated that sex was the only factor clearly associated with the variations, which occur in approximately half the sample. The higher frequencies of these variations among females, coupled with their faster maturation rates and smaller absolute skeletal sizes may reflect a hereditary sexual dimorphism, for which an explanation is suggested.

RESUME On a enregistrk les dkcomptes des centres de maturitk de l'ossification du cape et du mktacarpe et examine le rapport squelette/iige chronologique dans un kchantillonnage en section transversale de 3823 enfants japonnais innks ou non- innks ("de contr61eM).Tous les centres de l'ossification du poignet et de la main ont 6tk examinks pour des centres "manquants". Une analyse statistique approfondie des donnkes indique que le seul facteur clairement associk aux variations est sexuel; on le retrouve dans approximativement le moitik de l'kchantillon. La plus grande frkquence de cesvariations chez les femmes, associke ii leur taux de maturation plus &lev&et les dimensions absolues du gquelettes plus rkduites, semble refl6ter un dimorphisme sexuel hkreditaire, pour lequel on suggkrera une explication.

INTRODUCTION any atomic irradiation at all (Schull and Neel, In the 1950's, the Department of Human '65, p. 5). Hiroshima and Nagasaki had been Genetics of the University of Michigan Medical repopulated by Japanese immigrating from School undertook an extensive study of chil- nearby rural areas. They are essentially aggre- dren in Nagasaki and Hiroshima. Schull and gates of smaller villages, so they reflect much Neel ('65) have presented the results of the the same degree of consanguineous marriages as overall study in a comprehensive monograph on do the rural areas (Schull, '58). some 7,700 children between the chronological There are numerous socioeconomic, reli- ages of 5 and 11.5 years. The study was de- gious, and familial-traditional patterns which signed to assess the genetic effects of inbreeding reinforce the system of consanguinity even on the biological and mental attributes of a large human sample. 'A version of this paper was presented at the 7th annual meeting of the Canadian Association for Physi- About 60% of the children were the off- cal Anthropology, Ste.-Adile,QuBbec, 1979. spring of parents who had not been exposed to

26 CAN. REV. PHYS. ANTHROP. 2:26-52 JAPANESE HAND-WRIST OSSIFICATION 27 though the inbreeding patterns have largely tively common and "normal" variations. A broken down. The disruption and decrease in scrutiny of the standard comments about rural population densities accompanying the sexual dimorphisms in the adult human skele- repopulation of Hiroshima and Nagasaki did ton in the standard textbooks in physical not seriously affect the marriage and kinship anthropology indicated that there is relatively patterns of inbreeding, nor did these changes little attention paid to the meanings of these create any non-random marital patterns in the differences and to the reasons for their overlap three most common forms of consanguineous between the sexes. While there is a wealth of marriage. The frequencies of these mating data on skeletal maturation in male and female categories were essentially the same for both children, the meanings of the skeletal sexual cities. Approximately half the children, on the dimorphism and the between-sex overlap with basis of marital records extending back for regard to the more rapid female maturation are about 100 years, constitute a "control" sample still somewhat obscure. I thus began to ques- of non-inbred children (or at least non-inbred tion whether the patent sexual dimorphism in beyond fourth cousins). the overall human skeletal system partly fol- In the data collection, X-rays of left hands lows from the sexual dimorphism in skeletal and wrists were taken. There were a number of maturation rates among preadolescents. The purposes in assessing the hand-wrist X-rays. females were observed to have higher frequen- One was to attempt an assessment of the degree cies of variations than did the males. Applying of skeletal maturation so that the two cities several Chi-square tests to the data indicated could be compared with each other and with that the sex differences in these frequencies populations outside Japan; another was to were significant beyond the .05 confidence assess the possible effects of different degrees limits only when between-sex associations were of inbreeding. Yet another was to compare the examined. There seems to be no general pattern findings from the hand-wrist data with analyses for a significant association between the pres- of other areas of the body. ence of one or more variations with chronologi- It soon became apparent during analysis of cal age, consanguinity status, or city. the X-ray films that many showed "anoma- These and other results suggest that the pres- lous" or "variant" centers. My interest in the ence of a skeletal hand-wrist variation is associ- frequencies and distributions of the variations ated with the sex of the child, but it does not increased as I continued the scoring and analy- suffice to relegate the presence or absence of a sis of the films. Hence the focus of this paper is bony variation to simple sex influence alone. not on select areas of skeletal maturation in Similar data for some of the variations are preadolescent children, that is, on human known for other racial samples of both sexes, growth and development, nor with classifica- although the frequency of some of the observed tions of skeletal variations (gross skeletal path- variations may be higher among the Japanese ology). Rather, it is an attempt to present the than is the case for some other gene pools. result of investigations of alternative interpreta- Whether suppression or inhibition in the tions of skeletal hand-wrist variations in pre- mineralization of a particular cartilaginous area adolescent children within a particular gene is temporary cannot be determined here. The pool. literature favors the view that the child's nutri- The use of the term "variation" rather than tion, unless severely distrubed, does not seri- "anomaly" stems from the fact that during the ously affect the growth pattern and Japanese analysis it became apparent that these data data do not suggest that there were substantial included a comparatively high frequency of nutritional differences (Schull and Neel, '65). variations. The number of these variations was Thus, the basic question explored is the sufficiently large to justify abandoning the nature and distribution of skeletal hand-wrist term "anomaly", especially since almost all variations, with a resulting hypothesis that such the children in the sample were medically variations are associated with "short-cuts" "normal". Some of the variations are identifi- which aid the female skeleton in development able as medically pathological "anoma1ies", yet toward the size and shape necessary for there are twelve major variations that are rela- puberty. Evidence from other areas of the 28 C. E. EYMAN skeleton seems to support this hypothesis, in The open end of the notch is almost always part, in that "anomalies," "variations," "dis- oriented towards the shaft ulnar surface. crete traits'' of certain skeletal regions tend to Several instances were seen of two metacarpal show stronger female than male associations. shafts being affected. This involves 251 Examples are, metopism (Woo, '49); metopism children. and occipital wormiana (Torgerson, '5 1a, b); Variation 4: A "pseudofracture" distributed shortened toe digits (Venning, '56b, '61b); through all metacarpals, the fifth proximal "missing" toe phalanges (Garn, Rohmann, and phalanx, the second through the fifth middle Silverman, '65 ;Venning, '54, '56a); septa1 aper- phalanges and the first and fourth distal tures of the olecranon fossa (Anderson, '62, phalanges. Not noted in multiples for any given '63); shortened fifth middle phalanx of the shaft but occasionally noted to occur simul- hand (Garn, Fels, and Israel, '67; Herzog, '67); taneously in two or more phalanges or meta- and various cranial discrete traits (Ossenberg, carpals. Seventy-three children show this. '70; Corruccini, '74). Variation 5: A shaft dysplasia, present in the From the hand-wrist X-rays of 3,823 chil- metacarpals and middle and distal phalanges. dren (a 50% random sample), l 12,867 centers Shaft dysplasia is most frequent in the middle were examined for variations and 57,345 phalanges, especially that of the fifth digit. maturity scores were recorded. There are 2,477 Although by itself this category is not the most "variant" children in the total sample, and frequent kind of variation in the sample, it is a 6,230 "missing" centers. component of the most common kind of com- X-ray films were read using the Greulich and posite variation. Basically, this variation in Pyle ('59) atlas. Maturity scores were recorded combination with other kinds constitutes what only for the distal ulna and the distal radius, the has been described as "Geisha finger;" radial carpals, and the metacarpals. However, the flexion of the fifth middle phalanx gives that entire hand and wrist were assessed for the digit the appearance of a radial deviation, a presence of variations and "missing" centers. condition seen in some Japanese tapestries The sample involved ranking according to city, depicting Geishas (J. V. Nee1 and W. J. Schull, the degree of inbreeding, age in months, skele- personal communication). This term seems tal age, sex and presence of a variation. Further even more apt in that variations involving radial details of the design including controls for flexion, especially of the fifth middle phalanx, observer bias are available in the original study are more frequent in females than in males. (Eyman, '70). There are 166 children with this variation. Twelve major variations are described Variation 6: Essentially, this is a reciprocal below. A thirteenth observation was made in shaping of the phalangeal epiphysis and the connection with the presence of one or more diaphysis so that the epiphysis takes on a rather variations. If an ossification center or epiphysis characteristic "cone shape." This kind of varia- was "missing" it was not considered to repre- tion is the most common one found in the sent a variation but was recorded only for sta- overall Hiroshima-Nagasaki sample; by far the tistical comparison with similar centers for greatest number of occurrences are associated those children who manifested no variations with the fifth middle phalanx, rather closely and were thus classed as "normal." followed by the first distal phalanx. Cone- Variation 1: Presence of a single "growth shaped epiphyses are much more common in scar" or "Harris's line" (Harris, '3 1). This is the females than in the males, and are often located at the distal end of the radius some- associated with shaft dysplasia and other condi- where proximal to the diaphysis. One hundred tions; but like shaft dysplasia and other con- sixty-three children show this. ditions the cone shaping can occur as a discrete Variation 2: The presence of multiple variation. Also like shaft dysplasia, the cone "growth scars" or "Harris's lines." This involves shaping can occur in two or more digits simul- 10 1 children. taneously. When cone shaping was noted, it was Variation 3: "Notches" or "pseudoepiphy- easily recognizable in that the phalangeal proxi- ses" noted for all metacarpals, and recorded for mal diaphyseal surface was indented in a vault- the second middle phalanx in two instances. like appearance with the apex oriented towards JAPANESE HAND-WRIST OSSIFICIATON 29 the distal end of the phalanx. Usually, but not When present in the metacarpals and phalanges, always, this diaphyseal shape was accompanied the entire epiphysis is usually affected and is by reciprocal shaping of the diaphyseal surface markedly radiopaque. One or two children have of the adjacent epiphysis. It was often apparent small circular "stippled" (cf. Fairbank, '5 1) that the diaphyseal surface of the epiphysis had areas in the shaft diaphyseal region. Twenty- actually penetrated the diaphysis. Occasionally seven children manifest this variation. the apex of the cone was slightly to the radial Variation 10: Restricted to epiphyses and side of the midline, less often it was slightly to denoting "premature" fusion as defined by the ulnar side. Six hundred forty-six children Greulich and Pyle's ('59) standards. It rarely show this. occurred at the distal radial and ulnar epiphy- Variation 7: Almost totally restricted, ses; it is most frequently associated with the "capitate indentation" or the "lateral inter- fifth middle and the first distal phalanges. As carpal space" (OYRahilly,'53a, p. 401 ; fig. 1, previously mentioned premature fusion is asso- p. 403), found on the radial surface of the ciated with reciprocal or cone shaping of the capitate in its middle or distal portion. In the epiphysis and quite often with the diaphysis; two instances where the indentation occurred the overall impression is that the more pro- on the hamate ("medial intercarpal space," nounced the cone shaping, the more common ibid. and p. 408) the location was approxi- the simultaneous occurrence of premature mately the same but on the radial surface of the fusion. (Present in 26 children.) hamate. The radial tubercle often appears im- Variation II: Combines the presence of mediately above the indentation, but some- shaft dysplasia (Variation 5) and cone shaping times below the indentation or is absent. When (Variation 6) for the same short bone, appears present, the indentation was marked and much more frequently at the fifth middle definite with a semilunar outline. This is seen in phalanx than at any other. This variation 4 15 children. exemplifies the "Geisha finger" even more than Variation 8: A "secondary" or "multiple" does Variation 5 alone, and is more frequent ossification center, usually in the carpus. The among females. When Variation 11 was per- latitudinal nature of the study militates against ceived it was noted that both the radial shaft determining whether the ossification center is a dysplasia and the reciprocal cone shaping were multiple one or a true accessory bone in most usually quite marked. This variation was noted instances. Whether such secondary ossicles for 23 3 children. represent nucleation of the same cartilaginous Variation 12: Another combination of other precursor of a carpal bone or a metacarpal or variations already discussed. It connotes the phalangeal epiphysis in several areas, or simultaneous presence of premature fusion whether some other situation obtains, cannot (Variation 1O), radial shaft dysplasia (Varia- be answered since only one hand-wrist X-ray tion 5), and reciprocal cone shaping (Varia- was obtained for each child. In those cases tion 6). The radiographic appearance of this where the carpal ossicle did seem to represent a condition is merely a combination of those true accessory bone by reference to OYRahilly given for the three component variations. (88 ('53b, '57), it was so recorded. When present, children.) the "secondary" or "multiple" centers Other Combinations: Four other rather in- appeared to be definitely separated from the frequent combinations of variations were remainder of the bone (or other ossicles of the noted. These conditions were generally diffi- immediate area). This condition was noted in cult to assess from the radiographs and may 25 7 children. very well represent deviations from the kinds of Variation 9: A radiographically opaque non- variations discussed above. All of these are con- linear ovoid area quite separate from the radio- fined to the metacarpals and phalanges except graphic appearance of a nutrient foramen in a for one (Variation 7-8). Variation 5-10 (3 chil- phalangeal or metacarpal shaft, in that it dren) denotes radial shaft dysplasia and pre- appears as a very radiopaque area. Generally, it mature fusion, differing from Variations 11 is associated with the short bones, but there are and 12 in that no reciprocal cone shaping was two instances of its presence in the carpus. noted. Variation 6-1 0 (24 children) is the most 30 C. E. EYMAN

frequent of this group of combinations: it con- THE LITERATURE notes the simultaneous occurrence of recipro- In at least one area of the literature, the dual cal cone shaping with "premature" fusion, and aspect of sex and maturation in children's skele- differs from other combinations of variations in tons has not been investigated. that there is no radial shaft dysplasia. Variation Almost no attention has been paid to possi- 3-4 (3 children) indicates the combination of a ble effects of sex and maturation upon the pseudoepiphysis and a pseudofracture. Varia- appearance of skeletal variations in the hand tion 7-8 (one child) combines a radial capitate and wrist. Even those dealing with sex differ- indentation ("lateral intercarpal space") with ences in the skeleton, primarily with hand and an accessory ossicle; whether this represents a wrist variations, do not inquire into the prob- true os centrale or not (cf. 07Rahilly, '53a; lem of the meaning of skeletal variations with Virchow, '29) is unknown. sex and skeletal maturation. "Missing" Centers: A center was considered This paper therefore attempts to demon- as "missing" when it could not be perceived on strate that variations in the skeletal hand and the hand-wrist X-ray, not merely when the wrist of some healthy Japanese children reflect maturity status of other centers indicated that the interaction of sex and skeletal maturation. it should be present according to the Greulich and Pyle ('59) standards. These were recorded STATISTICAL ANALYSES for both "normal" and "variant" children and involve 6,230 centers. The overwhelming Statistical analyses included R-mode and majority of "missing" centers are associated Q-mode factor analyses, Chi-squares, and with the carpus; it is not surprising that the Pearsonian product-moment paired correlation majority of "missing" centers are associated coefficients. During the early stages of data with the triquetral, due to its normal late reduction, six groups of "variant" children appearance (Greulich and Pyle, '59; Johnston emerged. They were determined by relatively and Jahina, '65; Johnston, Whitehouse and high frequencies of variations associated with Hertzog, '68). The possible association between certain ossification centers singly and second- "missing" centers and the occurrence of varia- arily in combination with others. Therefore, tions will be discussed later; for the present, these anatomical regions are not mutually ex- suffice it to say that aside from the more stan- clusive in that a "variant" group primarily dard comments upon "missing" centers, there defined by variations at center 3 (capitate) can is also a growing body of literature dealing with and does include children who also manifest their significance. variations at centers 12 (MC2) through The literature dealing with the possible etiol- 29 (DP5). The center 3-29 "variant" group, ogies and causes for all of these variations is however, would not include children with varia- voluminous, and a more detailed discussion of tions at the distal radial and/or ulnar epiphy- them can be found in Eyman ('70). seal centers (centers 1 and 2). So, each "variant" child was placed in only one "variant" HYPOTHESIS group, with the placement determined by the order number of the "first" of the 29 centers Following analysis of the frequencies of the manifesting a variation. variations previously described for the 3,823 These "variant" groups are: children with Japanese children comprising this study, it is variations at centers 1-29, primarily Varia- tentatively concluded that the higher fre- tions 1 and 2 at center 1, the distal radius and quency of these variations among the females its epiphysis; children with variations at centers reflects biochemical and derivative "shortcuts" 3-29, primarily Variation 7 at center 3, the which function to put the female skeleton as capitate; children with variations at centers rapidly as possible into the basic size and shape 12-29, primarily Variations 3 and 4 at center 12, necessary for childbearing - thus effectively the second metacarpal and its epiphysis; chil- lengthening the female reproductive span at the dren with variations at centers 24-29, primarily juvenile end, regardless of cultural practices Variations 5, 6, 1 1 and 1 2 at center 24, the militating against early maternity (Schull, fifth middle phalanx and its epiphysis; children Yanase and Nemoto, '62). JAPANESE HAND-WRIST OSSIFICATION 31 with variations at centers 25-29, again primarily Another significant association between sex Variations 5,6, 11 and 12 at center 25, the first and frequency of "missing" centers is found in distal phalanx and its epiphysis; and children those children with variations located at with variations at "other" centers but in no "other" centers, indicating that there are more clear anatomical pattern or "field." These six "missing" centers among the boys than among groups and others were individually analyzed the girls of both cities. Further analysis of this by most of the statistical applications em- group indicated that there were more boys of ployed. relatively young chronological age than girls. The higher counts of ccmissing'7centers for the Chi-square Analyses boys compared with the girls is consistent with Analysis of the Chi-square matrix pertinent the generalized slower maturation rate of to females from first cousin matings indicates males. Sampling error becomes the more attrac- that there are more "variant" girls from tive explanation for the high frequency of Hiroshima than from Nagasaki. This may re- "missing" centers among Hiroshima girls in the flect a possible consanguinity or between-city 24-29 "variant" group. Thus the Chi-square effect, but there is little other evidence to bear analyses reinforce the overall impression that out either of these conclusions. These overall the variations present in the sample reflect a sex findings suggest that the significant differences influence rather than either a maturational one between pooled inbred girls and between girls as such or a marked difference in consanguinity of first cousin matings may well be "signifi- effects. cant" results which occur by chance alone due Certainly, the Chi-square tests do not indi- to the number of Chi-square tests undertaken, cate any consistent between- or within-sex con- enhanced by the sampling distribution of sanguinity effect other than the ones men- female children of first cousin inbred status. tioned here. It is possible however that the As noted earlier, "missing" centers were not "control" children tend to obscure the inbred considered as variations per se. Twelve of the 89 children with respect to the presence or Chi-square tests involved testing for the signifi- absence of variations, and that the tests just cance of association between "missing" centers discussed are significant for this reason rather in "normal" and "variant" children, and four of than for spurious ones. these indicated significance beyond the 2.5% confidence limit. There are higher frequencies Ranked Maturity Scores of "missing" centers associated with boys than Garn, Rohmann and Blumenthal ("66, with girls in both the control and the inbred p. 1 13) have observed an ossification sequence categories of the sample. This is in keeping with polymorphism which basically refers to a rank- the differences in skeletal hand-wrist matura- ordering of the ossification center maturation tion rates between the sexes; since girls gener- statuses in a given skeletal area, such as the hand ally mature more rapidly than males, the boys or foot. Since they feel that this approach can generally exhibit more "missing" centers than be applied to latitudinal data, the sample was do girls. ranked according to the fifteen maturity scores There is a significant association between available for each child. A listing was obtained sex and frequency of "missing" centers in those of all children, arranged into the six "variant" children with variations located at centers groups described previously plus the "control" 24-29; analysis of the Chi-square matrix indi- and the "inbred" groups. While the Q-mode and cated that more Hiroshima girls had "missing" R-mode factor analyses and their derivative in- centers than did Nagasaki girls or the boys from formation did not indicate any effect of con- both cities. Further analysis indicated a larger sanguinity status or presence or absence of vari- proportion of chronologically relatively young ations upon the maturity scores individually Hiroshima girls in this group; whether this is a and overall (as reflected in the skeletal age/ sampling error or not is difficult to resolve, but chronological age ratios), rank-ordering the no ancillary data are available to argue strongly maturity scores might show different ossifica- against this concept. tion sequence patterns between the sexes, 32 C. E. EYMAN between "variant" and "normal" children, of children indicated that there is a preponder- between control and inbred children, and/or ance of "fast" (index equal to or greater than between the six "variant" groups. 1.000) or "slow" (index less than 1.000) matur- As Garn, Rohmann and Blumenthal ('66) ing children, I am satisfied that the maturation have indicated, more indeterminate ossification rate, as based upon the centers scored, is not sequences are expectable among children with associated with the presence or absence of vari- rapid than with slow ossification rates; thus less ations by location or kind. There is no differ- variable ossification sequences are associated ence between "normal" and "variant" children with girls much more frequently than with in this light, and there seems to be no difference boys, since the former mature more rapidly between controls and the various inbred skeletally than do the latter. The accuracy of groups. Finally, there seems to be no apparent this observation was apparent after the fifteen relationship between the skeletal/chronological maturity scores had been ranked in terms of age ratio and the frequency of "missing" each other and the results examined. The girls centers for either "normal" or "variant" chil- in the sample manifested a more invariate rank- dren. This last set of inferences is in accord with ing difference between one center and the next the linear regression models for half the than the boys. Hiroshima sample as discussed by Schull and To determine if there was any definite pat- Nee1 ('65). tern of ossification sequences after the fifteen maturity scores were ranked, the data were Consanguinity Effects analysed again according to ossification Consanguinity status for the various groups sequences within each of the groups. Repeated of "normal" and "variant" children seem to scanning of the transformed data indicated that have no clear or prominent influence upon the the groups could not be distinguished from skeletal age or the presence or absence of varia- each other on the basis of ossification pattern, tions. Again, the consanguinity effect upon sex, city, inbred or control status, or chrono- hand-wrist skeletal maturation was not marked logical and skeletal age. Thus, there appears to in the Hiroshima sample reported upon by be no ossification sequence effect which can be Schull and Nee1 ('65), although an inbreeding linked with anatomic location of variations, effect may have been masked through assorta- kinds of variations, or with lack of variations. tive matings of control parents. There seems to be no preponderance of Skeletal/Chronological Age Ratios "control" or of any category of inbred children This consists of summing the fifteen matu- within any of the six "variant" groups, so the rity scores for each child (the "skeletal age") consanguinity status of the child does not seem and then dividing this total by the child's chron- to be a factor here. Otherwise, it would have ological age in months. These ratios are admit- been expected not only in different frequencies tedly crude since only about half of the total of such categories within the "variant" groups, maturity indicators entered the computation, but also in the Q-mode and R-mode factor and these were the ossification centers which analysis distributions of all "normal" and all are considered by many workers to be the least "variant" children. The numbers of control and reliable. However, the purpose of this ratio was inbred "normal" and "variant" children do not merely to give a rough estimate of whether a reflect any marked differences for these con- child was a "fast" or a "slow" maturer as judged sanguinity categories, nor indicate marked con- by these fifteen carpal and metacarpal ossifica- sanguinity effects between kinds of variations. tion centers. The only consanguinity effects which seem Inspection of the distribution of the skele- apparent are associated with the Chi-square tal/chronological age ratios indicated no analyses. These puzzling but significant Chi- marked effect upon the kind or distribution square associations were attributed to the possi- of any of the variations. The skeletal/ bility of sampling error within the distributions chronological age ratio did seem to be asso- of the matrices, also suggesting that the signifi- ciated with a "metacarpal - early appearing cance of some of these tests might be purely carpal" factor. Since none of the eight groups chance. JAPANESE HAND-WRIST OSSIFICATION 33

The contingency coefficient C (Siegel, '56), Numerous longitudinal studies of hand-wrist as well as the coefficient of association phi skeletal maturation have provided ample evi- (Spaulding, '60) were used to assess the dence that girls mature more rapidly during strength of the distributions within the pre-adolescence than do boys. This observation matrices for these five enigmatic and significant is in keeping with the Hiroshima-Nagasaki Chi-square tests. The contingency coefficients sample, for when the skeletal/chronological age for tests indicated that the strengths of these ratios are examined, by and large the girls are Chi-square associations were quite low. Simi- "faster" skeletal hand-wrist maturers than the larly, the phi values for these tests were quite boys although this is less clear in the distribu- low. In view of the results obtained from the tion of children by sex in the Q-mode and contingency coefficients and the phi values, R-mode factor analyses. Presumably this is a plus the overwhelming lack of consanguinity function of the expected more rapid growth effects from other analyses of the data, it seems rate of the bones involved in a "metacarpal - that there is no appreciable consanguinity early appearing carpal" factor as I have inter- factor operative upon the presence, absence, preted it. and/or distribution of the variations within the sample in spite of these five Chi-square tests. Normal and Variant Frequencies of "Missing" Centers City Effects There are more "missing" centers among the Any differences between Nagasaki and "normal" than among the "variant" boys, the Hiroshima are slight and are not statistically frequencies being approximately equal be- significant. The only apparent city differences tween the two cities. Since boys generally may be those indirectly indicated in certain of mature less rapidly than girls, not surprisingly the Chi-square matrices, where the major differ- there are higher frequencies of "missing" cen- ence is attributable to the number of "variant" ters associated with boys than with girls. How- and "normal" first cousin parentage girls in the ever, there are more "missing" centers among two cities, specifically the number of "variant" "variant" than among "normal" girls. The dif- first cousin parentage Hiroshima girls. This situ- ference in frequencies of "missing" centers is ation is best viewed as one of sampling error or relatively marked between the Hiroshima and of "chance" significance. Nagasaki girls, and it is the "variant" Hiroshima For "missing" centers, as already noted girls who have the highest frequencies of these there were more young Hiroshima girls than centers for all girls. boys and Nagasaki children. Hence this is attri- The frequencies of "missing" centers are not buted to sampling error also and it is concluded very different between the "normal" and "vari- that no marked between-city effect probably ant" children who comprise the sample. There exists. It is also worth noting that Schull and are 3,414 "missing" centers recorded for the Nee1 ('65) could not demonstrate any meaning- Hiroshima sample and only 2,8 16 such centers ful differences between the two cities in terms for the Nagasaki sample. This primarily reflects of their inbreeding patterns, anthropometric the fact that there are slightly more younger analysis, or for the overall analysis. children in the Hiroshima sample. There are more "missing" centers associated Sex Effects with "normal" males and females of both cities It thus appears that the only significant asso- than there are with "variant" children. There ciations between the presence or absence of are also more "missing" capitates among "vari- variations obtain for the sex of the child, with ant" children than among "normal" ones, pri- the few possible exceptions which have already marily young males, perhaps due to the slower been noted. Overall, females manifest the larger skeletal maturation of the hand and wrist number of variations, and generally females are exhibited by pre-adolescent boys. It is clear associated with the larger number of specific that "missing" centers in the sample are re- kinds of variations. stricted to the distal ulnar epiphysis, the carpals, and the first metacarpal epiphysis. 34 C. E. EYMAN

From highest to lowest frequencies, the centers there are more "variant" girls than there are involved are : pisiform, distal ulnar epiphysis, "variant" boys, the higher frequencies of such scaphoid, trapezium, trapezoid, lunate, tri- "missing" centers in "variant" girls as opposed quetral, capitate, first metacarpal epiphysis, to "normal" girls may be a reflection of this. hamate, and the fifth middle and first distal Nevertheless, it is difficult not to believe that epiphyses. This ordering is almost identical for the more numerous "missing" centers asso- each "variant" group. This suggests that ana- ciated with "variant" girls is in fact indicative of tomical concentrations of "missing" centers are the skeletal economy in mineralization asso- not influenced by the anatomical locations of ciated with girls in general. Here, it should be variations, at least in any patent manner. remembered that the frequencies of "missing" The frequencies of "missing" centers are centers indicate actual counts, not the number essentially the same for "normal" Hiroshima of affected children, since many children had and Nagasaki males and are not very different instances of multiple "missing" centers. Due to between "variant" boys from these cities, al- the cross-sectional nature of the data it is im- though in the former city the frequency is possible to say when these "missing" centers higher for "variants." The frequencies of ossified, if in fact all of them did. "missing" centers among "normal" females are higher for Hiroshima than they are for Frequencies of Variations Within Sexes Nagasaki; the "variant" girls indicate nearly Males: Table 1 gives the percentage frequen- twice as many "missing" centers for Hiroshima cies of these variations and their rank orders. It as there are for Nagasaki. The higher frequen- indicates that there are more "variant" cies of these centers among the Hiroshima chil- Nagasaki males than "variant" Hiroshima dren seem to be due to the sample from this city males, although there are fewer "missing" containing more younger-aged children than centers in the Nagasaki sample than in the does the Nagasaki sample. Hiroshima one. The probable explanation for It is interesting that there are more the between-city difference with respect to "missing" centers among the "normal" boys "missing" center frequencies for "variants" has than among the "variant" ones, while there are already been discussed. Pooling the males with more "missing" centers among the "variant" respect to city, Table 1 also indicates that the girls than there are for the "normal" girls. Since frequencies of different kinds of variations the boys' hand-wrist skeletons are known to ranked from highest to lowest are as follows: 6, mature less rapidly than the girls', one would cone shaped epiphysis; 7, lateral intercarpal expect more "missing" centers for boys than space; 3, pseudoepiphysis; 8, "multiple" or for girls; this is in fact true when "variant" and "secondary" ossification centers; 5, radial shaft "normal" children are lumped and the two dysplasia; 1 1, radial shaft dysplasia and cone sexes are compared. But why should this same shaped epiphysis; 1, single Harris's line; 2, mul- relationship not also obtain when the children tiple Harris's line; 4, pseudofracture; 12, pre- are divided into "normal" and "variant" chil- mature epiphyseal-diaphyseal fusion and radial dren within each sex? The skeletal age and shaft dysplasia with cone shaped epiphysis; hence the maturity scores do not seem to be 9, opaque area; 10, premature epiphyseal- affected by the presence or absence of varia- diaphyseal fusion; 6-1 0, cone shaped epiphysis tions; therefore, the frequencies of "missing" and premature epiphyseal-diaphyseal fusion; centers are not affected by the presence or 34, pseudoepiphysis and pseudofracture, and absence of variations, which has been borne out 7-8, lateral intercarpal space and a "multiple" by some of the Chi-square results. or "secondary" ossification center. Hence a sex effect seems operative in this Table 1 also shows that the Nagasaki boys case, at least for the "variant" females; and it have higher frequencies of certain kinds of vari- seems that the higher frequencies of "missing" ations than the Hiroshima boys, while the re- centers among "variant" as compared with verse is true for certain other kinds of varia- "normal" girls is further evidence for economy tions. Comparison between the Hiroshima and of mineralization of bone in girls as a group Nagasaki males of the frequencies of Varia- when considered against boys as a group. Since tion 8, the "multiple" or "secondary" ossifica- JAPANESE HAND-WRIST OSSIFICATION 35 tion centers, is somewhat intriguing. The higher of three simultaneously occurring variations. frequency obtains for Nagasaki, which is inter- The Hiroshima boys take the lead with regard esting in that there are more young Hiroshima to larger numbers of simultaneously occurring males in the overall sample. If these "multiple" variations. or "secondary" ossification centers were mere- Females: Table 1 indicates that although ly indicative of relatively late appearing centers there are a few more Hiroshima than Nagasaki beginning to ossify from several centers which girls, the overall frequencies are almost identi- would fuse together later, then it would be cal. When the girls are pooled by city, Table 1 expected that more of them would be asso- shows that the frequencies of kinds of varia- ciated with the Hiroshima males, since more of tions are as follows in order of decreasing fre- them are younger. Arimoto (1 95 2), Takahashi quency: 6, cone shaped epiphysis; 7, lateral (1966), and Schull and Nee1 (1965) do not intercarpal space; 1 1, radial shaft dysplasia and present data for prefecture differences between one cone shaped epiphysis; 8, "multiple" or the boys to nutrition in terms of frequencies of "secondary" ossification centers; 1, single "multiple" or "secondary" ossification centers. Harris's line; 3, pseudoepiphysis; 5, radial shaft Therefore, this kind of variation cannot be dysplasia; 12, premature epiphyseal-diaphyseal attributed only to nutrition, illness, sex differ- fusion with radial shaft dysplasia and cone ences in hand-wrist skeletal maturation, or to shaped epiphysis; 2, multiple Harris's lines; 4, skeletal age. pseudo-fracture; 6-1 0, cone shaped epiphysis The Nagasaki and Hiroshima males are about and premature epiphyseal-diaphyseal fusion; equal in terms of the number of single varia- 10, premature epiphyseal-diaphyseal fusion; 9, tions and similar in consanguinity distributions opaque area; 5-1 0, radial shaft dysplasia and for different kinds of variations. Somewhat premature epiphyseal-diaphyseal fusion; and more Nagasaki boys have two variations than 3-4, pseudoepiphysis and pseudofracture. Hiroshima boys. The boys from the two cities There are no occurrences of Variation 7-8 with- have about equal frequencies of the occurrence in the female sample, the lateral intercarpal

TABLE 1 Percentage frequencies of "variant" children by city, sex and kind of variation *

All Naga. Hiro. All Naga. Hiro. Variation d d d 9 9 9 Pooled 1 6.63 ( 7) 11.55 ( 4) 6.53 ( 5) 2.78 ( 6) 10.26 ( 4) 6.58 ( 7) 2 4.59 ( 8) 2.76 ( 9) 3.61 ( 9) 4.78 ( 5) 2.45 (11) 4.08 ( 8) 3 14.80 ( 3) 14.83 ( 3) 5.92 ( 6) 8.02 ( 4) 3.83 ( 7) 10.13 ( 4) 4 4.00 ( 9) 5.52 ( 8) 2.00 (10) 0.62 ( 8) 3.37 ( 8) 2.95 (10) 5 7.91 ( 5) 5.69 ( 7) 5.61 ( 7) 8.02 ( 4) 3.22 ( 9) 6.70 ( 6) 6 19.90 ( 1) 22.07 ( 1) 31.67 ( 1) 33.18 ( 1) 30.17 ( 1) 26.08 ( 1) 7 16.50 ( 2) 16.90 ( 2) 16.99 ( 2) 14.66 ( 2) 19.30 ( 2) 16.75 ( 2) 8 13.86 ( 4) 8.96 ( 5) 7.22 ( 4) 10.34 ( 3) 4.13 ( 6) 10.38 ( 3) 9 1.11 (11) 0.69 (12) 1.08 (13) 0.62 ( 8) 1.53 (12) 1.09 (11) 10 0.76 (12) 1.03 (11) 1.31 (12) 1.70 ( 7) 0.92 (13) 1-05 (12) 11 7.57 ( 6) 6.55 ( 6) 11.07 ( 3) 10.34 ( 3) 11.79 ( 3) 9.41 ( 5) 12 1.70 (10) 2.24 (10) 5.23 ( 8) 4.78 ( 5) 5.67 ( 5) 3.55 ( 9) 5- 10 - - 0.23 (i4) - 0.46 (14) 0.12 (14) 6- 10 0.51 (13) 1.03 (11) 1.38 (11) 0.15 ( 9) 2.60 (10) 0.97 (13) 3-4 0.08 (14) 0.17 (13) 0.15 (15) - 0.31 (15) 0.12 (14) 7-8 0.08 (14) - - - - 0.04 (15) Total 47.48 49.32 52.52 49.81 50.19 100.00 *Rank order position in brackets (1 = highest rank) 36 C. E. EYMAN

space and "multiple" or "secondary" ossifica- ters) account for 53.21% of the total kinds of tion centers. variations. Wnen Variations 3 (pseudoepiphy- Most marked between city differences in sis) and 11 (radial shaft dysplasia with cone frequencies occur for Variation 8, the "mul- shaped epiphysis) are included, these five kinds tiple" or "secondary" ossification centers; of variations account for 72.75% of the total approximately three times as many Nagasaki variations. Thus cone shaped epiphyses are the girls show this condition as do Hiroshima girls. primary kind of variation, intercarpal spaces Recall that this was also true for the boys, but secondary, "multiple" or "secondary" ossifica- that the Hiroshima overall sample contains tion centers third, and pseudoepiphyses are more young children than does the overall fourth. When combinations of variations are Nagasaki sample. Again this leads to the conclu- examined six groups are found to be involved. sion that these "multiple" or "secondary" ossi- The most common are premature epiphyseal- fication centers cannot be wholly attributed to diaphyseal fusion, radial shaft dysplasia, and the relatively young skeletal age of certain chil- cone shaped epiphyses alone and in combina- dren manifesting incomplete mineralization of tion with other variations (47.8 8 %). Radial the cartilaginous precursor through multiple shaft dysplasia and cone shaped epiphyses centers of ossification. alone and combined with other variations The Nagasaki and Hiroshima girls have about account for 42.19% of the variations. Cone equal frequencies of single variations. This also shaped epiphyses and premature epiphyseal- applies to the frequencies of two simultaneous diaphyseal fusions, alone and in combination occurrences. More Hiroshima than Nagasaki with yet other variations, comprise 28.10% of girls show three simultaneous variations. These the total variations. Pseudoepiphyses and samples have almost equal frequencies of girls pseudofractures comprise 13.20% of the total with four simultaneously occurring variations. variations. Harris's lines are responsible for 10.66% of the total, while radial shaft dys- Frequencies of Variations Between Sexes plasias and premature epiphyseal-diaphyseal Of the sample of 3,823 children, 2,477 fusions constitute 7.87% of the total. (64.79%) are "variant." It can be seen from Table 1 also shows that males have higher Table 1 that there are more "variant" children frequencies than females for certain kinds of by city and sex. The kinds of variations may be variations, and that the reverse is true for other ranked in order of decreasing frequency: 6, kinds of variations. Variations 6,7, and 8 (cone cone shaped epiphysis; 7, lateral intercarpal shaped epiphyses; lateral intercarpal space; space; 8, "multiple" or "secondary" ossifica- "multiple" or "secondary" ossification cen- tion centers; 3, pseudoepiphysis; 11, radial ters) contribute more than 70% of the variation shaft dysplasia and cone shaped epiphysis; 5, for both sexes. Variations 6 and 7 account for radial shaft dysplasia; 1, single Harris's line; 2, 36.40% of the male and 48.66% of the female multiple Harris's line; 12, premature epiphy- variation. For both sexes, Variation 6 is the seal-diaphyseal fusion with radial shaft major kind of variation followed by Varia- dysplasia and cone shaped epiphysis; 4, pseudo- tion 7, while Variation 8 is the fourth major fracture; 9, opaque area; 10, premature epiphy- contributor. The third and fifth kinds of varia- seal-diaphyseal fusion; 6-10, cone shaped epi- tions making up the remaining 30% are differ- physis and premature epiphyseal-diaphyseal ent between the sexes. Variation 3 or pseudo- fusion; 5-1 0 and 3-4, radial shaft dysplasia with epiphysis is the third major kind of variation in premature epiphyseal-diaphyseal fusion and males, but Variation 1 1, radial shaft dysplasia pseudoepiphysis with pseudofracture; and 7-8, with cone shaped epiphysis, is third among lateral intercarpal space and "multiple" or females. Variation 5 or radial shaft dysplasia is "secondary" ossification center. the fifth major kind of variation in males; Varia- To summarize the nature of the overall tion 1, single Harris's line, is the fifth major "variant" sample, Variations 6 (cone shaped kind of variation in girls. Females have almost epiphysis), 7 (lateral intercarpal space), and 8 twice as many cone shaped epiphyses as boys ("multiple" or "secondary" ossification cen- (Variation 6). The sexes have almost equal fre- JAPANESE HAND-WRIST OSSIFICATION 37 quencies of the lateral intercarpal space (Varia- tate; MC2; distal radius; DPI; lunate; MC5; tion 7). Not unexpectedly, the boys have al- distal ulna; trapezium; scaphoid; trapezoid; most twice as many "multiple" or "secondary" MC 1; MC3 and MP2; triquetral, pisiform, and ossification centers (Variation 8) as the girls. DP5; DP2; PP1; hamate, MC4, MP4, DP3 and The within-sex differences have already been DP4; PP2, PP3 and MP4. The sharp drop in commented upon. percentages between the first dist a1 phalangeal When different individual and combined epiphysis and the lunate is to be noted. No male variations of the same nature are compared has variations at the fourth or fifth proximal between sexes, males have over twice as many phalanges. involving pseudoepiphyses and pseudofractures The Nagasaki and Hiroshima male samples as females (Variations 3, 4, and 3-4). Boys have have certain different and identical positions a slightly higher frequency than girls of single for particular centers with reference to their and multiple occurrences of Harris's lines (Vari- percentage rank (Table 2). Both samples are ations 1 and 2), and slightly more combinations identical in the rank positions for MP5, capi- of radial shaft dysplasia and premature epiphy- tate, MC2, lunate, trapezoid, triquetral and seal-diaphyseal fusion (Variations 5, 10, and MC4. They differ in that the rank order of DP1 5-1 0). Girls lead boys in the frequencies of all and the distal radius are reversed; beyond this, remaining combinations, sometimes approach- the differences seem to follow no clear pattern ing quantities which are twice as high. The in the rank order positions of the remaining combinations involving the same kinds of varia- centers. tions which are higher in girls are: radial shaft As indicated earlier the "variant" children dysplasia and cone shaped epiphysis (Varia- were sub-divided into six groups based upon tions 5, 6 and 1 1); cone shaped epiphysis and anatomical regions of "variant" ossification premature epiphyseal-diaphyseal fusion (Varia- centers. These regions are not mutually exclu- tions 6, 10 and 6-10); and premature epiphy- sive in that a group having variations at centers seal-diaphyseal fusion, radial shaft dysplasia 1-29 (i.e., encompassing all centers) can and and cone shaped epiphyses (Variations 5,6, 10, does include children who also have variations 11, 12, 5-1 0 and 6-10). It seems obvious from at centers in another group (e.g., center 1 and this that while boys are more prone to pseudo- also centers 12-29). Nevertheless, each "vari- epiphyses and pseudofractures as combinations ant" child was placed in only one of these six than girls, for both the major kinds of variations groups. The groups were established on the involve shaft dysplasias, premature fusions, and basis of those centers which seemed to be the cone shaped epiphyses. Thus these three factors primary sites for variations such as the fifth are probably attributes of the Japanese gene middle phalanx (center 24, hence group 24-29 pool, although more intensely associated with with no variations patent for the first 23 girls than with boys. centers). Slightly more girls than boys have single and Comparison of the Nagasaki and Hiroshima double variations. Boys lead girls in the simul- boys in terms of "variant" centers indicates taneous occurrence of three variations, but that no city or inbreeding effects are apparent, these frequencies drop sharply from the single but the primary sites for variations are the dis- and double variation categories for both sexes tal radius, the capitate, the second metacarpal, (4.25% and 3.1 5%), where the frequencies for the fifth middle phalanx, and the first distal girls and boys were 5 1.78% and 5 1.04% and phalanx. The other affected centers reflect rela- 17.29% and 16.07% for the single and double tively small percentages of variations. categories. Females: Table 2 gives these data as percentages with rank order. Pooling the "variant" girls Frequencies of Affected Centers Within Sexes shows that almost half have variations at the Males: Table 2 gives the data as percentages fifth middle phalanx. The capitate is the next .nd by rank order. Ranked in descending order most frequently affected center. The remaining of "variant" center frequencies, the overall affected centers, ranked in decreasing percent- "variant" male sample is as follows: MPS; capi- age order, are as follows: DPl; distal radius; 38 C. E. EYMAN

MC2; lunate; distal ulna; MC5; MP2; MCl; centers, there are certain rank order similarities trapezium; MC3; pisiform; PP5; MP4; scaphoid, and differences. The fifth middle phalanx, capi- PP2, DP2 and DP5; trapezoid and MC4; PP3, tate, and second metacarpal occupy the same PP4, MP3 and DP3 ; and hamate, PP 1 and DP4. rank positions in both samples. Like the boys, There are no girls in the sample who have tri- the rank order positions of the first distal quetral variations. There is a sharp drop in phalanx and the distal radius are reversed be- percentages of variations between the fifth tween the two cities. The rank positions are middle phalanx, the capitate, the distal radius different for the remaining centers. Unlike the and the first distal phalanx, and the remaining "variant" males, the ordering of the "variant" centers. girls is identical according to these six arbitrary When the Hiroshima and Nagasaki "variant" groups except for the 1-29 and the "other" girls are compared according to decreasing fre- categories, which are reversed in position be- quencies of variations associated with specific tween the two female subgroups.

TABLE 2 Percentage frequencies of "variant" children by city, sex and center *

All Naga. Hiro. All Naga. Hiro. Center d d d 9 9 9 Pooled D. Radius 11.42 ( 4) 8.61 ( 5) 14.28 ( 4) 10.20 ( 4) 7.53 ( 4) 12.86 ( 3) 10.78 ( 3) D. Ulna 2.39 ( 8) 3.72 ( 7) 1.03 (10) 2.22 ( 7) 3.53 ( 6) 0.92 (10) 2.30 ( 7) Capitate 16.54 ( 2) 16.38 ( 2) 16.70 ( 2) 17.25 ( 2) 14.90 ( 2) 19.60 ( 2) 16.92 ( 2) Hamate 0.26 (17) 0.17 (16) 0.34 (14) 0.08 (19) - 0.15 (14) 0.16 (22) Triquetral 0.77 (14) 0.84 (12) 0.69 (12) - - - 0.36 (17) Pisiform 0.77 (14) 0.51 (14) 1.03 (10) 0.77 (13) 1.38 (10) 0.15 (14) 0.77 (14) Lunate 3.92 ( 6) 4.73 ( 6) 3.10 ( 6) 2.38 ( 6) 2.76 ( 7) 1.99 ( 6) 3.11 ( 6) Scaphoid 1.88 (10) 2.87 ( 8) 0.84 (11) 0.31 (16) 0.61 (12) - 1.05 (11) Trapezium 2.30 ( 9) 3.72 ( 7) 0.84 (11) 1.00 (11) 1.69 ( 9) 0.31 (13) 1.61 ( 9) Trapezoid 1.53 (11) 2.03 (10) 1.03 (10) 0.23 (17) - 0.46 (12) 0.85 (13) MC 1 1.11 (12) 0.51 (14) 1.72 ( 8) 1.07 (10) 0.46 (13) 1.68 ( 7) 1.09 (10) MC 2 14.41 ( 3) 14.36 ( 3) 14.46 ( 3) 5.14 ( 5) 6.61 ( 5) 3.68 ( 5) 9.53 ( 5) MC 3 0.94 (13) 0.68 (13) 1.20 ( 9) 0.92 (12) 0.31 (14) 1.53 ( 8) 0.93 (12) MC 4 0.26 (17) 0.34 (15) 0.17 (15) 0.23 (17) 0.15 (15) 0.31 (13) 0.24 (20) MC 5 2.56 ( 7) 2.36 ( 9) 2.75 ( 7) 1.61 ( 8) 2.00 ( 8) 1.22 ( 9) 2.06 ( 8) PP 1 0.60 (16) 0.51 (14) 0.69 (12) 0.08 (19) - 0.15 (14) 0.32 (18) PP 2 0.17 (18) - 0.34 (1.4) 0.31 (16) 0.31 (14) 0.31 (13) 0.24 (20) PP 3 0.17 (18) 0.17 (16) 0.17 (15) 0.15 (18) 0.15 (15) 0.15 (14) 0.16 (22) PP 4 - - - 0.15 (18) - 0.31 (13) 0.08 (23) PP 5 - - - 0.54 (14) 0.31 (14) 0.76 (11) 0.28 (19) MP 2 0.94 (13) 0.68 (13) 1.20 ( 9) 1.23 ( 9) 0.77 (11) 1.68 ( 7) 1.09 (10) MP 3 0.17 (18) - 0.34 (14) 0.15 (18) 0.15 (15) 0.15 (14) 0.16 (22) MP 4 0.26 (17) - 0.51 (13) 0.38 (15) 0.61 (12) 0.15 (14) 0.32 (18) MP 5 24.04 ( 1) 23.99 ( 1) 24.20 ( 1) 42.02 ( 1) 43.01 ( 1) 41.04 ( 1) 33.51 ( 1) DP 1 10.66 ( 5) 10.81 ( 4) 10.50 ( 5) 10.74 ( 3) 12.14 ( 3) 9.34 ( 4) 10.70 ( 4) DP 2 0.68 (15) 0.51 (14) 0.84 (11) 0.31 (16) 0.15 (15) 0.46 (12) 0.48 (16) DP 3 0.26 (17) - 0.51 (13) 0.15 (18) - 0.31 (13) 0.20 (21) DP 4 0.26 (17) 0.17 (16) 0.34 (14) 0.08 (19) 0.15 (15) - 0.16 (22) DP 5 0.77 (14) 1.35 (11) 0.17 (15) 0.31 (16) 0.31 (14) 0.31 (13) 0.52 (15) *Rank order position in brackets (1 = highest rank) JAPANESE HAND-WRIST OSSIFICATION 39

As with the boys, these comparisons of the order positions for the fifth middle phalanx, girls do not indicate that any city or con- the capitate, the distal radius, the lunate, the sanguinity effects are operative with regard to fourth metacarpal and the third proximal "variant" ossification centers. The girls are also phalanx. The rank order positions are reversed like the boys in that these "variant" centers are between the sexes for the second metacarpal essentially the same, except that the second and the first distal phalanx (third in boys, fifth metacarpal is included in the male category and in girls for the second metacarpal; third in girls the rank order sequence of the remaining four and fifth in boys for the first distal phalanx), centers differs between the sexes. Thus the four and this is also the case for the fifth metacarpal primary "variant" centers for the girls are the (seventh in boys, eighth in girls) and the distal fifth middle phalanx, the capitate, the first ulna (seventh in girls, eighth in boys). The distal phalanx, and the distal radius. ordered positions of the "variant" centers become more dissimilar between the sexes Frequencies of Affected Centers beyond these. Between Sexes The 1,802 "variant" children represent al- Since there seem to be no major city or most half (47.14%) of the sample. For both consanguinity effects which influence the fre- sexes, the two "variant" groups with the high- quencies of "variant" centers, it is worthwhile est frequencies are the 24-29 and 3-29 to consider a between-sex comparison of the categories. The frequencies for the other groups "variant" children. The percentages are pre- do not follow the same rank order between the sented in Table 2. sexes. A total of 2,477 variations are present with- The control and inbred boys in the "variant" in the sample; 1,173 (47.36%) are associated group represent 45.61% and 54.39% of the with boys and 1,304 (52.64%) with girls. Com- total male "variant" sample. The control and bining "variant" boys and girls, it is clear that inbred girls constitute 44.77% and 55.23% of all centers have one or more variations. Ranking the "variant" female sample. the overall "variant" centers by decreasing fre- For each of the six groups of "variant" cen- quencies (Table 2), the following order results: ters within sexes and by consanguinity status, MP5, capitate; distal radius; DPl ; MC2; lunate; the majority of boys and girls constituting the distal ulna; MC5; trapezium; MCl and MP2; 24-29 category are inbred. For the 25-29 and scaphoid; MC3; trapezoid; pisiform; DP5 ; DP2; 1-29 groups, the control females and the inbred triquetral; PPl and MP4; PP5 ; MC4 and PP2; males constitute the majority. The 12-29 and DP3 ;hamate, PP3, MP3 and DP4; PP4. "other" groups are similar in that the majority Certain "variant" centers have greater fre- of children in them are control males and quencies in boys than in girls; the reverse situa- females. The 3-29 category is primarily com- tion obtains for certain other centers. Table 2 posed of control males and inbred females. shows that for both sexes, the fifth middle In sum, it appears that neither city of resi- phalanx is the most frequently affected center. dence nor consanguinity affect the distribution Then there is a sharp .decrease in percentage of variations within the hand and wrist to any frequency for both sexes, with the capitate notable degree. While some sex differences being the second most frequently affected cen- obviously obtain in frequencies of affected center. Here the similarity between the sexes ends; ters, it is nevertheless clear that there is no in decreasing order, males show the second well-defined sex difference in patterns of cen- metacarpal, the distal radius and the first distal ters affected by certain kinds of variations. phalangeal centers affected by variations on the order of 10% or more, while for females the Skeletal/Chronological Age Ratios first distal phalanx and the distal radius have The overall sample reveals no association frequencies of variations of 10%or more. between chronological age and the presence or Comparison of the rank ordering between absence of variations, nor is there any patent sexes again indicates certain similarities and association between the skeletal/chronological dissimilarities. Both sexes have the same rank age ratio and the presence or absence of varia- 40 C. E. EYMAN tions relative to sex, city, or consanguinity. The kinds of variations involved with these se- sub-samples and the overall sample similarly quences are Variation 6, "cone-shaped" epi- indicate that no chronological age effect is sig- physis, for both males and females; Vari- nificantly operative upon any of ten anthro- ation 12, premature epiphyseal-diaphyseal pometric measurements during Schull and fusion with radial shaft dysplasia and "cone- Neel's (1 965) analysis; but when they analyzed shaped" epiphysis, for both boys and girls; and part of the Hiroshima hand-wrist maturity Variation 7, the lateral intercarpal space, for scores, they found a significant chronological Hiroshima girls. The two remaining "deviant" age effect, which indicates that at least the sequences involve complexes of variations; the children subjected to their statistical analysis "shaft dysplasia - cone epiphysis" complex mature similarly to other prepubescent chil- (Variations 5, 6, and 11) in Hiroshima boys, dren. and the "shaft dysplasia - premature fusion - Variation 8, the "multiple" or "secondary" cone epiphysis" complex (Variations 5, 6, 10, ossification center, is about twice as frequent in 11, 12,5-10,6-1O), also in Hiroshima boys. males as in females. Were this a function of a The frequencies for all variations by sex and delayed male maturation (later mineralization city indicate that about half the time those for of the cartilaginous model), then there should second cousins are greater than those for first have been a between-sex difference in the skele- cousins once removed. If the alleles associated tal/chronological age ratio since affected cen- with these variations are associated with the X ters would receive "zero" maturity scores com- chromosome, then they are not primarily pres- pared with unaffected centers. However, no sex ent in a hornozygous state. If they were, one differences were found for this variation and might expect the control category to have the skeletal/chronological age ratio. It became lower frequencies more often than the first even more apparent in the within-sex compari- cousin category at least, and this is not borne sons that Variation 8 is not associated with out by the evidence at hand. However, it should maturation rate; its frequency was slightly be recalled that the control children are the higher in the Nagasaki males than in the offspring of spouses who had no known genetic Hiroshima boys, and about three times higher relatedness, and since Japan is an inbreeding for the Nagasaki girls than the Hiroshima girls. isolate or has been until recently, the possibility Since more chronologically younger children of must be taken into account that the frequency both sexes obtain for Hiroshima than for of homozygous alleles is actually higher than Nagasaki, neither chronological age nor con- one might think. sanguinity status appear to be causative factors. Work has been done on gene involvement in Furthermore, no between-city nutritional dif- sex differences as seen in skeletal hand-wrist ferences are apparent which might be invoked maturation. If there are sex differences in those to explain these dissimilarities. factors coordinating skeletal growth and Therefore, differences in maturity scores or maturation, then the greater female variability in overall maturation rates do not have promi- in maturation rate shifts and ossification timing nent roles in the presence or absence of the may be associated with more female genotypes different kinds of variations within or between as represented by autosomal genes associated sexes. Maturation rate, especially as associated with the X chromosome. Whether the Y chro- with the child's sex, is an aspect to the meaning mosome does tend to retard skeletal matura- of variations; but it is not the only one. tion is a moot question. If the kinds of variations discussed here are associated with matura- DISCUSSION tion rates, it becomes difficult to understand There are seven deviations from the se- why the variation frequencies should not differ quence of decreasing frequencies given in more than they do between the sexes. Table 1. All involve the Hiroshima children; the Another problem here is the inbreeding sequence is first cousins, controls, then first effect in the sample. Regardless of the auto- cousins once removed followed by second soma1 genes associated with the X and Y chro- cousins or the reverse, and "other" inbreds. The mosomes and those directly and indirectly in- JAPANESE HANDWRIST OSSIFICATION 4 1

volved in producing the different variations, 38.35% in boys. These "complexes" indicate their frequencies might be higher in inbred than that while there is no definite sex-linkage, varia- in less inbred gene pools when data are com- tions of similar nature do indicate some sex pared for variation frequencies. influence and/or association when grouped If genes associated with the Y chromosome together. do tend to retard skeletal maturation while In addition one final and crude estimate of those associated with the X chromosome tend some alternate interpretations of these "male to produce "normal" or even "accelerated" complexes" and "female complexes" was maturation, one would expect the sex differ- attempted The overall "male complex" in- ence in maturation rates to be more marked volves seven discrete kinds of variations and than found for the Japanese sample. It is inter- three "complexes." From the literature, single esting to speculate whether the kinds of varia- and multiple Harris's lines may or may not be tions discussed herein do reflect a sex differ- associated with relatively slow skeletal matura- ence in hand-wrist maturation rates, albeit in- tion. Pseudoepiphyses, pseudofractures or directly and clouded by inbreeding. Such a sex notches, and "multiple" or "secondary" ossifi- difference in maturation rates, as these are com- cation centers would appear to be associated monly thought of, is not indicated by the with the relatively slower skeletal maturation variations. rate in males compared with females. Opaque Except for the "complexes" of variations, areas and radial shaft dysplasia do not appear to most of the treatment of the kinds of variations be associated with growth. The "pseudoepiphy- has involved dealing with each variation as a sis-pseudofracture complex" and the "Harris's discrete unit. However, it is quite possible that lines complex" seem to be associated with dif- one or more of the alleles associated with the ferential skeletal maturation rates between variations are pleiotropic and that many of the boys and girls; the "shaft dysplasia - premature variations are expressions of incomplete domi- fusion complex" does not. Hence, within the nance or incomplete recessiveness, or are main- overall "male complex," five of the seven single tained along with other characters having selec- kinds of variations and two of the three "com- tive advantage. plexes" seem to be associated with growth This possibly was evaluated crudely by factors. analyzing the percentage frequencies of each The overall "female complex" involves eight kind of variation between the sexes (Table 1). discrete kinds of variations and three "com- Those variations which had higher percentage plexes." Premature fusions and pseudoepiphy- frequencies in males than in females and the ses with pseudofractures are associated with reverse were separated and then put together as growth fairly clearly; premature fusions and "complexes." For males, one is the "pseudoepi- cone shaped epiphyses along with premature physis-pseudofracture complex" (Variations 3, fusions and radial shaft dysplasias, less clearly. 4, and 3-4); it has a frequency of 18.88% in Radial shaft dysplasia and cone shaped epiphy- boys and 8.07% in girls. Another is the "Harris's ses, radial shaft dysplasia with premature fusion lines complex" (Variations 1 and 2), with a fre- and cone shaped epiphyses, cone shaped epi- quency of 11.22% in boys and 10.14% in girls. physes, and the lateral intercarpal space are The "shaft dysplasia-premature fusion com- difficult to attribute to any appreciable skeletal plex" (Variations 5, 10 and 5-10) has a fre- hand-wrist maturation influence. The "pre- quency of 8.67%in boys and 7.1 5% in girls. The mature fusion - cone shaping complex" would "shaft dysplasia - cone shaping complex" seem to be partly influenced by a skeletal (Variations 5, 6 and 11) constitutes 48.35% of maturation factor, but this is not as clear for the the girls' and 35.3 8%of the boys' variation. The "shaft dysplasia - premature fusion - cone "premature fusion - cone shaping complex" shaping complex" nor especially for the "shaft (Variations 6, 10 and 6 -10) has a frequency of dysplasia - cone shaping complex." Thus 34.36% in girls and 21.17% in boys. The "pre- within the overall "female complex," only two mature fusion - shaft dysplasia - cone shaping (and at the most, four) of the eight discrete complex" (Variations 5, 6, 10, 1 1, 12, 5-1 0, kinds of variations may be attributed to skeletal and 6-1 0) has a frequency of 56.50%in girls and maturation factors, while only one of the three ''comp1exes" lends itself to that interpretation. 42 C. E. EYMAN

The kinds of variations and the "complexes" cance of pseudoepiphyses. Other investigators are by no means mutually exclusive between feel that pseudofractures and pseudoepiphyses the sexes. However, it does seem apparent that are associated with ligamentous stress pro- most of the overall "male complex" is asso- ducing delayed or inhibited ossification in the ciated with sexual dimorphism in relation to latter condition. If these investigators are cor- hand-wrist skeletal maturation; the overall rect in their interpretation, then the ligaments "female complex" cannot be similarly inter- and vascular supply to regions where pseudo- preted. The literature with regard to ossifica- epiphyses and notches appear would have to be tion sequence polymorphisms, sex ratios for positively charged in order to delay ossification various "anomalies," and the like, also fails to and there would have to be an absence of liga- demonstrate that such skeletal hand-wrist varia- mentous stress as well. This is the opposite of tions are due to nutritional differences, matura- the truth and argues for bone mineral tion rates, or illness susceptibility between the deposition, not its inhibition. sexes in any clear-cut fashion. Analysis of the It is further possible that vasodilation (with Japanese data has indicated that maturation an increase of negatively-charged ions, tending rates, inbreeding, or city of residence - along to increase ossification) is related to genetic with nutritional differences between the gene differences between human gene pools. It is pools and other data for the sample - can only interesting to speculate whether the piezoelec- be invoked with difficulty in interpreting the tric effect can be extended to differences be- meaning of these variations. Insofar as the data tween gene pools. Furthermore, the sequence and current research methods permit, the of electrical events in bone building and bone hypothesis of ccshortcuts"associated with get- destruction might be influenced by differences ting at least parts of the female skeleton into between different human gene pools in the the basic sizes and shapes necessary for child- anatomical orientation of ligaments and vascu- bearing cannot be demolished easily. It now lar systems. remains to investigate some possible mech- If the shape of the bone is related to the anisms whereby these shortcuts might be distribution of its electrical charges, then does a effected, and why. round bone (e.g., a carpal) have an electrical From the literature, it is apparent that some charge distribution different from a short bone but not all bony conditions are plastic re- (e.g., a metacarpal or a phalanx)? If so, it is sponses to biomechanical stresses. It is inter- possible that variations such as cone shaped esting to speculate as to how many bony epiphyses and the lateral intercarpal space are "anomalies" are not necessarily biomechanical influenced partly by the electrical charge dis- responses. Six come immediately to mind: devi- tributions in the affected bones. ations of the nasal suture to the left or to the It is also tempting to speculate that radial right of the sagittal plane; pterion shape; flexion, especially of the fifth middle phalanx, metopic suture; "wormian" bones; septa1 aper- is due to radial compressive and ulnar tensile tures of the olecranon fossa; relative index fin- stresses. In the X-rays showing this condition, ger length. The last four "anomalies" may be fine dense lines of bone mineral occurred at the sex-influenced (ex Woo, '49; Torgersen, '5 1a, site of radial (concave) flexion and extended b; Anderson, '63; Venning, '6lb). It appears fanwise into the phalangeal shaft toward the that certain bony joint surfaces shapes and ulnar (convex) surface. This is in keeping with synovial capsule tension are not necessarily bio- Glucksmann's ('42) observations. The question mechanical responses; hence the range of immediately arises as to why the fifth middle motion for certain joints. and to a lesser extent the first distal phalanges Similarly, bone bipartitions and accessory or should be subjected to flexions possibly asso- "multiple" bones and ossification centers do ciated with biomechanical stresses. not appear to be influenced substantially by It is possible that when partly flexed, strik- biomechanical factors. But what is their adap- ing the respective ulnar and radial surfaces of tive significance, if any? If pseudoepiphyses the little finger and the thumb upon hard sur- were due to mechnical stresses, one must again faces is partly responsible. When these two sur- speculate as to the possible adaptive signifi- faces are so struck, they might undergo tension JAPANESE HAND-WRIST OSSIFICATION 43 stress and the opposite surfaces would undergo The insights provided by biochemical inves- compressive stress. But such biomechanical tigations of bone are intriguing but they are also stresses would not constantly obtain; even if frustrating because of the extreme difficulty in they were responsible for the radial flexion of devising in vivo studies to help explain certain the fifth middle phalanx and the ulnar flexion phenomena on a macro-molecular level. There of the first distal phalanx, there should be no are methodological and procedural barriers appreciable frequency differences between associated with investigations of "how" and gene pools for this kind of variation nor should "why" phenomena such as wormian bones and the frequencies differ markedly between sexes. hand-wrist variations are produced. It might be possible that there are gene pool The literature clearly indicates that bone differences with regard to the proximal joint metabolism - whether it be concerned with surface shapes of one or both of these phalanges remodeling, pathology, or growth - is under and that the presence or absence of cone shaped fairly direct genetic control through various epiphyses is of importance also. The situation bone cells, with this genetic control mediated may be compounded by gene pool differences directly or indirectly by the composition of with regard to radial and ulnar synovial laxity at collagen, ground substance, extracellular fluid, these two joints. Kohler ('68, p. 76) calls this hormones, vascular supply, vitamins, and the "snapping finger" and noted that this condition like. This indicates that it is no longer sufficient is common in harp players, ball players, and in to refer obliquely to the "genetics" of bone, milkers of cows. and it also indicates that much future research The capitate lateral intercarpal space directed towards understanding human skeletal observed in the sample is equally puzzling, if variations, "anomalies" and pathologies must partly influenced by the bioelectric potential of concentrate upon a more detailed understand- bone. The anatomy of the capitate is such that ing of the genetic-metabolic processess involved it articulates with the scaphoid proximally and in their production. distally, and radially with the trapezoid by Certainly some of the literature dealing with means of a smooth flat facet. The interosseous the biochemistry of pathologies and variations intercarpal ligament is situated between these indicates that such studies are not impossible two articular surfaces on the radial side, attach- and that the classic phenomenological and mor- ing the capitate to the trapezoid. The lateral phological approaches to "e~plaining'~them intercarpal space, when present, is located in are no longer adequate. Nor is it sufficient any this region. If the interosseous intercarpal liga- more even to discuss bone pathologies in the ment were subjecting this region of the capitate "standard" medical literature if such discus- to ligamentous stress, an accumulation of sions and descriptions exclude what a certain negatively-charged ions and a buildup of bone pathology really means in terms of its produc- could be expected; however, the reverse is true, tion and subsequent changes. and so it does not seem that the lateral inter- Many non-serological, soft tissue pheno- carpal space can be interpreted wholly as a typic responses vary from one gene pool to plastic bone response to this ligament. another, and the genetic control seems to be Considerable research has been done on the clearly understood for many of them. If the biomechanical nature of bone metabolism, epiphyseal vascular supply is involved with modeling, growth, and crystal nucleation and cone shaped phalangeal epiphyses, then by inhibition. Baker and Angel ('65, pp. 105-106) extrapolation what are we to infer about them have stated the problem: when we know that there is gene pool variation Between the level of whole bones and the in autosomal genetic control of the mammary basic biochemistry and microscopic structure venous plexus vascular pattern, the palmaris there remains a gap . . . since we know very longus muscle, and the peroneous tertius little about the pathological or age variations in muscle (this last having a statistically significant structure and biochemistry between sections of sex difference at least among the Ramah the same bone. Navajo ;Spuhler, '5 1)? 44 C. E. EYMAN

It is possible that human gene pools differ in "variant" children (six separated according to terms of skeletal maturation and in steroid the ossification centers involved, plus all un- hormone production; it is also possible that related and all inbred children) with respect to male and female titration strengths of various the quarter-unit normalized standard deviation growth and steroid hormones differ from one distributions. Thus the "factors" involved are gene pool to another, while timing of onset, not different with respect to centers affected intensity, or duration need not be necessarily by variations or maturation rates, since the constant across gene pools. All of these factors "factors" are based upon the maturity scores. should be assessed, especially for their possible Within each group, the only clear distribution roles in producing skeletal "economy" within involves sex. From the R-mode factor analysis, and between gene pools. the child's sex is the major factor in the dis- It is possible that the rather high frequencies tributions; the presence or absence of varia- of some variations noted for the Japanese chil- tions, consanguinity status, or city of residence dren represent phenotypic expressions of do not markedly influence the maturity scores autosomal alleles in relatively high frequencies and the skeletal/chronological age ratios. due to inbreeding. But why certain bones such The Q-mode factor analyses of the eight as the captitate should be unossified in specific groups agree with the R-mode factor analyses areas does not seem to be a question answerable and again, sex has the most important effect on the strength of the data. upon the factor plane distributions and hence Some investigations have been carried out as upon the maturity scores and maturation rate to the possible effect of somatotype upon as assessed by the skeletal/chronological age skeletal maturation, but little if any literature ratios. deals with the possible association between The Pearsonian product-moment correla- somatotype and minor skeletal variations of the tion coefficients for the groups indicate that kinds considered here. The literature shows an the skeletal/chronological age ratio has a low even greater dearth of investigation into the correlation with the fifteen centers assessed for possible effects of skeletal maturation rates maturity status; especially the distal radial and upon hereditary mortality and longevity. The ulnar epiphyses, the triquetral, the pisiform, literature is intriguing but difficult to interpret and to a lesser extent the capitate and hamate. with respect to the possible effects of longevity This reflects the variability in carpal matura- and somatotype upon hand-wrist skeletal variation, well documented in the literature. The tions such as those discussed here. If there is a trapezium, trapezoid and distal radial epiphysis "skeleton-type" growth control mechanism have low correlations with the pisiform, con- and if different genotypes vary in response to sistent with its considerable ossification vari- common systemic agencies, these might also ability. Since these relationships are essentially partly account for the different frequencies of the same for each of the eight groups, it is fairly hand-wrist skeletal variations between gene clear that sex, consanguinity status, or the pres- pools and sexes. ence or absence of one or more variations do It is also possible that if the final adult male not affect the maturity score correlations; nor and female sizes and shapes reflect some genetic do they seem to affect the overall maturation control associated with sexual dimorphism rate as based upon the ossification centers. It is rather than with different maturation rates, also clear the the carpals and essentially the while shape differences reflect sex differences metacarpals tend to mature as an anatomical in certain periods of maturation; there might be "field ." a sexual dimorphism factor operating upon The communality scores for the eight groups hand-wrist variation frequencies and which is also indicate little between-group difference. affected by within- and between-sex variability For all groups, the communalities suggest the in specific maturation rates. patent existence of the same two "factors." One involves the distal ulnar epiphysis, tri- CONCLUSIONS quetral, lunate, scaphoid, trapezium and The R-mode factor analysis did not disclose trapezoid. The other involves distal radial epi- any marked differences between the groups of physis, capitate, hamate, second through the JAPANESE HAND-WRIST OSSIFICATION 45 fifth metacarpal epiphyses, and skeletal/ quences, in keeping with others' observations chronological age ratio. The pisiform and first attributing such lower female variability to metacarpal epiphysis are intermediate between their faster skeletal maturation rates. these two factors. Thus regardless of the pres- Because the eight groups were subdivided by ence or absence of variations, sex, consanguin- sex, consanguinity status, and (when present) ity status or city of residence, there appears to anatomical location of variations, analysis of be a "carpal" factor and a "metacarpal - early the ranked ossification sequences indicated appearing carpal" factor. The uniformity of the something else as well. No clear sequences are eight groups with respect to the components of apparent by sex, city, consanguinity status, these two factors indicates that maturation presence or absence of variations, chronological rates per se and consanguinity status do not age, or skeletal age. These findings also rein- affect the maturity scores of children who lack force the conclusion that the presence, kind or variations or who have them as specific centers. anatomical locations of variations or their Chi-square tests applied to the sample indi- absence are not associated with maturation cate that the few tests significant at or beyond rates or with ossification sequences. If a consan- the 5% confidence limit involved the sex of the guinity effect is present with respect to varia- child and the presence or absence of variations. tions, it certainly is not apparent with respect The few exceptions to this involved city and to the ossification sequence. consanguinity effects; but these results are Various investigators have found that the probably "significant" ones which have maturation rate seems to be associated with the occurred by chance alone, due to the number of presence or absence of certain variations dis- Chi-square tests performed. This conclusion is cussed here; others have noted inbreeding reinforced by the absence of similar findings effects upon anthropometrics. Maturation rates from other analyses of the data. as indicated by the skeletal/chronological age Other Chi-square tests indicate no signifi- ratios within each of the eight groups were cant association between frequencies of "miss- examined. One cannot conclude that matura- ing" centers and variations, providing ancillary tion rate is associated with the kind of anatomi- indications that maturation rate and probably cal location of variations or the lack thereof ossification sequence are neither affected by from this study. This analysis also indicated no nor associated with the presence or absence of clear difference between control and inbred variations. Those matrices involving "missing" children. There is no apparent association centers indicate more for males than females, between maturation rate and "missing" centers but this seems fairly in keeping with the overall for the "normal" and "variant" children. slower male maturation rates compared with Because no group has a preponderance of females. Other matrices involving frequencies "fast" or "slow' maturers based upon the of variations indicate that females have more skeletal/chronological age ratios and "missing" variations than males, again suggesting a sex centers, I cannot establish either an inbreeding effect. Interpretation of the Chi-square results effect or a presence-or-absence of variations repetitiously leads one to conclude that the effect upon maturity rate. Stated somewhat frequency of variations reflects a sex influence, differently, the maturation rate seems to have but not a consanguinity effect and almost no effect upon the presence or absence of surely not a maturation effect. variations. Because approximately half the children in There are some possible consanguinity and this sample reflect various degrees of inbreed- city effects upon skeletal age, the kind and ing, it is possible that some consanguinity effect location of variations, and "missing" centers, might be apparent in the ossification sequences apparent from a few Chi-square tests significant and that these sequences also might be asso- at or beyond the 5% confidence limit. I feel that ciated with the presence or absence or varia- these few significant effects are artifacts. If a tions. These influences were assessed by rank- consanguinity effect is operative, it is by no ing each of the eight groups according to the means clear from the data; the same applies order of ossification. Analyses of these rankings more strongly to possible city effects. Should suggested less variant female than male se- some such effects actually obtain, they should 46 C. E. EYMAN be interpreted as probably reflecting small dysplasias, premature epiphyseal-diaphyseal socioeconomic differences between cities and fusions, and cone shaped epiphyses; these com- consanguinity groups; this is in keeping with binations seem to apply more to girls than to Schull and Neel's ('65) conclusion for the over- boys. The majority of centers associated with all study. the majority of the kinds of variations are the The outcomes of the various statistical pro- same for both sexes; essentially fifth middle cedures applied to the data indicate fairly clear- phalanx, capitate, and first distal phalanx. The ly that there is some sex difference in the pres- sex differences in these percentages are not well ence or absence of variations. Girls have greater defined according to the kind and location of absolute counts of variations and generally the variations. Because city of residence and greater overall counts of specific kinds than consanguinity status do not seem to affect the boys. Perhaps this situation is associated with a anatomical distribution of the variations, the more rapid female skeletal maturation rate, major locations of the variations may reflect allied with "shortcuts" in the female skeleton one attribute of the Japanese gene pool. oriented toward developing it into basic adult Not having been able to establish associa- sizes and shapes. tions between chronological age and the skele- "Missing" centers seem to militate against tal/chronological age ratio with the presence this conclusion at first glance. There are more or absence of variations; or between "miss- "missing" centers for all boys than for all girls, ing" centers and maturation rate among to be expected with slower male maturation "normal" and "variant" groups do not indicate rates. But within-sex comparisons of "normal" any strong or clear effect upon the presence or and "variant" children present a somewhat con- absence of variations. Thus one cannot attrib- fusing picture. More "missing" centers were ute variations to maturation rate differences found for "normal" than for "variant" boys, either between or within the sexes, aside from the reverse for "variant" and "normal" girls. If noting that the overall maturation rates are greater numbers of "missing" centers indicate a generally faster for girls than for boys. slow skeletal maturation rate, then "normal" There seem to be some inbreeding effects males and "variant" females are slow maturers upon specific congenital malformation fre- while "variant" males and "normal" females quencies in Japanese infants, as well as a slight are fast maturers. This reversal in rates for retardation in hand-wrist maturation rates "normal" and "variant" children between the among part of the Hiroshima sample. Neverthe- sexes militates against a clear maturation rate less, the data do not indicate that inbreeding influence upon the presence or absence of varia- affects maturation rate, frequencies of overall tions, at least in any clear way. However, the variations, or of particular kinds. Yet if the reverse situation be tween sexes for "normal" variations are associated with sex, inbreeding and "variant" children reinforces my conclu- could increase the frequency of "variant" sions: that among the females at least, "missing" alleles which are presumably associated with centers are like the variations in reflecting a the X chromosome. This might increase such skeletal economy designed to orient skeletal allele frequencies among males, masking the sex development towards the basic adult hand- differences somewhat. Here one must ask what wrist size and shape, perhaps accompanying an the selective advantages of such variation- earlier onset of reproductive maturity. producing alleles might be within the Japanese Percentages of variations between sexes gene pool, aside from sexual dimorphism. I showed that for both, cone shaped epiphyses have few if any indications of how to answer (Variation 6) followed by the lateral inter- this. carpal space (Variation 7) and then by "multi- One conclusion seems fairly clear with ple" or "secondary" centers (Variation 8) were respect to the "meaning" of the different kinds the most common. Males seem to be more of variations discussed here: nutritional defi- prone to pseudo-epiphyses and pseudofractures ciencies cannot be invoked among the children than females, especially in combination. For comprising the sample. Several recent studies both sexes, the kinds of variations seem have demonstrated that the Japanese nutri- to involve major groupings of radial shaft tional status is quite high compared with other JAPANESE HAND-WRIST OSSIFICATION 47

Asian countries, and that Japan's post-war between-sex differences in variation frequen- nutritional status has increased steadily. Since cies compared with other gene pool samples. the great majority of the children were born Thus, compared to other samples, the after the end of World War 11, it is difficult to Japanese gene pool has higher allele frequencies conclude that nutritional deficiency is asso- responsible for the lateral intercarpal space, ciated with the variation frequencies. There cone shaped epiphyses, radial shaft dysplasia, may be some nutritional differences between and radial shaft dysplasia combined with cone the cities and consanguinity classes, but they shaped epiphyses. It appears intermediate to are not clearly apparent from the hand-wrist other pools with respect to the allele frequen- data. cies responsible for radial shaft dysplasia with The absence of a nutritional effect upon the premature epiphyseal-diaphyseal fusion and frequency of variations also manifests itself in cone shaped epiphyses. Comparatively, it has the comparative data for variations. The per- the lowest allele frequencies responsible for centage of single and multiple Harris's lines is single and multiple Harris's lines, pseudoepi- much lower among the Japanese sample than physes, notches or pseudofractures, and radial from some Caucasian samples; though the shaft dysplasia with premature epiphyseal- Japanese nutritionally lag behind U. S. Cau- diaphyseal fusion. The possible evolutionary casian standards, clearly a nutritional effect significance of these differences is not clear does not obtain. The same is true for lower to me. frequencies of notches or pseudofractures and Consider some possible effects of sex per se for pseudoepiphyses in the Japanese sample on the frequency and kinds of hand-wrist varia- compared with Caucasian samples. If these vari- tions. That the male skeleton is more suscep- ations are associated with comparatively tible to growth deterrents than the female "slow" hand-wrist maturation, then the skeleton; this shows up in higher notch and Japanese must mature more rapidly than U.S. pseudoepiphysis frequencies and in more Caucasians; and the difference in maturation "slow" than "fast" maturers for males than for rates could not be easily attributed to nutri- females. The frequency of cone shaped epi- tional differences. Since the percentage of pre- physes is generally higher in Japanese girls than mature epiphyseal-diaphyseal fusion seems to in boys, especially for the fifth middle phalanx. be lowest among the Japanese data compared It is also known that there are more "missing" with other gene pools, there seem to be differ- phalangeal foot centers in girls than in boys. ences between different gene pools in matura- Certain studies have shown that toe digit tion rates or maturation rate variability. length is shorter among Japanese than among Furthermore, the percentage of the lateral Caucasians, and that male index finger length is intercarpal space is higher in the Japanese data shorter than that of the female. It is also known than in Negro and Caucasian data, also true for that the length of the fifth middle hand phalanx the frequency of cone shaped epiphyses (with is shorter in females than males, possibly indi- the addition of South American Indian data), cating an adaptation to protein malnutrition but the Japanese between-sex ratios are much through an absolute size reduction. It is thus closer to equality than in the other samples. possible that the between-sex differences in the Percentages of radial shaft dysplasia, and speci- frequencies of variations noted for this sample fically radial shaft dysplasia associated with partly reflect smaller Japanese body builds as cone shaped epiphyses likewise differ between compared with other gene pools such as U.S. gene pool samples; these are comparatively high Caucasians; and that the variations are asso- in the Japanese sample, specifically the latter. ciated with some body size economy which One explanation for these gene pool differ- may or may not be an adaptation to some ences (aside from possible observer bias) is that environmental effect. there is an inbreeding effect reinforced by adap- It is difficult not to conclude that these tive variation in environmental responses, since differences are somewhat associated with in- nutritional differences can be discounted as breeding. If the degree of inbreeding has been major factors affecting the variation frequen- decreasing recently in some gene pools, then cies and since the Japanese data show some gene pool variations in inbreeding might in- 48 C. E. EYMAN crease patent differences in skeletal variations. I More specifically, the overall "male com- feel that the sex differences in the frequencies plex" seems to involve single and combined of some variations noted for this sample are variations which are associated with slower affected by inbreeding, and that some auto- male than female maturation. Contrasted to soma1 alleles associated with the X chromo- this is the overall '-female complex," which is some are responsible for these variations. If so, poorly associated with clearly maturational fac- should the penetrance of the "variation" alleles tors. The few growth-associated factors in the be greater in a homozygous state (incomplete overall "female complex" seem to be associated dominance or incomplete recessiveness), then with rapid rather than with slow maturation, so a long-term inbreeding effect might tend to that the "growth" effect is diminished even mask a sex difference. more. Therefore, I cannot conclude that matu- This is partly borne out by the general lack ration rates, inbreeding effects, nutrition, or of a sex difference in the frequency of vari- between-city differences of whatever nature ations for the unrelated and the various inbred suffice to "explain" the variations I have dis- groups. Because the frequency of specific vari- cussed. While I acknowledge that inbreeding ations is as often higher for second cousins than effects must be taken into account, I feel for first cousins once removed as not, I cannot that the variations primarily reflect a sexual conclude that the variation frequencies are de- dimorphism. Even though some of the possible pendent upon the frequency of homozygous mechanisms behind this sex difference are alleles. Otherwise, the unrelated children based upon speculations from the literature, should have the lowest variation frequencies, they may help to clarify the meaning of these with these increasing according to the degree of variations. inbreeding. This is not the case, but the general- If the kinds of variations discussed here in ized inbred character of the Japanese gene pool fact reflect a sexual dimorphism partly mani- somewhat clouds this issue. Also, one must festing itself in "shortcuts" or "economies" assume that one or more of the alleles pro- which organize the female skeletal system into ducing the variations may be pleiotropic; thus basic adult sizes and shapes, then what are some some of the variations might be retained due to of the mechanisms which might be involved? positive selection for other characters under the Here one must turn to literature somewhat partial control of the same gene. Since I feel ancillary to physical anthropology. that there is this possibility, the argument for I have speculated upon some biomechanical the association of variations with homozygous stresses which might produce the radial and loci is threatened even more. ulnar flexions seen in the middle and distal Even assuming that sexual dimorphism in phalanges and metacarpals of the first, second these variations may be masked somewhat by and fifth digital rays, and upon possible bio- inbreeding effects in the sample, one must con- mechanical forces which might inhibit minerali- clude that such sex differences do exist. This is zation of the radial surface of the capitate; but borne out by the "complexes' or variations for it should be apparent that in themselves, none Japanese boys on the one hand and for of these forces is an adequate explanation for Japanese girls on the other. Thus the "pseudo- such variations. No evidence from the literature epiphysis-pseudofracture complex" is more fre- or from the sample indicates noteworthy differ- quent in boys than in girls. The "Harris's lines ences in habitus patterns remotely sufficing to complex" has a slightly higher frequency in invoke such explanations for children from boys than in girls, as does the "shaft dysplasia - various human gene pools. However, with premature fusion complex." Conversely, the respect to biomechanical forces and the "shaft dysplasia - cone shaping complex," the anatomical locations of these rays and their "premature fusion - cone shaping complex," components, these forces probably influence and the "premature fusion - shaft dysplasia - the relatively high frequency of variations seen cone shaping complex," have higher frequen- at these sites. cies among girls than among boys. Thus it is The question still remains as to genetic dif- clear to me that these "complexes7~are defi- ferences between human groups in terms of nitely sex associated. bioelectrical potentials in bone; especially as JAPANESE HAND-WRIST OSSIFICATION 49 these potentials affect the shapes of specific in growth at about six years of age are due to bones and bony regions and their remodeling different gene mechanisms. For the last, it has cycles. The evidence for genetic differences in been suggested that sex differences in growth muscular and ligamentous orientation and that may be due more to size factors than to the suggesting that some skeletal variations are earlier completion of female growth. It has also associated with vascular supply would argue been noted that the number of toe phalanges, that human bone piezoelectric effects need not their lengths, and the presence of cone shaped be constant from one human population to epiphyses and "multiple" ossification centers another. are not associated with skeletal maturation Possibly there are human gene pool differ- rate, although metacarpal notches presumably ences in the genetic control of bone tissue (not are. These considerations suggest that if some to mention the gene pool differences in skeletal hand-wrist variations do represent a sexual "discrete traits"). Possibly many human skele- dimorphism (at least in Japanese children), tal tissue attributes reflect genetic differences, then they are probably poorly associated with such as the variations discussed here. There are skeletal maturation rate; but that something undoubtedly local differences in the rate of such as a size factor is also involved. collagen synthesis even within a given skeletal I suggest that the smaller size of the female system. Thus "pathological" bone collagen can skeleton compared with that of the male re- differ from "normal" bone collagen in terms of flects a skeletal "economy," at least partly asso- "basic amino acid content," "local failures in ciated with a shortcut to attain the minimum mineral supply" can produce clinodactyly, and size necessary for the biomechanics of child- inherited conditions such as familial exostoses bearing. Similarly, at least certain hand-wrist may possibly represent "delayed remodeling variations indicate to me that shape is primarily through a cartilage deficiency in inhibitor a function of size in this regard. Inhibiting or enzymes," to mention a few conditions dis- delaying mineralization of certain bony areas cussed in the literature. (e.g., the capitate, "multiple" centers, epiphy- Grant that bone cell activities differ geneti- sis and diaphysis of the fifth middle phalanx) cally between human populations, and recall would be one way to attain minimum skeletal that the structural genes within such cells are size, while "classic" or "norma1"rugose shape probably controlled by regulator genes re- (by male standards) would be developed later sponding to cellular microenvironments: then (or never). it is not unlikely that local bony regions and If autosomal genes associated with the Y sites within individual bones may undergo chromosome are partly responsible for male selective inhibition and/or configurational retardation in skeletal maturation compared changes which produce "variations." In this with females, I do not feel that they are asso- regard, the slower skeletal maturation of chil- ciated with the hand-wrist variations discussed dren with slight body builds compared with here. Loosely classifying males as "mesomor- children having heavier body builds suggests phic" and females as "ectomorphic" in having genetic control of skeletal configurations, medium and slight skeletal structures, females whatever the reasons might be. It is interesting would mature more slowly than males (which to note that the literature indicates less ecto- they do not) but females would be less variable morphic (slight body build) than mesomorphic than males in skeletal maturation rates ( which (heavier body build) variation in skeletal matu- they are). Because the females lack a Y chromo- ration rates - and that the hand-wrist ossifica- some but have significantly higher frequencies tion sequence variability is less in females than of hand-wrist variations than males, it is diffi- it is in males. cult to invoke within- and between-sex differ- It has been suggested that genotypes may ences in maturation rates or autosomal genes differ mainly in their tissue responses to essen- associated with the Y chromosome to interpret tially the same systemic control mechanisms; the variations. that bodily shape sex differences may be due to It seems increasingly clear that autosomal different growth rates for parts of the overall genes associated with the X chromosome influ- growth period, and that human sex differences ence skeletal hand-wrist ossification timings so C. E. EYMAN and patterns. Because females have less variabil- the X chromosome and a true sexual dimor- ity than males in this sense and also have more phism exists in the form of these variations, variations (at least of certain kinds), then the although the sex differences are somewhat ob- autosomal genes influencing these variations scured by the fact that much of Japan essen- are probably associated with the X chromo- tially represents an inbred human population. some. Females have a wider range of genotypes The adaptive significance of these variations than males with respect to the X chromosome; remains unknown; but there must be more posi- therefore, females might differ more widely tive selection for them because almost half of from each other than males in their skeletal the children in the sample have one or more tissue responses to generalized systemic control "variation," and approximately half of the chil- mechanisms. Thus a wider phenotype range dren in the sample are known to be inbred. might be expected to obtain for females than Because the capitate, hamate and distal would be seen for males with respect to hand- radial epiphysis are the earliest appearing and wrist variations. Repeatedly, for this sample perhaps are the most "stable" of the carpal and perhaps for other inbred gene pools, the sex bones in that they are almost always present difference in hand-wrist variations could be when variations appear elsewhere in the hand obscured somewhat if the alleles responsible for and wrist, there might be some justification for them had some selective advantage elsewhere in assuming that some widespread disturbance of the body. cartilage model mineralization would be partly From all of this, then, one interpretation of necessary to inhibit the nucleation sites of these hand-wrist variations (and perhaps others as three centers. Conversely, the most variable well, such as the frequency of metopism) is that carpal bones in terms of their absence when one they represent a sexual dimorphism expressed or more variations are present elsewhere are the by autosomal alleles associated with the X chro- pisiform, the distal ulna epiphysis and the mosome and which is allied with the general scaphoid. This suggests that possibly any minor earlier arrival of girls at puberty than boys. The disturbance suffices, or tends, to block their significance of many of these variations would mineralization. Centers tending to appear late be the approximate reduction of bony shape to in skeletal hand-wrist maturation tend to have a functional minimum so as to attain approxi- fewer variations associated with them, but this mately minimum absolute size at, or before, is not always the case; hence time of appearance puberty, so that the onset of puberty would not of centers does not strongly modify the previ- need to be delayed by the skeletal system. The ous statement. accuracy of this conclusion can only be tested The frequencies of variations by center sug- by a longitudinal study of variations and matu- gest that the fourth digital ray is the most ration rates in the skeletal systems of both stable, followed by the third; the second, first sexes, with a comparison of ages at which each and fifth digital rays are much more unstable. sex attains puberty. The frequencies of variations also indicate that the triquetral and the hamate are the least SUMMARY affected bones of the carpus. Because this sample is latitudinal rather than longitudinal, The frequency of certain hand-wrist skeletal there is no way to determine how realistic these variations noted in a sample of Japanese chil- impressions are. Also, since only the left hand is dren indicates higher frequencies for girls than represented, there is no way to determine for boys. These variations do not seem to be whether these conditions are symmetrical or associated with ossification patterns, city of not. residence, or nutrition. There is a tendency for When age group, sex, and consanguinity are the frequency of the variations to be associated taken into account the regularity of the obser- with maturation rates (e.g., more "missing" vations provides some measure of confidence. centers in boys than in girls, presumably indica- The most "variant" centers are the fifth middle tive of slower maturation in the former) and phalanx, capitate, distal radial epiphysis, first with consanguinity status. The autosomal distal phalanx and second metacarpal respec- "variation" alleles are probably associated with tively, exclusive of "missing" centers. The loca- JAPANESE HAND-WRIST OSSIFICATION 5 1 tion of these centers relative to the remainder rocal conical crater shaping of the diaphysis) of the hand and wrist provides a clue as to the suggest something else. Since the study is lati- relative stability of certain centers and the com- tudinal, there is a fair possibility that this "re- parative instability of others. This becomes ciprocal shaping" is actually part of a continu- more clear when one recalls that the carpus is ing process in which the reciprocal shaping indi- the region where "missing" centers are almost cates a trend toward premature fusion of the exclusively found, and that the highest frequen- epiphysis with the diaphysis. The conical crater cies of "missing" centers (excluding the pisi- indentation on the fifth middle and the first form) are associated with the radial side of the and second distal phalangeal diaphyses might hand. be other examples of nucleation inhibition de- Perhaps it is partly due to the dynamics of signed to economize upon the amount of bone the relatively wide range of motion in the mineral necessary for the approximately thumb that the radial group of carpals is absent minimal functional shape. Since these three so frequently. We may be observing a "sacri- centers are in exposed positions relative to the fice" of those centers less functionally neces- biomechanics and kinetics of the hand, pre- sary to the dynamics of hand-wrist movement mature fusion could serve to strengthen them. than others, in order to gain more rapid skeletal The stability of the hamate, triquetral, pisi- maturation of the hand and wrist (especially in form, and the fourth digital ray in terms of females). This may be reinforced by the almost absence of variations can also be explained by total restriction of "multiple" or "secondary" this reasoning. These three carpals are more ossification centers to the carpus. This argues intimately associated with the fourth (and that other nucleation areas are present in addi- fifth) digital rays than are the others; the tion to the "normal" or "major" nucleation hamate, triquetral, pisiform and the fourth digi- sites in the carpal cartilage models. The data tal ray thus seem to be relatively unaffected by suggest that a "multiple" or "secondary" ossifi- tangential biomechanical forces. Furthermore, cation center may indicate some inhibition in the fourth digital ray is somewhat protected by the ossification of parts of these cartilage the adjacent third and fifth digital rays while models. It is as if one or more areas of the the triquetral, pisiform and hamate are some- cartilaginous mold were either inhibited or what removed from the rotatory and kinetic totally blocked from calcdfication, with the un- effects of the first digital ray. blocked portions continuing their roles as If the pressure for earlier skeletal maturation nucleation sites. in at least the female hand can be satisfied This interpretation seems to be much more without a given ossification center exactly in keeping with what is known of the various replicating its adult configurations, it is possible nucleating mechanisms in bone tissue. It also that those functionally least important areas of provides a possible explanation for the indenta- a given center may be at least temporarily sacri- tion on the radial surface of the capitate. It is ficed to seduce the time required for minerali- suggested that within the cartilaginous model zation of the adult bone cartilaginous blue- of the capitate, a certain nucleation site or local print. It would thus appear that the higher area is consistently blocked and does not ossify. frequencies of hand-wrist variations, along with If the emphasis is upon developing a mature relatively faster maturation rates and smaller bone which approaches the "ideal" shape gen- absolute size in females compared with males, erally rather than exactly, it is possible that an reflect a sex effect where the more rapid female indentation on the capitate radial surface can maturation is partly manifested skeletally by meet the functional requirements of interosse- "sacrificing" overall mineralization (smaller ous ligaments as easily as a tubercle, thus saving sizes) in rather specific localities of certain a certain amount of bone mineral for use else- bones (shape variations). That there may be where or later. some selective advantage to the autosomal As for the first and second distal and the alleles responsible for these variations seems fifth middle phalanges, their prominent suscep- apparent in that they are also present in males, tibility to variations and the predominant kind and in that approximately half of the sample of variation (cone shaped epiphysis with recip- children manifest such variations and half are 52 C. E. EYMAN the results of consanguineous matings. This, O'Rahilly, R. 1953a Epitriquetrum, hypotriquetrum, and other, questions can only be resolved and lunatotriquetrum. Acta Radiologica 39:401- through future research. 411. -1953b A survey of carpal and tarsal anomalies. LITERATURE CITED J. Bone and Joint Surg. 35A:626-642. -1957 Developmental deviations in the carpus Anderson, J.E. 1962 The Human Skeleton: A Manual and the tarsus. Clin. Orthopaedics 10:9-18. for Archaeologists. National Museum of Canada, Ossenberg, N. S. 1970 The influence of artificial cra- Ottawa. nial deformation on discontinuous morphological -1963 The People of Fairty. National Museum traits. Am. J. Phys. Anthrop. 33: 357-37 1. of Canada, Bulletin No. 193. Schull, W. J. 1958 Empirical risks in consanguineous Arimoto, K. 1952 Nutrition in Japan. Nutrition marriages: sex ratio, malformation, and viability. Reviews 10: 321-323. Am. J. Hum. Gen. 10:294-343. Baker, P. T. and J. L. Angel 1965 Old age changes in Schull, W. J., T. Yanase and H. Nemoto 1962 Kuro- bone density: sex and race factors in the United shima: the impact of religion on an island's genetic States. Hum. Biol. 37: 104-12 1. heritage. Hum. Biol. 34:271-298. Corruccini, R. S. 1974 An examination of the meaning Siegel, S. 1956 Nonparametric Statistics for the Be- of cranial discrete traits for human skeletal biologi- havioral Sciences. McGraw-Hill, New York. cal studies. Am. J. Phys. Anthrop. 40:425-445. Spaulding, A. C. 1960 Statistical description and com- Eyman, C. E. 1970 Hand-Wrist Variability in Japanese parison of artifact assemblages. In: The Applica- Children. Ph.D. Dissertation, University of Cal- tion of Quantitative Methods in Archaeology. gary. R. F. Heizer and S. F. Cook (eds.). Viking Fund Fairbank, T. 1951 An Atlas of General Affections of Publications in Anthropology, No. 28. the Skeleton. Williams and Wilkins, Baltimore. Spuhler, J. N. 1951 Some genetic variations in Ameri- Garn, S. M., S. L. Fels and H. Israel 1967 Brachymeso- can Indians. In: The Physical Anthropology of the phalangia of digit five in ten populations. Am. J. American Indian. W. S. Laughlin (ed.). Fourth Phys. Anthrop. 27:205-209. Viking Fund Summer Seminar in Physical Anthro- Garn, S. M., C. G. Rohmann and T. Blumenthal1966 pology, New York. Ossification sequence polymorphism and sexual Takahashi, E. 1966 Growth and environmental factors dimorphism in skeletal development. Am. J. Phys. in Japan. Hum. Biol. 38: 112-130. Anthrop. 25: 147-151. Torgersen, J. 195la Hereditary factors in the sutural Garn, S. M., C. G. Rohmann and F. N. Silverman 1965 pattern of the skull. Acta Radiologica 36: 374-382. Missing secondary ossification center of the foot: -195 1b The developmental genetics and evolu- inheritance and developmental meaning. Annales tionary meaning of the metopic suture. Am. J. de Radiologie 8:629-644. Phys. Anthrop. 9: 193-210. Glucksmann, A. 1942 The role of mechanical stresses Venning, P. 1954 Sib correlations with respect to the in bone formation in vitro. J. Anat. 76:231-239. number of phalanges on the fifth toe. Annals of Greulich, W. W. and S. I. Pyle 1959 Radiographic Eugenics 18: 232-254. Atlas of Skeletal Development of the Hand and -1956a Radiological studies of variations in the Wrist, 2nd. ed. Stanford University Press. segmentation and ossification of the digits of the Harris, H. A. 1931 Lines of arrested growth in long human foot. I. Variation in the number of pha- bones in childhood: correlation of histological and langes and centers of ossification of the toes. Am. radiographic appearances .in clinical and experi- J. Phys. Anthrop. 14: 1-34. mental conditions. Br. J. Radiol. 4:56 1-588. -1956b Radiological studies of variation in the Hertzog, K. P. 1967 Shortened fifth middle phalanges. segmentation and ossification of the digits of the Am. J. Phys. Anthrop. 27:113-117. human foot. 11. Variation in length of the digit Johnston, F. E. and S. B. Jahina 1965 The contribu- segments correlated with differences of segmentation of the carpal bones to the assessment of skele- tion and ossification of the toes. Am. J. Phys. tal age. Am. J. Phys. Anthrop. 23:349-354. Anthrop. 14: 129-151. Johnston, F. E., R. H. Whitehouse and K. P. Hertzog -196 1 Radiological studies of variation in ossifi- 1968 Normal variability in the age and first onset cation of the foot. IV. The length and growth of of ossification of the triquetral. Am. J. Phys. bones of the foot in relation to morphology. Am. Anthrop. 28:97-99. J. Phys. Anthrop. 19: 137-145. Kohler, A. 1968 Borderlands of the Normal and Early Virchow, H. 1929 Das os centrale carpi des menschen. Pathologic in Skeletal Roentgenology, 3rd. Ameri- Morphol. Jahrbuch, vol. 63, part 2, pp. 480-530. can ed. Grune and Stratton, New York. Woo, J. K. 1949 Racial and sexual differences in the frontal curvature and its relation to metopism. Am. J. Phys. Anthrop. 7:215-227. The Estimation of Maximum Crown Dimensions in Human Permanent Teeth with Severe Attrition M. McKEOWN' and K. ISOTUPA2 ' Department of Oral Biology, University of Saskatchewan, Saskatoon, Saskatchewan S7N OWO, and 252,Lemminkaisenkatu 2, Turku, Finland

KEY WORDS Human Tooth Wear Quantitative Variation Prediction

ABSTRACT A study of 589 unworn human permanent teeth shows that strong statistical correlations are present between linear measurements of maximum crown dimensions and similar measurements in the cervical region in human teeth. These relationships provide a basis for estimation of previously present crown dimensions in teeth severely worn by attrition. Correlation values for different teeth suggest that some teeth are more morphologically stable than others.

RESUME LYCtude de 589 dents humaines permanentes, et sans usure, a dCmontr6 la forte corrClation statistique qui existe entre les mensurations lin6aires des dimensions maximum de la couronne et les mensurations semblables de la rCgion cervicale de la dent humaine. Ces rapports offrent la base de 17estimedes dimensions originales de la couronne dans le cas de dents extrkmement usCes par attrition. La valeur des corrClations obtenues pour des dents diffkrentes suggkre que certaines dents sont morphologiquement plus stables que d'autres.

INTRODUCTION teeth were examined and there were between The presence of severe attrition, by which 45 and 50 teeth in each tooth group. The the major portion of the crown is removed, is a material was classified into tooth types individ- substantial handicap in studies that seek to ually by each author. Teeth of dubious classifi- examine the size of the unworn tooth in skele- cation were excluded. No distinction was made tal material. An investigation was carried out to between left and right sides, nor was the sex of establish the mathematical relationship be- the material known. The 1st and 2nd molars in tween certain root and crown measurements in both upper and lower jaws were grouped unworn human teeth, in the hope that valuable together, as positive identification was not relationships could be established that would possible. permit an accurate estimate of the previously Measurements were made of maximum present crown dimensions to be made from the mesio-distal and bucco-lingual crown dimen- remaining tooth substance. sions and mesio-distal and bucco-lingual cervical root dimensions on the root surface. In MATERIAL AND METHODS addition, the root length was measured along the buccal root face in all teeth except the A collection of unworn human permanent molars (Fig. 1). Measurements were made using teeth from the cities of Winnipeg and Saska- a vernier caliper calibrated to 1/20 mm. Each toon, Canada, was examined. The material was measurement was carried out by one observer primarily of white Caucasian origin, although on three separate occasions. The mean values the presence of some North American Indian were calculated and used in the study. The material could not be excluded. A total of 589

CAN. REV. PHYS. ANTHROP. 253-56 M. McKEOWN and K. ISOTUPA

shows the values for the mesio-distal crown diameter of each group. Table 2 shows the linear correlation matrix for the four and five variables in each tooth group. The relationship between two of the best correlations in the study was then examined. A test of the ability to estimate an unknown dimension, e.g., mesio-distal crown diameter from a known value such as mesio-distal cervical diameter, was then carried out. The actual values of each measurement for all the teeth were set out accompanied by the value predicted by the equation for the unknown crown dimension. The actual and predicted values were then compared and the percentage by which the predicted measure deviated from the actual one recorded for each set of measurements (each tooth). The range of variation of these percentage differences together with the mean value on both positive and negative sides of the actual values is shown in Table 3. If we consider the upper molars, we can say with Fig. 1 Linear measurements as taken on an upper second premolar tooth (see Table 2). Two-rooted certainty that the actual value of any prediction first premolars were also measured on their buccal within the range of the teeth studied lies be- root. tween -8.2% and +9.1% of the predicted value. One can of course substantially reduce that group mean values and standard deviations range by accepting an element of chance in the were calculated and a correlation matrix ob- prediction. A range of variation of approxi- tained with the use of a computer program. The mately 8.5% for avalue of 10 mm is 20.85 mm. mathematical significance of the correlations This is a moderately large range of variation. was then calculated. The primary purpose of The distribution of the data, however, showed the correlations was to identify associations that 60% of the material lay inside the mean best suited for establishing a prediction tech- limits of -2.41 and +2.87%. Thus, if one is nique. The accuracy of prediction of a crown willing to accept this level of probability, the measurement from a root measurement was range of variation is substantially smaller. calculated for two of the best correlations obtained, using a curve fitting procedure. The DISCUSSION correlations chosen were those between the The highly significant correlations seen be- mesio-distal crown measurements and mesio- tween many parameters in the tooth groups distal cervical measurements of the upper suggests a cohesive structural variation pattern canines and molars. (see Table 2). A series of in overall form. It is apparent (Table 2) that curve fits was carried out by computer in which root length has the poorest relationship with a least squares curve fit of the linear transforma- any of the other measurements, in all the teeth. tions was carried out. Using the index of deter- It is also apparent that within each tooth group, mination a best curve was chosen for the data. the level of correlation varies considerably, Table 3 shows the curves chosen for upper being highest in the canines and lowest in the molars and upper canines. second premolars. It appears that some teeth are more consistent in form than others and RESULTS may prove more valuable for measurement stud- The means and standard deviations for each ies related to variability. It is also interesting parameter of each tooth was calculated. Table 1 to observe that the correlations for the crown dimensions of upper and lower teeth in the ESTIMATION OF TOOTH CROWN DIMENSIONS

TABLE 1 Mesio-distal crown diameter (mm)

S.D. 0.62 0.52 0.5 2 0.45 0.47 0.78 Mean 8.70 6.58 7.62 6.83 6.74 9.88

Mean 5.61 6.09 6.87 6.99 7.06 11.31 Lower S.D. 0.33 0.39 0.39 0.34 0.4 3 0.67

TABLE 2 Correlation coefficients

Variables Centrals Laterals Canines 1 st Premolar 2nd Premolar: Molars 0.69* 0.69* 0.86* 0.80* 0.79* 0.71* Upper MDCRIBLCR 0.66 0.54* 0.76* 0.75* 0.80* 0.75* Lower 0.65* 0.82* 0.87* 0.67* 0.34 0.90* Upper MDCEIMDCR 0.68* 0.45 * 0.81* 0.76* 0.33 0.7 1 * Lower 0.59* 0.59* 0.87* 0.77* 0.49 0.61 * Upper MDCEIBLCR 0.69* 0.43 0.86* 0.68 0.35 0.69* Lower 0.60* 0.62* 0.75* 0.76* 0.47* 0.72* Upper BLCEIMDCR 0.66* 0.39 0.77* 0.52* 0.50* 0.58 Lower 0.86* 0.86 * 0.88* 0.$4* 0.59* 0.61* Upper BLCEIBLCR 0.92* 0.79* 0.98 * 0.58* 0.63* 0.80* Lower 0.59* 0.59* 0.82* 0.73* 0.35 0.65* Upper BLCEIMDCE 0.75* 0.40 0.84* 0.36 0.31 0.79* Lower -0.03 0.11 0.59* 0.32 0.25 - Upper RLIMDCR 0.6 1 * 0.31 0.44 0.09 0.12 - Lower 0.07 0.19 0.55* 0.43 0.28 - Upper RLIBLCR 0.44 0.42 0.45 0.17 0.22 - Lower -0.02 0.11 0.59* 0.38 0.27 - Upper RLIMDCE 0.33 0.22 0.52* -0.01 -0.16 - Lower 0.21 0.3 1 0.52* 0.47* 0.08 - Upper RLIBLCE 0.45 0.4 1 0.48" 0.33 0.12 - Lower

* P<0.001; MDCR = mesio-distal crown diameter; BLCR = bucco-lingual crown diameter; MDCE = mesio-distal cervical diameter; BLCE = bucco-lingual cervical diameter; RL = root length (buccal face). M. McKEOWN and K. ISOTUPA

TABLE 3 The curves used and the relationship between predicted and actual data

Teeth Correlation Best Curve Type Mean difference between calculated Range of Percentage and actual values (%) Difference Negative Values Positive Values Upper canines 0.87 Y=A+BlogX -2.49 +2.38 -10.5 t0+7.7 (A =-6.37, B = 5.85) Upper first and 0.90 Y= 1 second molars A+BX (A= 0.17, B =-0.01) same groups are surprisingly similar, with the may be more advantageous to examine com- exception of upper and lower lateral incisors. plete dentitions as opposed to individual un- This may be a reflection of phylogenetic related teeth. Interrelationships using multiple origins. linear regression procedures could be estab- The fitting of curves to the data is, of course, lished which would enhance estimative capa- a mathematical exercise designed for practical bility. purposes and of limited biological significance. An accurate estimation technique would be It does show, however, that it is possible to valuable in anthropological studies seeking to predict crown dimensions from cervical root examine crown size, in cases where attrition has measurements with some degree of accuracy. It removed the major part of the crown. This is not, however, claimed that the mathematical study suggests that such a technique can be expressions used here are necessarily the best or developed. only method of estimation, nor should the data presented here be extrapolated beyond their ACKNOWLEDGEMENTS mathematical limits. The results should, of The authors would like to thank Dr. F. course, be applied with caution to other Chebib and his staff, Biostatistics, College of material. It does seem, however, that further Dentistry, University of Manitoba and Mr. investigations, using more diverse large samples, Dennis Duncan of the College of Dentistry, would be worthwhile. Studies using multiple University of Saskatchewan for their assistance associations derived from measures on 'un- with statistics. related' teeth are probably of limited value. It L'Utilisation de 1'Analyse de Variance Multiple en Anthropologie de la Sante ' F. FOREST et D. BERTHELETTE Dgpartement d'Anthropologie, Universitt de Montreal, Montrtal, P.Q. H3C 3J7

MOTS CLES Anthropologie Mkdicale Ethno-psychiatrie Niveau de Santk Environnement Industrialisk

RESUME L'anthropologie mkdicale est un domaine qui tend de plus en plus a se dkvelopper. Des anthropologues se sont dkjA penchks sur les problbmes relatifs l'ethno-psychiatrie et it la morbiditk comparke de diffkrentes populations. I1 semble cependant que les recherches anthropologiques considkrant l'impact de l'environnement socio-culture1 sur les maladies physiques et psychiques soient limitkes en nombre. Ce fait est d'autant plus frappant en ce qui concerne l'ktude des sociktks industrialiskes. Nous nous proposons donc, dans cet article, de presenter une mCthode statistique permettant d'analyser l'influence de plusieurs facteurs environnementaux sur la variation des niveaux de santk d'une population de travailleurs.

ABSTRACT Medical anthropology is emerging as a major subfield within the discipline of anthropology. Already important contributions have been made in ethno-psychiatry and the comparative study of morbidity in various populations. Nonetheless, there is a derth of studies dealing with how specific sociocultural environmental factors influence physical and mental disease patternsespecially in modern industrialised societies. In order to promote and facilitate research activities in this area, we present a statistical method capable of analysing the influence of an array of environmental variables on health standards in an industrialised setting.

INTRODUCTION dhles complexes. Le chercheur doit donc se La variabilitk humaine demeure l'intkrgt doter d'un minimum d'outils statistiques afin fondamental de l'anthropologue physique. Ce de faciliter ses recherches et d'engager le phknombne soulbve des questions primordiales dialogue avec d'autres disciplines. quant a l'effet des conditions environnemen- Le but de cet article est de presenter un tales sur la variation phknotypique. De plus en modele statistique utilisk afin d'analyser la vari- plus de chercheurs concernks par cette problk- ation d'indices morphologiques et physio- matique, dksirent s'impliquer dans leur milieu pathologiques en fonction de nombreuses en identifiant les causes de variation, afin sources de variations environnementales dif- d'amkliorer la qualitk de vie. fkrentes. La santk est certes le lieu privilkgik des intkr- actions en tre la biologie e t le milieu. L'iden tifi- ' La m6thode statistique decrite dans cet article a 6t6 adapt6e et opBrationnalis6e pour des fins de re- cation des facteurs de risque, dans une popula- cherche~en anthropologie, par M. F. Forest. Le pr6- tion, peut nous permettre de les diminuer. L'en- sent texte, redig6 par Mme. D. Berthelette, fait suite A vironnement est toutefois constituk d'un en- une communication qu'elle a presentee A l'occasion du semble de variables matkrielles, psychologiques septieme congres de 1'Association pour 1'Anthropolo- gie Physique au Canada, Ste.-Adele, QuBbec, 1979. et culturelles qui interagissent selon des mo-

CAN. REV. PHYS. ANTHROP. 25765 5 8 F. FOREST et D. BERTHELETTE

Cette Ctude s'inscrit B l'interieur d'un projet somme du nombre de ses niveaux, moins un. de recherches portant sur la variation des Ainsi, la profession possbde cinq categories et niveaux de santk d'une population de 3000 les quatre premieres servent de pseudo- travailleurs quCbCcois. Le le cteur pourra se variables. referer B des articles dCjB parus (Bastarache et Dans la matrice, la valeur +1, assignee a une al., '78; Auger et al., '79)' afin de prendre note pseudo-variable d'un vecteur, signifie que le des mCthodes relatives au recueil des donnCes et sujet considCrC appartient B cette catkgorie; la B la mise au point des indices morphologiques et valeur zero est alors donnCe aux autres prC- physio-pathologiques. dicats. On attribue la valeur - 1 it toutes les pseudo-variables d'un vecteur, lorsque le sujet METHODE considCr6 se situe dans le dernier niveau du Notre modble comporte 39 critbres de classi- critbre, absent de la matrice. fication subdivisks en 12 groupes (tableau 1). I1 est possible d'inclure les effets d'inter- Ces variables ont CtC considerees simultane- action dans le modble. Nous prbfkrons cepen- ment dans une analyse de variance multiple dant les delaisser au profit d'autres effets basCe sur la regression multiple (Overall et principaux, l'espace en mCmoire centrale &ant Klett, '72). Cette mCthode statistique possbde limite. l'avantage d7Ctreplus fidble B la realit6 qui veut Les pseudo-variables x, au nombre de p, font qu'une cause n'agisse pratiquement jamais par la suite l'objet du calcul de leur matrice de seule. I1 est en effet important de tenir compte correlation et de l'inverse de celle-ci (R-1). On des correlations existant entre les critbres, afin Ctablit enfin les correlations entre la variable de ne pas mesestimer leurs effets. dependante y, et chacune des pseudo-variables Ceci peut Ctre assure par le calcul de rCsultats x, pour obtenir un vecteur v. ajustes representant l'effect reel d'un critbre, A partir de ces deux renseignements, R-1 et comme s'il n'ktait aucunement corrCle aux v, nous procCdons au calcul du vecteur des autres crithres prCsents dans le modble. Ces coefficients p, de mCme qu'B celui du coeffi- resultats sont d'autant plus reprksentatifs de la cient de rCgression multiple ~2~. realitC, que le nombre total de crithres signifi- catifs est 6levC. La premibre Ctape de cette mCthode consiste A construire une matrice de prkdicats con- Le ~2~ reprksente la proportion de la vari- tenant l'information relative a la dispersion des abilite totale de la variable dependante y, qui sujets A l'intkrieur des differents niveaux des est expliquCe par sa regression sur les pseudo- critbres. Chacun des critbres y est represent6 par variables representant l'ensemble des critbres de une quantitk de pseudo-variables &gale B la classification et leurs niveaux. On peut verifier

TABLEAU 1 Crit8res de classification

Groupes et Critbres Niveau 1 Niveau 2 Niveau 3 Niveau 4 Niveau 5 A- Stratification socio-profe~sio~~elle manuel non manuel employe 1- Profession specialise spCcialis6 contremaitre de bureau cadres 2- Scolarite (cours) primaire secondaire technique collBgial universitaire 3- Employes h charge aucun manuels de bureau B- Socio-demographic 1- Age 20-29 30-39 40-49 50-59 60 et plus 2- Etat civil cklibataire mari6 separe veu f 3- Nombre d'enfants aucun 183 486 7 et plus VARIANCE MULTIPLE ET SANTE

C- Famille d'origine 1- Taille de la fratrie nulle 183 496 7 et plus 2- Rang de naissance premier 2e i 4e 5e i 7e 8e et plus D-Habitation 1- Proprietaire oui non 2- Nombre de pibces 3 ou moins 485 6 et plus 3- Autres personnes vivant 8 la maison aucune parents amis chambreurs E- Travail moins 1- MCme travail depuis d'un an 1 9 loans plus de 10 2- Nombre d9emplois plus d'un un seul 3- Degr6 d'activit6 au travail nu1 modere important F- Facteurs psychologiques lies au travail 1- Absence dYint6rCtau travail nullement legkrement modCrCment pleinement 2- Sous-utilisation des capacites nullement 16ggrement moderement pleinement 3- Travail trop dur physiquement nullement 16gGrement mod6rBment pleinement 4- Manque d'autonomie nullement legkrement mod6rCment pleinement G-Facteurs psychologiques extrinseques 1- Tension dCveloppBe au travail nulle legere modCree importante 2- Mauvais climat de travail nullement legbrement moderement pleinement 3- Absence de possibilitCs de pramotion nullement legbrement moderement pleinement 4- Trop de pressions au travail nullement legeremen t moderement pleinement 5- Absence de reconnaissance nullement legeremen t moderement pleinement 6- Difficultds avec les superieurs nulles mod6rCes importantes H-Fateurs psychologiques extrinseques lies a l'aspect r6muneration 1- Salaire insuffisant nullement legbrement modBrBment pleinement 2- Manque de sCcurit6 d'emploi nullement leggrement moderement pleinement I- Anomie 1- Anomie 1 nulle modBr6e importante 2- Anomie 2 nulle moderee importante J- Antecedents hkrkditaires 1- Cancer presence absence 2- Maladies cardio-vasculaires presence absence 3- Maladies du sy stCme nerveux presence absence 4- ObBsit6 presence absence K-Habitudes de vie 1- DegrC d'activitk physique hors travail nu1 moderC important 2- Utilisation de mddicaments nulle rnod616e importante 3- Nombre de cigarettes par jour aucunes 1 9 15 16 25 25 et plus 4- Quantite d'onces d'alcool par mois moins de 3 4 8 15 16 et plus L- Digte 1- Quantitt? de lipides faible moyenne forte 2- Quantite de glucides faible moyenne forte 3- Quantite de fruits et legumes faible moyenne forte 60 F. FOREST et D. BERTHELETTE la non kquivalence a zero de la valeur ~2~ par le tionner la somme des carrks d'kcarts inter- test de signification de rkgression suivant, n groupe en ses componsantes d'effets principaux ktant le nombre total de sujets: et d'effets d'intkraction. Nous identifions ainsi les impacts respectifs de chacun des critbres sur la variabilitk de la variable dkpendante. Le coefficient de rkgression du critbre "A", Les coefficients pi au nombre de p, doivent par exemple, est calculk a partir d'un modble par la suite &re transformks en valeurs bi, afin qui ne contient pas les pseudo-variables de ce de permettre le calcul des moyennes ajustkes critbre "A". On obtient donc un nouveau ~2 de nos critbres. L'kquation suivante doit Ctre total excluant le critbre "A" dont nous dksirons utiliske : connailre l'effet, et qui sera not6 ~2~~.La mkthode de partition de la somme des carrks d'kcarts inter-groupe que nous utilisons, ajuste i, variant de 1 a p, Sy reprksentant l'kcart-type chaque effet principal aux autres effets princi- de la variable dkpendante y et SX~,lY6cart-type paux du modble. du prkdicat xi. Le R2 correspondant A l'effet ajustk du cri- Les coefficients bi sont d'abord utilisks pour tbre "A", not6 ~2~'sera calculk de la faqon calculer le moyenne gknkrale ajustke, qui cor- suivante : respond a la valeur K dans lYCquationde rCgres- sion multiple et qui sera la rCfkrence partir de laquelle toutes les deviations seront kvaluCes: et cet effet ajustk du critere "A" pourra &re kprouvk selon la prockdure habituelle :

Dans cette expression, est la moyenne empiri- F = R2A (1-a;a-l;n-p-l) (1-R~T) (a- 1) que de la variable dkpendante et les diffkrentes moyennes Xi, sont les moyennes des pseudo- dans laquelle a correspond au nombre de niv- variables. eaux du critbre "A". Le calcul des diverses moyennes ajustbes des Toutefois, lorsque plusieurs criteres de clas- niveaux des critbres se fait en utilisant les coeffi- sification sont trbs fortement corrklks entre cien ts de rCgressions multiples bi: eux, une telle prockdure peut correspondre 6 une sousdvaluation certaine de l'effet de chacun d'eux. Le regroupement de ces critbres Cependant, a cause de l'abandon, dans la et le calcul de coefficients de rkgression ajustks, matrice des pseudo-variables, d'un niveau par pour chacun des groupes peut remkdier a ce critbre, il y aura au moins autant de coefficients probleme. Lorsque des criteres fortement cor- bi qui manqueront, qu'il y a de critkres. Pour rkles, ont approximativement le mCme effet sur chacun de ces derniers, la deviation bi man- la variable dkpendante, dans le mdme sens et quante, peut &re calculke en sachant que la avec la mCme amplitude, la probabilitk ajustke somme des dkviations ajustkes bi de tous les de chacun d'eux est gkneralement assez faible, niveaux d'un critbre est kgale a zero : alors que leur probabilitk de groupe, ajustke a Pour tout critere: Cbi = 0 tous les autres criteres, est beaucoup plus importante. Ainsi, au niveau de l'interprktation Dbs lors, pour le critbre "A", par exemple, des rksultats nous pouvons supposer que la vari- qui compor terait 5 niveaux diffkrents, le vec- able dkpendante puisse Ctre like a des phk- teur b comporterait les bi correspondant aux nombnes dont les critbres du groupe sont les deviations des quatre premiers de ces niveaux. reflets, bien qu'ils n'aient pas d'effet ajustk Le dernier, absent du vecteur b, peut Ctre cal- significatif lorsqu'ils sont considCrks isolkment, culk en sachant que la somme de ces cinq dkvia- a cause de la prksence des autres criteres du tions est Cgale a zkro. m6me groupe. Le ~2 d'un groupe est calculb de Dans un deuxibme temps, nous devons pro- la faqon suivante : ckder au calcul d'un ~2 brut et ajustk pour chacun des critbres. Ceci nous permet de parti- VARIANCE MULTIPLE ET SANTE 6 1

I1 s'agit dans un premier temps, de calculer un tant de pr6ciser la morphologie des individus R2 ne contenant aucune des pseudo-variables qui composent notre population, car cet aspect des critbres du groupe (~2-G) et de soustraire est fortement lie B plusieurs tests physio- ce RzdG du ~2~.La contribution d'un groupe logiques. Ainsi, le pourcentage de tissus adi- 1 la regression sur la variable dependante est peux, indice d'obesitk, a des effets sur la mor- ensuite 6prouvee par les test de F suivant : bidite de plusieurs pathologies et notamment sur les maladies cardio-vasculaires.

RESULTATS la lettre g representant la somme du nombre de La moyenne g6nCrale de pourcentage de pseudo-variables des critbres faisant partie du tissus adipeux est de 18.34%' dans notre popu- groupe . lation, 1'Ccart-type est de 6.50 %. L'ensemble de Afin de faciliter l'interpretation des rksultats nos critbres de classification rdussit A expliquer decoulant de ces diverses operations mathemat- 15.9 1% de la variation totale de cette variable. iques, nous u tilisons des procCdCs graphiques La stratification socio-professionnelle ex- que nous nous devons d'expliquer quelque peu. plique 0.54% de la variation, mais de faqon Les resultats qui sont prCsentCs B titre d'exem- non significative, la probabilite Ctant dgale B ples dans cet article, concernent les trois vari- 0.871. La profession, premier critbre de ce ables suivantes: pourcentage de tissus adipeux, groupe, possbde un R2 brut de 0.42% et les tension diastolique, sympt6mes neurologiques. differences apparentes entre les moyennes de Chaque diagramme peut 6tre interpret6 ses cinq niveaux sont significatives. On note comme suit: sur 1'6chelle de droite sont notees cependant que les 6carts entre les moyennes les valeurs centrales de 15 classes de rksultats de brutes disparaissent lorsque les calculs sont faits la variable dependante, en unites normalisees de faqon ajustCe, le R2 Btant alors Cgal B O.l2%, (Y = 100, Sy = 20). La classe "1 14" contient avec une probabilite de 0.483. Ceci nous aussi toutes les valeurs qui lui sont superieures indique donc que le critbre serait fortement et la classe "86", toutes les valeurs inferieures. correle B un autre critbre ayant un impact sur la Ces classes ont une Ctendue de 0.1 kart-type. variation du pourcentage de tissus adipeux. L'tchelle de gauche rapporte les m6mes valeurs La scolaritk demeure le seul critkre de ce centrales de classes, mais dans les unites origi- groupe qui prCsente des differences significa- nales de la variable dependante. L'avant- tives entre les moyennes ajustees (P = .950). dernikre ligne du bas indique les diffkrents niv- L'iige (fig. 1) est le critbre dont la probabilite eaux du critbre, alors que la dernibre ligne iden- ajustee est la plus 6levee. On peut toutefois tifie le critbre, donne le rang d'importance remarquer que les moyennes brutes dont le R2 obtenu par sa probabilitk ajustee et rapporte les est 6gal B 9.13%, ont tendance B surestimer probabilites brutes et ajustees (P(1- 09). l'importance de l'effet du critbre sur la variation Les lettres apparaissant dans le tableau ainsi de notre variable dependante. De faqon ajustke, delimite, indiquent les valeurs des moyennes l'iige explique 2.5 5 5% de sa variation. brutes (B) et des moyennes ajustees (A) de On observe des differences significatives chacun des niveaux du critbre. La presence d'un entre les moyennes des niveaux du critkre con- C signifie que la moyenne brute de ce niveau est cernant la presence (niveau 1) et l'absence dans la meme classe que sa moyenne ajustke. (niveau 2) de personnes souffrant ou ayant Les diffkrentes moyennes ajustCes sont relikes souffert d'obbsitk dans la famille du sujet. Les par un trait continu et representent l'effet resultats bruts possbdent un R2 de 1.45 %alors ajuste du critbre aux autres critbres presents que les resultats ajustes expliquent 2.18% de la dans le modkle. Les moyennes brutes jointes variation. On note une tendance selon laquelle par des traits en pointilles, ne representent que le pourcentage de tissus adipeux serait plus l'effet apparent du critbre . eleve chez les individus du premier groupe. Les variables dependantes que nous avons La consommation de cigarettes est respon- analyskes sont de trois types. Dans un premier sable de 1.65 % de la variation de no tre variable temps, nous avons 6tudie la variation de vari- dependante, et ce de faqon significative, en ce ables anthropometriques. I1 est en effet impor- 62 F. FOREST et D. BERTHELETTE qui a trait aux resultats ajustks. La relation sont significatives, P ( 1- 6)atteignant le seuil de entre les niveaux du critQreet la variable depen- 0.999. On observe des differences significatives dante semble &re curviligne. On peut observer entre les professions, ce critgre expliquant, de de ce fait, que les non-fumeurs et les fumeurs de faqon ajustee, 0.36% de la variation. D'apr6s le plus de 25 cigarettes par jour semblent avoir des graphique, les manuels specialises de notre resultats plus ClevCs que les fumeurs de 1 a 15 et population possbdent les resultats les plus de 16 B 25 cigarettes par jour. 6leves, alors que les contremaitres se situent Enfin, le degre d'activite physique, dont le dans la classe de resultats la plus faible. Les trois ~2 ajuste est de 0.2476, serait en relation inverse autres categories professionnelles ont des rksul- avec le pourcentage de tissus adipeux. On tats comparables. observe en effet (fig. I), une diminution des Le degre de responsabilite au travail, ex- rCsultats like B l'augmentation du degr6 d'acti- plique 0.27% de la variation et on observe des vite physique et ce de faqon significative. differences significatives entre les moyennes La deuxihme s6rie de variables dependantes (P > 0.999. . .); ce meme critQre prksente des qui a 6t6 analy.de, porte sur les resultats relatifs differences significatives tant au niveau des aux divers tests physiologiques. Ceux-ci sont moyennes brutes qu'ajustees. On note cepen- des indicateurs importants quant a 1'6tat de dant que les resultats bruts (~2= 1 1.69%) ont santC de notre population. tendance a surevaluer l'importance de l'effet de L'une de ces variables, la tension diastolique, ce critQre par rapport aux resultats ajustes qui obtient une moyenne generale de 81.44 n'expliquent plus que 3.5 1% de la variation. Les mm.Hg. et un Ccart-type de 12.08. Notre mo- deux courbes (fig. 2) nous indiquent que la dQleparvient a expliquer 18.76% de sa variation. tension diastolique tend a augmenter de 20 59 La stratification socio-professionnelle possQde am, pour ensuite diminuer lkgkrement chez les un ~2 de 0.95% et les differences observCes individus plus ages.

I I I N I VEAU 1 2 3 4-ACTIVITE PHYSIQUE PB- .999978 PA- .964618

Fig. 1 Variation du pourcentage de tissus adipeux en fonction du degr6 dYactivit6physique. VARIANCE MULTIPLE ET SANTE

89.948- 4 14 88.740- 412 87.531, 110 86.323- 408 85.115- 106 83.907 4 04 82.699 102 81.491 100

7g.07[80. 283- - 9896 77.867 - 94 76.65& B= - 92 75.450- - 90 74.2421 A - 88 73.0341 -86

I I I I I NIVEAU 1 2 3 4 5 1- Age PB - 1.000000 PA - 1.000000 Fig. 2 Variation de la tension diastoloque en fonction de l'ige.

89.948- -1 14 88.740, 212 87.531, dl0 86.323- A08 85.115- -1 06 83.907- 204 82.699, A02 81.491 100 80. 283- A - 98 79.075- - 96 77.867, - 94 76.658- - 92 75.450- - 90 74.242- - 88 73.034, - 86

N IVEAU f 2! 3 3-ALCOOL PB - ,999955 PA - .999917 Fig. 3 Variation de la tension diastoloque en fonction de la consommation d'alcool. 64 F. FOREST et D. BERTHELETTE

Le nombre de cigarettes fumkes par jour est De mCme, les individus tendus au travail un critkre posskdant des R2 brut et ajust6 de prksentent des rksultats plutbt dkfavorables en 1.19%. On observe des differences significatives ce qui a trait aux sympt6mes consid6rCs ici. Ce entre les moyennes des quatre niveaux. deuxikme critkre nous permet d'expliquer La quantitk d'alcool consommk. par mois, 0.90% de la variation et ce, de faqon significa- posshderait un effet posifit sur l'augmentation tive (fig. 5). Les antbckdents hkrbditaires rela- de la tension diastolique. Ce critkre est respon- tifs au systhme nerveux sont responsables de sable de 0.63% de la variation de la variable 0.58% de la variation. On note 6galement une dbpendante (fig. 3). On note enfin, une tension tendance a l'augmentation des sympt6mes neu- diastolique possbdant une tendance 6 aug- rologiques chez les individus dont les parents menter chez les individus ayant des antkcbdents souffrent ou ont souffert de ces sympt6mes. hkrkditaires en ce qui a trait a l'obksitk. Ce Enfin, un indice d'anomie explique 0.62% dernier critkre parvient i expliquer, de faqon de la variation. Les diffkrences entre les ajustbe, 0.3 1% de la variation. moyennes sont significatives et on observe une La dernikre catkgorie de variables d6pen- prkpondkrance des symp tames neurologiques dantes analyskes concerne des sympt6mes chez les sujets dont l'anomie est importante. rattachbs des systhmes particuliers. Les symp- t6mes du systdme neurologique par exemple, CONCLUSION ont 6tb calculks ?I partir de nombreuses rk- sponses aux questions obtenues dans un ques- Les rCsultats qui ont 6tC prksentds A titre tionnaire prkexamen m6dical. C'est en fonc- d'exemple, d6montrent bien l'importance de tion de la relation entre chacun de ces symp- l'impact des facteurs environnementaux sur la t6mes et de la cote d'apprkciation attribuke variation des niveaux de santb. Une cinquan- par un mkdecin B l'ensemble du systkme, que taine d'autres variables physiologiques ont 6tk chaque sympt6me a pu 6tre pondkrk. L'indi- analyskes selon le technique prkconisee ici et cateur global du systkme neurologique reprb- feront l'objet de publications dans un futur sente le cumul de ces diverses pondkrations. La rapproch6 . moyenne des rksultats obtenus dans notre Cette methode tire son importance du fait population est de 11.54 et l'ecart-typ:, de 1.63. qu'elle nous permette de diffkrencier les effets La stratification socio-professionnelle ex- bruts ou apparents d'un critkre, de ses effets rbels ou ajust6s. Ces derniers, 6tant plus repre- plique de faqon significative (P > 0.999. . .) 0.94% de la variation de la variable concernant sentatifs des relations existant entre l'environ- les symptames neurologiques. La scolarit6 et le nement et l'incidence des maladies, pourront degrk de responsabilitk au travail, ont respec- par la suite Ctre la source de campagnes d'inter- tivement, des R2 ajustbs de 0.52% et de 0.1 5%. vention dans le milieu et favoriser ainsi l'amkli- Les diffkrences entre les moyennes sont signifi- oration de la qualit6 de vie de notre population. cative~. I1 n'en est pas de meme pour la profession. OUVRAGES CONSULTES L'utilisation de medicaments dktient la pro- Auger, F., F. Forest and E. Bastarache 1979 An anth- babilit6 ajustbe la plus klevke (P > 0.999 . . .) et ropological project on French Canadian workers: a explique 2.29% de la variation (fig. 4). On peut progress report. Can. Rev. Phys. Anthrop. 1:5-9. observer qu'une tendance h l'augmentation des Bastarache, E., F. Auger, F. Forest et R. Bastarache 1978 Utilisation d'un questionnaire pr6-examen sympt6mes neurologiques serait libe B l'accrois- medical auprbs d'une population de travailleurs sement de l'utilisation de mbdicaments. quCbCcois. L'Union Medicale du Canada 107: 1-12. Overall, J. E. and J. C. Klett (eds.) 1972 Applied Multivariate Analyses. McGraw-Hill, New York. VARIANCE MULTIPLE ET SANTE 65 12.681 - . *B J14 12.519, m..*m** J12 ...** 12.356 , .*b'" 210 12.193 , .em J08 12.031 - . -106 11.868 , .m* J04 11.705 , m. J02 . ** /A 11.543 . 2 00 11.380 - - 98 11.217 , - 96 11.054, ABo/A - 94 10.892 - - 92 10.729 , - 90 10.566, - 88 10.404, - 86

I I I N IVEAU 1 2 3 1-MEDICAMENT PB - 1.000000 PA - 1.000000 Fig. 4 Variation des sympt6mes neurologiques en fonction de l'utilisation de mkdicaments.

I I I I NIVEAU 1 2 3 4 2-TENSION AU TRAVAIL PB - 1.000000 PA - .999989 Fig. 5 Variation des sympt6mes neurologiques en fonction de la tension au travail. Diffe'rences entre les Sexes et Lateralisation au Niveau de Performances Manuelles YVON DESJARDINS Dkpartement d 'Anthropologie, Universitk de MontrCal, Montrkal, P.Q. H3C 357

MOTS CLES Differences Sexuelles Lateralit6 Tests Moteurs Performance Manuelle

RESUME Les differences entre sexes et la lateralisation ne sont pas des questions nouvelles, si l'on tient compte des ideologies qui entourent ces questions. Nous avons cherch6 dans un petit bchantillion d'etudiants sous-graduks si nous pouvions repondre B ces questions. Nous avons trouve qu'effectivement il y avait des differences entre les sexes au niveau de trois tests manuels et que l'hypothkse de lateralit6 ne s'applique qu'i l'un d'entre eux. Nous suggkrons en fonction de nos resultats qu'il existe des diffbrences entre les sexes au niveau des performances manuelles de certains tests moteurs. Ces differences ne seraient que partiellement dues a des facteurs morphologiques. Pour un de nos tests, le test de l'alphabet kcrit, les sujets ferninins de notre 6chantillon sont plus lateralises que les sujets masculins.

ABSTRACT Sex differences and laterality are not new questions, inasmuch as the research upon these questions are fulfilling new designs. We have surveyed a small sample of undergraduate students to see if we could answer them. We have found that there are sex differences in three manual tests and that the laterality hypothesis applied only to one of them, the written alphabet. We suggest, in view of our results, that there is a sex difference in terms of manual performance on certain motor tasks due in part to morphological factors and that, for one of the tests, the females in our sample are more lateralized than the males.

INTRODUCTION ences entre les sexes en terme de performances L'hypothkse de lateralit6 de mbme que celle manuelles. de difference entre les sexes ne sont pas nou- I1 va de soi que 1'6tude de ces performances velles (Garai et Scheinfeld, '68). Depuis long- serait incomplb te si on se limitait 21 la seule main temps, les chercheurs se sont penches sur ces predominante, puisque I'hypothGse de late- questions. La lateralisation peut btre definie ralite implique un rapport differentiel entre comme une asymetrie fonctionnelle que l'on les deux mains. L'analyse portera sur les perfor- retrouve dans les segments corporels de manihre mances tant de la main droite que de la main prevakente, sans pour autant exclure le r81e gauche ainsi que sur leurs differences. Nous compl6men taire des segments con tra-lateraux sommes en presence, effectivement, de deux (Azemar, '75). Cette asymetrie bilaterale questions dans cette recherche .Premihrement, couvre l'ensemble du corps humain (Levy et existe-t-il des differences entre les sexes au Levy, '78). Nous nous sommes cependant niveau des performances manuelles?, ce qui est limit6 B une region du corps: les mains. C'est Le present texte fait suite a une version presentbe au dans cette opthue que nous organis' une septi&meeongr&s de l'Associationpour i'Anfhropolo- recherche ayant pour but 1'6tude des differ- gie Physique au Canada, Ste.-Adele,QuCbec, 1979.

66 CAN. REV. PHYS. ANTHROP. 2:66-74 LATERALISATION ET PERFORMANCE MANUELLE 67 notre premiere question. La deuxi6me s'6nonce mances entre celles-ci afin de tester l'hypothese en deux volets. Existe-t-il une lat6ralisation de lateralit6 diffhrentielles entre les sexes. effective chez les individus, c'est-&dire une dif- Ajoutons toutefois que, pour la latCralisation, f6rence systkmatique entre le cot6 gauche et le chacune des Cpreuves n'aura pas la meme sig- cot6 droit de leur corps? Si oui, pouvons-nous nification ni le meme micanisme de base dire que cette difference est dissemblable chez (Maccoby et Jacklin, '74; Fairweather et Hutt, les deux sexes? '72; Buffery et Gray, '72). L'argument de depart de notre 6tude, pour les deux questions en cause, est qu'il n'existe Description des kpreuves pas de difference entre les sexes. Ce choix Le test de frappe est simple (Provins et comme point de depart n'en est pas un de Cunliffe, '72). I1 s'agit pour le sujet de frapper facilit6, mais plutat d'ouverture d'esprit et sur une clC le plus grand nombre de coups dYobjectivit6(Maccoby et Jacklin, '74). possible pour une p6riode totale de 30 secondes, alternativement de la main droite et de la METHODE main gauche; ces coups sont enregistrks sur un LY6chantillonde 122 sujets sur lequel porte compteur par p6riode de 10 secondes. Une pra- notre Btude est constitu6 en tres grande major- tique est allou6e aux sujets. I1 est peu probable it6 d'6tudiants sous-gradu6s du d6partement que les sujets aient eu la possibilitk de pratiquer d'Anthropologie de l'Universit6 de Montreal. souvent ce type d'activit6 (Provins et Cunliffe, L'Bge moyen des sujets, tant f6minins que man- '72). Le test de l'alphabet 6crit consiste, pour le sulines, est de 23 ans. Mentionnons que 1'6chan- sujet, A 6crire l'alphabet en lettres d6tach6es le tillon est particulikrement homogene, tant par plus rapidement possible. L76preuveest r6pbt6e la provenance des individus que par leur situa- quatre fois, d'une main et de l'autre alternative- tion prbsente. Les sujets ont subi trois 6preuves: ment. Les temps d'ex6cution sont pris A l'aide le test de frappe, le test de l'alphabet 6crit et la d'un chronombtre, B la fin de chaque alphabet dynamomktrie. Les deux premiers tests mesur- compl6t6; un temps d'arrbt suit. Aucune pra- ent grosso-mod0 la motricit6 et'l'habilet6 man- tique n'est allou6e pour cette bpreuve, 1'6criture uelle. Bien que le test de l'alphabet 6crit soit Ctant une activitb qui siirement est celle la plus plus 6laborb que le test de frappe, il fait aussi pratiqube par le main pr6dominante (Provins et appel B la rapidit6 du mouvement des mains. Le Cunliffe, '72). Pour la dynamombtrie, il est dernier test, la dynamom6trie, mesure la force demand6 au sujet d'appliquer la plus grande musculaire. traction possible avec sa main sur l'appareil de Le choix de ces tests est tout design6 par mesure (Preston hand-dynamometer). Le meil- notre problbmatique. Nous recherchions - en leur de deux essais de la main droite et de la tout premier lieu - les differences au niveau des main gauche sont re tenus pour l'analyse. performances manuelles et, en second lieu, nous cherchions une difference, un biais sys- Methodes statistiques tkmatique entre la main droite et la main Pour d'analyse de donnQesrecueillies, nous gauche, ce qui est en fait une des definitions de avons utilis6 trois mbthodes statistiques: l'anal- la latbralitt. Ce n'est peut-&re pas le meilleur yse de variance simple, l'analyse de covariance indicateur de lateralit6 (Colbourn, '78); mais simple (Snedecor et Cochran, 1967) et l'analyse notre indicateur, c'est-&dire les performances de variance par regression multiple ou l'analyse de la main droite "moins" les performances de de variance multiple (Overall et Klett , 1973). la main gauche, nous offre la possibilit6 d'apprbcier A un degr6 reel la lateralit6 dite RESULTATS fonctionnelle. La majorit6 des individus sont droitiers et ce, Nous avons deux types de variables opera- dans une proportion de 85%. Le reste de toires: les performances de la main gauche et lY6chantillonse partage entre gauchers et ambi- puis celles de la main droite pour lesquelles dextres, respectivement 8% et 7% de 1'6chan- nous tenterons de voir s'il y a des diffbrences tillon total. entre les sexes, et les differences de perfor- 68 YVON DESJARDINS

On retrouve dans notre khantillon une dis- ferentielle. Nous avons retenu parmi ces infor- tribution de preference manuelle (le fait de se mations 35 critbres tels le sexe, la fratrie, l'bge, servir le plus souvent de la main droite ou des indicateurs morphologiques, la preference encore de la main gauche, ou encore Cgalement manuelle, etc. des deux mains) semblable B celle que l'on L'analyse de variance multiple confirme en retrouve dans les autres populations humaines partie les rksultats mentionnes ci-dessus, une (Annett, 1972). difference en fonction du sexe tant pour la main gauche que pour la main droite (tableau 1 ). Les Test de frappe moyennes presentees B l'analyse de variance Les performances des sujets, pour le test de multiple peuvent differer legerement de celles frappe, ont 6te standardisees en nombre de qui sont obtenus Q l'analyse de covariance en frappe par seconde. L'analyse de covariance raison du processus d'ajustement de chaque (fig. 1 et 2) nous montre des moyennes tr6s crit6re B tous les autres du modQle. Notre differentes en fonction du sexe. Les sujets modble explique de 50 Q 60% de la variation de masculins frappent en moyenne 6,l coups par ces deux performances alors que le critere sexe seconde de la main gauche et 6,8 de la droite, n'explique que 1,5076 de la variation de la main tandis que les sujets feminins frappent moins de gauche et que 2,09% de celle de la main droite. coups par seconde: 5,7 coups de la main gauche Notre indicateur de lateralit6 n'est plus sensible et 6,3 de la droite. Ces resultats representent les au critkre sexe. 11 semble vraisemblable que le moyennes des trois periodes prises ensemble et sexe lui-meme, si on le considere Q bge, B mor- non &paremen t. Les moyennes de la difference phologie et Q d'autres conditions Qquivalentes, entre la main droite et la main gauche, indica- ne soit pas la source la plus importante de teur de lateralit&,different peu; 0,61 coups de variation des trois variables operatoires du test difference pour les sujets fCminins et 0,70 pour de frappe. Les facteurs morphologiques sem- les sujets masculins (fig. 3). I1 peut donc exister blent les plus importants dans notre modele. une diffbrence appreciable entre les mains chez Mentionnons aussi un autre critere qui pourrait les deux groupes, si l'on tient compte de l'ordre expliquer, beaucoup plus que le sexe, la vari- de grandeur de la variable en cause. ation consideree : la preference manuelle ou L'estimation par la periode de temps, c'est-6- main preftrentielle (dextralite, ambidextralite dire la regression en fonction des durees et sinistralite). moyennes, est significative tant pour la main gauche que pour la main droite. Nous trouvons Test de l'alphabet ecrit cependant pour la difference droite-gauche une Les resultats du test de l'alphabet 6crit sont estimation non-significative. La regression dans inverses au niveau des performances des deux ce cas n'est pas de bon aloi et la relation entre groupes, relativement au test de frappe (fig. 4 l'indicateur de lateralit6 et la duree n'est pas et 5). Les sujets ferninins ont des temps significative. Bien que la regression de la main d'execution meilleurs, tant de la main gauche gauche et de la main droite le test de frappe soit que de la main droite, comparativement aux significative, il n'y a pas de difference de pente suje ts masculins. en fonction du sexe. Les temps moyens pour completer l'alpha- A l'analyse de variance multiple, nous avons bet par les sujets ferninins sont respectivement constitu6 un modele (variables de contr6le ou 22,ll secondes pour la main gauche et l3,32 independantes) Q partir d'une banque d'infor- secondes pour la main droite. Les sujets mascu- mations contenant des renseignements multi- lins sont plus lents: 25,88 secondes pour la ples sur les individus que nous avons examines. main gauche et 14,38 pour la droite. La differ- Ces renseignements comprennent des donnkes ence entre les deux mains, la main droite socio-demographiques, d'habitudes corpo- "moins" la main gauche, serait plus grande chez relles et alimentaires, d'image de soi (person- les sujets masculins que chez les sujets feminins, nalite et perception de soi), morphologiques et respectivement - 11,50 secondes et -8,79 un relev6 de la main le plus utilisee selon plu- secondes (fig. 6). I1 semble que la plus grande sieurs activitks, ce qui nous donne une indi- difference qui existe entre les sexes soit au cation de la preference manuelle ou main prC- niveau de la main droite. La relation simple LATERALISATION ET PERFORMANCE MANUELLE 69 Source Var. Rap. F I DL1,2 1 P (1-a) source var. ;;: :: DL1,2 P (1-a) Regression 47.91 1,349 Rggression 1,349 0.999988 Df Pente 0.00 1,347 Df Pente 1,347 0.447453 Df Position 19.89 1,348 Df Position 48.82 1,348 0.999999

-Y 6.7 4

5 15 25 Total 5 15 2 5 Total Sec. Sec . Fig. 1 Test de frappe main gauche; : nombre de Fig. 2 Test de frappe main droite; : nombre de coups 3 la seconde; sexe : F - sujet ferninin, M - coups B la seconde; sexe : F - sujet ferninin, M - sujet masculin, T - sujet feminin et masculin. sujet masculin, T - sujet ferninin et masculin. Source Var. I Rap. F I DL1,2 I P (1-a) Source Var. Ra:; yi DL1,2 P (1-a) Regression 1,258 0.999730 Df Pente 1 0.50 1 1,347 1 0.488070 Df Pente 1,256 0.324423 Df Position 1 46.52 1 1,348 1 0.999999 ~f Position 26.39 1,257 0.999982

r 5 15 2 5 Total 1 > \ ~oial Sec . Essai Fig. 3 Test de frappe difference entre main droite Fig. 4 Test de l'alphabet Ccrit main gauche; et main gauche; : nombre de coups de difference; Y : secondes d'exkcution; sexe : F - sujet ferninin, sexe : F - sujet ferninin, M - sujet masculin, T - M - sujet masculin, T - sujet feminin et masculin. sujet feminin et masculin. 70 YVON DESJARDINS

Source Var. DL1,2 P (1-a) Source Var. Raii DL1,2 P (1-a) Regression 1,258 0.999851 Rggression 1,258 0.904422 Df Pente R;!i :; 1,256 0.223541 Df Pente , 1,256 0.496001 Df Position 186.82 1,257 1.000000 Df Position 69.98 1,257 1.000000

-Y

15.5

11.512*0 0 1 2 3 4 Total 13.0 4 , , , Essai 1 2 3 4 Total Fig. 5 Test de l'alphabet Bcrit main droite; Essai Y : secondes d'execution; sexe : F - sujet ferninin, Fig. 6 Test de l'alphabet 6crit difference droite- M - - sujet rnasculin, T sujet ferninin et masculin. gauche; Y : secondes; sexe : F - sujet ferninin, M - sujet masculin, T - sujet feminin et masculin. entre le temps dYex6cutionet le nombre d'essais entre les deux groupes, pour la main droite, a est statistiquement significative pour la main 6t6 augment6 Q la suite de l'ajustement aux gauche et pour la main droite, mais non pour la autres criteres. Lors de cet ajustement, on diffbrence entre les deux. De plus, pour les trois observe une baisse de temps chez les sujets variables, il n'y a pas de diff6rence de pente en f6minins tandis qu'il y a une augmentation de fonction du sexe. Cette absence de diffkrence temps chez les sujets masculins, respectivement de pente chez les deux groupes indiquerait un 13,l secondes Q 7,6 secondes et de 14,4 sec- meme mode de fonctionnement chez les deux ondes Q 20,2 secondes (tableau 2). L'impor- sexes. tance du critkre sexe semble quand mdme toute L'analyse de variance multiple ne nous relative, puisque le R2 est infbrieur Q 1 %. donne pas les mdmes indications que l'analyse La diff6rence entre les deux mains est aussi de covariance (tableau 2). I1 n'y aurait pas de significative en fonction du sexe. Le pourcen- diffCrence entre les deux sexes en ce qui con- tage de variation li6 au critere sexe demeure cerne la main gauche. Bien que notre modele de cependant trds faible. Les r6sultats les plus critbres tende Q expliquer, pour notre 6chantil- 6tonnants que l'on observe avec l'analyse de lon, plus de 90% de la variation de ces trois variance multiple sont ceux de la diff6rence indicateurs, il semble que le sexe des sujets n'en droite-gauche apres ajustement. Nous avions soit responsable que pour un trks petite part. avant l'ajustement une diff6rence de -8,79 Les performances de la main droite montrent, secondes pour les sujets f6minins et de - 1 1,50 apres ajustement, un Ccart encore plus pro- pour les sujets masculins; apres ajustement, on nonce en faveur des sujets ferninins, et ce de obtient - 16,6 chez les femmes et -3,7 chez les fa~onsignificative. LYCcart que l'on observe hommes, indiquant un plus grand kcart entre les LATERALISATION ET PERFORMANCE MANUELLE

TABLEAU 1 Liste des variables optratoires du test de frappe en fonction du sexe d'un tchantillon d'gtudiants x (n=lI 7) (rtsultats ajustts)

Var. oper. FBminin (65) Masculin (5 3) R 2 (%I P(1- 6) R 2 Total (%) Main gauche 5.67 6.16 1.50 0.9982 59.3242 (coups/sec) Main droite 6.26 6.87 2.09 0.9992 (coups/sec) Droite-gauche 0.59 0.7 1 0.09 0.4958

TABLEAU 2 Liste des variables optratoires du test de 1 'alphabet tcrit en fonction du sexe d'un tchantillon d 'ktudiants x (n=65) (rtsultats ajustts)

Var. oper. FCminin (34) Masculin (31) R 2 (%I P ( 1-d) R 2 Total (%) Main gauche 24.2 23.9 0.00 0.0489 93.2623 (sec/es) Main droite 7.6 20.2 0.78 0.9999 93.3571 (secles) Droite-gauche -16.6 -3.7 0.19 0.9753 92.8646 (set)

TABLEAU 3 Liste des variables opkratoires de la dynamomktrie en fonction du sexe d'un tchan tillon d Vtudiants x (n =120) (rtsultats ajustks)

Var. oper. FBminin (63) Masculin (5 7) R 2 (%I P(1- Or) R 2 Total (%) Main gauche 26.0 48.5 14.62 0.9999 91.2489 (kg) Main droite 28.3 53.2 16.67 0.9999 (kg) Droite-gauche 2.26 4.62 1.44 0.8643 YVON DESJARDINS deux mains chez les sujets ferninins que mas- est-il lui-m6me lie A un ensemble de caracter- culins. istiques qui expliquerait cette difference entre les sexes. On ne peut, par conskquent, attribuer Dynamometrie A une lateralisation differentielle les per- Pour la dynamometrie, nous trouvons par formances respectives des deux groupes dans le l'analyse de variance simple des differences sig- cas du test de frappe, c'est-A-dire une fonction nificative~entre les sexes. Tant pour la main motrice assez grossikre (Garai e t Scheinfeld, gauche que pour la main droite, les sujets mas- '6 8). culins sont physiquement plus forts que les L'analyse de variance multiple confirme les sujets feminins. Les performances moyennes resultats obtenus par l'analyse de covariance, des sujets masculins sont, pour la main droite et mais elle nous renseigne sur l'importance des la main gauche respectivement, de 53,3 kg et facteurs responsables de la variation de nos 49 1 kg, et pour les sujets ferninins, de 29,3 kg variables operatoires. Comme nous l'avons fait et 26'6 kg. Pour la difference entre les deux remarquer, l'importance du crithre sexe, mdme mains, nous ne trouvons pas de difference sig- si significative est assez reduite. Les facteurs nificative entre les sexes. Ces differences sont qui semblent les plus importants sont les indi- de l'ordre de 2,7 1 kg pour les sujets ferninins et cateurs morphologiques et la main preferen- de 4,25 kg pour les sujets masculins. tielle. Ces tendances semblent confirmer l'ex- Nous avons obtenu de meme, par l'analyse plication la plus souvent invoquee par les de variance multiple (tableau 3) des differences auteurs, qui veut que les meilleures perfor- significative entre les sexes en ce qui concerne la mances des sujets masculins seraient partielle- main gauche et la main droite. Le critkre sexe ment le fait d'une masse masculaire plus impor- explique en trhs grande partie la variation des tante chez l'homme que chez la femme. Un scores de la main gauche, le R2 ajust6 &ant de autre facteur, anterieur au precedant, serait un 14,62%, ce quiest trhs blevk. Il en est de m6me taux maturationel plus lent chez les sujets mas- pour la main droite pour laquelle le R2 ajuste lie culins (Garai et Scheinfeld, '68). Ce taux plus au sexe est 16,67%. Nous voyons au tableau 3 lent permettrait aux individus de sexe masculin que l'ensemble de notre modkle explique plus de developper, pour ce genre d'activite, des de 90% de la variation des performances des voies neurales plus efficaces, c'est-A-dire une mains droite et gauche. La difference droite- meilleure ou plus grande myklination des gauche par contre, n'est pas statistiquement lie neurones (Connolly , '70; Buffery et Gray, '72). au sexe. I1 semble que le crithre sexe ne soit pas De plus, il determine en partie une masse mas- prkpond6rant dans l'explication de la variation culaire et osseuse plus importante. Suite A ces de la difference droite-gauche. Les facteurs constatations, on ne peut veritablement parler morphologiques semblent beaucoup plus lies A dans le cas du test de frappe d'une latkralisation la variation de cet indicateur, leur R2 groupe differentielle en fonction des sexes. ktant supkrieur A 1 1%. Test de l'alphabet ecrit DISCUSSION La tendance qui se degage des rksultats Test de frappe obtenus pour le test de l'alphabet 6crit repre- sente l'inverse de ce qui a 6te observe pour le L'absence d'une difference de pente nous test de frappe. Nous ne trouvons aucune differ- suggkre, pour les variables du test de frappe, que ence de pente, ce qui suggQre fortement un la dkcroissance de la performance serait kqui- valente chez les deux sexes pour cette tranche rappel de l'alphabet semblable chez les deux sexes. Au niveau de la main gauche, nous ob- d'iige (Garai et Scheinfeld, '68;Fairweather et servons des performances sem blables chez les Hutt '72). Comme nous l'avons mentionn6, il deux sexes, tandis que pour la main droite les semble qu'on ne peut trouver de difference en sujets feminins ont des performances plus terme de lateralit6 pour cette fonction motrice. rapides que celles des sujets masculins, et ce Ceci &ant dit, on observe tout de mdme une A difference de performances des mains gauche et Gge et A morphologie equivalents. On retrouve au niveau de l'indicateur de latkralitk une plus droite entre les sexes. Sans doute, le critkre sexe LATERALISATION ET PERFORMANCE MANUELLE 73 grande difference chez les sujets fCminins que Dynamometrie chez les sujets masculins, ce qui semble suggCrer Nous trouvons, a la dynamomktrie, des dif- une plus grande latkralisation chez les femmes. fCrences entre les sexes tant pour la main droite Toutefois, selon nous, le test de l'alphabet que pour la main gauche. Le sexe des sujets comporte trois composante: la memoire de semble un facteur important dans la variation l'alphabet , l'alphabet en tant que structure des performances. Toutefois, d'autres facteurs verbale, et la dextCritC manuelle. I1 a CtC montrC tels la morphologie, habitudes corporelles, etc, par de nombreux auteurs (Garai et Scheinfeld, manifestent aussi leur influence. I1 semble que '68; Maccoby et Jacklin, '74) que les femmes la diffCrence droite-gauche, soit sensiblement la obtiennent de meilleurs rCsultats au niveau de la mdme chez les deux sexes. mbmoire (en terme de rappel), et de la dextBritC Les meilleures performances des sujets mas- manuelle. Dans ces deux types d7activitCs,il ne culins pour la dynamomktrie semblent Btre une semble pas que l'hypothbse de lateralit6 puisse question partiellement morphologique. Le fait s'appliquer. Si l'on reconnait que l'alphabet d'appliquer une meilleure pression est fonction constitue une structure verbale, l'hypothkse de de l'ossature et de la musculature. L'effet de 1atCralitC diffkrentielle pourrait s'appliquer. I1 maturation diffkrentielle des sexes peut en Btre semble que le sexe fCminin ait de meilleures une des causes principales, ce qui implique que performances en ce qui concerne la verbal- les sujets masculins ont plus de chances d'un isation, c'est-a-dire une certaine facilit6 verbale plus fort d6veloppement osseux et musculaire (Maccoby et Jacklin, '74; Garai et Scheinfeld, que les sujets ferninins (Garai et Scheinfeld, '68). La verbalisation serait fortement later- '68). Pour la dynamomktrie, il n'est pas pos- alis6e. Le sibge de la parole ou de 1'Clocution de sible, selon nos rCsultats, de parler de latbralisa- la parole se situe dans l'hbmisphbre cCrCbral tion diffCren tielle. gauche (Oppenheimer, '77). Ces dernibres annCes une hypothbse a CtC mise de l'avant CONCLUSION concernant la latkralisation chez les sexes. I1 est dit que les femmes prksenteraient une latkralisa- I1 a 6tC montrC que les sujets de sexe mascu- tion moindre, voire une bilatkralisation hemis- lin, au seul niveau de la vitesse montrice, ont des ph6rique des fonctions linguistiques, alors que performances plus rapides que les sujets fCmin- les hommes montreraient une latbralisation ins. Toute Cpreuve de rapidit6 neuro-motrice, plus nette de ces rrlCmes fonctions (De Agostini, d'une activitC peu spCcialisCe, favorise les hom- '79). Cette hypothbse est en contradiction avec mes de l'adolescence a la senescence. L'exemple nos rCsultats, qui semblent indiquer que les le plus adCquat de ce type d7activitCest le test sujets de sexe fCminin seraient plus 1atCralisCs de frappe. Le test de l'alphabet Ccrit est un test que les sujets de sexe masculin. Nos rCsultats, plus complexe. I1 fait appel a plusieurs fonc- cependant, sont obtenus par la technique de tions, entre autre la plus importante selon variance multiple, qui "ajuste" l'effet principal nous, le langage Ccrit. Plusieurs Ctudes ont du critbre sexe a l'ensemble des effets princi- montrC que les sujets fCminins Ctaient favorisbs paux des autres critbres du modble (iige, mor- au niveau de langage (Maccoby et Jacklin, '74). phologie, socio-dkmographie, etc). Sans cet Le lien entre le contr6le de la main droite et ajustement, tel que par exemple obtenu par le les fonctions linguistiques a aussi CtC mis en technique plus conventionnelle de l'analyse de Bvidence (Annett, '77). I1 n'est donc pas 6ton- covariance et tel quYillustrC dans nos propres nant d'avoir trouvC de meilleures perform- rbsultats utilisant cette dernikre, ce sont les ances chez les femmes dans notre Cchantillon. hommes qui "paraissent" manifester une datCr- Les masses musculaire et osseuse plus impor- alisation plus prononp5e. Ce revirement com- tantes chez l'homme expliquerait en grande plet de la situation suite a la technique plus partie les meilleures performances de ces sophistiquC de l'analyse de variance multiple, derniers A la dynamombtrie (Garai et en plus de jeter un Cclairage nouveau sur le Scheinfeld, '68). problbme de la 1atCralisation entre les sexes, Dans lYCtudedes performances manuelles, il nous invite a une plus grande prudence lors de y a deux notions qui dkcoulent de la lateralisa- lYinterprCtation de mCthodes univarikes e t tion: la latkralisation fonctionnelle, c'est-a-dire bivariCe s. 74 YVON DESJARDINS l'organisation specifique au niveau des hemis- ~ze'mar,G. 1975 Lateralit6 et differenciation qualita- pheres cCrCbraux et la latCralisation usuelle, tive des conduites motrices. Revue de Neuro- c'est-a-dire la main pr6fCrentielle sur laquelle psychiatrie Infantile 23: 13-2 1. nous n'avons pas ax6 notre etude. Nos rksultats Buffery, W. H. and J. A. Gray 1972 Sex differences in peuvent nous aider prbciser notre principale the development of spatial and linguistic skills. In: a Gender Differences: Their Ontogeny and Signifi- question: la lateralisation fonctionnelle des cance. C. Ounsted and D. C. Taylor (eds.). sexes. Le test de l'alphabet bcrit tend A montrer Churchill, London. une plus grande latbralisation chez les femmes Colbourn, C. J. 1978 Can laterality be measured? et une certaine bilateralisation chez les hom- Neuropsychologia 16: 283-289. mes. Cette plus grande lateralisation a l'hbmis- Connolly, K. (ed.) 1970 Mechanisms of Motor Skill phhre gauche, tient au test lui-m6me mettant en Development. Academic Press, New York. evidence le lien entre le contr6le de la main De Agostini, M. 1979 Hommes et femmes ont-ils la droite et le langage bcrit. Les autres tests ne meme organisation cerbbrale? La Recherche montrent pas cette tendance. 10:77-79. Fairweather, H. and S. J. Hutt 1972 Sex differences in Les Ctudes sur les differences entre les sexes perceptual-motor skill in children. In: Gender Dif- deviennent de plus en plus frCquentes. Les ferences: Their Ontogeny and Significance. C. reticences d'un cote comme de l'autre sont Ounsted and D. C. Taylor (eds.). Churchill, grandes. Afin de minimiser les inbgalites entre London. les sexes, il faut connaftre 1'6tendue et les Garai, J. E. and A. Scheinfeld 1968 Sex differences in limites des capacites de chacun des sexes pour mental and behavioral traits. Genetical Psychology en apprecier A leur juste valeur ce qui est com- Monographs 77: 169-299. mun. Les sexes diffkrent, effectivement, mais Levy, J. and J. M. Levy 1978 Human lateralization n'en manifestent pas moins une communalitb from head to foot: sex related factors. Science 200: 1291-1292. de capacitk. Maccoby, E. E. and C. N. Jacklin 1974 The Psychology of Sex Differences. Stanford Univer- REMERCIEMENTS sity Press. Nous tenons remercier le Dr. Francis Oppenheimer, J. M. 1977 Studies of brain asymmetry: A historical perspective. Ann. N.Y. Acad. Sci. Forest de ses conseils et de ses critiques genti- 299:4-17. ment apportes a l'elaboration de cet article. Overall, J. E. and C. J. Klett 1972 Applied Multi- variate Analysis. McGraw-Hill, New York. OUVRAGES CONSULTES Provins, K. A. and P. Cunliffe 1972 The reliability of some motor performance tests of handedness. Annette, M. 1972 The distribution of manual asym- Neuropsychologia 10: 199-206. metry. Br. J. Psych. 63:343-358. Snedecor, G. W. and W. G. Cochran 1967 Statistical 1977 Handedness and the cerebral representa- Methods. Iowa State University Press, Ames. tion of speech. In: Physiological Variation and its Genetic Basis. J. S. Weiner (ed.). Halsted Press, New York. Book Review cal apparatus into the field in 1971, particularly when the skeletal material itself was scattered and in poor condition. In fact, the author has AN EARLIER POPULATION OF HESQUIAT done an efficient job of wringing as much infor- HARBOUR, BRITISH COLUMBIA. By mation (and speculation) as possible from what Jerome S. Cybulski. Cultural Recovery might appear to be a rather "miserable" collec- Paper No. 1 ; Physical Anthropology Contri- tion of bones which, even 20 years ago, would bution No. 1, Hesquiat Cultural Committee; have been virtually ignored by a field investiga- British Columbia Provincial Museum, tor to the exclusion of skulls only. Neverthe- Victoria. 1978. 80 pp., 12 figures, tables, less, a report of this type is notably lacking in references, appendices. $2.00 (paper). one area that would be expected for bones we will never see again, and that is photography. This work represents the first in a series of Although the included drawings by Brian D. volumes which are intended to be summaries of Seymour are of excellent quality, I would have a scientific investigation known as the Hesquiat liked to see more photos of the general Project. Hesquiat Harbour, located on the Hesquiat area, the burial sites, and specific southwest coast of Vancouver Island, is the bones. traditional territory of the Hesquiat peoples, a General organization of the report is logical localized group of the Nootka ethnolinguistic and concise. In chapter I the site and sample are division of the Northwest Coast culture area. introduced and the aims of the analysis set out. Because of increasing vandalism of historic Chapter 11, basically an apologia, discusses the burial sites the Hesquiat Band formed a Cul- problems the investigator had to deal with tural Committee in 1970 to prevent the com- because of the condition of the skeletal remains plete destruction of the burials and to attempt and the additional disturbing influence of the to reconstruct and preserve as much of their practice of cranial deformation. Chapter 111 heritage as possible. This physical anthropologi- includes the description of individual burials by cal study represents Dr. Cybulski's attempt to site, a necessary exercise, it seems, in all site reconstruct information from the skeletal reports that makes for tedious reading and material by recovery (begun in 197 1) and on- which I for one have often wished to be rele- site data analysis before reburial in a crypt at gated to an appendix. Chapters IV to XI pre- the village of Hesquiat. Consequently, the work sent the results of analyses of age and sex esti- represents one of a present few in an increasing mations, further details about cranial deforma- number of future reports where skeletal mater- tion, cranial morphology, dental morphology, ial is expected to be completely analysed on- non-metric postcranial morphology, stature, site before reburial. and dental and skeletal pathology. Chapter XI1 Having gone through the same experience attempts interpretive analyses reconstructing with a skeletal sample which could not be re- demography, making physical comparisons moved to a professional lab for analysis, I sym- with other groups, and assessing health status. pathize and congratulate Dr. Cybulski on his My criticisms of various details in Chapters I efforts at the skeletal reconstruction, and here- to XI are few but bear some mention. Some of in lies the publication's strong point. It is an the aging and sexing methods seem a bit out- admirable model for what physical anthro- moded. The Schour and Massler dental erup- pologists can accomplish while working under tion aging chart could have been updated but, the enforced restrictions set by native peoples of course, was still in wide use in 197 1. No regarding ancestral burials. Any conscientious mention is made of the Phenice method of investigator would recognize the "after the sexing hip bones which was published in 1968. fact" regrets one has in thinking about what Reliance was said to have been placed on the information might have been missed and what form of the sciatic notch, a feature known to be one could have done, but didn't, while in the quite variable in terms of sex determination. field. Therefore, I cannot criticize the author The problems associated with aging females for errors of omission, such as neglecting to from pubic bones are not discussed. take portable x-ray equipment or other techni- 76 BOOK REVIEW

In the section on non-metric post-cranial a mortality profile. The profile for Hesquiat morphology, the trait descriptions for the atlas Harbour compares well with published mortal- vertebra are somewhat confusing between the ity profiles of other prehistoric groups and the text and Table 9. In addition, the author con- author concludes that the burials represent tinues to follow the tradition set by other individuals who were living under relatively physical anthropologists in simply mentioning favorable conditions just prior to the decimat- certain traits and their frequencies (e.g., Fossa ing influences of white contact. Cybulski also of Allen) without considering some of the diffi- presents a number of arguable but feasible culties in subjective assessments of trait pre- interpretations for an apparent lower mortality sence. One should note, though, that Cybulski rate for Hesquiat females. at least considers how available sex, side, and The attempt to assess the biological features age differences might be affecting the presence of the Hesquiat sample compared to other of some traits. Nootkan groups using cranial non-metric traits More details and traits might have been is interesting but seems to me, fruitless. Cybul- examined for dental morphology, and this is ski's statement that the MD~distance statistic one area where future field investigators would is sophisticated and sensitive rather than robust due well to consider taking dental cast impres- has been contested elsewhere, despite the fact sions in the field. The statement that osteo- that the MD~equation seems to adequately arthritis is by far the most common of patho- reflect biological separation of populations. logical manifestations in prehistoric skeletal Cybulski himself admits that his cranial samples material is self-evident and expected but it are very small (6 from Hesquiat!) and this in would have been interesting to see a compari- itself reduces the confidence in his conclusions. son of the degree and distribution of degenera- He concludes that the lack of linearity in MD~ tive joint disease in this sample compared to distances when Hesquiat Harbour is compared other populations such as Jurmain (1 977) has to the four other Nootkan samples indicates done. that Hesquiat Harbour was not a distinct breed- Of far more interest than these small points ing population while the four other samples are the author's summary statements about this were. It seems to me he is pushing his conclu- skeletal sample in Chapter XII, the Interpretive sions a little too far. Nevertheless, this report is Analyses. However, in all sections of this chap- a useful adjunct to skeletal studies of prehis- ter it is apparent that the basic problem in toric and early historic Northwest Coast popu- making general interpretations is the incom- lations. pleteness of the sample. Nevertheless, Cybulski presents some cogent arguments, backed by S. R. Saunders University of Toron to ethnohistoric references to say that the skeletal remains represent members of the prehistoric local lineage group, the ma'apiath. He then LITERATURE CITED reconstructs population size from what was Jurmain, R. D. 1977. Stress and etiology of osteo- known of the size of the ma'apiath to calculate arthritis. Am. J. Phys. Anthrop. 46:353-366. ABSTRACTS OF PAPERS PRESENTED AT THE SEVENTH ANNUAL MEETING OF THE CANAD IAN ASSOCIATI ON FOR PHYSICAL ANTHROPOLOGY/ ASSOCIATION POUR L' ANTHROPOLOG IE PHYSIQUE AU CANADA

STE ,-ADELE 0 QUEBEC 0 NOVEMBER 7-10 0 1979

VARIATION DES NIVEAUX DE SANTE DANS UNE POPULATION DE TRAVAILLEURS QUEBECOIS - I FRANKLIN AUGER Unioersitk de MontrkaZ La santg et la securitg au travail sont devenues des th'emes importants dans nos soci- 6t6s industrialisges. A ce stade-ci de nos analyses, nous ne pouvons prgsenter que quelques glgments descriptifs de la variation morphologique de dew Echantillons de travailleurs qu6b$cois, de meme que de la variation de certains indicateurs des niveaux de santg d'un seul de ces deux 6chantillons. Le premier gchantillon regroupe 438 individus (employ& municipaux de la Ville de Montrgal). Cet 6chantillon a dtg constitug par des chercheurs de llInstitut de Cardiologie de Montrdal au d6but des annges soixante. Notre deuxisme dchantillon est form6 des employ& d'une importante entreprise de la rdgion de Sorel-Tracy. C'est cet dchantillon qui constitue llessentielde la recherche que nous poursuivons prg- sentement: 3 000 travailleurs ont dtg examin6s. Le poids, la stature et le pourcentage de tissus adipeux ont dt6 analysgs succinte- ment en fonction de lt8ge et de la catggorie socio-professionnelle chez les travailleurs de Sorel-Tracy et en fonction de 1'Bge et des niveaux de scolaritg chez les employ& municipaux de la Ville de Montrgal. Chez les travailleurs de Sorel-Tracy, nous avons retenu divers sympt6Bes rattach& 5 des systsmes particuliers et calcul6s 3 partir de nombreuses r6ponses contenues dans le questionnaire pr6-examen mgdical, r6ponses pondkrges par le mgdecin lors de l'examen clinique. Quelques r6sultats ophtalmomgtriques, audiomgtriques et spiro- mgtriques, de mEme que la tension systolique et diastolique sont donngs. Les variables physio-pathologiques sont consid6r6es come des indicateurs de sant6 et ce, pour les diverses catggories socio-professionnelles. Les dgviations prgsentkes par les travailleurs d'une catggorie et ce, pour l'ensemble des variables consid6rges, le sont par rapport 3 la moyenne ggngrale de notre population de travailleurs. VARIATION DEX NIVEAUX DE SANTE DANS UNE POPULATION DE TRAVAILLEURS QUEBECOIS - I1 DIANE BERTHELETTE Universit6 de MontrkaZ (Complete paper published in this number)

LES CONNAISSANCES ANATOMO-PHYSIOLOGIQUES DES ANGBANDI DU ZAIRE GILLES BIBEAU UniversitS Lava2 Depuis Ackerknecht (1946) , historien de la mddecine et anthropologues mddicaw rdpk- tent que les mgdecines traditionnelles se caractdrisent par une absence de connai- sances anatomo-physiologiques. Cela expliquerait dtailleurs pourquoi leur thdrapie est dlabord symptomatique et leur 6tiologie magic-religieuse. Aprds avoir mis en 6vidence la base qui a permis ltdmergence dlune vgritable science du corps chez les Angbandi, je prdsenterai les structures organisatrices de leur systgme de connaissances physiologiques (0s et organes internes principalement). Je terminerai en donnant les caract6ristiques de la science du corps chez les Angbandi.

INTER-RELATION DE LA CROISSANCE DES DIFFERENTES COMPOSANTES DU CORPS0 MUSCLES ET TISSUS ADIPEUXt ET LA CONSOMMATION ALIMENTAIRE CHEZ DES ENFANTS CANADIENS-FRANCAIS DE 6 A 19 ANS MICHELINE BRAULT-DUBUC et ART0 DEMIRJIAN Universit6 de Montrgal Dans le cadre d'une 6tude longitudinale, dew cohortes dtenviron200 garsons et 200 filles ages de 6 ans et de 10 ans en 1967 ont dtd examin6es annuellement jusqu'en 1976. Les paramstres utilisgs pour l'analyse prdsent6e ici sont: la stature (squelette), la circonfdrence du bras maigre (muscles), la sommejde trois plis cutangs (tissus adi~eux), la valeur de la diste en calories, protdines, lipides, glucides et au calcium, bas& sur un relevd alimentaire de 7 jours. Les courbes de grandeur et de vdlocitd ont dtd utilisges sur des kchelles dtbge chronologique et dt&ge de vdlocitd maximale de la stature (VMS). On a pu ddcrire les changements dans la composition du corps pendant l'adolescence partir de simples mesures anthropomdtriques. Le dimorphisme sexuel observd est plus important que pour d'autres populations. La croissance staturale se fait parallglement la croissance musculaire chez les garscons et leur vdlocitd maximale concide; tandis que chez les filles, il y a un an de diffdrence. Les corr6lations dtablies des ages donnks chez les garscons entre la croissance staturale et musculaire ntexistent plus lorsqu'on retire le facteur de la maturation, ce qui n'est pas le cas chez les filles. On a pu, d'autre part, observer une augmentation peu importante mais rdelle de la consommation alimentaire, jusqut$ 12 ans chez les filles et jusqut$ 15 ans chez les garqcons; mais aucune corrdlation pour les grandeurs ou les vdlocitds n'a pu etre 6tablie entre les diffsrentes composantes du corps et l'ingestion alimentaire. (Octrois nos. 605-1388- 41 et 360-29-000-305 du ~inistkrede la Santd et Bien-Stre Social)

LATERALISATION ET DIFFERENCES ENTRE LES SEXES WON DESJARDINS Universit& de MontrSaZ (Complete paper published in this number) PREDICTION DE,LA TENSION ARTERIELLE PAR ANALYSE DISCRIMINANTE MULTIVARIEE D ECHELLES PSYCHOSOCIALES COGNITIVES CHEZ L HOMME SHIMON DOHLAN et ANDRE ARSENAULT UniversitS de MontrkaZ Dans le cadre d'un projet de recherche sur le stress au travail r6alis6 auprgs des travailleurs du milieu hospitalier, les auteurs ont test6 une hypothsse gdn6rale voulant que la perception consciente du contenu et du contexte de la tbche, de m6me que les sympt8mes et des signes ressentis par l'individu, pouvaient permettre de pr6- dire la r6ponse du systsme cardio-vasculaire, telle que mesurge objectivement par la mesure de la tension artgrielle. En utilisant un modgle multi-dimensionnel, on a cherchs a discriminer les individus ayant une tension artsrielle diastolique au repos infsrieure b 75 ml de mercure des individus dont la diastolique 6tait supgrieure & ce meme seuil. Des rgsultats prgliminaires obtenus b partir de l'gtude de 60 cas indiquent que l'on peut discriminer correctement dans 82% des cas les individus qui se situent au-delb et en-desa du seuil b partir de quatre facteurs dont dew sont li6s au contexte de la tgche et les dew autres b la symptomatologie psychologique consci- emment persue. Si ces rssultats sont confirm& dans les analyses ultdrieures, ils indiqueraient que la rgaction physiologique objective et inconsciente chez l'humain pourrait &re prsdite partir de la psychs consciente.

PROGRESSION ORDER IN BONNETS AT BANNERGHATTA GEORGE ELLIS Unioersitb de MontrgaZ The spatial and temporal patterning of food resources in Bannerghatta National Park provide the bonnet macaque with a productive mosaic environment which they utilize in a fine-grained manner. Distinctive patterns of group dispersion and social behavior appear a8 the monkeys exploit alternative resource commodities. Group progression order is one such pattern. This paper examines changes in one-dimensional group geo- metry associated with movements within and to the outskirts of the troop's home range. The social and communicative matrix of such progressions is described. The implications of this study for the "protective function1' hypothesis of spatial organization of moving troops are discussed.

IMPLICATIONS OF CERTAIN HAND-WRIST VARIATIONS FOR SEX AND GROWTH C. E. EYMAN University of CaZgary (Complete paper pubZished in this number) FRANCIS FOREST et URSULA FOREST-STREIT ~niversitsde MontrSaZ (Complete paper pubZished in Vol. 2, No. 21

ETUDE ANTHROPOBIOLOGIQUE D' UNE POPULATION FEMI NINE CANADI ENNE-FRANCAISE DU QUEBEC GLORIA FRAPPIER UniversitS de MontrkaZ Au Canada, il existe relativement deu d'dtudes anthropobiologiques de la population, et notamment en ce qui concerne la population fgminine canadienne-frangaise. Au cows des anndes 1978-79, nous avons recueilli de nombreuses informations auprds d'un dchantillon de 206 femmes canadiennes-franscaises de la rkgion de Sorel-Tracy, dont l'6ge variait entre 18 et 65 ans. Les informations obtenues 2 l'aide d'un questionnaire socio-dsmographique et mkdical ainsi qu'a partir de mesures anthropo- mgtriques et physiologiques, nous ont permis de crder une banque de donndes qui constitue, 2 notre connaissance, la premihe et la seule disponible 3 ce jour pour ce type de population. Dans un premier temps, nous avons analysk de fqon descriptive l'dchantillon fdminin qu6bdcois tant au niveau des variables socio-d6mographiques qu'anthropomdtriques et physiologiques. Par la suite, nous tenterons d16valuer l'6tat de sant6 g6n6ral de la population feminine de la r6gion Sorel-Tracy ainsi que d'en 6tudier les facteurs de variation.

ETUDE DE L' AUDIOMETRI E CHEZ UNE POPULATION DE TRAVAI LLEURS NORMAND LAPLANTE Unioersit6 de MontrkaZ L'dtude portera sur les r6sultats au test d'audiomgtrie effectu6 auprss d'une population de travailleurs en milieu industriel. Les frgquences analys6es vont de 250 Hz $ 8000 Hz et le seuil d'audition est 6valu6 par intervalles de 5 dB. Une premisre analyse dhontre une forte variabilitg des donndes, caractgristique d'une population soumise 2 des niveaw de bruit assez 6lev6s. L'analyse en composante principale des huit frdquences ktudi6es permet d'identifier deux facteurs principaw responsables de la plus grande partie de la variation. Le premier facteur, expliquant 66% de la variation semble dtre un facteur d'exposition au bruit. Le second facteur, responsable de 16% de la variation serait un facteur de vieillissement physiologique. L'dtude du seuil d'audition en fonction de 1'Bge ddmontre des diffgrences importantes entre les classes socio-professionnelles. Ces diffdrences sont les plus marqudes pour les frdquences supdrieures 2 1000 Hz. La classe la plus 6levEe, soit celle des cadres, se rapproche d'une population standard non-exposge 2 des sources de bruits 6levds. Les employks manuels et les contremaftres prgsentent les rgsultats les moins bons tandis que les employks de bureau se situent dans une position intermgdiaire. Ces diffdrences entre classes peuvent stexpliquer par une exposition 2 des niveaw de bruits diffgrents et sans doute ggalement par un vieillissement physiologique diffg- rentiel. RAPPORT PRELIMINAIRE SUR UNE ETUDE DES CRANES DU SITE DAWSON ROBERT LAROCQUE Universitd de Montr6aZ (Complete paper published in this number)

DEVELOPPEMENT DE L' INTELLIGENCE CHEZ LE CHI MPANZE MIREILLE MATHIEU Universitd de MontrSaZ (Abstract not submitted)

EVOLUTION DE LA STRUCTURE GENETIQUE D'UNE POPULATION OUVERTE FRANCINE MAYER Universitd de MontrdaZ L16tude de la migration et de ses m6canismes op6rationnels constitue un d6fi: la migration impliquant une mobilit6 constante des individus, elle ne peut &re que difficilement mesurge. Cette recherche a tent6 d16valuer la nature et ltintensit6 du flux migratoire d'une communaut6 des Cantons de 1'Est fond6e en 1875 en faisant l'analyse socio-demographique et l'analyse g6n6tique. Une m6thode de recoupements des sources dlinformation a permis une analyse plus com- plhe de la migration. L'approche m6thodologique utilis6e dans l'analyse g6n6tique est bas6e sur le calcul des probabilitgs dlorigine des individus estim6es partir des contributions g6n6tiques respectives des ancetres fondateurs.

PRELIMINARY EVIDENCE FOR AN ALTERNATIVE MODEL OF RHESUS MONKEY SOCIAL STRUCTURE ON CAY0 SANTIAGO CAROL McMILLAN State University of New York, BuffaZo Although much excellent research has been done on the rhesus macaques of Cayo Santi- ago, most of it has been fit into a model of social structure based on a dominance hierarchy. Even after kinship was seen to be a primary organizing factor, as much emphasis has been placed on how matrilineages fit into the dominance hierarchy as on a more branching view of kinship itself. Many studies have been primarily e7?ic, based on categories defined by researchers previous to their contact with the monkeys. The model I propose in no way invalidates previous research, but merely offers an alternative framework for the organization and analysis of data. Using Gearing's (1958) human ethnographic model and his current (1979) methodology, video tapes were made on the monkeys in a month-long pilot study on the island. Behavioral routines were sought which, by their patterning, could be seen to define emit categories of the monkeys. Preliminary evidence indicates that rhesus monkey social structure on Cayo Santiago may be explained by a nesting series model based on em& categories of activities and kinds of individuals. DEMOGRAPHY AND PATHOLOGY OF THE GANJ DAREH POPULATION: EARLY NEOLITHIC OF IRAN C. MEIKLEJOHN University of Winnipeg P. LAMBERT, C. BYRNE, and M. H. REED University of Manitoba The site of Ganj Dareh lies in the Kermanshah district of the Inner Zagros zone of Iran at an altitude of 1400 m. It is a small mound of ca. 40 m diameter with five levels, dating primarily from the 7th and 8th millennia B.C. It appears to span the boundary from hunting and gathering to a food producing economy and has yielded among the earliest dated ceramics in the Middle East. The site has particular interest in the recovery of the skeletal remains of 49 individuals, one of the largest reported collections of early Neolithic remains from the eastern flank of the fertile crescent. Both sexes and all age classes are represented, though it is evident that few individuals reached old age. Study of the postcranial skeleton suggests a relatively low degree of sexual dimorphism in the sample, as compared to other regional groups. The dentition is characterized by a relatively large surface area, as especially seen in large maxillary molars and mandibular premolars. There is general size equivalence to contemporary Natufian samples. Human interference is noted in the presence of artificial cranial deformation. Pathology is most evident in hyperostosis eymnetrica, found in over 50% of the adult series. Postcranial changes and traumatic injury occur infrequently. A radiological study of the 26 most complete individuals confirms the generally healthy status of the population. Harris lines are present but in relatively low numbers. One individual, a 9 mo. old child, shows evidence of generalized metabolic disorder. The relatively young age at death and good health profile is similar to that of hunter-gatherers and differs from later proto-urban and urban populations.

MANDIBULAR TORUS: A SYNTHESIS OF NEW AND PREVIOUSLY REPORTED DATA AND A DISCUSSION OF ITS CAUSE N. S. OSSENBERG Queen's University Since publication of Hrdlicka's (1940) treatise on mandibular torus, theories concerning its cause have shifted emphasis from masticatory stress and mucosal irrita- tion to heredity. Pedigree studies have clearly shown that genetic factors influence torus development. Yet despite numerous investigations, this continues to be one of the most controversial and poorly understood of skeletal traits. Using original data on 1,500 mandibles, but mainly previously published data, I present an overview of the distribution characteristics of mandibular torus and a hypothesis concerning its cause. Extrinsic factors are strongly implicated by: prevalence among Arctic peoples, effect of dietary change, age regression, preponderance in males and on the right side, influence of cranial deformation, concurrence in individuals with palatine torus and maxillary alveolar exostoses, and clinical evidence. I propose that the primary factor is masticatory stress. According to a mechanism suggested by orthodontic research, the horizontal component of bite force tips the lower C to M, so that their root apices exert pressure on the periodontal membrane, causing formation of new bone on the lingual cortical plate of the alveolar process. Once formed, the hyperostosis is vulnerable to trauma and its periosteal covering becomes bruised, causing additional deposition of bone. Genes influence torus development indirectly through their effect on occlusion. A pattern of increased expressivity with incidence suggests that a quasi-continuous model msy provide a better fit to pedigree data than single-locus models previously tested. PASSIVITY AS AN ADAPTIVE RESPONSE ANDREW J. PETTO University of Massachusetts Various ethnographers of Canadian Native Peoples have noted that a common response to new stressors is passivity or inaction. Explanations of the frequency of this behaviour have centred upon the psychological make-up, the need for cooperative behaviour for survival, or the restraints of the native belief systems. Environmental factors have generally been treated by noting the harshness of the conditions the Native Peoples must face. Focusing especially upon the Athapaskan peoples of the MacKenzie Drainage, this study illustrates the high predictability of resources throughout the year based upon native reports and observations of ethnographers. Using both least risk and expected cost to benefit ratio models, it is the very predictability of the environment which makes nonaction a viable strategy in the face of unfamiliar and/or short-term stressors. The organisational and behavioural aspects of such a generalist strategy can be contrasted with those of anadromous fish specialists. Responses similar to the latter would be predicted as the focus for the subsistence strategy in response to external demands for intensive harvesting of single, seasonally available resources.

A STUDY OF POSSIBLE INFLUENCE OF ENVIRONMENTAL POLLUTANTS ON CHILD GROWTH SUSAN PFEIFFER University of GueZph Members of the St. Regis (Akewasasne) Mohawk band, St. Regis, Quebec, have solicited the co-operation of the scientific community to quantify what they believe to be a serious problem of high pollution levels. Of the three local industrial pollutants, PCB's, Mirex and air-borne fluorides, the latter has received the most attention. A study of cattle on Cornwall Island substantiated the presence of chronic fluoride poisoning on at least that part of the reserve. Early in September, 1979, I collect- ed anthropometric and dental emergence data from 235 randomly chosen children, aged 2 to 14 years, who live on the reserve. Subsamples from high-, moderate- and low- pollution areas of the reserve were compared to one another. Preliminary results indicate that the external indicators of fluoride poisoning seen in the cattle (short stature, delayed dental emergence) are not apparent in Cornwall Island children.

INFANT MORTALITY AMONG THE SEPHARDIC JEWS OF GIBRALTAR: 1869-1977 L. A. SAWCHUK and L. E. FLANAGAN University of Toronto (CompZete paper pubZished in VoZ. 2, No. 2)

DENTAL ERUPTION SEQUENCES IN FOSSIL HOMINIDAE MARK SKINNER Simon Fraser University Recently, it has been suggested by a number of palaeoanthropologists that comparative dental eruption sequences amongst recent and fossil Hominoidea, particularly in re- gards to molar retardation, may be utilized as an indirect means of documenting when, during hominid phylogeny, there occurred a prolongation in the absolute time required for our ancestors to mature. Because of the significance in humans of prolonged maturation for the acquisition of culture, it was decided to undertake a study of the fossil evidence for dental eruption sequences in the Hominidae (N = 133). Comparison was made to 90 Pan and 71 CoriZZa as well as to published data on various primates. Results strongly suggest that within the Hominoidea, despite striking differences in the dental maturation interval (ca. 12 years for pongids and 20 years for recent humans), molar retardation is trivial. It is concluded that dental eruption sequences cannot be used to infer maturation rates in fossil hominids and that, rather, molar retardation reflects the genetic consequences of posterior molar reduction in a lineage evidencing diminution of the craniofacial complex. Support for this research ws provided by Natwnaz Research Council, Boise Fund (Oxford), and Canada CounciZ ShaZZ Grants.

DEMOGRAPHIC CONSIDERATIONS IN HUMAN BIOLOGICAL FIELDWORK LOUISE K. STEIN State University of New York, BuffaZo Thorough demographic analysis is critical to many areas in human biological research. Much of the research in human biology, however, does not give sufficient weight to the subjects' responses as indicators of the cultural rules that structure the human biological framework. As an example of this problem, Polish village women (N = 120) were asked birthdate, age at marriage, and age at birth of first child. In order to obtain more complete information, village records were used to supplement informants' responses. During this process, a significant number of discrepancies between the informants' replies and the village records were discovered. The discrepancies indicate the villagers' varying reactions in reporting the dates of these events to different institutions. The discrepancies also indicate that a scientific researcher should check written records (if they exist) of vital events. This check will provide additional information for scientific research. Both sources of information, that from informants and that from documents, give valuable information. What anthropologists bring to the study of human populations is an understanding of the legitimacy of data from either source, and anthropological insight allows for the interpretation of the varying responses in the light of the cultural context. FieZdwork ws supported by a DHEW/FuZbright-Hays DoctoraZ Dissertation FeZZowship. Most of the data were gathered while the author was a member of the fieM research group of the Department of Human EcoZogy, PoZish Academy of Soiences, Warsaw.

GLUCOSE INTOLERANCE AMONG DOGRIB INDIANS OF THE NORTHWEST TERRITORIES EMOKE J. E. SZATHMARY MoMaster Universitg Maturity-onset diabetes has reached epidemic proportions in North American native populations of the United States. The rise in the incidence of this disease is sud- den, for diabetes appears to have been unknown 2 to 3 generations ago. Studies suggest that both environmental and genetic factors are involved in the onset of the disease. Of these, obesity and the inheritance of a major gene have been singled out as important variables in the etiology of MOD. This study reports preliminary findings of a survey designed in part to determine the current status of glucose tolerance in an Indian population of the Northwest Territo- ries, and to determine the association of obesity and glucose tolerance. Approximately 25% of adult Dogrib Indians (159 persons) were given oral glucose tolerance tests. Of these, 3 presented with fasting hyperglycemia (plasma glucose 2 140 mg/dl). Thirty-five persons were abnormal on two-hour values (plasma glucose 2 160 mgldl), yielding a prevalence estimate of 22% for glucose intolerance. Clinical diabetes is unknown in the Dogrib, and glucose intolerance among northern Athapaskans has until now been reported to be negligible. Skinfold measurements were obtained on 159 persons at eight different body sites. Sig- nificant correlation occurs with 2-hour glucose values > 160 mgldl in men at subscapu- lar, suprailiac, abdominal, forearm and lower neck sites, and in women at all of the former plus triceps, medial calf and midaxillary sites. No correlations occur in men between any skinfold and 2-hour glucose values I 160 mgldl. Similar findings occur for the women, except that the lower neck skinfold value was still significant. These results indicate that in this group, as in many though not all others, obesity and hyperglycemia are associated.

VARIATIONS IN THE VERTEBRAL COLUMNS OF AUSTRALIANS AND POLYNESIANS WILLIAM D. WADE University of Manitoba

The vertebral columns of 454 native Australians and 73 Polynesians were examined for anomalies associated with the inter-segmental borders. These are hypothetically attributed to the general phenomenon of "caudal shifting", a model that serves to re- late a profusion of variations in the vertebral column that have been interpreted in rather disparate and unrelated ways, hopefully serving the end of explanatory simpli- fication. Several consequent conclusions of the proposed model are supported, in varying degree, by the reported data from the Pacific and other skeletal samples. Other propositions of the model are essentially untested and require verification.

CUSP SIZE IN PONGIDS LAUREN R. WILLIAMS University of Toronto (Comptete paper pubzished in VoZ. 2, No. 21

198 1 ANNUAL MEETING The ninth annual meeting of the Canadian Association for Physical Anthro- pology/l'Association pour 1'Anthropologie Physique au Canada will be held 10-13 December 1981 at the Banff Centre. Registration will cost $20, and room rates will be $39/day (single) or $30/day (shared). These rates will include three meals a day. Session information, titles and abstracts of papers, and requests for further information should be addressed to: Dr. James D. Paterson Department of Anthropology University of Calgary 2500 University Drive Calgary, Alberta T2N 1N4

NEW JOURNAL Dr. P. Dash Sharma, Executive Secretary of the Sarat Chandra Roy Institute of Anthropological Studies, Ranchi, India, has informed us of the imminent appearance of a new journal, South Asian Anthropologist. It will publish on South Asian regional topics in all fields of anthropology and will appear biannually in March and September. It has an international editorial board. Annual subscription rates are $8.00 (U.S.), $11.00 (U.S.) for institutional subscriptions, including surface mailing. All correspondence should be directed to Dr. P. Dash Sharma, Executive Secretary, SRIANS, 18 Church Road, Ranchi - 834001, Bihar, India. CANADIAN REVIEW OF PHYSICAL ANTHROPOLOGY REVUE CANADIENNE D'ANTHROPOLOGIE PHYSIQUE

Contents Vol. 2, Nos. 1-2, /ROBERT I. SUNDICK. The Skeletal Remains from the Taber Child Site, Taber, Alberta ...... x,"' ROBERT LAROCQUE et NORMAN CLERMONT. Etude PrCliminaire de Certains Iroquoiiens PrChistoriques du QuCbec ...... JEROME S. CYBULSKI. Possible PreColumbian Treponematosis on Santa Rosa Island, California ...... C. E. EYMAN. Sexually Dimorphic Variation in the Ossification of the Hand and Wrist in Japanese Children ...... M. McKEOWN and K. ISOTUPA. The Estimation of Maximum Crown Dimensions in Human Permanent Teeth with Severe Attrition ...... F. FOREST et D. BERTHELETTE. L'Utilisation de 1'Analyse de Variance Multiple en Anthropologie de la SantC ...... YVON DESJARDINS. DiffCrences entre les Sexes et LatCralisation au Niveau de Performances Manuelles ...... BookReview ...... Canadian Association for Physical Anthropology/l'Association pour 1'Anthropologie Physique au Canada. Abstracts of Papers Presented at the Seventh Annual Meeting, November 7-10, 1979, Ste.-Adble, QuCbec

sprinting Services The University of Manitoba