Ancestral dental traits in recent Sub-Saharan Africans and the origins of modern humans

18 Pages·1998·0.12 MB·English

by  Irish, J. D.

Joel D. Irish Ancestral dental traits in recent
DepartmentofAnthropology, Sub-Saharan Africans and the origins of
UniversityofNewMexico, modern humans
Albuquerque,NM
87131-1086
Assuming that phenetic expression approximates genetic variation,
E-mail:[email protected]
previous dental morphological analyses of Sub-Saharan Africans by
Received6January1997 the author show they are unique among the world’s modern popu-
Revisionreceived lations. Numerically-derived aYnities, using the multivariate Mean
3September1997and Measure of Divergence statistic, revealed significant diVerences
accepted8September1997 between the Sub-Saharan folk and samples from North Africa,
Europe, Southeast Asia, Northeast Asia and the New World,
Keywords:Africa,dental Australia/Tasmania, and Melanesia. Sub-Saharan Africans are char-
anthropology,morphological acterized by a collection of unique, mass-additive crown and root
variation,biologicalaYnity, traitsrelativetotheseotherworldgroups.Recentworkfoundthatthe
pheneticsimilarity,ancestral mostubiquitousofthesetraitsarealsopresentindentitionsofearlier
andderivedcharacters, hominids, as well as extinct and extant non-human primates; other
humanorigins. ancestraldentalfeaturesarealsocommonintheseforms.Thepresent
investigation is primarily concerned with this latter finding. Quali-
tative and quantitative comparative analyses of Plio-Pleistocene
through recent samples suggest that, of all modern populations,
Sub-SaharanAfricansaretheleastderiveddentallyfromanancestral
hominid state; this conclusion, together with data on intra- and
inter-population variability and divergence, may help provide new
evidenceinthesearchformodernhumanorigins.
(cid:63)1998AcademicPressLimited
JournalofHumanEvolution(1998)34,81–98
Introduction morphologically simple, mass-reduced den-
tal traits with Europeans (e.g., molar cusp
In several other studies (Irish, 1993, 1994, number and size reduction, M3 agenesis,
1995, 1997, 1998), dental morphological etc.). Sub-Saharan Africans, however, diVer
variation in 28 late Pleistocene through markedly from all others in their recurrent
recent Sub-Saharan and North African expression of such complex, mass-additive
sampleswasanalyzedtoestimatesynchronic traitsasBushmanCanine,two-rootedUP1,
anddiachronicintra-andinter-regionalbio- UM1 Carabelli’s trait, three-rooted UM2,
logical relatedness. Dental morphological LM2 Y-groove pattern, LM1 cusp 7, LP1
homogeneity was found among both Sub- Tome’s root, two-rooted LM2, and UM3
Saharan and, most notably, North African presence. They are also characterized by a
samples. However, significant diVerences verylowprevalenceofUI1doubleshoveling
wererevealedbetweensamplesfromthetwo andUM1enamelextension.Thisdiagnostic
regions.Subsequentcomparativeanalysesof suite of nine high- and two low-frequency
these data with those recorded by Turner features was termed the ‘‘Sub-Saharan
(1985, 1987, 1992a,b), in samples from African Dental Complex’’ (Irish, 1997).
Europe, Southeast Asia (Sundadonts), Further analysis (Irish, 1997) revealed
Northeast Asia/New World (Sinodonts), that the high-frequency traits (along with
Melanesia, and Australia/Tasmania, helped severalotherdentalfeatures)arealsopreva-
identifyAfrica’splacefromaglobalperspec- lent in the dentitions of many extinct
tive. North Africans share a number of hominids, from australopithecenes through
0047–2484/98/010081+18$25.00/0 (cid:63)1998AcademicPressLimited
82 j. d. irish
archaic Homo sapiens, as well as extinct and Sakuma & Ogata, 1987; Haeussler et al.,
extantnon-humanprimates.Thus,thetraits 1988; Turner & Markowitz, 1990; Irish &
apparentlyrepresentveryancientcharacters. Turner, 1990; Irish, 1993, 1994, 1997,
The goal of the present study is to investi- 1998). Only each specimen’s highest
gate this finding in greater detail. First, to antimereexpressionswereanalyzedforeach
providearequisitebackgroundforthisgoal, trait;thismaximizesthegeneticpotentialfor
themethodologyemployedandresultsfrom these polygenic features (see Turner &
previous dental analyses are summarized. Scott,1977).Becauseofadocumentedlack
Second, a list of extinct and extant homi- of sexual dimorphism for the traits (Scott,
noids that exhibit the diagnostic Sub- 1973, 1980; Turner et al., 1991, personal
Saharantraitsisprovided.Third,totestthe communication, 1993; Hanihara, 1992;
degree to which Sub-Saharan Africans are Irish, 1993, 1997; among others), it is stan-
characterized by these seemingly ancient dard System protocol to pool the sexes
dental traits, qualitative and quantitative (Irish, 1997); this protocol is followed here.
comparisons are undertaken between the For a complete description of ASU System
seven modern groups (Sub-Saharan and procedures and traits see Turner et al.
North Africa, and five other world samples) (1991).
andthreesamplesofPlio-Pleistocenehomi- After recording, trait frequencies were
nids,including:apooledsampleof‘‘robust’’ determined for each sample and C. A. B.
African hominids (Australopithecus robustus Smith’s Mean Measure of Divergence
and A. boisei), a pooled sample of ‘‘gracile’’ (MMD) statistic, using the Freeman and
African hominids (A. africanus and Homo Tukey angular transformation (Berry &
habilis), and a sample of archaic Homo Berry, 1967; Sjøvold, 1973; Green &
sapiens—the Krapina Neandertals. Finally, Suchey, 1976) to correct for small sample
the ramifications of this apparent African sizes, was employed. This multivariate
traitantiquityarediscussedregardingacon- technique provides a quantitative estimate
tributiontobetterunderstandingtheorigins of biological divergence among samples
of modern humans. based on the degree of phenetic similarity
for all traits. An MMD calculated between
each sample pair is a dissimilarity measure,
Previous dental analyses
so a lower value indicates greater aYnity,
Methodology and vice versa. It is assumed that phenetic
For all studies, the Arizona State University similarityapproximatescladisticrelationship
(ASU) Dental Anthropology System was (see Sokal & Sneath, 1963; Scott et al.,
employed; it comprises a set of rank-scale 1983). To determine if samples signifi-
referenceplaquesandvariousproceduresto cantly diVer from one another, an MMD is
helpstandardizetheobservationandrecord- compared to its standard deviation. If
ing of >40 common crown, root, and intra- the MMD>2(cid:35)SD, the null hypothesis
oral osseous morphological traits of the (P1=P2, where P=sample population) is
human permanent dentition (e.g., incisor rejected at the 0·025 significance level
shoveling, molar cusp number, groove pat- (Sjøvold, 1977).
tern, root number, etc.). Procedures used It is recommended that as many discrete
in the ASU System are based on well- traits as possible be used in determining
established criteria for scoring intra-trait MMDs (see Sjøvold, 1977). More traits
variation, and have proven to be reliable in providemorepreciseaYnityestimates;con-
many preceding studies (see Scott, 1973, versely, a low number of traits (e.g., <10)
1980; Turner, 1985, 1987, 1990, 1992a,b; may provide questionable results. In no
ancestral sub-saharan dental traits 83
instance should any of these traits be corre- (1993, 1997, 1998), and are not repeated
lated, otherwise diVerential weighting of the here for the sake of brevity.
underlying dimensions can lead to erron- MMD analyses revealed a significant
eousresults(Sjøvold,1977).Thus,alltraits dichotomybetweenSub-SaharanandNorth
were tested for undesired correlation with African samples. However, measurable
the Kendall’s tau-b and Spearman’s rho sample homogeneity within regions is
rank-order correlation coeYcient statistics; apparent, particularly in North Africa. The
no significant correlations were found (see average North Africa intra-region MMD is
Irish, 1993). The MMD statistic also 0·019 (sum of all pairwise MMD compari-
requires that the rank-scale traits be dichot- son values divided by the total number of
omized into categories of present or absent comparisons),whichindicatesahighlevelof
(Sjøvold, 1977); this was eVected based dental homogeneity. The Sub-Saharan
on each trait’s appraised morphological samples also show dental phenetic simili-
threshold (Scott, 1973; Haeussler et al., tude, but to a somewhat lesser degree. The
1988),accordingtostandardprocedure(see average MMD among these samples is
Turner, 1985, 1987). On occasion, these 0·051,roughlythreetimesthatoftheNorth
standardized breakpoints may be varied, up Africans. Lastly, there is a sizable diver-
or down, if a particularly ubiquitous trait gence between the Sub-Saharan and North
in a population (e.g., Sub-Saharan African Africans; the average inter-region MMD is
cusp 7 LM1) needs to be emphasized or 0·125. These findings support those based
de-emphasized for more accurate intra- ongenetic,skeletal,anthropometric,linguis-
regional comparisons (see Irish, 1993, tic, cultural, and various other data (see
1997). Such variation is unavoidable when Mourant, 1954, 1983; Greenberg, 1959,
usingpublisheddataandmustbeaccounted 1966; Murdock, 1959; Hiernaux, 1975;
for when attempting comparisons between Nurse et al., 1985; Roychoudhury & Nei,
studies (refer to traits in Tables 1 and 2). 1988; Howells, 1989).
Becauseofobviousintra-regionhomogen-
African sample comparisons eity, the 28 samples, plus four samples
Twenty-eight Late Pleistocene through (n=106 individuals) with less certain cul-
recent samples were originally assembled turalaYliationfromknowngeographicareas
andcompared(Irish,1993);theycomprisea (i.e., west, east, south, and north Africa),
representative cross-section of Sub-Saharan were pooled into two large groups—Sub-
(n=16 samples) and North African (n=12) Saharan (n=976 individuals) and North
cultures.1 A total of 1529 dentitions were Africa(n=659).Thiscombiningofdatawas
analyzed, and up to 36 of the most diag- donetofacilitatecomparisonwithpublished
nostic ASU System traits (refer to Table 1 dental data from several large, pooled
for a partial listing) were observed in each. non-African regional samples: Europe,
The samples, dental traits, and rationale for Sundadonts, Sinodonts, Melanesia, and
their study are described in detail in Irish Australia/Tasmania (see Turner, 1985,
1987, 1992a,b).
1The 28 samples incorporate the following tribes/
culture groups: Sub-Saharan Africa—Teke, Kongo,
Fang, Nkomi, Lumbo, Mpongwe, Ashanti, Fanti, African and non-African sample comparisons
Kikuyu, Swahili, Chaga, Pare, Nama, Korana, Efik,
Table 1 lists dental trait frequencies for
Ibibio, Boki, Anyang, Binga, Bongo, Nguni, Zulu,
Xhosa, !Kung, Naron, Tshakwe, Mkaukau, Wolof, Sub-Saharan and North Africa, Europe,
Balante,Serer,Sotho,Pedi,Nyamweze,Ngindo,Hehe, Southeast Asia (Sundadonts), and North-
Gogo, Ewe, Fon, Tukulor; North Africa—Berber,
east Asia and the New World (Sinodonts).
Bedouin, Guanche, Carthaginian, Toubou, Masalit,
Kanembu,NubianandEgyptian. Table 2 lists frequencies for both
84 j. d. irish
Table1 Comparison of 27 dental traits in Sub-Saharan African (SSA), North African (NAF),
European(EUR),Sundadont(SUN),andSinodont(SIN)*samples(fromIrish,1997)
Sample
Trait† SSA NAF EUR SUN SIN
WingingUI1 % 6·6 7·4 9·0 22·8 41·4
(+=ASU1) n 742 460 129 219 1954
ShovelUI1 % 28·1 19·5 17·0 79·2 98·8
(+=ASU2–6) n 413 154 141 202 1922
DoubleShovelUI1 % 1·1 8·6 23·3 14·9 71·0
(+=ASU2–6) n 437 175 137 201 1781
TuberculumDentaleUI2 % 61·2 58·7 38·1 58·1 64·2
(+=ASU1–6) n 454 189 152 186 1990
BushmanCanineUC % 18·1 6·1 4·8 2·0 1·2
(+=ASU1–3) n 586 261 146 245 2280
DistalAcc.RidgeUC % 71·8 34·9 51·7 65·0 73·9
(+=ASU1–5) n 483 195 89 139 1073
HypoconeUM2 % 99·0 95·7 79·4 92·0 90·2
(+=ASU1–5) n 772 446 228 414 3639
Cusp5UM1 % 32·8 18·5 13·7 30·0 19·0
(+=ASU1–5) n 619 357 212 370 2817
Carabelli’sTraitUM1 % 51·2 54·7 47·4 30·6 32·1
(+=ASU2–7) n 683 331 230 427 3194
ParastyleUM3 % 2·0 1·2 4·5 1·3 4·8
(+=ASU1–5) n 550 332 134 153 2400
EnamelExtensionUM1 % 9·4 6·8 43·7 57·5 68·5
(+=ASU1–3) n 574 503 286 388 5135
RootNo.UP1 % 58·9 57·1 34·0 40·8 12·2
(+=ASU2+) n 570 468 238 441 4757
RootNo.UM2 % 83·7 78·6 53·8 73·9 51·8
(+=ASU3+) n 503 374 197 333 3718
OdontomeP1–2 % 0·4 0·2 1·2 1·4 4·6
(+=ASU +) n 756 441 171 146 2738
Peg-reducedabsentUM3 % 5·4 18·3 20·7 44·0 22·5
(+=ASUP,R,C) n 708 545 241 300 4623
LingualCuspNo.LP2 % 68·5 72·6 62·9 79·1 47·2
(+=ASU2–9) n 530 270 159 278 2393
GroovePatternLM2 % 52·4 30·6 22·9 19·6 10·9
(+=ASUY) n 617 402 214 342 3783
CuspNo.LM1 % 16·6 7·7 7·9 35·5 47·8
(+=ASU6+) n 561 352 178 282 2947
CuspNo.LM2 % 24·1 66·4 65·1 40·4 7·9
(+=ASU4) n 585 381 189 317 3583
DeflectingWrinkleLM1 % 30·1 24·7 30·9 55·3 70·7
(+=ASU1–3) n 432 267 149 161 1817
C1–C2CrestLM1 % 1·3 3·3 8·6 7·3 5·4
(+=ASU +) n 447 276 185 96 2825
ProtostylidLM1 % 21·0 32·5 20·0 30·0 34·7
(+=ASU1–6) n 556 351 200 337 3739
Cusp7LM1 % 38·5 9·4 5·8 7·4 9·8
(+=ASU1–4) n 598 414 223 367 3998
Tome’sRootLP1 % 22·4 8·6 7·4 9·8 8·2
(+=ASU3–5) n 361 372 148 133 2883
RootNo.LC % 0·0 2·3 4·8 0·0 0·5
(+=ASU2+) n 333 347 207 207 3722
RootNo.LM1 % 1·7 1·2 0·8 9·3 13·8
(+=ASU3+) n 409 337 254 343 5192
RootNo.LM2 % 93·3 88·3 71·4 81·5 65·5
(+=ASU2+) n 388 333 224 297 4346
*†DataforSUN(S.E.Asia)andSIN(N.E.AsiaandNewWorld)andstandardrank-scaledbreakpointsfrom
Turner (1985). Data for EUR from Turner (1985) and from Turner’s unpublished Poundbury data. All
frequenciesareofindividualcounts.
ancestral sub-saharan dental traits 85
Table2 Comparison of 14 dental traits in Sub-Saharan African (SSA), North African (NAF),
Australia-Tasmanian(AUT),andMelanesian(MEL)*samples(fromIrish,1997)
Sample
Trait† SSA NAF AUT MEL
ShovelUI1 % 5·3 3·2 15·9 9·3
(+=ASU3–6) n 413 154 44 118
DoubleShovelUI1 % 1·1 8·6 7·0 5·0
(+=ASU2–6) n 437 175 43 118
InterruptionGrooveUI2 % 13·4 36·1 21·6 18·4
(+=ASU +) n 471 208 51 147
BushmanCanineUC % 18·1 6·1 5·0 3·2
(+=ASU1–3) n 586 261 60 156
Cusp5UM1 % 32·8 18·5 68·2 62·4
(+=ASU1–5) n 619 357 88 221
EnamelExtensionUM1 % 0·3 3·0 8·1 4·4
(+=ASU2–3) n 574 503 208 250
RootNo.UM2 % 83·7 78·6 82·6 73·3
(+=ASU3+) n 503 374 178 240
Peg-reducedabsentUM3 % 5·4 18·3 6·4 12·6
(+=ASUP,R,C) n 708 545 220 270
CuspNo.LM2 % 24·1 66·4 7·2 53·0
(+=ASU4) n 585 381 97 219
DeflectingWrinkleLM1 % 2·3 0·4 17·1 17·6
(+=ASU3) n 432 267 35 136
Cusp7LM1 % 38·5 9·4 8·2 11·7
(+=ASU1–4) n 598 414 97 240
Tome’sRootLP1 % 22·4 8·6 21·6 12·1
(+=ASU3–5) n 361 372 83 116
RootNo.LM1 % 1·7 1·2 5·7 3·2
(+=ASU3+) n 409 337 157 222
RootNo.LM2 % 93·3 88·3 91·4 94·7
(+=ASU2+) n 388 333 152 209
*†DataforAUTandMEL,andstandardrank-scaledbreakpointsfromTurner(1987,1992a).Allfrequencies
areofindividualcounts.
African samples, Australia/Tasmania, and MMD comparisons among the seven
Melanesia. Just 27 and 14 traits are worldsamplesare,accordingly,basedon27
presented in each table, respectively, and 14 traits (with appropriately adjusted
becausethesearetheonlydatapresentedin breakpoints in the African samples). As
Turner’s various articles. In each table, the such,MMDvaluesarenotdirectlycompar-
percentage of individuals per sample with able between the resulting matrices in
a particular trait is presented, along with Tables 3 and 4. Such complexities are un-
the total number of individuals for whom avoidable when using published summary
the trait was scored. The dichotomized data.
presence/absence breakpoints are listed As seen, the North Africans are most
under the trait names. Several trait break- similar to Europeans; the 27-trait MMD
points also vary between tables (i.e., UI1 between samples is 0·154. Many trait simi-
shovel, UM1 enamel extension, LM1 larities are evident (see Table 1), including
deflecting wrinkle) because of variation in low frequencies of UI1 winging, UM1
thepublisheddata;thisshouldbetakeninto cusp-5, UM1 deflecting wrinkle, LP1
consideration when comparing frequencies. Tome’sroot,andthree-rootedLM1,aswell
86 j. d. irish
Table3 Meanmeasureofdivergencevaluesamongfivemodernworldsamples*based
on27dentaltraits
SSA NAF EUR SUN SIN
Sub-SaharanAfrica —
NorthAfrica 0·166† —
Europe 0·244 0·154 —
Sundadonts 0·279 0·279 0·166 —
Sindodonts 0·610 0·671 0·406 0·172 —
*Seetextforsampledetails.
†AllMMDsaresignificantlydiVerent.
Table4 Meanmeasureofdivergencevaluesamongfourmodernworldsamples*based
on14dentaltraits
SSA NAF AUT MEL
Sub-SaharanAfrica —
NorthAfrica 0·216† —
Australia/Tasmania 0·152 0·403 —
Melanesia 0·143 0·265 0·085 —
*Seetextforsampledetails.
†AllMMDsaresignificantlydiVerent.
as high frequencies of UI2 interruption ties may instead be due to the retention
groove and UM3 reduction/absence. Of the of ancestral traits in the two populations
27 traits, only UI1 double shoveling and (seeStringeretal.,1997,andbelow).Again,
UM1 enamel extension incidences vary it should be remembered that only 14
noticeably between samples. Any North traits are available for use in these MMD
African deviations away from the mass- comparisons; thus, they are not directly
reducedEuropean-liketraitsaregenerallyin comparable to the 27-trait aYnities. For
the direction of mass-additive Sub-Saharan example, the 14-trait MMD between
Africans—suggestingsomegeneticinfluence Sub-Saharan and North Africa is 0·216,
from these southern peoples (27-trait compared to 0·166 above. Nevertheless,
MMD=0·166). North Africans are wholly general population aYnity trends can be
unlike the remaining samples (MMD assessed.
range=0·265–0·671) (Tables 3 and 4). Sub-Saharan Africans are least like
Sub-Saharan Africans show some re- Sinodonts (27-trait MMD=0·610) (Table
semblance to the Australia/Tasmania and 3). The Sub-Saharan sample displays the
Melanesia samples, particularly regarding highest frequencies of Bushman Canine,
root traits (MMDs=0·152 and 0·143) two-rooted UP1, three-rooted UM2, LM2
(Table 4). Other workers also noted some Y-groove,LP1Tome’sroot,andtwo-rooted
aYnity between Sub-Saharan Africa and LM2; Sinodonts have among the lowest
Australia populations (Giblett, 1969; Nurse frequencies of these traits. Further, Sub-
etal.,1985;Howells,1989;Brace&Tracer, Saharan Africans have very low occurrences
1990; Turner, 1992a). However, no direct of UI1 winging, UI1 double shovel, UM1
connectionhasbeenfound,andthesimilari- enamel extension, UM3 reduction/absence,
ancestral sub-saharan dental traits 87
premolar odontomes, LM1 deflecting KNM-WT 15000 Homo erectus (Robinson,
wrinkle, and three-rooted LM1; Sinodonts 1956; Wood & Engleman, 1988; Aiello &
exhibit among the highest (Irish, 1993, Dean, 1990; Tobias, 1991: Plates 93 and
1994, 1995, 1997, 1998). 97; Brown & Walker, 1993); it is also
Based on this seven sample comparison, present in living gibbons, chimpanzees, and
I proposed (Irish, 1997) that the ‘‘Sub- gorillas (Swindler, 1976, personal com-
Saharan African Dental Complex’’ (those munication, 1994; Turner, 1992a). Dry-
traits which best characterize Sub-Saharan opithecus and many later hominoids display
Africans) comprises a suite of two low- and the diagnostic LM2 Y-groove (Gregory,
nine high-frequency traits. This Complex 1922; Gregory & Hellman, 1926; Johanson
includes among the world’s lowest fre- etal.,1982;Wood&Abbott,1983;Woodet
quenciesofUI1doubleshovelingandUM1 al., 1983; Brown & Walker, 1993). The
enamel extension, and among the world’s LM1 cusp 7 trait is present in A. robustus,
highest frequencies of Bushman Canine, OH 7 Homo habilis, KNM-WT 15000, and
two-rooted UP1, UM1 Carabelli’s trait, the anatomically-modern Klasies River
three-rooted UM2, LM2 Y-groove, LM1 Mouth hominid (Wood & Abbott, 1983;
cusp7,LP1Tome’sroot,two-rootedLM2, Tobias, 1991: Plates 66–67; Turner, per-
and M3 presence (or correspondingly, the sonal communication, 1993); it also occurs
lowest frequency of M3 absence). Some of in living ape species (Swindler, 1976). LP1
these diagnostic traits are also found in Tome’sroothasbeenfrequentlyrecordedin
comparable frequencies in other world australopithecines, the OH 4 Homo habilis,
samples,buttheappropriatecombinationof Homoerectus,andlivingapes(Dubois,1924;
all 11 traits clearly identifies a Sub-Saharan von Koenigswald, 1968; Ward et al., 1982;
pattern (refer to Tables 1 and 2). Wood et al., 1988; Tobias, 1991: Plate 87;
Turner&Hawkey,1991).BushmanCanine
(ASU grades 1–2) has been observed in
The present investigation
Krapina Neandertals (Stringer, personal
Ancestral Sub-Saharan dental traits communication, 1996; personal observation
In the dental studies summarized above, by author) and is bilaterally expressed
phylogenetic relationships of the samples (ASUGrades2and3)inAmud1(personal
were not a concern. However, a subsequent observation by author), although, it has not
literature review (e.g., Gregory, 1922; yet been observed in any African fossils
Gregory & Hellman, 1926; Weidenreich, (Tobias, personal communication, 1997).
1937; Hillson, 1986; Le Gros Clark, 1960; Theremainingfourhigh-frequencytraits,as
Dahlberg, 1945, 1968; Wood & Abbott, wellasothercommonSub-Saharanfeatures
1983; Wood et al., 1983, 1988; Wood & (i.e., UI2 tuberculum dentale, UM2
Engleman, 1988; Tobias, 1991; Brown & hypocone, UM1 cusp 5, LM1 protostylid,
Walker, 1993, among others) revealed that and 5-cusped LM2), are also present in
the same nine high-frequency traits are various Plio-Pleistocene forms, including
also ubiquitous in the dentitions of extinct australopithecines, Homo habilis, Homo erec-
hominids and many extinct and extant tus, and archaic Homo sapiens (Dahlberg,
non-human primates (Irish, 1997). 1947; Ward et al., 1982; Wood & Abbott,
The UM1 Carabelli’s trait is present in 1983; Wood et al., 1983, 1988; Wood
Australopithecus afarensis (LH-17 has a large & Engleman, 1988; Tobias, 1991; Scott,
protoconal cingulum), Australopithecus afri- personal communication, 1992; Brown &
canus, Australopithecus robustus (SK-13), Walker, 1993; personal observation by
Homo habilis (OH 21 and 44), and the author).
88 j. d. irish
The presence and, indeed, prevalence ancestral features since very remote times’’
(see next section), of high-frequency Sub- (Zubov, 1992a: 6).
Saharan dental traits in fossil and recent A final ancestral feature found with some
hominoids—some of which are probably regularity in Sub-Saharan Africans, relative
direct ancestors of modern humans, sug- to other modern groups, is polydontia.
geststheyhavebeenaroundforalongtime. Numerous cases of extra incisors, third
This is not a cladistic study,2 but to borrow premolars, and fourth molars have been
a few terms and concepts—the traits are noted (e.g., Randell, 1925; Shaw, 1931;
apparently ancestral or plesiomorphic (see DeVilliers, 1968; personal observation by
Hennig, 1966; Mettler et al., 1988). As author). In one study (Watters, 1962) the
Delson(1977:436)states‘‘...thepresence incidence reached 2·5–3% in several hun-
of one of them [a character] in an ancient dred west Africans; many of the extra teeth
fossil would be suYcient grounds to accept were fully formed and erupted. ‘‘Typical’’
that state as ancestral ...’’ Moreover, mammals exhibit three incisors and four
because the low-frequency UI1 double premolars (Jordan et al., 1992). Polydontia
shovel and UM1 enamel extension traits is also found in living non-human
have not been observed in fossil forms, they primates—often in notably higher fre-
are likely derived or apomorphic. quencies than most modern human popu-
Zubov (1992a,b, personal communi- lations (Lavelle & Moore, 1973); the most
cation,1992)describestwoadditionalhigh- common supernumerary teeth in primates
frequency ancestral African traits not are fourth molars (Jungers & Gingerich,
includedintheASUSystem.Thefirstisthe 1980).
Type 1 upper incisor of Mizoguchi (1985).
It is a form of shoveling where weak mar- Statistical comparisons of modern and early
ginal lingual ridges converge toward the hominid dental samples
tooth cervix. This form diVers from Type 2 In most of the above examples, the fre-
(strong non-converging ridges and large quencies of the ancestral dental traits are
lingual tubercle) and Type 3 (markedly unknown. This paucity of published data
developed,convergingridgeswithareduced inhibits a quantitative comparison to that of
lingual tubercle) incisors found in some the seven modern groups to formally test if
early (e.g., Neandertal) and modern non- Sub-Saharan Africans do indeed display a
African groups, respectively. The second higher incidence of ancestral traits. More-
trait is the LM1 epicristid of Hershkovits over,itisthetraitfrequenciesinearlyhomi-
(1971).Itischaracterizedbyamiddletrigo- noids,notsimplythepresenceofsuchtraits,
nid crest between the protoconid and meta- that are important in determining whether
conid(see LiuWu&Turner,1993).Zubov a character is ancestral or derived. Fortu-
observed both traits in australopithecines, nately, there are a few cases where applica-
Homo habilis, Homo erectus, archaic Homo ble data, that are adaptable to the ASU
sapiens, and reports that they have been the system, are available. For example, studies
predominant forms in Africans during the byRobinson(1956);Sperber(1974),Wood
past 10,000 years. He goes on to say that & Abbott (1983), and Wood & Engleman
the presence of such traits in Sub-Saharan (1988) for Plio-Pleistocene samples of
Africans ‘‘... suggests retention of local Australopithecus robustus, Australopithecus
boisei, Australopithecus africanus, and Homo
habilis provide some trait frequencies.
2See Stringer et al. (1997)) for a formal cladistic Although there are methodological diVer-
analysisusingmuchofthesamedentaldatapresented
herein. encesandlikelyinter-observervariationthat
ancestral sub-saharan dental traits 89
may aVect comparability between the pub- siomorphicrelationshipbetweenAustralians
lished and present data (Haeussler et al., and Africans (see Stringer et al., 1997, and
1988), I was able to approximate ASU Sys- above).
tem grades (same breakpoints as Table 1) A 13-trait MMD comparison among the
for 13 Plio-Pleistocene traits that exhibit Plio-Pleistocene and five of seven modern
roughly equivalent expressions in modern samples produced the distance estimates in
humans. These data are presented in Table Table 6. Australia/Tasmania and Melanesia
5. As would be expected, the number of were not included in this analysis because
specimens in the four articles is low. Thus, they possess too few comparable traits
to maximize sample size I was compelled to (n=7) to provide meaningful results. The
pool frequencies from A. robustus and A. Plio-Pleistocene robust sample diVers
boisei to create a ‘‘robust’’ sample, whereas markedly from all five modern groups
A. africanus and H. habilis, who many (MMDs=0·965–1·75),althoughitisclosest
believe are direct human ancestors (see to the Sub-Saharan Africans; it is also sig-
Skelton et al., 1986; Delson, 1987; Grine, nificantly diVerent from the gracile sample
1993; among others), were combined into (MMD=0·361). The latter sample is much
a ‘‘gracile’’ sample. Table 5 also lists fre- more akin to the modern groups, especially
quenciesfortheKrapinaNeandertalswhich Sub-Saharan Africa (MMD=0·180). The
will be discussed shortly. gracilehominidsareleastliketheSinodonts
It stands to reason that the modern (MMD=0·740), who possess high fre-
group exhibiting the highest frequencies of quencies of derived dental features (see
‘‘ancestral-like’’ traits should show the below).Again,theseMMDsarenotdirectly
closest aYnity to the pooled gracile sample; comparable to those using more traits.
suchanaYnitymaynotnecessarilyreflecta I also have trait frequency information
close biological relationship but rather, at a from a sample of 33 archaic Homo sapiens—
minimum, a sharing of ancestral characters the 130,000 year-old Krapina Neandertals
orsymplesiomorphy(Hennig,1966;Mettler (Rink et al., 1995). Although European
et al., 1988) (see discussion below). There Neandertals often possess derived skeletal
wouldlikelybealesseraYnitytotherobust (thin cranial bones, cold adapted features,
hominids, whose herbivorous diet resulted etc.)anddental(Type2shoveling,taurodon-
in reduced anterior teeth and an increase in tism, etc.) characters (Mizoguchi, 1985;
posteriortoothsizeandmorphologicalcom- Trinkaus & Shipman, 1993; Stringer, 1994;
plexity. By comparing Tables 1, 2, and 5 it Franciscus, 1995; Holliday, 1995), and are
can be seen that, relative to other modern generally not considered direct ancestors to
groups, Sub-Saharan Africans share the modern humans, the sample (like the robust
greatest number of trait similarities with hominids) can yield some comparative data.
the gracile sample. Six of these traits Moreover, because (1) a number of dental
are included in the Sub-Saharan African traits have been shown to be plesiomorphic
Dental Complex (UM1 Carabelli’s, UM3 [see Stringer et al. (1997)], (2) frequencies
presence, LM2-groove pattern, LM1 cusp (see Table 5) are similar to those of the
7, LP1 Tome’s root, 2-rooted LM2). Sub- gracile sample (9-trait MMD is 0·202), and
Saharan Africans are less like the dentally- (3) it predates known divergence times of
derived robust australopithecines. Australia/ moderndentalpatterns(Stringeretal.,1997)
Tasmania also shares several comparable (see Hanihara, 1967, 1969; Zubov, 1979;
trait frequencies with the Plio-Pleistocene Turner, 1984, 1985, 1990), the sample may
hominids(esp.gracile);thisfindingprovides provide a reasonable ‘‘ancestral-like’’ model
additional evidence for a purported symple- forsimple,pheneticanalyses.
90 j. d. irish
Table5 Dental trait percentages and frequencies of occurrence in PlioPleistocene
AfricanandEuropeanfossilhominidsamples
Sample*
Trait Robust† Gracile† Krapina‡
WingingUI1 0·0
(+=ASU1) 0/9
ShovelUI1 100·0
(+=ASU2–6) 11/11
DoubleShovelUI1 0·0
(+=ASU2–6) 0/11
InterruptionGrooveUI2 21·4
(+=ASU +) 3/14
TuberculumDentaleUI2 100·0
(+=ASU1–6) 14/14
BushmanCanineUC 46·2
(+=ASU1–3) 6/13
DistalAcc.RidgeUC 55·6
(+=ASU1–5) 5/9
HypoconeUM2 100·0 100·0 100·0
(+=ASU1–5) 11/11 17/17 10/10
Cusp5UM1 100·0 28·6 42·9
(+=ASU1–5) 9/9 2/7 3/7
Carabelli’sTraitUM1 80·0 50·0 100·0
(+=ASU2–7) 8/10 5/10 9/9
ParastyleUM3 33·3
(+=ASU1–5) 3/9
RootNo.UP1 100·0 93·8
(+=ASU2+) 20/20 15/16
Peg-reducedabsentUM3 0·0 0·0 0·0
(+=ASUP,R,C) 0/18 0/8 0/10
LingualCuspNo.LP2 100·0
(+=ASU2–9) 11/11
GroovePatternLM2 96·0 69·2 69·2
(+=ASUY) 24/25 9/13 9/13
CuspNo.LM1 80·9 0·0 23·1
(+=ASU6+) 17/21 0/16 3/13
CuspNo.LM2 0·0 0·0 0·0
(+=ASU4) 0/19 0/13 0/14
DeflectingWrinkleLM1 54·5
(+=ASU1–3) 6/11
C1–C2CrestLM1 10·0
(+=ASU +) 1/10
ProtostylidLM1 47·4 56·3 7·7
(+=ASU1–6) 9/19 9/16 1/13
Cusp7LM1 25·0 55·5 50·0
(+=ASU1–4) 6/24 10/18 6/12
Tome’sRootLP1 100·0 83·3
(+=ASU3–5)§ 23/23 10/12
RootNo.LM1 0·0 0·0
(+=ASU3+) 0/17 0/14
RootNo.LM2 100·0 100·0
(+=ASU2+) 19/19 10/10
*Robust=Australopithicus robustus and A. boisei; Gracile=Australopithicus africanus
andHomohabilis;Krapina=ArchaicHomosapiens(KrapinaNeandertals).
†TraitdataderivedfromRobinson(1956),Sperber(1974),Wood&Abbott(1983),
andWood&Engleman(1988).
‡Traitdatafrompresentstudy(allfrequenciesareofindividualcounts).
§Tome’srootfrequenciesalsocounttwoseparaterootsaspresent.