The Late Peopling of Africa According to Craniometric Data. A Comparison of Genetic and Linguistic Models

HUMAN  EVOLUTION Vol.  28    n.1-­2  (1-­12)  -­  2013 Terrazas  Mata,  A.   The  Late  Peopling  of  Africa  According   Serrano  Sánchez,  C. to  Craniometric  Data.  A  Comparison  of   Instituto  de  Investigaciones   Antropológicas,   Genetic  and  Linguistic  Models Universidad  Nacional  Autónoma   This  paper  shows  the  results  of  a  multivariate  study  on  a  large   de  México,   Circuito  Exterior  s/n,   craniometric  database  collected  from  literature  to  examine  the   04510,  México  DF.   late  peopling  of  Africa.  Our  study  includes  cranial  samples  of   E-­mail:  [email protected] peoples   as   well   as   individuals,   from   the   terminal   Pleistocene   to  historic  times,  for  most  of  the  African  continent.  We  discuss   the  implications  of  such  results  in  comparison  with  the  genetic   Benavente,  M.   Área  de  Prehistoria  y  Evolución   /  linguistic  model.   Instituto  de  Investigaciones   The  results  show  the  primitive  morphology  of  the  early  Homo   Antropológicas.   sapiens  fossils  from  Middle  Pleistocene  and  the  possible  con-­ Universidad  Nacional  Autónoma   tinuity   of   this   conservative   pattern   in   some   Late   Paleolithic   de  México,   populations.   The   great   diversity   of   Pigmy   populations   across   Circuito  Exterior  s/n,   East   and   West  Africa,   suggesting   different   origins   is   showed   04510,  México  DF. too.  Finally,  the  late  expansion  of  relatively  homogenous  Ban-­ tu-­speaker  populations  is  suggested. KEY WORDS:    Modern  Homo   sapiens,  African  continent,  classic   craniometrics,  microevolution  and   migrations. A  series  of  major  craniological  studies  have  been  carried  out  on  the  people  of  Af-­ rica.  However,  the  explanations  for  the  late  peopling  of  the  African  continent  based  on   a  raciologic  approach,  as  well  as  the  technical  limitations  of  treating  large  quantities  of   data,  have  limited  considerably  the  scope  of  such  studies.  The  subsequent  abandonment   of  the  raciologic  paradigm  has  led  to  the  rejection  of  large  databases,  which  also  include   valid  measurements  and  insights. The   development   of   a   new   paradigm   in   physical   anthropology   encouraging   the   analysis  of  population  variability,  and  the  dynamics  of  change  in  time  and  space  have   EHQH¿WHGIURPWKHGHYHORSPHQWRIQHZWHFKQRORJLHVZKLFKDOORZWRSHUIRUPDFFXUDWH multivariate  calculations  in  the  short  term  (cf.  Howells,  1974;;  Keita,  2004;;  Pucciarelli,   2008).  This  has  given  rise  to  new  hypothesis  on  the  population  dynamics  of  Africa  over   the  last  10,000  years.   On  the  other  hand,  the  analysis  of  linguistic  families  in  Africa  have  led  to  several   PRGHOVRQWKHSRVVLEOHPLJUDWRU\ÀRZVDFURVVWKHFRQWLQHQW'LDPRQGDQG%HOOZRRG 2003;;  Holden,  2002).  Genetics  has  also  provided  a  large  set  of  highly  consistent  data   about  the  genetic  diversity  throughout  the  territory  (Reed  and  Tishkoff,  2006).  The  inter-­ action  of  these  two  disciplines  has  produced  a  general  model  of  population  movements   over  the  last  thousands  of  years  in  Africa. 2 TERRAZAS  MATA,  SERRANO  SÁNCHEZ,  BENAVENTE Figure  1.  Linguistic  families  in  Africa.  (From  Diamond  and  Bellwood,  2003). 9 TABLE  1  -­  Populations  and  bibliographical  references. Table 1 - Populations and bibliographical references POPULATION REFERENCE Fish Hoek This study (cast) Akoa from Cap Lopez Marquer, 1972 Ba Binga from Binga Marquer, 1972 Ba Binga from Bayanga Marquer, 1972 Ba Binga from Bouetou Marquer, 1972 Ba Binga from Zomie Marquer, 1972 Ba Binga from Ouesso Marquer, 1972 Ba Bongo from Franceville Marquer, 1972 Ba Bongo from Zanaga Marquer, 1972 Ba Bongo from Haut Ogooué Marquer, 1972 Bushman A Broom, 1923 Bushman Cape Colony Broom, 1923 Bushman Kalahari Broom, 1923 Bushman (Howells) Howells, 1973 Dogon Howells, 1973 Egypt Howells, 1973 Herto White, et al., 2003 Kef Oum Toviza Balout y Briggs s/d Korana Broom, 1923 Mali Oliver, 1989 Rabat Odano and Riquet, 1978 Sahara Chamla, s/d Southeast Cameroon Chamla, s/d Taforalt Odano and Riquet, 1978 Teita Howells, 1973 Zulu Howells, 1973 Eastern Pygmies Beningthon and Peason, 1912 Nazlet Khater Crevecoeur, 2006 Basuku Ribot, 2003 Bahutu Ribot, 2003 Eastern Africa Ribot, 2003 Horn of Africa Ribot, 2003 Table 2 - Craniological measurements GOL Glabello-occipital length BNL Basion-nasion length THE  LATE  PEOPLING  OF  AFRICA  ACCORDING  TO  CRANIOMETRIC  DATA 3 TABLE  2  -­  Craniological  measurements. GOL Glabello-occipital length BNL Basion-nasion length BBH Basion-bregma height XCB Maximum cranial breadth ZYB Zygomatic breadth NLB Nasal breadth OBH Orbit height OBB Orbit breadth This  paper  shows  the  results  of  a  multivariate  study  on  a  large  database  collected   from  literature,  including  cranial  samples  of  populations  as  well  as  isolated  skulls,  from   the  terminal  Pleistocene  to  historic  times  (some  fossil  individuals  of  Middle  Pleistocene,   like  Herto  and  Rabat  or  Kebibat,  were  included),  for  most  of  the  African  continent.  We   will  discuss  the  implications  of  such  results  in  comparison  with  the  genetic  /  linguistic   model. Traditional  craniometry  has  proved  to  be  very  useful  for  the  study  of  the  biologic   diversity  of  human  beings,  as  well  as  for  the  reconstruction  of  migration  histories  and   microevolutionary  processes  throughout  the  globe.  We  now  propose  its  use  as  a  research   tool,  due  to  its  primary  importance  for  our  discipline  in  providing  new  insights  on  the   past  and  present  of  human  groups  in  a  continent  that  represents  the  cradle  of  mankind.   Our  aim  is  to  test  certain  hypothesis  about  the  possible  origins  of  the  earliest  people   in  Africa,  their  evolution,  and  their  mutual  relations. An   important   point   here   is   the   arrival   of   human   groups,   possibly   adapted   to   the   semi-­arid  environments  of  Eastern  and  Southern  Africa,  to  the  main  forest  ecosystems   in  Central  and  West  Africa,  as  well  as  the  peopling  of  the  Mediterranean  region,  and  the   great  deserts  of  Sahara  and  Sahel.   Studies   on   the   morphologic   diversity   of  Africa   usually   accept   the   idea   that   lan-­ guages  of  the  Niger-­Kordofanian  family,  including  the  Bantu  languages,  recently  spread   RXWIURPWKH&HQWUDODQG:HVWHUQSDUWVRIWKHFRQWLQHQW6XFKH[SDQVLRQZRXOGUHÀHFWD series  of  great  migrations  by  groups  carrying  a  whole  set  of  technological  innovations,   including  metal  working,  farming,  and  cattle  breeding.  These  people  would  have  dis-­ placed  other  groups  already  settled  in  the  continent,  who  apparently  spoke  several  lan-­ guages  of  the  Khoisan  family,  some  of  which  are  still  present  in  Southern  Africa.  Other   groups,  such  as  the  Pygmies  of  Central  Africa,  would  have  lost  their  former  language,   and  adopted  the  incoming  shepherds  and  peasants’  language,  but  without  mixing  with   them  genetically  (Cavalli-­Sforza  and  Cavalli-­Sforza,  1994). Moreover,   people   in   Northern   and   Eastern  Africa   would   have   been   the   result   of   interbreeding  between  local  people  and  immigrants  from  Asia.  This  process  would  have   led  to  the  presence  of  the  Afro-­Asiatic  languages  in  these  areas.  Finally,  the  Nilo-­Saharan   ODQJXDJHIDPLO\LVSUHVHQWDORQJWKHXSSHUFRXUVHRIWKH1LOHDQG6RXWKHUQ6DKDUD'LD-­ mond  and  Bellwood,  2003). 4 TERRAZAS  MATA,  SERRANO  SÁNCHEZ,  BENAVENTE Materials The  data  for  this  study  were  drawn  from  the  bibliographic  references  in  Table  1.   Since  Howells’  database  (1973)  has  been  largely  used  in  recent  studies,  we  have  relied   on  it  as  a  starting  point  for  the  selection  of  the  measurements  employed.  For  other  pub-­ lications,  we  have  used  equivalent  measurements  to  those  referred  by  Howell.  This  has   VLJQL¿FDQWO\UHGXFHGWKHQXPEHURIYDULDEOHVEXWLWKDVDOVRDGGHGDJUHDWQXPEHURI populations,  which  reinforces  the  geographic  representativeness  of  our  materials.  The   total  number  of  applied  measurements  is  eight  (see  Table  2). We   have   used   data   from   a   total   of   1103   individuals,   geographically   distributed   across   the   Northern,   Eastern,   Central,  Western,   and   Southern   parts   of  Africa.   From   a   chronological  point  of  view,  the  samples  were  divided  into  Paleolithic,  Epipaleolithic,   Neolithic,  Recent,  and  Historic.   Methodology Metric  data  from  each  skull  were  concentrated  in  an  Excel  database.  We  also  re-­ corded  the  geographic  origin  and  the  serial  number  for  each  skull.  Whenever  only  the   DYHUDJHIRUDSRSXODWLRQZDVDYDLODEOHZHUHFRUGHGWKLV¿JXUH Individual  skulls  were  also  included  in  the  sampling,  because  in  most  cases  they   represent   periods   at   the   end   of   Pleistocene   and   the   beginning   of   Holocene,   or   poorly   represented  geographic  areas. 'XH WR WKH IDFW WKDW LQ PDQ\ FDVHV WKH HWKQLF RULJLQ LV XQFHUWDLQ ZH KDYH EDVHG the  analysis  essentially  on  geographic  criteria,  without  taking  into  account  any  a  priori   conditions  other  than  geography  and  chronology.  Only  for  the  interpretation  of  results,   we  have  examined  the  linguistic,  ethnic,  and  historic  information  available  to  evaluate   which  hypothesis,  aimed  at  describing  the  causes  for  the  distribution  of  certain  traits,  are   most  likely  to  be  applied  to  each  region  or  period  of  time. Contrary  to  other  studies,  we  didn’t  divide  the  analysis  according  to  the  traditional   cranial   regions   (base,   facial   region,   and   cerebral   region),   because   this   procedure   iso-­ lates   functional   regions   which   are   closely   interrelated.   Therefore,   we   have   made   our   functional  interpretation  of  skulls  following  Pucciarelli’s  proposal  (2008):  neurocranial   components  (anteroneural,  midneural,  posteroneural),  and  facial  components  (otic,  op-­ tic,  respiratory,  masticatory,  alveolar).  It  was  not  possible,  however,  to  make  a  complete   cranial-­functional  study  because  not  all  the  required  measurements  were  available.   ,QWKHFDVHRIGDWDDQDO\VLVZHKDYHDSSOLHGWZRPXOWLYDULDWHSURFHGXUHV¿UVWD cluster  analysis  for  the  construction  of  dendrograms  according  to  UPGMA  (Unweighted   3DLU*URXS 0HWKRG ZLWK$ULWKPHWLF 0HDQ ZKLFK UHÀHFWV SRVVLEOH SKHQHWLF RU PRU-­ phometric  distances  between  subjects.  This  method  could  be  used  in  the  construction  of   phylogenies  if  the  evolution  rate  is  similar  among  cases,  and  provides  for  the  formation   TABLE  3:  0HWULF'DWDIURP(DFK6NXOO8VHGLQ7KLV6WXG\'RFXPHQWHG:LWK*HRJUDSKLF2ULJLQDQG6HULDO1XPEHU:KHQHYHURQO\WKHDYHUDJH ZDVDYDLODEOHIRUDSRSXODWLRQZHUHFRUGHGWKLV¿JXUH 1#"" 2"#$-3450$6 3-778 #.93#.9 #.934.9 #.93#.9 #.93-$- #.93:# #.93- .931.0'# .93.9 .93- -$9-; -":. : -":.7.4 : -":.="# : -":. >9. 94$ $ !"#$$%&& =/-:2'#8 -$4#99?1 =. : THE  LATE  PEOPLING  OF  AFRICA  ACCORDING  TO  CRANIOMETRIC  DATA # #'( )$ 3.$ #@-$( & * * * +%* * " ":?3 -3$3-:-. ":?3 2/$ 3.$ #@-$( (% & % % %* ,% %( +% 2#$ -- 49:;#.$-A .#.9$".4. *% 8$="$ '0-&&, ( , +( , +& & -- #)$&& +(% +%+ +%+ %, ,%( % ( "-$- #)$&& (%* *% +%, %+ (% , % %+ $./#0 #)$&& %+ ,% ,%+ % *%* ,% % .//#0 #)$&& % ,% &%, % +% % (%( 5 6 TERRAZAS  MATA,  SERRANO  SÁNCHEZ,  BENAVENTE Figure  2.  Dendrogram  of  Cranial  Samples of  methodologically  related  groups.  The  program  Mesquite  (Maddison  and  Maddison,   2010)  was  used  here  for  the  generation  of  dendrograms. Next,  we  carried  out  a  Principal  Components  Analysis  (PCA)  by  covariance  matrix,   which   uses   an   orthogonal   transformation   to   convert   a   set   of   observations   of   possibly   correlated  variables  into  a  set  of  values  of  uncorrelated  variables  called  principal  compo-­ QHQWV7KLVSURFHGXUHHQDEOHVWKHLGHQWL¿FDWLRQRIYDULDEOHVGLPHQVLRQVZLWKDJUHDWHU loading  for  the  generation  of  dendrogram  clusters.  We  used  the  program  Past  (Paleonto-­ logical  Statistics)  for  the  construction  of  plots  and  covariance  tables. 7KXVVSHFL¿FYDULDEOHVVXFKDVWKHOHQJWKDQGEUHDGWKRIVNXOOVFRXOGEHLQGLYLGX-­ ally  compared,  and  the  possible  causes  of  similarities  (common  ancestry,  physical  activ-­ ity,  way  of  life,  geographical  location,  etc.)  could  be  evaluated.   A  major  limitation  to  this  study  is  the  unequal  sample  size  for  each  region,  and  the   overrepresentation  of  certain  regions,  such  as  Northern  and  Southern  Africa.  In  spite  of   this,  we  believe  that  the  results  provide  a  good  starting  point  for  the  reconstruction  of  the   continent’s  peopling,  and  to  formulate  new  hypothesis  on  the  subject. Findings We  have  generated  a  dendrogram  showing  the  relationships  among  samples.  This   dendrogram  contains  11  large  branches.  Basically,  there  is  a  tendency  to  regional  or  geo-­ graphic  clustering  of  populations;;  however,  a  more  detailed  analysis  of  each  segment  is   needed  to  analyze  the  probable  causes  for  each  clustering. THE  LATE  PEOPLING  OF  AFRICA  ACCORDING  TO  CRANIOMETRIC  DATA 7 )LUVWRIDOOWKH¿UVWWKUHHPDMRUEUDQFKHVLQWKHGHQGURJUDPFRUUHVSRQG almost  entirely  to  individuals  or  populations  from  Middle  Pleistocene,  Upper  Paleolithic   and  Early  Holocene.  Only  two  recent  populations,  both  from  Western  Africa,  have  been   integrated  into  the  third  branch,  and  they  fall  near  the  Paleolithic  sample  from  Taforalt,   Algeria.  These  populations  are  still  too  far  away  from  the  Neolithic  and  historic  samples,   so  they  could  not  represent  direct  ancestors  of  any  present  human  population.  However,   as  we  will  see  from  the  principal  components  analysis,  the  fossil  from  Fish  Hoek  may  be   more  closely  related  to  the  Khoisan  (Bushman)  groups  in  South  Africa  than  to  any  other   African  people.   The  cranium  from  Herto  (Ethiopia),  which  is  dated  to  160,000  years  ago,  is  included   in  the  most  remote  branch,  with  the  remains  of  Rabat,  considered  from  the  Middle  Pleis-­ tocene.  This  cluster  is  at  the  root  of  the  dendrogram,  and  suggests  a  historic  and  genetic   continuity  along  the  Pleistocene,  from  the  earliest  populations  of  Homo  sapiens  to  the   last  groups  of  the  Upper  Paleolithic.   Branch  4  contains  only  one  Bantu-­speaking  population  from  Central  Africa,  and  we   cannot  yet  explain  such  isolation  (perhaps  related  to  the  size  of  the  sample).   Branch  5  contains  two  Pygmy  groups,  from  Central-­West  and  Central-­East  Africa,   ZKRPRYHGDSDUWIURPWKHUHVWRIWKH3\JP\SRSXODWLRQVUHÀHFWLQJWKHKLJKGLYHUVLW\ of  these  groups,  perhaps  unduly  clustered  under  the  same  label  of  “Pygmies”,  Vigilant  et   al.  1989,  propose  an  early  time  for  the  separation  of  the  pigmy  populations  from  east  and   west  Africa  of  ca.\HDUVPD\EHUHÀHFWHGLQWKHGLVWDQFHVLQWKHEUDQFKDQGWKH UHVWRIWKHSLJP\SRSXODWLRQV,QWKHRWKHUKDQGUHFHQWVWXGLHV¿QGDVSOLWVHSDUDWLRQRI about  to  18,000  years  for  the  east  and  west  pigmy  populations  (Anagnostou  et  al.,  2010).   7KLVVWXG\PD\EHVXSSRUWWKH¿UVWK\SRWKHVLV Next,   in   branch   6,   we   see   a   cluster   of   three   Khoisan-­speaking   populations   from   Southern  Africa,   with   a   hunter-­gatherer   way   of   life   characteristic   of   the   inner   desert   zones.  Interestingly,  the  fourth  population  ascribed  to  these  groups  lays  too  far  away,  in  a   cluster  assigned  to  groups  with  different  origins  (branch  10).  The  later  sample  was  clas-­ VL¿HGE\5REHUW%URRPDVD³%XVKPDQ´JURXSWKRXJKOLYLQJRQWKHFRDVWRIWKH&DSH region.  Apparently,  this  ascription  is  inaccurate,  and  the  skulls  could  probably  belong  to   another  Bantu  population.   The  next  two  branches  (7  and  8)  cluster  only  groups  characterized  as  Pygmies  from   Central  and  Western  Africa.  They  consist  of  very  widespread  Bantu-­speaking  popula-­ tions,  who  apparently  acquired  these  languages  through  an  acculturation  process  follow-­ ing  their  contact  with  other  Bantu-­speaking  cattle  breeders  and  farmer  groups.  Basically,   their  large  morphologic  homogeneity  could  be  explained  by  the  common  origin  of  these   people,  prior  to  the  expansion  of  the  Niger-­Congo  language  family,  including  the  Bantu   branch. %UDQFK LV GLYLGHG LQWR WZR ODUJH FOXVWHUV WKH ¿UVW FOXVWHU LQFOXGHV D ³:HVWHUQ Pygmy”  population  with  two  population  groups  speaking  languages  of  the  Bantu  family   (Korana  and  Zulu)  in  South  Africa,  and  two  Niger-­Congo  language-­speaking  populations   8 TERRAZAS  MATA,  SERRANO  SÁNCHEZ,  BENAVENTE from  Western  Africa.  This  apparently  confusing  situation  could  probably  be  explained   LQWZRZD\VWKH\UHÀHFWWKHVDPHZD\RIOLIHRIQRPDGLFKHUGHUVRUZHFRXOGDF-­ cept  the  hypothesis  according  to  which  the  Bantu  people  of  South  Africa  originated  in  a   region  located  between  Central-­Western  Africa  and  South  of  the  Sahara.  The  subsequent   principal  components  analysis  may  support  either  one  of  these  hypotheses.   Branch  number  12  clusters  population  groups  from  the  Horn  of  Africa  (Somalia)   and  Howells’  samples  from  dynastic  Egypt.  These  groups  seem  to  represent  migrations   of  groups  from  the  Middle  East,  who  spoke  languages  of  the  Afro-­Asiatic  family,  prior   to  the  expansion  of  Islam,  and  mixed  with  the  local  African  populations.   7KHODVWEUDQFKLQFOXGHVSRSXODWLRQVIURP:HVW&HQWUDO$IULFD'RJRQIURP Mali,   and   Babiga   from   Cameroon),   who   are   not   culturally   related,   clustered   with   the   sample  of  Central-­Eastern  Africa’s  Pygmies.  This  clustering  seems  to  make  little  sense.   In   any   case,   it   seems   to   justify   the   hypothesis   that  Western   and   Eastern   Pygmies,   al-­ though  having  very  ancient  common  ancestors,  reached  their  cranial  morphology  sepa-­ rately,  leading  to  similarities  in  height,  through  a  process  of  convergent  evolution  caused   by  living  in  similar  environments  and  having  almost  identical  ways  of  life,  as  the  genetic   studies  suggest.   2QWKHRWKHUKDQGWKH'RJRQSHRSOHOLYHGLQFORVHYLFLQLW\ZLWK3\JP\JURXSVLQ the  South  of  Mali  for  centuries,  until  the  expulsion  of  Pygmies  in  the  twelfth  century.   It  is  possible,  therefore,  that  a  certain  degree  of  interbreeding  took  place,  which  might   H[SODLQWKHPRUSKRORJLFDOUHODWLRQVLGHQWL¿HG Principal  Components  Analysis   Principal  Components  Analysis  (PCA)  provides  supplementary  information  on  the   morphological  models  of  studied  populations  and  can  yield  to  a  phenetic  pattern  recogni-­ tion  by  projecting  the  cases  onto  the  PC  axes.  Figure  3  shows  the  relation  between  the   ¿UVWDQGVHFRQGFRPSRQHQWZKLFKDFFRXQWVDVDZKROHIRURIWKHVDPSOHYDUL-­ ance.  The  symbol  9  refers  to  the  Pygmy  populations  refers  to  the  Khoisan  (Bushman)   groups  in  South  Africa;;   Ŷ  refers  to  the  dynastic  Egypt  and  the  Horn  of  Africa  popula-­ tions;;   6UHIHUVWRWKHSHRSOHIURP6DKDUDDQG0DOLLQFOXGLQJWKH'RJRQ   refers  to   the  Bantu-­speaking  populations  in  Eastern  Africa,  while        refers  to  Bantu  people  from   Central  Africa,  and  +  the  sample  from  Pleistocene  Africa. 7KHPRVWHORQJDWHGHOOLSVHFRUUHVSRQGVWRRIWKH3\JP\YDULDELOLW\ZKLOHWKH oval  group  corresponds  to  the  Bushman  variability.  As  we  can  see,  almost  every  African   population  falls  clearly  within  the  two  ellipses,  which  indicates  that  the  variability  of  the   FRQWLQHQWWKRXJKVLJQL¿FDQWLVZLWKLQZHOOGH¿QHGOLPLWV7KHJUHDWH[FHSWLRQVKHUHDUH the  Paleolithic  remains  from  the  Pleistocene  and  the  early  Holocene,  which  fall  outside   the  variability  range  or  are  too  close  to  the  limits. THE  LATE  PEOPLING  OF  AFRICA  ACCORDING  TO  CRANIOMETRIC  DATA 9 Figure  3.  Representation  of  the  First  and  Second  Components  in  PCA  (see  text  for  interpretation). Certainly,   there   is   a   tendency   across   the   continent   for   a   “primitive”   morphology,   characterized  by  the  large  size  of  the  cranium,  and  the  extreme  anteroposterior  length,   compared  to  shorter  and  more  rounded  forms.   The  Fish  Hoek  skull  is  clustered  almost  exclusively  with  the  samples  from  South   Africa,  which  suggests  that  this  fossil,  from  12,000  to  20,000  years,  could  represent  an   ancestral  branch  to  the  Khoisan  people,  either  if  it  is  at  the  limit  of  its  variability.   If  we  draw  a  graphic  with  the  loadings  of  each  measurement  comprised  in  the  prin-­ cipal  components,  we  observe  the  following  distribution  pattern: )RU &RPSRQHQW ¿JXUH HYHU\ FRUUHODWLRQ LV SRVLWLYH ZKLFK PHDQV WKDW WKH skulls  to  the  right  of  the  graphic  are  long  (GOL),  and  high  (BNL,  BBH),  the  facial  region   is  broad  (ZYB),  and  the  orbits  are  broad  (OBB).  In  comparison,  the  skulls  to  the  left  are   short,  low,  with  a  narrow  face.  In  resume:  big  skulls  fall  in  the  right  side  of  the  graphic. 7KHJUDSKLFRI&RPSRQHQW¿JXUHVKRZVPL[HGFRUUHODWLRQV This  means  that  the  skulls  to  the  top  of  the  graphic  are  short  (GOL),  narrow  (-­XCB),   with  a  broad  nose  (NLB),  and  high,  narrow  orbits  (OBH  »,  OBB);;  while  the  skulls  to   10 TERRAZAS  MATA,  SERRANO  SÁNCHEZ,  BENAVENTE Figure 4. Analysis of Component 1. Figure 5. Analysis of Component 2. the  bottom  are  round,  broad,  with  a  narrow  nose,  and  low,  broad  orbits.  In  short,  PC2   account  for  residual  nasal  and  orbital  height  variation.   In  short,  the  Pleistocene  skulls  from  across  Africa  tend  to  be  broad,  long,  with  a   broad  face,  and  broad,  short  orbits.   In  contrast,  the  skulls  of  the  Khoisan  (“Bushman”)  population  are  relatively  short,   low,  broad,  narrow,  with  a  comparatively  intermediate  nose. Pygmies  are  characterized  by  their  great  variability,  but  they  usually  have  small-­ size  round  skulls,  and  a  balanced  face.  Their  dispersion  degree,  however,  is  in  contradic-­ WLRQZLWKWKHFRQFOXVLRQVIURPRWKHUVWXGLHVZKLFK¿QGDVWURQJKRPRJHQHLW\DPRQJ Pygmy  populations,  even  if  they  support  the  hypothesis  that  the  typical  features  of  these   populations,  including  their  short  stature,  took  place  after  their  geographical  separation   through  convergent  evolution.  Like  other  recent  studies  suggest  (Ramírez  Rossi  y  Sardi,   2010;;  Anagnostou,  2010;;  Vigilant,  1989). The  Bantu-­speaking  populations  are  mostly  at  the  center  of  the  graphic,  which  rep-­ resents   a   common   morphologic   tendency,   but   with   a   strong   variability,   whether   they   come  from  Southern,  Eastern  or  Central  Africa.  This  supports  the  idea  of  a  common,   more  recent  ancestor  than  for  the  Pygmy  and  Khoisan  groups,  as  well  as  similar  ways   of  life  based  on  cattle  breeding  and  farming,  independently  from  their  surrounding  en-­ vironment.   THE  LATE  PEOPLING  OF  AFRICA  ACCORDING  TO  CRANIOMETRIC  DATA 11 Conclusion Although  this  comparison  exercise  is  at  an  early  stage,  we  want  to  highlight  how   GLI¿FXOWLWLVWRH[SODLQWKHPRUSKRORJLFGLYHUVLW\RIWKH$IULFDQFRQWLQHQWH[FOXVLYHO\E\ one  cause  or  a  simple  set  of  variables,  such  as  the  expansion  of  the  Bantu  languages  or   the  migration  of  people  from  one  part  to  another  of  the  continent,  or  from  abroad. Microevolutionary  processes,  such  as  the  selection  pressure,  genetic  drift,  or  gene   ÀRZKDYHFRPELQHGLQWLPHWRSURGXFHWKHFXUUHQWGLVWULEXWLRQPRGHOREVHUYHGLQUH-­ cent  and  modern  populations.  We  need  to  create,  however,  larger  databases,  and  to  apply   complex  theoretic  models,  which  incorporate  geographic  and  historic  variables,  in  order   to  fully  understand  the  continent’s  demographic  history  from  the  most  ancient  and  long-­ lasting  human  occupation  in  the  world.   Note.  Authors  wish  to  express  our  gratefulness  to  the  journal’s  lector  for  the  useful  ob-­ servations  to  the  manuscript. References $QDJQRVWRX 3DROR9DOHQWLQD &RLD *DEULHOOD 6SHGLQL DQG *LRYDQQL 'HVWUR%LVRO 0LWKRFRQ-­ drial,  Y-­Chromosomal  and  Autosomal  Variation  in  Mbenzele  Pygmies  from  the  Central  African  Re-­ public.  Coll.  Antropol.  34,  2  ,pp.  535-­543. Balout,  Leonet  and  Lloyd  Cabot  Briggs.  s/d.  Tête  Osseuse  Du  Kef-­Oum-­Touiza.  Gouvernement  Général   GHO$OJpULH'LUHFWLRQGHV$QWLTXLWpVVG Benington,  Robert  Crewdson  and  Karl  Pearson.  1912,  A  Study  of  the  Negro  Skull  with  special  Refer-­ ence  to  the  Congo  and  Gaboon.  Crania.  Biometrika,  Vol.  8,  No.  3/4,  pp.  292-­339.  http://www.jstor. org/stable/2331584. Broom,  Robert.  1923,  A  Contribution  to  the  Craniology  of  the  Yellow-­Skinned  Races  of  South  Africa.   The  Journal  of  the  Royal  Anthropological  Institute  of  Great  Britain  and  Ireland,  Vol.  53  (Jan.  -­  Jun.),   1923,  pp.  132-­149. Cavalli-­Sforza,  Luca  and  Francesco  Cavalli-­Sforza.  1994,  Quiénes  somos.  Historia  de  la  diversidad   humana%DUFHORQD'UDNRQWRV&UtWLFD Chamla,  Marie-­Claude.  s/d.  Les  populations  anciennes  du  Sahara  et  des  régions  limitrophes  étude  des   restes  osseux  humains  néolithiques  et  protohistoriques.  Mémoires  du  centre  de  recherches  anthro-­ pologiques  préhistoriques,  Laboratoires  d’Anthropologie  du  Musée  de  l’Homme  et  de  l’Institut  de   Paléontologie  humaine,  vol.  IX,  Paris. Crevecoeur,   Isabelle.   2006,   Etude   anthropologique   des   restes   humains   de   Nazlet   Khater   (Paléolith-­ ique  supérieur,  Egypte.  Paris  :  Thèse  en  co-­tutelle  entre  l’Université  Bordeaux  1-­(FROH'RFWRUDOH   SVGSE,  2006. 'LDPRQG-DUHGDQG3HWHU%HOOZRRG)DUPHUVDQGWKHLUODQJXDJHVWKH¿UVH[SDQVLRQVScience,   Vol.  300,  No.  5619,  2003,  pp.  597-­603. 'XWRXU2OLYHUHommes  fossiles  du  Sahara,  peuplements  holocènes  du  Mail  septentrional.  Paris:   &HQWUH1DWLRQDOGHOD5HFKHUFKH6FLHQWL¿TXH 12 TERRAZAS  MATA,  SERRANO  SÁNCHEZ,  BENAVENTE +ROGHQ &ODUH -DQDNL %DQWX ODQJXDJH WUHHV UHÀHFW WKH VSUHDG RI IDUPLQJ DFURVV VXE6DKDUDQ Africa:  a  maximum-­parsimony  analysis.  Proceedings  of  the  Royal  Society  of  London,  B,  269,  pp.   793-­799.  http://rspb.royalsocietypublishing.org/content/269/1493/793#related-­urls. Howells,  William  W.  1973,  Cranial  Variation  in  Man.  A  Study  by  Multivariate  Analysis  of  Patterns  of   Difference  Among  Recent  Human  Populations.  Cambridge:  Harvard  University,  Peabody  Museum   of  Archaeology  and  Ethnology. Keita,  Shomarka,  O.Y.  2004,  Exploring  northeast  African  metric  craniofacial  variation  at  the  individual   level:  A  comparative  study  using  principal  components  analysis.  American  Journal  of  Human  Biol-­ ogy.  Vol.16(6),  2004,  pp.  679-­689. 0DGGLVRQ:LQH3DQG'DYLG50DGGLVRQ0HVTXLWH$PRGXODUV\VWHPIRUHYROXWLRQDU\DQDO\-­ sis.  Version  2.74,  http://mesquiteproject.org. Marquer,   Paulette.   1972,   Nouvelle   contribution   a   l’étude   de   squelette   des   pygmées   occidentaux   du   centre  africain  comparé  a  celui  des  pygmées  orientaux.  Mémoires  du  Muséum  National  d’  Historie   Naturelle,  Série  A.  Zoologie,  tome  LXXII,  Paris. Odano,  Ann  Marie  and  Raimond  Riquet.  1978,  Etude  anthropologique  des  restes  post-­atériens.  Bulletin   D’Archéologie  Marocaine,  Tome  XI,  pp.  25-­45. Pucciareli,   Héctor   Mario.   2008,   (YROXFLyQ \ GLYHUVL¿FDFLyQ ELROyJLFD KXPDQD GHVGH OD SHUVSHFWLYD craneofuncional.  México:  UNAM,  Instituto  de  Investigaciones  Antropológicas/Instituto  Nacional   de  Antropología  e  Historia. Reed,  Floyd  A.  and  Sarah  A.  Tishkoff.  2006,  African  human  diversity,  origins  and  migrations.  Current   Opinion  of  Genetics  Development,  Vol.  16  (6),  pp.  597-­605. Ribot,  Isabelle.  2003,  Craniometrical  analysis  of  Central  and  East  Africans  in  relation  to  history.  A  case   VWXG\EDVHGRQXQLTXHFROOHFWLRQVRINQRZQHWKQLFDI¿OLDWLRQAnthropologica  et  Praehistorica,  vol.   114,  pp.  25-­50. Ribot,  Isabelle,  Marta  Lahr  and  Camila  Storto.  2007,  Populations  sénégalaises  de  l´Age  du  fer:  une   approche  craniométrique  dans  une  perspective  historique.  Anthropologie  et  Archéologie,  pp.1-­10. 9LJLODQW/LQGD5HQHH3HQQLQJWRQ+HQU\+DUSHQGLQJ7KRPDV'.RFKHUDQG$OODQ&:LOOVRQ 0LWRFKRQGULDO '1$ VHTXHQFHV LQ VLQJOH KDLUV IURP D VRXUWKHUQ$IULFDQ SRSXODWLRQ Proc.   Natl.   Acad.  Sci.  Vol.  86,  pp.  9350-­9354. :KLWH7LP'%HUKDQH$VIDZ'DYLG'H*XVWD+HQU\*LOEHUW*DU\'5LFKDUGV*HQ6XZD)DQG Clark  Howell.  2003,  Pleistocene  Homo  sapiens  from  Middle  Awash,  Ethiopia.  Nature.  Vol.  423  No.   6941,  pp  742-­747.