university of groningen analysis of the dental morphology ... · hominidmandibularpostcanine roots...

University of Groningen

Analysis of the dental morphology of Plio-Pleistocene hominids. IV. Mandibular postcanineroot morphologyWood, B A; Abbott, S A; Uytterschaut, H

Published in:Journal of anatomy

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:1988

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):Wood, B. A., Abbott, S. A., & Uytterschaut, H. (1988). Analysis of the dental morphology of Plio-Pleistocenehominids. IV. Mandibular postcanine root morphology. Journal of anatomy, 156, 107-39.

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 01-06-2020

https://www.rug.nl/research/portal/en/publications/analysis-of-the-dental-morphology-of-pliopleistocene-hominids-iv-mandibular-postcanine-root-morphology(b35da576-aeae-405f-831d-8a7498f5d89e).html

https://www.rug.nl/research/portal/en/persons/hilde-uytterschaut(51b21c91-d754-45cb-8964-7757bc6cec05).html



https://www.rug.nl/research/portal/en/journals/journal-of-anatomy(a34fbd13-004c-43f3-ad82-5b22149388b2).html

J. Anat. (1988), 156, pp. 107-139 107With 9 figures

Printed in Great Britain

Analysis of the dental morphology of Plio-Pleistocene hominidsIV. Mandibular postcanine root morphology

B. A. WOOD, S. A. ABBOTT AND H. UYTTERSCHAUT

Department of Human Anatomy and Cell Biology, The University of Liverpool,P.O.Box 147, Liverpool L69 3BX

(Accepted 26 March 1987)

INTRODUCTION

The first three papers in this series surveyed and documented the morphology ofhominid mandibular postcanine tooth crowns (Wood & Abbott, 1983; Wood, Abbott& Graham, 1983; Wood & Uytterschaut, 1987). These studies have demonstratedconsistent patterns of morphological variation both within and between fossiltaxonomic categories, and these patterns have been used to assess the affinities ofspecimens whose taxonomic designation is unknown and/or problematic. However, atooth includes its root, or roots, as well as the crown. Indeed, the root form of hominidteeth assumes a particular importance because although a significant number of earlyhominid mandibles from East African sites have lost the crowns of the postcanineteeth, such specimens often preserve a relatively complete record of the rootmorphology. It is in an attempt to utilise these data, for both taxonomic andfunctional analysis, that the subocclusal morphology of early hominid mandibularpremolar and molar teeth will be considered in this the fourth paper in the series.The subocclusal morphology of the mandibular molars and premolars has been

considered together for two main reasons. Firstly, whereas molar roots are generallymonomorphic in primates, following the generalised eutherian pattern (Butler, 1941),mandibular premolar roots show considerable variation in extant higher primates,both within and between taxa (Goh, 1957; Abbott, 1984). Variation in fossil hominidmandibular premolar root form has also been noted for more than half a century (e.g.Dubois, 1924; Weidenreich, 1937), and its particular importance among aus-tralopithecines has been recognised for at least three decades (Robinson, 1952).However, an important part of this variation tends towards so-called 'molarisation'of the premolar roots, and the consideration of molar and premolar root formtogether provides a useful context for the variations in the form of the latter teeth.Secondly, it avoids repeating detailed and extensive information about radiographicand measurement techniques, which are similar for both multi-rooted premolar andmandibular molar teeth.

It is a compelling hypothesis, if not an axiom, that the sequence of morphologicalmanifestations of a structure which go to make up a phylogenetic series is the resultof sequential modifications of that structure's ontogeny. The digital morphology ofamphibians has provided a particularly good theoretical and experimental model todemonstrate the mechanisms and effects of such changes (Oster & Alberch, 1982;Alberch, 1983; Alberch & Gale, 1983, 1985). However, evidence has also beenprovided to suggest that dental ontogeny may also be a fruitful area in which therelationship between ontogeny and evolution might usefully be explored (Butler,

108 B. A. WOOD, S. A. ABBOTT AND H. UYTTERSCHAUT

1956, 1982, 1986). The dental studies referred to have particularly emphasised changesin tooth crown morphology but it is evident that any informed analysis of theevolution of the form of tooth roots must also be set in a sound ontogeneticcontext.Mammalian tooth development is most probably the result of an interaction

between mesenchymal cells of neural crest origin and the ectoderm overlying the jaws(Gaunt & Miles, 1967) and there is evidence that the form of a tooth (i.e. incisor ormolar) is determined by the mesenchyme (Kollar & Baird, 1969). Whereas the formof the crown of a tooth is determined by the contact between the covering ectodermalenamel organ and the mesenchymal dental papilla (and more specifically by the shapeof the basement membrane of the inner enamel epithelium), it is the behaviour of theencircling edge of the enamel organ, the cervical loop, which determines the form ofthe roots (Schour & Massler, 1940; Diamond & Applebaum, 1942; Kovacs, 1963,1967). The edge of the enamel organ is a double-layered structure, Hertwig's epithelialsheath, the two layers corresponding to the external and internal enamel epithelium ofthe enamel organ. The inturned edge of the sheath is called the epithelial diaphragmand it is the existence and nature of any differential cellular proliferation around themargin of the epithelial diaphragm (see below) that determines whether the tooth hasa single, or multiple roots. This proliferation and 'mapping' of the root apparentlyprecedes the formation of the root dentine from the inner of the two layers of the rootsheath (Grant & Bernick, 1971; Bhaskar, 1980).

In single-rooted teeth Hertwig's epithelial sheath maintains a single basal opening,the primary apical foramen. Multiple tooth roots are produced when the primaryforamen is subdivided into two or more secondary apical foramina (Jorgensen, 1950;Gaunt, 1960). The number and site of secondary apical foramina are apparentlyanticipated by the location of pulpal blood vessels (Butler, 1956; Gaunt, 1960). Thefirst sign of such a subdivision is the definition, in the margin of the primary foramen,of the areas that will give rise to the inter-radicular processes (tongues or projections)(Orban & Mueller, 1929). The processes grow inwards towards the centre and it isthe fusion of such processes that eventually forms the subpulpal, or bifurcal, wall of amultiple root system (Alexandersen 1962, 1963; Bhaskar, 1980). The site of fusion ofinter-radicular processes is represented by a ridge, known as either the intermediatebifurcational (Everett, Jump, Holder & Williams, 1958), or the inter-radicular(Kovacs, 1971) ridge; it is a combination of dentine and cementum (Everett et al.1958; Carlsen, 1967). While it is the site (or sites) of proliferation on the edge of theprimary apical foramen that determines the location of any additional roots(Jorgensen, 1950; Bhaskar, 1980), the final form of the root system also depends onthe degree of development or penetration of the inter-radicular processes (Alex-andersen, 1962; Carlsen, 1967). If the processes only minimally penetrate thepapilla, a superficial root groove is the result. Progressive degrees of penetration resultin a deepening groove and fusion of the processes leads to complete separation of theroot moieties.As well as being dependent on the site, number and degree of development of the

inter-radicular processes, the eventual form of the roots, and in particular the heightof bifurcation, depends on the timing of fusion of the inter-radicular processes(Carlsen, 1967). If fusion occurs early in relation to growth in root length, then theroots are separated close to the crown; if they fuse much later during rootdevelopment the only sign of multiple roots may be apical bifurcation. The directionof growth of the inter-radicular processes is the final influence on the form of the roots,for it determines the contours of the subpulpal wall (Alexandersen, 1962, 1963;

Hominid mandibular postcanine roots

Carlsen 1967; Kovacs, 1971). When they grow more or less horizontally, then the wallwill be similarly orientated and the inter-radicular recess more rounded (viz.Gregory & Hellman, 1939, p. 359).

In this study the root morphology of early hominids from sites in Kenya andTanzania has been assessed both radiographically and by direct observation of theexposed root form. These new data will be integrated with existing publishedinformation about the root morphology of hominids from sites in southern Africa, aswell as remains recovered from Pliocene East African sites. A small sample of modernhuman postcanine mandibular teeth has been included in the study to provide areference for the variations seen within the fossil hominid material. These data onfossil hominids will then be set in comparative context by reviewing previous studiesof the root form of modern humans and extant higher primates. The final part of thepaper will attempt to interpret the variations in root form in the light of what is knownabout the ontogeny of root formation. Evidence will be presented for two sequencesof root evolution, one of root reduction and the other of root elaboration. Thesemorphological trends will then be examined to see to what extent, if any, they can berelated to phylogenetic hypotheses based on other evidence.

This analysis uses the same informal taxonomic categories that were adopted for theprevious analyses of mandibular postcanine crown morphology. Details of the fossilhypodigms subsumed within each category are given in an earlier paper (Wood &Abbott, 1983) and extra information is included in the current paper only if theallocations deviate from the principles and practice set out in the earlier publication.

MATERIALS AND METHODS

Fossil and extant samplesThis study was based on a sample of 168 hominid mandibular permanent

postcanine teeth (33 P3s; 44 P4s; 35 Mls; 33 M2s; and 23 M3s), from at least 46individuals (Table 1). Twenty nine specimens were allocated to one of three informaltaxonomic categories, designated EAFROB, EAFHOM and EAFHER. The categoryEAFROB includes specimens allocated to Australopithecus boisei, EAFHOMsubsumes material allocated to Homo habilis (Leakey, Tobias & Napier, 1964), Homocf. habilis (e.g. Leakey, 1974) and Homo ergaster (Groves & Masak, 1975) and the twospecimens included in EAFHER were attributed to Homo cf. erectus (Tobias, 1969;Rightmire, 1980). Specimens were only placed in one of these three categories if therewas sufficient non-dental evidence from the form of the mandible, or an associatedcranium, to suggest a relatively reliable taxonomic allocation. Fragmentary specimensand specimens whose taxonomic affinity is problematic or contentious, were classed as'unknown', and 17 of the specimens used in this analysis were placed in this category.Specimens common to this analysis, and the earlier studies in this series, retain theiroriginal taxonomic allocations. When specimens have the same tooth preserved on theright and left sides, both sets of measurements have been recorded.The modern human sample consists of 34 adult specimens (16 males and 18 females)

comprising 20 Romano-British and 14 Australian Aboriginal mandibles from thecollection at the British Museum (Natural History), London.

Root formAlthough the molar roots of modern human mandibular molar teeth may show

considerable variability, especially of the M3 (e.g. Drennan, 1929; Pedersen, 1949;

109


Table 1. List of Plio-Pleistocene hominid premolar and molar teeth from EastAfrican sites, together with their taxonomic category

TaxonomicSpecimen no. P3 P4 Ml M2 M3 category

Site: Koobi ForaKMN-ER 403KMN-ER 404KMN-ER 725KMN-ER 726KMN-ER 728KMN-ER 729 RKMN-ER 729 LKMN-ER 730KMN-ER 733KMN-ER 801KMN-ER 810KMN-ER 818KMN-ER 819KMN-ER 820 RKMN-ER 820 LKMN-ER 992 RKMN-ER 992 LKMN-ER 1468KMN-ER 1482KMN-ER 1483KMN-ER 1501KMN-ER 1502KMN-ER 1506KMN-ER 1507KMN-ER 1508KMN-ER 1801KMN-ER 1802 RKMN-ER 1802 LKMN-ER 1803KMN-ER 1805KMN-ER 1806KMN-ER 1808KMN-ER 1811KMN-ER 1812KMN-ER 2597KMN-ER 3229KMN-ER 3230 RKMN-ER 3230 LKMN-ER 3729KMN-ER 3731KMN-ER 3734KMN-ER 3889KMN-ER 3954

Site: Olduvai GorgeO.H. 7 RO.H. 7 LO.H. 13O.H. 16 RO.H. 16 LO.H. 22O.H. 23O.H. 37O.H. 51

Site: PeninjPeninj RPeninj L

1 1 1

1* 1 1

1* 1

1 1

1

1* 1* 1

1* 1* 1

1* 1 1

1* 1*

1* 1

1 1

1 1

1 1 1

1 1 1

1 1

1 1 1

1 1

1 1 1

~~~~~~1~~~~~~1

~~~~~~1

1 1 1

1 1 1

1 1

1 1 1

1*

1 1 1

1 1 -

1*

1 1 1

- 1 1

1 1 1

1* 1*

1 1*

1 1 1

1* 1*

1* 1

1 1 1

1 1 1

1 1 1

1 1 1

1

1 1 1

1 1

1 1

1 1

1 1 1

1 1

* Only root form recorded.

1 - EAFROB1 1 EAFROB1 1 EAFROB1 1 EAFROB1 1 Unknown1 1 EAFROB1 1 EAFROB1 1 EAFHOM1 1 EAFROB1 1 EAFROB1 1 EAFROB

EAFROBUnknownEAFHOMEAFHOM

1 EAFHOM1 1 EAFHOM1 1 EAFROB1 - Unknown

Unknown1 1 EAFHOM1 - EAFHOM1 - Unknown

EAFHOM1 - Unknown

1 Unknown1 - Unknown1 - Unknown

EAFROBUnknown

1 EAFROB1 EAFHER

UnknownUnknown

1 - UnknownEAFROB

1 1 EAFROB1 1 EAFROB

Unknown- - Unknown1 - EAFHOM

Unknown-- Unknown

EAFHOM1 - EAFHOM1 1 EAFHOM1 - EAFHOM

EAFHOM1 1 EAFHER1 - EAFHOM1 - EAFHER- - Unknown

1 1 EAFROB1 1 EAFROB


Lateralview

A. Slngle root 1 R

B. Two roots: Tomes' root 2T

Root cross-sectionCervical to Apical tobifurcation bifurcation

C. Two roots: mesiobuccal and distal 2R: MB + D

D. Two roots: mesial and distal 2R: M + D

,f~Distal * * Mesial

Mesial

Lingual + Buccal

Distal

Fig. 1. Categories of mandibular premolar root form.

Kraus, Jordan & Abrams, 1969), they are relatively monomorphic in extant non-human primates (Abbott, 1984; Abbott & Wood, in preparation). Preliminaryobservations made in the course of the present analysis, together with evidenceincluded in other detailed studies of early hominid tooth morphology (e.g. Robinson,1956), suggest that variation in early hominid mandibular molar root form can beadequately expressed by root measurements without recourse to shape categories.However, for the mandibular premolar roots, the situation is more complex. There areconsiderable differences between the root form of modem humans and those of theextent non-human higher primates (e.g. Tomes, 1923; Peyer, 1968; Scott & Symons,1974; Abbott, 1984; Abbott & Wood, in preparation), and the premolar rootmorphology of A. boisei is distinctive, if not unique, among higher primates (Wood,1981; Abbott, 1984). Thus, information about the number and the form of the rootsof the premolars from both radiographs (see below) and direct observation were

III


Table 2. Definitions of the four categories of hominid mandibular premolar root form

(1) Single root (I R) (Fig. 1A)The radiographic image shows a single, tapering root with a single (main) canal. Single roots of Tomes'root form are also included in this category, i.e. a groove or cleft on the mesiolingual root surface;additionally there may also be one on the distobuccal root surface. Such roots were considered to be singlewhere the grooves did not lead to bifurcation, or if they did, where the bifurcation was restricted to theperiapical region. The cross section of a single root is more or less circular with a centrally placed rootcanal. If one or both of Tomes' grooves/clefts are present, these indent the mesiolingual and distobuccalmargins of the section (the appearance is the same as the section cervical to the bifurcation shown in Fig.1 B).

(2) Two roots: Tomes' root (2 T) (Fig. 1 B)These are teeth with Tomes' root form in which the characteristic grooves/clefts led to bifurcation. Theroots are mesiobuccal and distolingual, and appear superimposed on the radiographic image - the mesialarea of the distolingual root overlaps the distal area of the mesiobuccal root. Each root has a separatepulp canal. Cervical to the level of bifurcation, the root cross section shows mesiolingual and distobuccalclefts, of which the mesiolingual is typically the more developed. Apical to the bifurcation, the obliquebifurcation cleft separates the triangular-shaped cross sections of both roots.

(3) Two roots: mesiobuccal and distal (2R: MB and D) (Fig. 1 C)Although the radiographic image is similar to that for two roots the cross section of the root is distinctive.Cervical to the bifurcation, the section is triangular, showing greater mesiodistal development buccally.Apical to it there is a plate-like distal root and a buccolingually constricted mesiobuccal root.

(4) Two roots: mesial and distal (2R: M and D) (Fig. 1 D)The radiographic image shows clear bifurcation into mesial and distal roots, each of which possesses itsown pulp canal(s). The buccal and lingual clefts apparent on the root cross section cervical to the bifur-cation lead to the separation of parallel, plate-like mesial and distal roots. Each root commonly possessestwo pulp canals.

integrated to produce the simple four-category scheme which is illustrated in Figure1; details of each category are given in Table 2. Morphometric data about the rootsof the hominid mandibular premolar and molar teeth take the form of a series of linearand angular measurements taken from radiographs.

Radiographic techniqueAll the specimens were radiographed by S. A. and B. W. at the National Museum of

Kenya, Nairobi. The power and voltage were fixed at 25 MA and 90 kV, the onlyvariable being the exposure time. All radiographs were taken at a focus-film distanceof 150 cm.

Radiography was arranged so as to follow, as far as possible, the parallel-filmtechnique (McCormack, 1937; Barr & Gron, 1959; Gron, 1960). Mandibles that couldbe aligned so that the plane of the film was parallel to the longitudinal axis of theroot(s) were set up for radiography as in Figure 2(A). Specimens in which thesymphysis was intact could not be so orientated and in these cases the mandible wasangled so that the X-ray beam passed through both mandibular bodies withoutsuperimposing the images of the two sets of premolars (Fig. 2 B). In this arrangementthe plane of the film was no longer parallel to the long axis of the roots and a meansof correcting the resulting image for distortion had to be devised. Several methods areused in dental radiography (Bregman, 1950; Benkow, 1957; Hamernik, 1957; Best et al.1960; Bramante & Berbert, 1974), but a two dimensional correction was needed forthis study. We adapted a method used by Cave (1929), who devised a grid of fine wiresembedded in cellophane. Everett & Fixott (1963) used a grid made of enamelledcopper wire and Plexiglas but in this study the grid was made by sandwiching fine

112

Hominid mandibular postcanine roots 113

X-raysLongitudinal

axis ofpremo larsor molars - - -- -- -

Film

Perspex

A

X-rays

( f 1 -. | Longitudinalaxis of

l a premolars--- Grid

B ~~~~~~~~~~~FilmComplete'premolar

image

Fig. 2. Distal view of the orientation arrangements for radiography of the right mandibularpremolars, or molars using (A) the parallel-film technique, and (B) angled radiography.

copper wire between two layers of perspex. The basic grid consisted of 1 mm squares,but a slightly thicker gauge was used for every fifth wire so that 5 mm squares wereeasily distinguishable. The grid was placed between the specimen and the film, as closeto the former as possible, and orientated parallel to the long axis of the tooth root(s).Depending on the size of the film required, either 'Kodirex', Kodak NS-2T non-screen, or Kodak Dental Occlusal film, was used for the radiographs.Measurements were not taken directly on the radiographs because of the risk of

damaging the emulsion. Instead, the root outlines were traced onto clear acetate paperwith a fine pen; the best definition was obtained by viewing the radiographs in adarkened room with the light box masked except for the pertinent area of theradiograph. The mesial and distal cemento-enamel junctions were identified on eachtracing. The tracings were used to construct root diagrams similar to those used byKovacs (1966, 1971) except that in this study the tracings were in relation to the'cervical axis' rather than to the occlusal plane, as suggested by Kovacs. The 'cervicalaxis' was defined as a straight line joining the most mesial and distal locations of thecemento-enamel junction.A maximum of eighteen reference points was marked on each root diagram; these

are illustrated in Figure 3, and defined in Table 3. Up to nine linear measurements and


Mesial Distal

a e b a b a b Cervical

1 Rpremolar 2T premolar 2R premolar/molar

Fig. 3. Lateral root diagrams of the mandibular postcanine teeth, illustrating the locations of thereference points used to define the measurements.

Table 3. Definitions of the reference points on mandibular premolar and molar rootdiagrams

(a) The most mesial point on the cemento-enamel junction.(b) The most distal point on the cemento-enamel junction.(c) The point of bifurcation between the mesial and distal roots or, for premolars with a Tomes' root

form (see below), the most cervical point on the mesial border of the distolingual root.(c') The most cervical point on the distal border of the mesiobuccal root (premolars with Tomes' root

form only).(d) Projection of point c onto the cervical axis.(dt) Projection of point c' onto the cervical axis.(e) The midpoint between a and b on the cervical axis (single-rooted premolars) or, the midpoint

between a and d' (Tomes' root form) or, the midpoint between a and d (remaining teeth).(e') The midpoint between d and b.(I) A point on the longitudinal axis of the root which passes through point e and which is approxi-

mately two-thirds of the way from the cervical axis to the root apex.(f') As forf, but related to point e'.(g) The point of intersection between the longitudinal root axis through point e and the root apex.(g') As for g, but related to point e'.(h) The midpoint on the longitudinal root axis between points e and g.(h') The midpoint on the longitudinal root axis between points e' and g'.(09 The point of intersection between the mesial root border and the line through point h which is

perpendicular to the longitudinal root axis.(i') As for i, but related to point h'.(D) The point of intersection between the distal root border and the line through point h which is

perpendicular to the longitudinal root axis.(j') As for]j, but related to point h'.

one (or more) angular measurements were taken directly off each root diagram andthese are defined in Table 4, which also includes details of a compound variable andthree indices. One of the latter reflects root robusticity, and two relate the site of rootbifurcation to the whole root system. Where root measurements were taken fromangled radiographs, the raw measurements were corrected using the wire grid. At afocus-film distance of 150 cm the X-rays may be assumed, for practical purposes, tobe parallel. Thus, no correction for mesiodistal measurements was- necessary.However, measurements in the direction of the axis of the roots, i.e. root heights andheight of bifurcation, did require correction. This was done by measuring the size of

114


Table 4. Definitions of the hominid mandibular premolar and molar rootmeasurements, compound variable and indices

Measurements(1) Neck mesiodistal diameter (NMD): the distance along the cervical axis between points a and b.

(2 and 3) Root mesiodistal diameter (RMD - mesial and distal): the distance perpendicular to the rootaxis between points i and j, or i' and fj.

(4 and 5) Actual root height (ARH - mesial and distal): the distance along the longitudinal root axis be-tween points e to g, or e' to g'.

(6 and 7) Projected root height (PRH - mesial and distal): the perpendicular distance between thecervical axis and the root apex.

(8) Location of bifurcation (LB): the distance between points a and d (not applicable to premolarswith Tomes' root form).

(9) Height of bifurcation (HB): the distance between points c and d.(10 and 11) Root angulation (RA - mesial and distal): the angle in degrees between the longitudinal root

axis (of the mesial root, or both mesial and distal root) and the cervical axis.

Compound variables and indices(12) Root divergence (RD): RAm-RAd.

(13 and 14) Root robusticity (RRI - mesial and distal): RMD/ARH x 100.(15) Location of bifurcation (LBI): LB/NMD x 100.(16) Height of bifurcation (HBI): HB/PRHm x 100.

the image of 20 mm of grid in the occluso-apical direction and correcting themeasurement as follows:

lMeasured root measurement\Corrected root measurement =

Measured root measurementx 20.Measured size of grid image

The corrections required ranged up to 35% of the measured value.The precision of the root measurements of the hominid teeth was assessed by taking

root tracings and preparing root diagrams on three separate occasions. The averageerror of all root measurements was 4%, with the largest error, 8 %, being for theheight of bifurcation. This larger error was due to two factors, the smallness of the di-mension being measured and the error resulting from estimating the location of thecervical axis in specimens in which the crown was missing or damaged.The mean value, standard deviation, and range were computed for each of the

relevant measurements and indices for the two main taxonomic categories, EAFROBand EAFHOM. The difference between the mean values of the two categories wascompared using the two-tailed Student's t test. Univariate comparisons using the rawdata were supplemented by multivariate methods, specifically by Principal Com-ponents Analysis (PCA). The principal components were derived from a covariancematrix based on the untransformed data. The affinities of the 'unknown' specimenswere assessed using posterior probabilities calculated using Canonical VariatesAnalysis (CVA). The result is expressed as a percentage, which is the likelihood thatthe specimen belongs to a particular reference category. All multivariate computationswere based on SPSS programs.

RESULTS

The specimens in which it was possible to determine the category of premolar formare listed in Table 5, and the results for the two main taxonomic groups aresummarised in Table 6. The specimen numbers are too small to merit detailedstatistical treatment, but the root morphology is evidently different in the two

115


Table 5. List of mandibular premolar specimens and their root form as determined inthe present study

TaxonomicSpecimen no. category P3 p4

Site: Koobi ForaKMN-ER 403KMN-ER 725KMN-ER 726KMN-ER 729KMN-ER 730KMN-ER 733KMN-ER 810KMN-ER 818KMN-ER 819KMN-ER 992KMN-ER 1468KMN-ER 1482KMN-ER 1483KMN-ER 1501KMN-ER 1801KMN-ER 1802KMN-ER 1803KMN-ER 1805KMN-ER 1806KMN-ER 1811KMN-ER 1812KMN-ER 3229KMN-ER 3230KMN-ER 3729KMN-ER 3731KMN-ER 3734KMN-ER 3889KMN-ER 3954

Site: Olduvai GorgeO.H. 7O.H. 13O.H. 16O.H. 22O.H. 23O.H. 37O.H. 51

Site: Peninj

EAFROBEAFROBEAFROBEAFROBEAFHOMEAFROBEAFROBEAFROBUnknownEAFHOMEAFROBUnknownUnknownEAFHOMUnknownUnknownEAFROBUnknownEAFROBUnknownUnknownEAFROBEAFROBUnknownUnknownEAFHOMUnknownUnknown

EAFHOMEAFHOMEAFHOMEAFHEREAFHOMEAFHERUnknownEAFROB

2R: MB+D?2R: MB+D?2R: M+D

2R: MB+D?2R: MB+D?2R: M+D?2R: M+D2T2T

2TI RI R2T?2R: M+D

?2R: MB+D(DL)2R: M+D?2 T?1 R2R: M+D

?2R: MB+D2R: M+D2T?1 R2R: M+D

2TI RI RI RI RI R

2R: M+D

Table 6. Incidence of mandibular premolar root form in the two major taxonomiccategories of East African hominids

Categories

A B C D(I R) (2 T) (2R: MB+ D) (2R: M + D)

P3EAFROB 0 0 3 6EAFHOM 4 4 1 0

P4EAFROB 0 0 0 15EAFHOM 7 0 0 2

2R: M+D2R: M+D2R: M+D2R: M+D?2R: M+D?2R: M+D2R: M+D?2R: M+D2R: M+DI R2R: M+D2R: M+D2TI R2R: M+D?2R: M + D

2R: M+D2R: M+D

2R: M+D2R: M+D2R: M+D?2R: M+D2R: M+D?I R2R: M+D

I RI R?1 RI R?1 R?1 R?I R

2R: M+D2R: M+D

Hominid mandibular postcanine rootstaxonomic categories for both premolar types. The parameters of the eleven originalvariables, root divergence and the three indices for the two major taxonomiccategories are listed in Table 7, and the subsequent table (Table 8) records the resultsof using Student's t test to examine the significance of the differences in univariate rootmetrics between the samples of two main taxonomic categories.

All eleven original variables (for M1) and subsets of seven (M2 and M3) and fivevariables (P3 and P4) were used to construct covariance matrices for each tooth typefrom which two principal components have been computed. The percentage varianceaccounted for by the two components and their eigenvector scores are given in Table9. For the small sample of P3s for which five variables are available, root height andneck mesiodistal diameter (NMD) account for most of the discrimination containedin the first principal component, which itself accounts for approximately 90 % of thevariance. Root height similarly dominates the PCI for the P4 data, with NMDdominating the second, and orthogonal, component (Fig. 4).The arrangements of specimens along the first principal components of the M1 and

M2 root variables are also dominated by the effect of variables which reflect rootheight. In the M1 plot (Fig. 5), the height of bifurcation also influences theirdisposition. For the M2s (Fig. 6) root heights are the major influence on PCI, butNMD is the dominating effect on the distribution along the axis of the secondprincipal component in both M1 and M2. The sample size for EAFHOM for M3 istoo small to make any useful deductions about the discriminative value of multivariateanalysis, but it is noteworthy that the pattern of eigenvector scores for M3 is unlikethose for M1 and M2, with the distribution along PCI in the former being dominatedby neck mesiodistal diameter, the location of bifurcation and the mesiodistal diameterof the mesial root. These differences in variable effect between the molar tooth typesmay presage the greater within-species variability in M3 morphology which is noted inextant modern human populations (Kraus, Jordan & Abrams, 1969).The same data were analysed by Canonical Variate Analysis (CVA) to provide an

assessment of the affinities of the 'unknown' specimens. The posterior probabilityvalues are given in Table 10, which also lists the affinities of the 'unknown' premolarteeth as assessed from the multivariate analysis of root form. The results of all themethods used to classify the 'unknown' teeth are summarised in Table 11.

DISCUSSION

Comparative context

The form of the roots of primate mandibular molars corresponds with thegeneralised eutherian pattern, i.e. they have two roots, mesial and distal (Gregory,1920-1; Butler, 1941; Clark, 1971). However, variation in root form does occur, andthis is most easily comprehended in its developmental context. In modern humanmandibular molars, for example, suppression of one or more of the inter-radicularprocesses, particularly in the second and third molars, results in root reduction(Kovacs, 1963, 1967). Reduced penetration of a process results in a root groove orcleft, and when normal development occurs in only one process it leads to 'fusion'with a semilunar root form (Kovacs, 1963). When there is no development of inter-radicular processes, the resulting single molar root is referred to as pyramidal. Whenboth processes develop, but there is a delay in their fusion, this leads to varying degreesof apical bifurcation. This is the basis of taurodontism, as defined by Keith (1913), ananomaly which Shaw (1928) was to divide into three categories, though there is littleagreement about the most suitable nomenclature to describe the essential continuum

117


t~~~~~~~~~~~~~~~e It WI W) WI W) o0 C4 0 r W

[E~~~~~~~~~~~~oN_e tn - t- U) oo o 61 on'c r- C m o

tX N F ,_w~~~~c W00 W 00r-_-O 00

xi~ ~ ~ ~~~~~~~oo o _- en W)O) W)C4000 W0_ C a-"o I_ 00-

to~~~~~0<- en _7 a o C4 o oo oo rt mooo '.to ; t tn oo C; a

S~~~~~~~~~~~~~e r- In l%IC°nI as en 0 enEt X Ontoo NXeNX XFOX t^£e oXe>_~~~~~~~~~~~~~~~~~--4C4 clXLm-Nbx mx N o ot.;S~~~~~~~~~W en_ n en oo 00 C4 00 e _ Nt~tE~~~~~~~~~~~~~C Cq e I>,14 0 nc 0 k

~~~~~~~~~~en C4 kn I o ,X o n '_IO Q -X0 0 X m..-Ien

°t g I v I < I z - z m b ^ ^̂ N̂£N....__X~~~~C

PL ^ V N t- X - £ °o t N O

,~~~~~~ 00outmu m£

°~~~~~~0 Io "- C; en WI eq '. 4 tV-oe) as,N _ft tn eoov4aN

2~~~~~~L -.4 - . 7 e W O C e 0 ) en l0o

q6 < : f :>>fiau

Q ~ ~ ~ ~~~ en en sD~~~~~~x - - ,r


00xo Nr 0 WI) 9 9t:: a, 'I C1 08 en WI -

O No ta-U^eoo ooo~~ 6~bor co e oo

00C4W( Q as asas000(N -O 0 0O0 enr0_eN00r£e WIO _ _ _ 00 _ '-_ 00 N O m 0 _ >

- t m- eI

o -h -

t N o

99

oN-6 q W Iom~ - as~ O6Nt- e6 C; , ;rt C)o0O

-en r- en en

OiN b O 00o

oL _9 _P 9 9 9o } F t o t - N \

> £ ~ ONO oo0~m0AC - oooO t -ci

en Ot --0 0ew w It mNO

C_ -o t_ _

O o t ^ci --99 o 999

O en 00en oo oo 00c -t - C4 N p- e WI- a r- '- N0tW T l

O0 W)--r- kn I*_ } oy_IOensr- :sooom°°g 0.

O o4tM e I 'leqa s'O1 00 'l .0IC .0. t o~ooN~~~~~~~~0 0 "roen 'IOn 00 cr0N °° sn °xw

-cici Ci

00 W) 99C 9 9

0 N _ I

8-o-r: -n en o) o)W sa - n4 00 o odo6 e; o; o;-

0 en _oo 00 00 00 en en en _ A ON 'II

00 9~~~~~~~~~I 1O991? 15 999?r-~o"oen O000 O~Na, 00NN-r-00 00 OrNO 'f "itfWt -' I W WIC4 O

N 0~~~~~~~r00.r I ''~~~~~~~n0 Q 0 -N-, e- Ci400fIenI0ci ~ ~ ~ f

- %-cgs > + > / Xi4 % I it

o 9 - _en iN- W 00 en en00 I 00 'IcO0ON C4t nk -ciC4 hen ~ e~O

'- ci Ci4 cO~VI) IN IC4 Cci ci cci

C

-

C

00o


L- 00 N 00t-N e No ON O-- O'NO 00eeC o N--4-- V-- r-N 00 - -N00 00 (D00 as

00 0 1 W n O nW 0

0oo N e 0oo oo N_0ON 0 oo #

(N 00 ~ N00Or- 000 e 00 0W t- 00 O%0 W

-~ ~O0Q ',~ c en cNI C00 en -Oass

tn~ Q: r: . eC 00N " ) C106 ; 4h C tr-6I en

oqoo nC~r- 6oo o5 C: C; C; oo(5 64e ^ oo6 4 tr Z ooo en

=NF $ o X m £ ^ F t N t N ^ F Xff-4 t- i t

N z41-N;-N c9OO'-tlC- Nt1O-Q (Th e0- C ()1

0 O0 0 oOo en CNs 00 N %OO 0F 08

0 Oa 'xg 00 00 00 en enenN -7OO2 00 00 00 en cne

= N o t eXF m N_X _ o 6 et_

0 F F t _ F N £ oW NNo

N0Nt- r 0fr-N f 00o

ea - 00 00 0 It 05)9s O~ -oo o o 00 c 64 CIoo WO64 '. O OONt eoO.-O

O0 O 00t O00 - 9"

m~~~~~~Z u~ o Z ~o C£ 6 t o CZ v}C:nO

-00a.,

(NIO% 00 00~IC4r )enaseF r N0\0 0O t N £ O N ^ N90 00O e0 0 -Q en Io en

C'0

D~~~~~ £

O(NU- o o - - -o--oONN s Fgt


Table 8. Between-group comparisons (EAFROB and EAFHOM) of rootmeasurements. Significance values are based on Student's t test, using pooled variancesand two-tailed model

P3 P4 M1 2 M3

NMD N.S. < 0-001 < 0-001 < 0-001 <0001RMD(m) N.S. N.S. N.S. < 0-001 < 0-05RMD(d) N.S. < 0-001 < 0-001 < 0-001 < 0-05ARH(m) N.S. N.S. N.S. N.S. N.S.ARH (d) N.S. N.S. N.S. < 0-05 N.S.PRH (m) N.S. N.S. N.S. N.S. N.S.PRH (d) N.S. N.S. N.S. < 0-05 N.S.LB N.S. < 0001 < 0-001 < 0001HB N.S. N.S. N.S. N.S.RA(m) N.S. N.S. < 0-05 < 0-01 N.S.RA (d) N.S. N.S. < 0-05 < 0-01 N.S.RD N.S. < 0-01 N.S. N.S. N.S.RRI (m) N.S. N.S. N.S. < 0-05 N.S.RRI (d) N.S. < 0-05 < 0-05 < 0-05 N.S.LBI N.S. < 0-01 < 0-01 N.S.HBI < 0 01 N.S. < 0-05 N.S.

between mild degrees of taurodontism and the pyramidal root form (e.g. Pedersen,1949; Tratman, 1950; Brabant & Kovacs, 1961). Conversely, the development ofadditional inter-radicular processes results in the three-rooted molars reported insome series (Tratman, 1938; Pedersen, 1949; Turner, 1967).

Previous comparative studies of mandibular premolar root form suggest that, withthe notable exception of Homo sapiens, such teeth in non-human primates are usuallytwo-rooted (Owen, 1840-5; Duckworth, 1923; Tomes, 1923; Colyer, 1936; Senyurek,1953; James, 1960; Peyer, 1968). However, variations have been noted. In someprosimians and New World monkeys the lower premolars may be single-rooted, withroot reduction being concentrated anteriorly, so that P2 has one root, and P3 and P4two roots (Senyurek, 1953). A similar trend is observed in some fossil prosimiangenera (Quinet, 1966). Degrees of root fusion have also been reported in themandibular premolars of Pan (Duckworth, 1923; Tomes, 1923; Colyer, 1936; James,1960), and buccolingual rotation of the roots may result in such roots appearing to besingle on a radiograph (Gregory & Hellman, 1926). In contrast, the P3 of Hylobates larwas reported to have an additional distolingual root (i.e. a total of three roots) in 44%of a sample of 96 dentitions (Frisch, 1963, 1973). A more recent survey of both the sizeand form of the premolar roots of higher primates confirmed that mandibularpremolars are typically two-rooted, the P3 having mesiobuccal and distal roots(2R: MB and D) and the P4 mesial and distal roots (2R: M and D) (Abbott, 1984;Abbott & Wood, in preparation). Intraspecific variation was recorded for P3 only.Occasionally in Pongo (3 %), and more frequently in Pan (29 %), the P3 appeared (atleast radiographically) to be single-rooted, but such a variation was not seen in theGorilla sample. The little evidence that has been published about fossil hominoidssuggests that their root form was similar to that of the majority of extant higherprimates, i.e. P3= 2R: MB and D, and P4= 2R: M and D (Gregory & Hellman,1926; Woo, 1962; Leakey, 1968, 1970; Tekkaya, 1974).

In modern human populations mandibular premolars are usually described assingle-rooted (Duckworth, 1923; Sprinz, 1953; Barker, Parsons, Mills & Williams,1973; Scott & Symons, 1974), but published evidence suggests that modern human

121

5 ANA 156


o~~~~~~

E { > ; O : O ~~>I:lOC4 0f0 4

O O O <>~~~~~~~~> 0s C> oC,3N>

e.00 r 0 ~e

C> 0 e r c-' N

Ec C;'N ON ° -Nr en

^Q~~~~~~~~~~~ C; o1 66: 6 z ,

~~~~~~~~~~E 8 g , - £ ,°I e

Q3 N O O%N0ON- No

O EX O O O ° | ° °l X >>~~I I

1~~~0~

| ;! cL xs >00 0 0 000° __e.

00 ~ ~ ~ ~ 0

10 0

E>B B M § SSC£sI s

.5

122


* KNM-ER 3954*KNM-ER 1483

0 KNM-ER 819*KNM-ER 1801

* KNM-ER 1802 LA * KNM-ER 1802 R

* KNM-ER 1482* OH 51

A EAFROB* EAFHOM* Unknown specimens

.

I

-1 0PC

2RH

Fig. 4. Plot of the first (PCI) and second (PC II) principal components generated from thecovariance matrix of five root variables for early hominid P4s from East Africa.

AL A

* KNM-E R 1802 L* KNM-ER 1801 * KNM-ER 1802 RA

. *OH 51


-2 -1 0 2PC RH

HB

Fig. 5. Plot of the first (PCI) and second (PCII) principal components generated from thecovariance matrix of five root variables for early hominid M1s from East Africa.

samples show several consistent variants. The root of a human mandibular premolarmay be marked by longitudinal root grooves which may be associated with apical, or

more extensive, bifurcation (Goh, 1957). The most common site for a groove is themesiolingual aspect. This is usually located on the mesial root surface, some one thirdof the distance from the lingual to the buccal border. The groove may be superficial,or it may penetrate to varying degrees into the root along the mesiolingual-distobuccalaxis, when it is better described as a cleft (De Jonge-Cohen, 1919; Goh, 1957; DeJonge, 1961; Gher & Vernino, 1980). When such a cleft is present it defines a Tomes'root and when the cleft becomes more pronounced apically it leads to bifurcation intomesiobuccal and distolingual apices (De Jonge-Cohen, 1919; Tomes, 1923; Diamond,1952; Goh, 1957; Scott & Symons, 1974), although De Jonge (1961) has describedthem as mesiobuccal and distal. In some specimens, a second longitudinal groove isfound on the distal or buccal root surface (Goh, 1957; Kraus, Jordan & Abrams,1969). In cross section, Tomes' root is C-shaped with the open part of the 'C'corresponding to the mesiolingual cleft. The pulp usually divides at some level within

5-2

2-

NMD

123

1 -

PC II 0-

-1-

-2-

A

-2

2-NMDLB

RMD(d)RA(m)RA(d)

OPC 11

1-

-2-

-3-

..

1

1

1


A

* KNM-ER 1802 RA~~~~~~~~~~~~~~~~~~*KNM-ER 1802L A

A A

* KNM-ER 1506 A

.

.

-1

PC I

* KNM-ER 1482

U

U

0RH


2

Fig. 6. Plot of the first (PCI) and second (PC II) principal components generated from thecovariance matrix of five root variables for early hominid M2s from East Africa.

Table 10. Taxonomic affinities of teeth in the unknown category as suggested fromposterior probabilities derived by canonical analysis. Design group: EAFROB/EAFHOM (both left and right side included)

Specimen Type

Design group

Site: Koobi ForaKMN-ER 728KMN-ER 819KMN-ER 1482

KMN-ER 1483

KMN-ER 1506KMN-ER 1801

KMN-ER 1802 (R)

KMN-ER 1802 (L)

KMN-ER 3954

Site: Olduvai GorgeO.H. 51

P3: 63 % correctly classified (EAFROB, n = 2; EAFHOM, n = 6)P4: 62% correctly classified (EAFROB, n = 6; EAFHOM, n = 7)M1: 62% correctly classified (EAFROB, n = 6; EAFHOM, n = 7)M2: 69% correctly classified (EAFROB, n = 8; EAFHOM, n = 8)M3: 80% correctly classified (EAFROB, n = 8; EAFHOM, n = 2)

M. EAFROB 98P4 EAFHOM 79P. EAFHOM 58P4 EAFHOM 55M2 EAFHOM 62P. EAFHOM 62P4 EAFHOM 70M2 EAFHOM 74P3 EAFHOM 76P4 EAFHOM 75M1 EAFHOM 84P. EAFROB 50P4 EAFROB 52M1EAFHOM 67M2 EAFHOM 75P4 EAFROB 52M1 EAFHOM 62M2 EAFHOM 75P4 EAFHOM 73

P4 EAFHOM 55M1 EAFROB 64

124

NMDLB

RMD(m)

2-

1*

0-

PC 11

-1-

-2-

-3- r-3 -2 1


0 0 00

40 ;4 4

0

"O LL 44

I w w_-,

000 0

I.I 1-1 I -=

2 m-[O 40

I-I

7--00

0 0104

LL4 _S

-1 1- 1- I-I 1-

"O 11 [J o;L L4 .o;Lw w w w O=I.- 1 -1I-,1- I-

e N O s

oo4N0 oo o o o o o o -. . a-,

N Zdwmooooxooooo

m0co

m0

I--

00

00xz[.1#L

I ==

S:. .

x

.

S: .4X ;s

* ,

X Re

oo

Cq zu Co.X °. ._ _

o ks o

^oo oXo

11 t:C^:

*e Cy

R oz

o tgo Q

11 +u :D. *,* zo

*

X svX .o

s tqoo AbW_$ .Cqo <tq.N X. Cqs XsNs z4Y .t

2 .^-

125


the root to form two canals in mesiobuccal and distolingual positions (Fig. 1 B).Rarely, Tomes' root may lead to bifurcation in the middle, or even cervical, thirds ofthe root so that two well-formed roots are present; examples of two-rooted humanmandibular premolars have been reported (De la Parra, 1952; Brabant et al. 1953;Pearlman, 1972; Barker, 1973; Shapira & Delivanis, 1982). Three-rooted modernhuman mandibular premolars are extremely rare, but examples have been documented(De Jonge-Cohen, 1919; De la Parra, 1953; Brabant, Sahly & Bouyssou, 1961; DeJonge, 1961; Armitage, 1971; Barker, 1973; Shapira & Delivanis, 1982).The root form of the modern human P3 is more variable than that of P4 (Brabant

et al. 1953; Sprinz, 1953) and there is some doubt about the assertion that two rootsare more common in P4s than in P3s (Tratman, 1950; Goh, 1957). However, in all theseries investigated the incidence of Tomes' root form has been greater for P3 thanP4 (Taylor, 1899; Campbell, 1925; Shaw, 1931; Nelson, 1938; Abrahams, 1946-7;Pedersen, 1949; Selmer-Olsen, 1949).

Root form of early hominid taxaThe subocclusal morphology of the mandibular postcanine teeth of hominids from

Hadar attributed to Australopithecus afarensis has been summarised and describedelsewhere (Ward, 1979; Ward, Johanson & Coppens, 1982). The form of 15 P3s isassessed in the latter publication, and approximately 25 % (from three individuals) areclassified as single-rooted, but it is noted that even in such teeth two root canals arefound, the buccal one being subcircular and the lingual one triangular in cross section.The majority of Hadar P3s are two-rooted with a larger buccal and a smallerdistolingual root component.The sample of A. afarensis P4s from Hadar numbers 15 teeth from 13 individuals.

Three teeth show evidence of root fusion, with two doing so to the extent that theirroots are described as 'pyramidal'. The remaining teeth have distinct mesial and distalroots. Each is mesiodistally compressed with root canals at the buccal and lingualedges of the root. The mesial root is 2-3 mm longer than the distal, and the bifurcationis described as " high, occurring immediately below the cervix ", and " arch-like, rathernarrow or cleft-like" (Ward, Johanson & Coppens, 1982, p. 610).

Details are provided of the root morphology of 45 mandibular molars from Hadarattributed to A. afarensis (M1 = 13; M2 = 19; M3 = 13). All teeth have mesial anddistal roots. The roots are subequal in length, with the mesial root being moremesiodistally compressed, with a dumb-bell-shaped cross section and a charac-teristically bifid apex. The distal roots of M1 and M2 are inclined buccally and possesseither one or two root canals; in the latter case the distobuccal pillar is dominant. Thedistal root of M3 contains a single root canal and is strongly inclined buccally anddistally. The root angles are commented on and illustrated (Ward, Johanson &Coppens, 1982, Figs. 8 and 16), but no measurements are given. While it is evidentfrom the tracings of the lateral radiographs that root inclination is variable in themandibular postcanine teeth, there is a clear trend for the degree of distal inclinationof the roots to be greater in M2 andM3, with the distal root ofM3 showing the greatestdegree of distal inclination. No root length measurements are given but " M2 appearsto be the dominant tooth in terms of root length and robusticity" (Ward, 1979, p. 490).The relatively few specimens from Laetoli for which root data are available suggestthat the root forms of these teeth are similar to those of the Hadar material (White,1977, 1980). However, the description of the P3 roots of LH4 and LH24 suggest thatthey may be of the 2R: MB and D variety. Specific mention is made of the mesiodistalcompression of the molar roots and buccolingual width measurements suggest that themesial root is the larger of the two.

Hominid mandibular postcanine rootsEvidence about the root form of Australopithecus robustus (including Australo-

pithecus crassidens) and Australopithecus africanus comes from the studies ofGregory & Hellman (1939), Broom & Schepers (1946), Robinson (1956) and Sperber(1974). Of these, Robinson (1956) provides the most comprehensive morphologicalinformation. The collections from Swartkrans and Kromdraai provide a total of 14P3s belonging to A. robustus from which some information can be gleaned. None wassingle-rooted, four had separate mesial and distal roots and in the remainder the rootswere partly separated with a "well-developed" cleft on the mesiolingual face of theroot system (Robinson, 1956, p. 71). Thus, the majority of the P3s of A. robustusshowed the Tomes' root form (Broom & Robinson, 1952; Robinson, 1952). Sperber(1974) does not give a breakdown of root types, but does provide data on root lengthsfor nine teeth, with site sample means of 17-2 mm for Swartkrans, and 17-5 mm forKromdraai. Gregory & Hellman (1939) commented that the form of P3s ofParanthropus (i.e. A. robustus) corresponds to that of the P4s of Plesianthropus (i.e. A.africanus) (see below).

Fifteen specimens of the P4s of A. robustus are available, but information about rootform is only available for nine teeth (Broom & Schepers, 1946; Robinson, 1956). Allare described as having a mesial and distal root. At least three are described asbeing wide and "flattened mesiodistally" (Robinson, 1956, p. 75). Sperber (1974) citesthe average length of the mesial and distal roots of six P4s from Swartkrans as 16-2mm.The sample of molar teeth of A. robustus comprises 31 M,s, 13 M2s and 21 M3s, but,

of these, details of the root system are only available for a much smaller number ofisolated teeth, (4-M1s, 3 M2s-and 2 M3s) (Robinson 1956). All mandibular molars-havetwo roots, mesial and distal. The mesial root is larger than the distal in M,s, but thesituation is apparently reversed in M2, with the distal root of M3 being peg-like andvariable in size. Sperber (1974) comments on the form of the roots of six M1s fromSwartkrans. The teeth all have mesial and distal roots which are noted to be broadbuccolingually. The mesial root is, on average, slightly longer than the distal (N = 5),with more tendency to apical bifurcation. For M2, Sperber reports that the roots of A.robustus are subequal in length, and longer in the mesiodistal direction than those ofA. africanus. The M3 root form is two-rooted for the majority of teeth, with one havingthree roots. Average root lengths are M1: 17-5 mm (N = 5); M2: 17-0 (N = 5) andM3: 16-4(N=6).Few premolar teeth attributed to A. africanus, and able to provide information

about root structure, were availabole to Robinson (1956). Of the three P3s available,Robinson's descriptions suggest that three possessed Tomes' root form, with twoadditionally having a distobuccal cleft (Robinson, 1952, 1956). There is no unequivocalevidence that these teeth had two separate roots. The single P4 specimen (TM 1523) istwo-rooted, with a mesiobuccally situated mesial root; Gregory & Hellman (1939)comment on the similarity of such roots to those of primitive apes. Sperber (1974)makes no specific comment about the root form of A. africanus P3s, but infers that theP4s were two-rooted. The root lengths given for the single Sterkfontein P4 are at thelower end of the range for the six Swartkrans teeth. There are also a few mandibularmolars from which root evidence can be gleaned (Robinson, 1956). All seven are two-rooted, and Robinson does not suggest that they depart significantly from the rootform of A. robustus. Sperber (1974) records root length for M3 only; the sample of 4has an average length of 15 1 mm.Root form in the early representatives of Homo, e.g. Homo habilis, has received little

or no attention. Single-rooted P3s and P4s are reported in material usually attributedto H. erectus from Swartkrans, Sangiran, Zhoukoudian, Rabat and Sidi Abderrahman

127


(Weidenreich, 1937; Arambourg & Biberson, 1956; Howell, 1960; Arambourg &Hofstetter, 1963; von Koenigswald, 1968, 1969; Jacob, 1973). Anterior premolarswith Tomes' root form have been described in H. erectus specimens from Olduvai,Kedung Brubus, Trinil, Sangiran, Zhoukoudian and Ternifine (Dubois, 1924;Weidenreich, 1937, 1945; Arambourg & Hofstetter, 1963; von Koenigswald, 1968,1969; Jacob, 1973; Rightmire, 1980).

Comparison of root form and size in early hominidsAlthough this study and review have illustrated that relatively few early hominid

teeth have their root system either preserved or available for assessment, nonethelesssufficient data are now available to attempt to establish whether there are significantmorphological differences between the relative taxonomic categories considered in thisstudy and in the review. The distribution of the four sub-classes of mandibularpremolar root form is evidently different in EAFHOM and EAFROB (Table 6). Whencompared to what is proposed as the primitive condition for the ape/hominid clade,i.e. P3= 2R: MB and D and P4 =2R: M and D (Abbott & Wood, 1985, inpreparation), EAFHOM and EAFROB appear to show divergent trends. InEAFHOM root separation in the P3s is mostly incomplete, (2 T), or absent (1 R),whereas in EAFROB the majority of the nine P3s show evidence of 'molarisation' ofroot form, i.e. the possession of two mesiodistally flattened root plates. In the sampleof P4s, the polarisation is more exaggerated for EAFROB, but in EAFHOM, althoughthe majority are single-rooted, two specimens have the ' molarised' condition.However, in one of the two specimens (KNM-ER 730), the P4 root images on theradiographs suggest apical displacement of the root bifurcation (see below). Thecombination in EAFROB of 'molarisation' of both the anterior and posteriorpremolars is apparently unique among higher primates. The specialisation ofEAFROB is also evident when details of root form are considered, for the plate-likemesial and distal root plates are differentially thickened, at the buccal sides on themesial root, and at the lingual side of the distal root. Two other taxa, A. afarensis andA. robustus, also have a high incidence of 2R: M and D P4s, but their P3 root formdiffers from A. boisei. In A. robustus, a Tomes' root form (not seen in A. boisei) is morethan twice as common as 2R: M and D, and in A. afarensis the P3 root form isrelatively unchanged from the primitive condition for the ape/hominid clade, i.e.2R: MB and D. Premolar root form in EAFHOM is most closely matched by that inH. erectus among the fossil taxa, and H. sapiens among the extant taxa. Like H.sapiens, the EAFHOM sample shows a greater tendency for Tomes' root form in P3than in P4. The A. africanus sample shows an unusual combination of premolar rootform, with root reduction in both P3 and P4 relative to the hypothetical ancestralcondition, with Tomes' root form predominating in P3 and apparently 2R: MB andD in P4, but the sample is a small one.

Despite these differences in root form, the metrical evidence (admittedly availableon fewer specimens) shows no significant differences between EAFHOM andEAFROB P3s, and few differences for P4 (Tables 7, 8). The differences that do existcorroborate the morphological classification for they concentrate on the greatermesiodistal diameter and robusticity of the distal root (Table 8). It is noteworthy thatthere are no significant differences in root height between EAFHOM andEAFROB.When the root heights of the two fossil samples are related to values obtained from

20 samples of modern human teeth (Abbott, 1984), the greater height of the hominidroots is emphasised. However, this difference in root height is accompanied by


proportional increases in root thickness so that root robusticity indices in the twofossil hominid samples, extant apes and modem humans, are similar (Abbott, 1984;Abbott & Wood, in preparation). The limited data for root height available for A.robustus and A. africanus (Sperber, 1974, Tables 24 and 27) suggest that the roots arenot only longer in both the East African taxa, but that the differential between rootheights in P3 and P4 may differ between A. boisei and A. robustus. In the former P4roots are longer than, or subequal with, those of P3, whereas in A. robustus it is P3roots which have the greater average height. Individual values for the latter are notgiven, so the significance of the differences in sample means cannot be examinedstatistically. The relatively few significant univariate differences between the premolarmeasurements of the two taxonomic categories clearly influence the degree to whichmultivariate statistical methods can discriminate between the two samples. Thepercentage of correct classification for the premolars in the design groups (Table 10)is poor by other standards (Wood & Uytterschaut, 1987).The usual primate mandibular molar root form, with mesial and distal moieties, is

apparently ubiquitous across the two hominid categories considered in this study. Themorphology of the M3 roots of hominids has been noted to be different, and morevariable, than those of M1 and M2 (Robinson, 1956), but there is no consistentinterspecific variation. For M1, the major differences between EAFHOM andEAFROB are the more distal inclination and more robust distal roots of theEAFROB M1s, whereas for M3 it is the greater mesiodistal thickness of both themesial and distal roots of the EAFROB sample. The greatest contrast between the twosamples is seen in M2s. All but four variables are significantly different (Table 8), withthe roots of EAFROB being both larger overall and more robust. The differences inroot heights are greater (and significant) for the distal roots. Only Gorilla and H.sapiens among extant samples approach the root robusticity index values forEAFROB (Abbott & Wood, in preparation). A relatively short M3 root system is theusual pattern in extant higher primates, but what is unusual about the rootcharacteristics of EAFROB mandibular molar teeth is the relative shortness of boththe M2 roots as well (Abbott & Wood, in preparation). Sperber (1974) provides rootheight data for site samples of 'robust' and 'gracile' australopithecines from southernAfrican sites, and it is noteworthy that despite their crowns being smaller the averageroot length of EAFHOM exceeds that for A. robustus from Swartkrans. When all theavailable variables are considered together in the multivariate analyses, M2, andparticularly M3 can be more effectively discriminated than P3-M1.

Functional interpretation of root morphologyAn important function of tooth roots is to support the tooth crown in the jaw.

Analyses of chewing and biting suggest that tooth crowns are subject to obliqueshearing, as well as vertical compressive forces, and it is evident that changes in the sizeand shape of the crown have to proceed paripassu with modification to the root form.Comparative studies of mandibular molar and premolar crown morphology ofEAFHOM and EAFROB (Wood, Abbott & Graham, 1983; Wood & Uytterschaut,1987) have suggested that the greater crown area of teeth attributed to 'robust'australopithecine taxa is mainly related to selective expansion of the talonid, or distal,component of the premolar and molar crowns. However, an allometric analysis byJungers & Grine (1986), using many of the same data as the studies cited above, hassuggested that the pattern of M1 crown expansion may differ between A. robustus andA. boisei, with the latter showing distal expansion and the former more buccolingualexpansion of the crown area. In the light of these conclusions, it is worthwhile

129


T,r

I1

-T

-F

/,

I Jr

EAF ROB

j ,m s sl ,m s d ,m d m J m,P3 P4 Ml M2 M3

Fig. 7. Plot of root mesiodistal diameters for EAFROB and EAFHOM.

T

130

8-

7 -

6-

5-

4-

5-

4

3,

2-

EAFROB

IEAFHOM

m s d m s d m d m d m d

P3 P4 Ml M2 M3

Fig. 8. Plot of root robusticity indices for EAFROB and EAFHOM.


Table 12. Comparison between the size and shape of the mesial and distal rootmoieties of EAFROB and EAFHOM

EAFROB EAFROB EAFHOM EAFHOM(one-tailed) (two-tailed) (one-tailed) (two-tailed)

RMDP N.S. N.S. N.S. N.S.3

P4 S N.S. N.S. N.S.

Ml N.S. N.S. N.S. N.S.M2 s N.S. N.S. N.S.M3 s N.S. S N.S.

RAPa N.S. N.S. N.S. N.S.P4 S S N.S. N.S.

ml s s s sm2 s N.S. S SM3 s s N.S. N.S.

RRIP3 N.S. N.S. N.S. N.S.P4 S N.S. N.S. N.S.

ml N.S. N.S. N.S. N.S.

M2 s s s sM3 s s s s

s = P < 0-05

examining whether there is evidence of any differential enlargement of the distalmoiety of the root systems of the A. boisei sample in order to support the distallyexpanded crown.

Reference to Table 7 suggests that such differences may be evident in both thepremolars and molars. Root mesiodistal diameter and robusticity follow very differentpatterns in the premolars of the two taxonomic categories. Whereas in EAFHOM thedistal root is narrower and more gracile than the mesial, in EAFROB the reverse is thecase (Figs. 7, 8). The significance of the mesial/distal difference has been examined fortwo root variables (RMD and RA) and the root robusticity index (RRI). Values forboth one and two-tailed tests are given in Table 12, and these suggest that significantmesial/distal root dimorphism is more common in EAFROB then in EAFHOM, withthe P4 being the tooth where the sharpest distinctions lie. Thus, the root morphologyofEAFHOM P4s, with its larger and more robust distal root element, seems to mirrorthe distinctive talonid enlargement seen in its crown morphology (Wood &Uytterschaut, 1987).

Classification of unknown specimensThe morphological and metrical evidence collected in other papers in this series

have demonstrated the utility of such comparative data for assessing the affinities ofproblematical or isolated specimens. In this study, 17 specimens were classed as'unknown'. The affinities of premolars are based on their root form and on rootmorphometric data; for the molars only the latter are available. The detailedinformation contributing to the assessments is given in Tables 5 and 10, and theaffinities are summarised in Table 11.

Relatively unequivocal evidence of taxonomic affinity based on premolar root formis provided for only five specimens (KNM-ER 819, 1482, 1483, 1801 and 3731). Forall except one of these specimens, KMN-ER 3731, the allocation is corroborated bythe morphometric data. The affinities suggested by the postcanine root form ofKMN-


ER 1482 fit with those suggested on the basis of the molar and premolar crownmorphology and enamel thickness. These all suggest that this specimen has few, if any,of the derived traits associated with A. boisei. The allocations of KMN-ER 1483 and1801 also seem clear, but that of the latter conflicts with the evidence of M3 crownarea, which suggests that this tooth is too large to belong to a Homo taxon (Wood,Abbott & Graham, 1983). The root form of the P3 of KMN-ER 1805 (2R: MB andD/2R: MB and DL) is indicative of a change from the proposed primitive conditionfor P3 root morphology towards eventual root reduction, but the vital evidence onP4 root form cannot be properly discerned on the specimen. However, the small molarcrown area is strong evidence against its inclusion in EAFROB. The mandible KMN-ER 1802 shows a mixture of traits associated with both EAFROB and EAFHOM.This conclusion is similar to that reached after the analyses of the characters whichrelate to detailed molar and premolar morphology (Wood, Abbott & Graham,1983; Wood & Uytterschaut, 1987), and the implications of these results will beexplored in a separate paper (Dean & Wood, in preparation). Suffice it to say that themixture of traits shown by such specimens may either be interpreted as evidence fora close phyletic association between early Homo and A. boisei, or as examples ofhomoplasies in dental morphology between A. boisei and early Homo.

Evolutionary trends in hominid premolar root morphologyThe relative homogeneity of the root form of the great apes (Abbott, 1984; Abbott

& Wood, in preparation) suggests that the primitive root form for the African ape/hominid clade is P3= 2R: MB and D, and P4 =2R: M and D. In such a pattern themesiobuccal component of the P3 root is oblique and contains a single pulp canal,whereas the distal root is transversely orientated, and contains two pulp canals. TheP4 has mesial and distal roots which are not noticeably displaced buccolingually withrespect to each other. Each is transversely orientated, symmetric and contains twopulp canals. In both P3 and P4 the level of bifurcation is high, usually being within thecervical third of the root height, with the roots showing slight to moderate divergence.The appearance of the two root forms in cross section, above (left) and below (right)their bifurcation, are given in Figure 9(A).The results of this study, together with the evidence derived from the review, suggest

that a similar premolar root form to that described above is commonly seen in teethattributed to A. afarensis from Hadar and Laetoli (Ward, 1979; Ward et al. 1982;White, 1977, 1980). Four specimens from Koobi Fora also appear to have a similarmorphology, but the quality of the evidence is good enough in only one of these(KMN-ER 403) to be confident of its premolar root form. The root morphology of theother three specimens (KMN-ER 725, 733 and 3729) can only be regarded assuggestive of the apparently primitive root form. The P3s ofMLD 29 and TM 1518 arereported to have the same root form (Broom & Schepers, 1946).The range of premolar morphology observed in the hominids considered in the

present study, together with a knowledge of the variations and ontogeny of root formin modern humans, and extant higher primates suggest that there were at least twotrends in root morphology leading, respectively, to the apparent derived conditions ofroot reduction, seen in modern man and Homo erectus on the one hand, and rootelaboration in A. boisei on the other. The trend towards root reduction will bedescribed first.The first stage for P3 is the alteration from 2R: MB and D to 2R: MB and DL. This

requires both a change in the shape of the primary apical foramen and migration ofthe inter-radicular processes. Teeth with this root form are known from Hadar (Ward


C D

P3 1R P3 2R: M+D

K4 1 R B2R:M+

P3 2T P4 2R:M+D

P4 2R:M+D A

3\ 2R: MB + D

P4 2R:M+D

Root reduction Root molarisation

Fig. 9. Proposed evolutionary trends in mandibular premolar root form in fossil hominids.

et al. 1982), Laetoli (White, 1977), Kromdraai (TM 1600) (Sperber, 1974), Sterkfontein(Sts 7 and Sts 5) (Robinson, 1956) and Koobi Fora (e.g. KMN-ER 819, 1482, 1801and 3734). The next transition, from 2R: MB and DL to Tomes' root form requiresthe suppression of the distobuccal inter-radicular process (Fig. 1 B), to producecontinuity distobuccally between the mesiobuccal and distolingual root components,yet normal development of the mesiolingual inter-radicular process, so that themesiolingual cleft remains (Fig. 9 B). Periapical bifurcation combined with amesiolingual cleft also produces the appearance of Tomes' root form. Such a rootmorphology is found in a minority of Hadar P3s as well as in P3s from Swartkrans andSterkfontein (Robinson, 1956), Laetoli (LH 14), Olduvai (OH 16 and 23), Koobi Fora(e.g. KMN-ER 1483 and 1501), Trinil (Dubois, 1924) and Sangiran (Weidenreich,1945; Jacob, 1973). Reduction of the mesiolingual cleft and the progressive apicalmigration of the bifurcation site, so that the root tip is bifid only, are the final stagesin the pathway towards single-rooted P3s (Fig. 9C). KMN-ER 992 and OH 13, areexamples of specimens which lie on the continuum between P3s with Tomes' root andthose which have a simple, single root.The trend towards root reduction in P4s is less complex, apparently just involving

the progressive apical displacement of the bifurcation, with a concomitant reductionin the size of the roots and a shift from a mesiodistally flattened to a more roundedroot form and finally to a single root (Fig. 9 B, C). Such progressive apical migrationis seen in the series KMN-ER 3734, 1483, 1501; OH 16, 23 and 37 and L 75i-1255from the Omo (Coppens, 1973). Single-rooted P4s are found, for example, in KMN-ER 992 from Koobi Fora, OH 13 from Olduvai and mandible B from Sangiran (vonKoenigswald, 1968). Robinson (1952) long ago suggested a scheme for P3 rootreduction in hominids and his scheme corresponds to the basic alterations required topass from 2R: MB and DL, to 1 R. However, he did not attempt to relate his schemeto any presumed ancestral hominid root morphology.

In addition to one (or more) trends in premolar root reduction, there is goodevidence for a separate trend in premolar root elaboration, or 'molarisation'. In itsmost developed expression P3 and P4 are both supported by mesial and distal roots (i.e.2R: M and D) which resemble those of mandibular molars (Fig. 9D). The distal rootis displaced and better developed lingually, where it provides support for the talonidof the asymmetric premolar crown. The opposite applies to the mesial root; its buccalportion is better developed than the lingual. These features are more accentuated in

133


P4 than P3, but in both tooth types the level of bifurcation is high, i.e. within thecervical third of the root height. The morphology described is best seen in KMN-ER 729, 1468, 1806, 3229 and 3230 from Koobi Fora, and in the Peninj mandible. Allthese mandibles are attributed to A. boisei. The greatest shape change from theapparent primitive condition is the transformation and reorientation of the P3mesiobuccal root, presumably by a simple enlargement of the primary apical foramencombined with the migration of the mesiolingual inter-radicular process. The P3 rootform of KMN-ER 3731 probably represents the transition from 2R: MB and D to2R: M and D, and the P4 roots of KMN-ER 726 are only moderately asymmetric,thus lacking the fully 'molarised' specialisation of the P4. Full molarisation of themandibular premolar roots is seen in specimens from other sites, such as the P3 (L 18-33) from Member D of the Omo (Coppens, 1970) and lingual displacement of theflattened distal root of both P3s and P4s has been described from Swartkrans (e.g.P4s SK 7, 88 and 876), as well as in a P4 from Sterkfontein (TM 1523) (Robinson,1956; Sperber, 1974).As has been suggested above, the special root morphology of the mandibular

premolars of A. boisei can be linked with the preferential enlargement of the talonidcomponent of the crown, especially of the P4 (Wood & Uytterschaut, 1987). Thedominance of the P4 in A. boisei is also seen in differences in root length between thetwo mandibular premolars, where the condition P4 > P3 is apparently derived withrespect to the hypothetical ancestral condition.The nature of the relationship between A. afarensis, A. africanus, A. robustus and A.

boisei has been the subject of considerable interest and recent debate. Leaving aside thediscussions about whether A. afarensis and A. africanus are distinct taxa (Tobias,1980; Johanson, 1985), much of the debate has centred on the relationship between A.africanus and the 'robust' australopithecines. Do A. africanus, A. robustus and A.boisei represent successively derived specialisations along a trend towards hypertrophyof the postcanine masticatory apparatus (Johanson & White, 1979; White, Johanson& Kimbel, 1981; Rak, 1983; Kimbel, White & Johanson, 1984), or is A. africanus bestregarded as a potential common ancestor of all known hominids (Tobias, 1980) or ofjust Homo and the 'robust' australopithecines (Skelton, McHenry & Drawhornl1984)? If A. afarensis is accepted as the most primitive known hominid, and A.africanus placed at the base of a 'robust' australopithecine clade, A. africanus and A.robustus would be expected to lie along the trend of root elaboration for A. boisei.What is the evidence that this is the case? Although the data for A. africanus areparticularly sparse, the evidence from the two taxa does not support such aplacement.The majority of the P3s of A. robustus (9 out of 13), and the P3s of A. africanus,

apparently have Tomes' root form which, with respect to the hypothetical ancestralcondition, is part of a trend towards the reduction, not the elaboration, of the P3 rootform. Likewise, the evidence from the P4 roots suggests that A. africanus P4s, with2R: MB and D, may be derived not in the direction of root elaboration but of rootreduction.

Clearly, mandibular premolar root morphology carries a priori no more, or less,weight than other variables in any phylogenetic analysis. However, the fact thatvariations in mandibular premolar root form can be set in such a firm ontogeneticcontext makes them particularly relevant for assessing the likelihoods of differentphylogenetic schemes. Suffice it to say that the most parsimonious interpretation of thedistribution of premolar morphology presented in this paper does not support theconcept of an australopithecine clade containing both A. africanus and the 'robust'

134


australopithecines, nor do details of premolar root morphology support the conceptof a simple hypermasticatory 'trend' extending from A. afarensis, through A. africanusand A. robustus, to A. boisei (Hunt & Vitzthum, 1986).

SUMMARY

The subocclusal morphology of 168 permanent mandibular premolars (N = 77) andmolars (N = 91) of Plio-Pleistocene hominids has been investigated. The taxonomicallocation of the teeth, which represent at least 46 individuals, was based on non-dental evidence. Specimens were allocated to one of two major taxonomic categories,(EAFROB or EAFHOM), East African Homo erectus (EAFHER), or their taxonomicaffinity was regarded as 'unknown' (N = 17). Information about the root system wasderived from radiography and direct observation. Morphometric data were in theform of nine linear and two angular measurements based on eighteen reference points.Root form was also assessed using a scheme which recognised four classes of rootmorphology. Data were compared using both univariate and multivariate techniques,including Principal Component and Canonical Variate analysis. Posterior prob-abilities derived from the latter were used (in a two-taxon design model) to assess theaffinities of the 'unknown' specimens.The variation in hominid mandibular premolar root form was interpreted as two

morphoclines, based on the presumed primitive condition of the P3 (with mesiobuccaland distal roots, 2R: MB and D) and P4 (with mesial and distal root, 2R: M and D)root systems. One trend apparently leads towards root reduction (i.e. P3= 1 R;P4 = 1 R), and the other to root elaboration (i.e. P3 and P4 = 2R: M and D). Theextreme form of the latter is the 'molarisation' of the premolar roots seen inEAFROB. Despite major differences in root form there was relatively little taxonomicvariation in root metrics, except for a more robust distal root system in EAFROB.Molar root form showed little interspecific variation except for M2 in which the rootsin EAFROB were larger and more robust, with differences in root height being greaterfor the distal than for the mesial roots.Root form and metrics enable four of the 'unknown' specimens (KMN-ER 819,

1482, 1483 and 1801) to be tentatively allocated to EAFHOM, and a single specimen,KMN-ER 3731, to EAFROB. Published assessments of the root morphology of the'robust' australopithecines from Swartkrans suggest that the premolar root form ofAustralopithecus (Paranthropus) robustus is not obviously intermediate between thepresumed ancestral condition, and the 'molarised' mandibular premolar root systemsof Australopithecus (Paranthropus) boisei.

We thank the Custodians of all the fossil collections used in this analysis and theDirectors, Trustees and Staff of the institutions in which they are housed forpermission to study them and for assistance and hospitality while the work wasbeing undertaken. Particular thanks are due to Richard Leakey, the Trustees of theKenya National Museum, and the Government of Kenya for the opportunity tomake a detailed study of the Koobi Fora remains. We thank Craig Engleman andJill Dorman for their help with the preparation of the illustrations and the text andPercy Butler for commenting on the manuscript. This research was supported by agrant from the N.E.R.C.

135


REFERENCES

ABBorr, S. A. (1984). A comparative study of tooth root morphology in the great apes, modem man and earlyhominids. Ph.D. thesis, University of London.

ABBOTT, S. A. & WOOD, B. A. (1985). Mandibular premolar root form and evolution in the Hominoidea.Journal of Anatomy 140, 536-537.

ABRAHAMS, L. C. (1946-7). The masticatory apparatus of the people of Calvinia and Namaqualand in theNorth Western Cape of the Union of South Africa. Official Journal of Dental Association of South Africa1(4), 5-51.

ALBERCH, P. (1983). Morphological variation in the neotropical salamander genus Bolitoglossa. Evolution 37,906-919.

ALBERCH, P. & GALE, E. A. (1983). Size dependence during development of the amphibian foot. Colchicine-induced digital loss and reduction. Journal of Embryology and Experimental Morphology 76, 177-197.

ALBERCH, P. & GALE, E. A. (1985). A developmental analysis of an evolutionary trend: digital reduction inamphibians. Evolution 39, 8-23.

ALEXANDERSEN, V. (1962). Rodforholdene hos menneskts underkaebehhometaender med saerligt henblik patorodede hjornetaender, I. Den morfologiske variation. Tandlaegebladet 66, 53-80.

ALEXANDERSEN, V. (1963). Double-rooted human lower canine teeth. In Dental Anthropology (ed. D.Brothwell), pp. 235-244. Oxford: Pergamon Press.

ARAMBOURG, C. & BImERSON, P. (1956). The fossil human remains from the Palaeolithic site of SidiAbderrahman (Morocco). American Journal of Physical Anthropology 14, 467-489.

ARAMBOURG, C. & HoFsTETrER, R. (1963). Le gisement de Ternifine. Archives de l'Institut de PaleontologieHumaine, Memoire 32. Paris: Masson et C" Editeurs.

ARMITAGE, B. (1971). Three-rooted first premolars and their genetic and anthropological significance. BritishDental Journal 130, 245-246.

BARKER, B. C. W. (1973). Dental anthropology: some variations and anomalies in human tooth form.Australian Dental Journal 18, 132-140.

BARKER, B. C. W., PARSONS, K. C., MILLS, P. R. & WILLIAMS, G. L. (1973). Anatomy of root canals. I.Permanent incisors, canines and premolars. Australian Dental Journal 18, 320-327.

BARR, J. H. & GRON, P. (1959). Palate contour as a limiting factor in intraoral X-ray technique. Oral Surgery,Oral Medicine and Oral Pathology 12, 459-472.

BENKOW, H. H. (1957). A new principle and appliance for radiographic tooth measurements. Journal ofDentalResearch 36(4), 641-643.

BEST, E. J., GERVASIO, W., SOWLE, J. T., WINTER, S. & GuRNEY, B. F. (1960). A new method of tooth lengthdetermination for endodontic practice. Dental Digest 66(10), 450-454.

BHASKAR, S. N. (1980). Orban's Oral Histology and Embryology, 9th ed. St Louis: Mosby & Co.BRABANT, H., KLEES, L. & PHILIPPART, R. (1953). Contribution clinique et histologique a 1e'tude des dents et

des racines surnumeraires. Archives de stomatologie 8, 1-24.BRABANT, H. & KOVACS, I. (1961). Contribution i l'etude de la persistance du taurodontisme dans les racesmodernes et de sa parente possible avec la racine pyramidale des molaires. Bulletin du GroupemenE.international pour la recherche scientifique en stomatologie 4(2, 3), 232-286.

BRABANT, H., SAHLY, A. & Bouyssou, M. (1961). Etude des dents prehistoriques de la station archeologiquedes Matelles, department de l'Herault, France. Bulletin du Groupement international pour la recherchescientifique en stomatologie 4, 382-448.

BRAMANTE, C. M. & BERBERT, A. (1974). A critical evaluation of some methods of determining tooth length.Oral Surgery, Oral Medicine and Oral Pathology 37(3), 463-473.

BREGMAN, R. C. (1950). A mathematical model of determining the length of a tooth for root canal treatmentand filling. Journal of the Canadian Dental Association 16(6), 305-306.

BROOM, R. & ROBINSON, J. T. (1952). Swartkrans Ape-man: Paranthropus crassidens. Transvaal MuseumMemoirs, No. 6.

BROOM, R. & SCHEPERS, G. W. H. (1946). The South African fossil Ape-men: the Australopithecinae.Transvaal Museum Memoirs, No. 2, 1-272.

BUTLER, P. M. (1941). A theory of the evolution of mammalian molar teeth, American Journal ofScience 239,421-450.

BUTLER, P. M. (1956). The ontogeny of molar pattern. Biological Reviews 31, 30-70.BUTLER, P. M. (1982). Some problems of the ontogeny of tooth patterns. In Teeth: Form, Function, and

Evolution (ed. B. Kurten), pp. 44-51. New York: Columbia University Press.BUTLER, P. M. (1986). Problems of dental evolution in the higher primates. In Major Topics in Primate andHuman Evolution (ed. B. Wood, L. Martin & P. Andrews), pp. 89-106. Cambridge University Press.

CAMPBELL, T. D. (1925). Dentition and Palate of the Australian Aboriginal. Adelaide: Hansell Press.CARLSEN, 0. (1967). Odontogenetic morphology. Sartryck odontologisk tidskrift 75, 499-516.CAVE, E. H. P. (1929). A new method of root measurement by X-rays. British Dental Journal 50, 245-248.CLARK, W. E. LE GROS. (1971). The Antecedents ofMan, 3rd edition. Edinburgh: Edinburgh University Press.COLYER, F. (1936). Variations and Diseases of the Teeth of Animals. London: John Bale & Danielsson.COPPENS, Y. (1970). Les restes d'Hominides des series inferieures des formation plio-villafranchiennes del'Omo en Ethiopie. Comptes rendus des seances de l'Academie des sciences 271, 2286-2289.

Hominid mandibular postcanine roots 137COPPENS, Y. (1973). Les restes d'Hominides des series superieures des formations plio-villafranchiennes del'Omo en Ethiopie (recoltes 1970, 1971, et 1972). Comptes rendus des se'ances de l'Acadimie des sciences 276,1981-1984.

DE JONGE, TH. E. (1961). Gedanken uiber die Wurzeldifferenzierung des vorderen unteren Praemolaris.Zeitschrift fur Morphologie und Anthropologie 51(2), 220-235.

DE JONGE-COHEN, TH. E. (1919). A contribution to the morphogeny of the 'Dens in Dente'. Dental Cosmos61(3), 224-227.

DE LA PARRA, C. (1952). Een geval van bilaterle volledige wortelsplitsing bij twee eerste onder-praemolaren inmesio-distale richting. Tijdschrift van tandheelkunde 59, 805-807.

DE LA PARRA, C. (1953). Een eerste onderpraemolar met drie wortels. Tijdschrift van tandheelkunde 60, 287-289.

DIAMOND, M. (1952). Dental Anatomy, 3rd ed. New York: MacMillan Co.DIAMOND, M. & APPLEBAUM, E. (1942). The epithelial sheath: histogenesis and function. Journal of Dental

Research 21(4), 403-411.DRENNAN, M. R. (1929). Dentition of a Bushman tribe. Annals of the South African Museum 24, 61-87.DuBois, E. (1924). On the principal characters of the cranium and the brain, the mandible, and the dentition

of Pithecanthropus erectus. Proceedings. Koninklijke nederlandse akademie van wetenschappen te Amsterdam27, 265-278.

DUCKWORTH, W. L. H. (1923). Morphology and Anthropology. Cambridge: Cambridge University Press.EvERETT, F. G. & FixoTT, H. C. (1963). Use of an incorporated grid in the diagnosis of oral roentgenograms.

Oral Surgery, Oral Medicine and Oral Pathology 16, 1061-1064.EVERETT, F. G., JUMP, E. B., HOLDER, T. D. & WILLIAMS, G. C. (1958). The intermediate bifurcational ridge:

a study of the morphology of the bifurcation of the lower first molar. Journal of Dental Research 37,162-169.

FRISCH, J. E. (1963). Dental variability in a population of Gibbon (Hylobates lar). In Dental Anthropology (ed.D. Brothwell), pp. 125-128. Oxford: Pergamon Press.

FRISCH, J. E. (1973). Taxonomic and phylogenetic uses of the study of variability in the hylobatid dentition.Symposia of the Fourth International Congress of Primatology 3, 128-147.

GAUNT, W. A. (1960). The vascular supply in relation to the formation of roots on the cheek teeth of themouse. Acta anatomica 43, 116-136.

GAUNT, W. A. & MILES, A. E. W. (1967). Fundamental aspects of tooth morphogenesis. In Structural andChemical Organization of Teeth, vol. 1 (ed. A. E. W. Miles), pp. 151-197. London: Academic Press.

GHER, M. E. & VERNINO, A. R. (1980). Root morphology - clinical significance in pathogenesis and treatmentof periodontal disease. Journal of American Dental Association 101 (4), 627-633.

GOH, S. W. (1957). Variations in the morphology of mandibular premolar roots. British Dental Journal 102,311-314.

GRANT, D. & BERNICK, S. (1971). Morphodifferentiation and structure of Hertwig's root sheath in the cat.Journal of Dental Research 50, 1580-1588.

GREGORY, W. K. (1920-1). The origin and evolution of the human dentition. Journal of Dental Research 2,89-193; 215-283; 357-426; 607-717; 3, 88-228.

GREGORY, W. K. & HELLMAN, M. (1926). The dentition ofDryopithecus and the origin ofman. AnthropologicalPapers of the American Museum of Natural History 28(1), 1-123.

GREGORY, W. K. & HELLMAN, M. (1939). The dentition of the extinct South-African man-ape Australopithecus(Plesianthropus) transvaalensis Broom. A comparative and phylogenetic study. Annals of the TransvaalMuseum 19, 339-373.

GRON, P. (1960). A geometrical evaluation of image size in dental radiography. Journal of Dental Research39(2), 289-301.

GROVES, C. P. & MAZAK, V. (1975). An approach to the taxonomy of the Hominidae: Gracile Villafranchianhominids of Africa. Casopis pro mineralogu a geologu 20, 225-247.

HAMERNIK, F. J. (1957). A ruler for measuring radiographic magnification. Medical Radiography andPhotography 33, 19.

HOWELL, F. C. (1960). European and North West African Middle Pleistocene hominids. Current Anthropology1(3), 195-232.

HUNT, K. & VITZTHUM, V. J. (1986). Dental metric assessment of the Omo fossils: implications for thephylogenetic position of Australopithecus africanus. American Journal of Physical Anthropology 71, 144-155.

JACOB, T. (1973). Palaeoanthropological discoveries in Indonesia with special reference to the finds of the lasttwo decades. Journal ofHuman Evolution 2, 473485.

JAMES, W. W. (1960). The Jaws and Teeth of Primates. London: Pitman Medical Press.JOHANSON, D. C. (1985). The most primitive Australopithecus. In Hominid Evolution: Past, Present and

Future (ed. P. V. Tobias), pp. 203-212. New York: Alan R. Liss, Inc.JOHANSON, D. C. & WHITE, T. D. (1979). A systematic assessment of early African hominids. Science 203,

321-330.JORGENSEN, K. D. (1950). Macroscopic observations on the formation of the subpulpal wall. Odontologisk

tidskrift 58, 83-103.JUNGERS, W. L. & GRINE, F. E. (1986). Dental trends in the australopithecines: the allometry of mandibular


molar dimensions. In Major Topics in Primate andHuman Evolution (ed. B. Wood, L. Martin & P. Andrews),pp. 203-219. Cambridge: Cambridge University Press.

KEITH, A. (1913). Problems relating to the teeth of the earlier forms of prehistoric man. Proceedings of theRoyal Society of Medicine (Odontological Section) 6(3), 103-124.

KIMBEL, W. H., WHITE, T. D. & JOHANSON, D. C. (1984). Cranial morphology of Austalopithecus afarensis: acomparative study based on a composite reconstruction of the adult skull. American Journal of PhysicalAnthropology 64, 337-388.

KOLLAR, E. J. & BAIRD, G. E. (1969). The influence of the dental papilla on the development of tooth shapein embryonic mouse tooth germs. Journal of Embryology and Experimental Morphology 21, 131-148.

KOVACS, I. (1963). Contribution a l'etude de la morphologie ontogenetique des racines des dents humaines.Bulletin du Groupement international pour la recherche scientifique en stomatologie 6, 95-124.

KOVACS, I. (1966). Le radico-diagramme: characteristique de l'inclinaison, de la diverge, et de la courbure desracines dentaires. Bulletin du Groupement international pout la recherche scientifique en stomatologie 9, 59-90.

KOVACS, I. (1967). Contribution to the ontogenetic morphology of roots of human teeth. Journal of DentalResearch 46(5), Suppl. 865-874.

KOVACS, I. (1971). A systematic description of dental roots. In Dental Morphology and Evolution (ed. A. A.Dahlberg), pp. 211-256. Chicago: University of Chicago Press.

KRAUS, B. S., JORDAN, R. E. & ABRAMS, L. (1969). Dental Anatomy and Occlusion. Baltimore: Williams andWilkins Co.

LEAKEY, L. S. B. (1968). The lower dentition of Kenyapithecus africanus. Nature 217, 827-830.LEAKEY, L. S. B. (1970). 'Newly' recognised mandible of Ramapithecus. Nature 225, 199-200.LEAKEY, L. S. B., TOBIAS, P. V. & NAPIER, J. R. (1964). A new species of the genus Homo from Olduvi Gorge.

Nature 202, 7-9.LEAKEY, R. E. F. (1974). Further evidence of Lower Pleistocene hominids from East Rudolf, North Kenya,

1973. Nature 248, 653-656.MCCORMACK, D. W. (1937). Dental Roentology: a technical procedure to furthering the advancement toward

anatomical accuracy. Journal ofthe California State Dental Association and Nevada State Dental Association13, 89-116.

NELSON, C. T. (1938). The teeth of the Indians of Pecos Pueblo. American Journal of Physical Anthropology23, 261-293.

ORBAN, B. & MUELLER, E. (1929). The development of the bifurcation of multi-rooted teeth. Journal of theAmerican Dental Association 16, 297-319.

OSTER, G. & ALBERCH, P. (1982). Evolution and bifurcation of developmental programs. Evolution 38, 444-459.

OWEN, R. (1840-5). Odontography: or a Treatise on the Comparative Anatomy ofthe Teeth. London: HippolyteBailliere.

PEARLMAN, C. (1972). Mesio-distal two-rooted premolars. Oral Surgery, Oral Medicine and Oral Pathology 33,144.

PEDERSEN, P. 0. (1949). The East Greenland Eskimo dentition. Meddelelser om Gronland 142, 1-244.PEYER, B. (1968). Comparative Odontology. Chicago: University of Chicago Press.QUINET, G. E. (1966). Le gradient morphogenetique responsable de la formule dentaire mammalienne. Bulletin

de l'Institut royale des sciences naturelles de Belgique 42, 1-13.RAK, Y. (1983). The Australopithecine Face. New York: Academic Press.RIGHTMIRE, G. P. (1980). Middle Pleistocene hominids from Olduvai Gorge, Northern Tanzania. American

Journal of Physical Anthropology 53, 225-241.ROBINSON, J. T. (1952). Some hominid features of the Ape-man dentition. Official Journal of the Dental

Association of South Africa 7, 102-113.ROBINSON, J. T. (1956). The dentition of the Australopithecinae. Transvaal Museum Memoirs, No. 9.SCHOUR, I. & MASSLER, M. (1940). Studies in tooth development. The growth pattern of human teeth. Part 1.

Journal of the American Dental Association 27, 1778-1793.SCOTT, J. H. & SYMONS, N. B. B. (1974). Introduction to Dental Anatomy, 7th ed. London: Churchill

Livingstone.SELMER-OLSEN, R. (1949). An Odontometrical Study on the Norwegian Lapps. Oslo: Det Norske VidenskapsAkademi.

SENYOiREK, M. S. (1953). A study of the pulp cavities and roots of the lower premolars and molars of Prosimii,Ceboidea, and Cercopithecoidea. Belleten, Ankara 17, 321-365.

SHAPIRA, Y. & DELIVANIS, P. (1982). Multiple-rooted mandibular second premolars. Journal of Endodontics8(5), 231-232.

SHAw, J. C. M. (1928). Taurodont teeth in South African races. Journal of Anatomy 62, 476-499.SHAw, J. C. M. (1931). The Teeth, the Bony Palate and the Mandible in Bantu Races ofSouth Africa. London:John Bale, Sons & Danielsson.

SKELTON, R. R., McHENRY, H. M. & DRAWHORN, G. M. (1984). Phylogenetic analysis of Plio-Pleistocenehominids. American Journal of Physical Anthropology 63, 219.

SPERBER, G. (1974). Morphology of the cheek teeth of early South African hominids. Ph.D. thesis, Universityof the Witwatersrand.

Hominid mandibular postcanine roots 139SPRINZ, R. (1953). The linking tooth: a rare anomaly of a mandibular premolar. British Dental Journal 95,

108-109.TAYLOR, A. E. (1899). Variations in the human tooth-form as met with in isolated teeth. Journal ofAnatomy

and Physiology 33, 268-272.TEKKAYA, I. (1974). A new species of Tortonian Anthropoid (Primates, Mammalia) from Anatolia. Bulletin

of the Mineral Research and Exploration Institute of Turkey 83, 148-165.TOBIAS, P. V. (1969). Notes on the Baringo mandible. In An Acheulean Industry with prepared core techniqueand the discovery of a contemporary hominid mandible at Lake Baringo, Kenya (ed. M. Leakey, P. V.Tobias, J. E. Martyn & R. E. F. Leakey). Proceedings of the Prehistoric Society 35, 48-76.

TOBIAS, P. V. (1980). 'Australopithecus afarensis' and A. africanus: Critique and alternative hypothesis.Palaeontologica africana 23, 1-17.

ToMws, C. (1923). Manual ofDental Anatomy, 8th ed. (ed. H. W. Marrett Tims). London: J. & A. Churchill.TRATMAN, E. K. (1938). Three-rooted lower molars in man and their racial distribution. British Dental Journal

64, 264-274.TRATMAN, E. K. (1950). A comparison of the teeth of people: Indo-European racial stock with the Mongoloid

racial stock. Dental Record 70(2-3), 31-53; 63-88.TURNER, C. G. (1967). Dental genetics and microevolution in prehistoric and living Kariag Eskimo. Journalof Dental Research 46, 911-917.

VON KOENIGSWALD, G. H. R. (1968). Observations upon two Pithecanthropus mandibles from Sangiran,Central Java. Proceedings. Koninklijke Nederlandse akademie van wetenschappen te Amsterdam 71, 99-107.

VON KOENIGSWALD, G. H. R. (1969). Java: prae-Trinil man. Proceedings of the 8th International Congress ofAnthropological and Ethnological Science, 1968, TokyolKyoto 1, 104-105.

WARD, S. C. (1979). Subocclusal dental morphology of Pliocene hominid mandibles from the HadarFormation. American Journal of Physical Anthropology 50, 490.

WARD, S. C., JOHANSON, D. C. & COPPENS, Y. (1982). Subocclusal morphology and alveolar processrelationships of hominid gnathic elements from the Hadar Formation: 1974-1977 collections. AmericanJournal of Physical Anthropology 57, 605-630.

WEDENREICH, F. (1937). The dentition of Sinanthropus pekinensis: a comparative odontography of thehominids. Palaeontologica Sinica, New Series D, No. 1. Whole series 101.

WEIDENREICH, F. (1945). Giant early man from Java and South China. Anthropological Papers ofthe AmericanMuseum of Natural History 40(1), 1-134.

WHITE, T. D. (1977). New fossil hominids from Laetoli, Tanzania. American Journal ofPhysical Anthropology46(2), 197-230.

WHImTE, T. D. (1980). Additional fossil hominids from Laetoli, Tanzania: 1976-79 specimens. AmericanJournal of Physical Anthropology 53, 487-504.

WmTE, T. D., JOHANSON, D. C. & KIMBEL, W. H. (1981). Australopithecus africanus: Its phyletic positionreconsidered. South African Journal of Science 77, 445-470.

Woo, J-K. (1962). The mandibles and dentition of Gigantopithecus. Palaeontologia Sinica, New Series D, No.11. Whole number 146.

WOOD, B. A. (1981). Tooth size and shape and their relevance to studies of hominid evolution. PhilosophicalTransactions of the Royal Society, London, B Series 292, 65-76.

WOOD, B. A. & ABsorr, S. A. (1983). Analysis of the dental morphology of Plio-Pleistocene hominids. I.Mandibular molars - crown area measurements and morphological traits. Journal of Anatomy 136, 197-219.

WOOD, B. A., ABBorr, S. A. & GRAHM, S. H. (1983). Analysis of the dental morphology of Plio-Pleistocenehominids. II. Mandibular molars - study of cusp areas, fissure pattern and cross sectional shaped of thecrown. Journal ofAnatomy 137, 287-314.

WOOD, B. A. & UYmRscHAur, H. (1987). Analysis of the dental morphology of Plio-Pleistocene hominids.III. Mandibular premolar crowns. Journal of Anatomy 154, 121-156.

university of groningen analysis of the dental morphology ... · hominidmandibularpostcanine roots...

Documents