Turner 2005.4


55

This note is based on observations made on a small 
sample of prehistoric human teeth from excavations 
on Rotuma Island submitted to the author for analysis 
by Richard Shutler, Jr. He and Jamie Evrard directed 
test excavations in June and July, 1981, one resulting 
in the recovery of human remains from ROT 2-9, an 
archaeological site in the Oinafa District location called 
Risumu on the east end of Rotuma Island. Rotuma is 
remotely located in the mid-Pacific. Volcanic in origin 
and only 25 km2 in area, it is in the western Polynesian 
Outlier culture division of Oceania at approximately 
12˚ 25’ S and 177˚ 5’ E. The Risumu site is the legendary 
landing place of the first immigrants, supposedly 
from Tonga, who are said to have arrived about one 
thousand years ago (Shutler and Evrard, 1991:136). The 
people of Tonga are Polynesians, and the present-day 
Rotuma islanders speak a language that is classified 
as Polynesian, although its exact genetic relationship 
to other Polynesian languages is unclear (Shutler, 1998:
252).  Melanesian populations occupy the Solomon 
Islands to the west and Fiji to the south, whereas 
Polynesians are settled on Samoa to the east. The human 
remains were found in a burial mound (Rot 2-9, test 4, 
level 4) at a depth of 90-100 cm. There were no cultural 
remains associated with the human teeth and bones. 
This small but geographically rare assemblage has since 
been reburied after study. However, before reburial 
Shutler had a sample of the human bone dated in the 
carbon 14 laboratory on his campus at Simon Fraser 
University, Burnaby, British Columba, Canada. The 
assay (SFU-118) produced an uncorrected date of 1,000 
BP + 100 radiocarbon years (Shutler, 1998).

Dental Indications of Polynesian Affinity for Prehistoric 
Rotuma Islanders, South Pacific

Christy G. Turner II*

Department of Anthropology, Arizona State University, Tempe

*Correspondence to: Christy G. Turner II, Department 
of Anthropology, Arizona State University, Tempe, AZ 
85287-2402
E-mail: christyturner@aol.com

ABSTRACT    Human skeletal reburial, reasonable from 
a religious and personal point of view, nevertheless 
diminishes the physical record of human evolution. The 
present study preserves some information for a small 
but rare Pacific Basin skeletal assemblage. Prehistoric 
human tooth-bearing cranial and jaw fragments and 
loose teeth of probably 19 individuals excavated on 
Rotuma Island were examined for crown and root 
morphology. The purpose of the examination was to 
assess whether these individuals were morphologically 
more like Melanesians or Polynesians. Rotuma is in the 
Polynesian culture area north of the Fiji group, which 

exhibits archaeological and ethnographic evidence of 
colonists from both Oceanic populations. Polynesians 
belong to the Malayo-Polynesian language family, so 
if the Rotuma teeth are similar to Polynesians they 
should also be more similar to Southeast Asian teeth 
than to those of linguistically different Melanesians or 
Australians. Indeed, this seems to be the case, although 
the small Rotuma sample size reduces confidence 
somewhat in this finding of Rotuma similarity with 
Polynesians and Southeast Asians. Dental Anthropology 
2005;18(2):54-60.

MATERIALS AND METHODS

 The number of Rotuma individuals based on 
maxillary teeth is 14; mandibular teeth, 17; maxillary 
and mandibular, 18; probable total, 19.   Following 
standardized observation and scoring procedures for 
non-metric dental traits (Turner et al., 1991), crown 
and root morphology was analyzed by univariate and 
multivariate statistics to estimate Rotuma’s phenetic 
dental relationships with selected comparative 
populations. The regions chosen for comparison were 
(1) South Pacific, because of geographic proximity; (2) 
Southeast Asia, because of ultimate linguistic homeland; 
and (3) Native America because of T. Heyerdahl’s 
(1952) hypothesis that Polynesians originated from the 
Americas.  Although large samples are always desired 
in assessing affinity for archaeologically-derived and 
usually incomplete and fragmentary skeletal samples, 
it appears that the Rotuma series is adequate for 
moderately confident inferences about probable past 
inter-group relationships. The ten comparative dental 
series used to identify Rotuma relationships are part 
of the published and unpublished data base in the 
author’s computer and other files. The traits selected 
for comparison are those that occur most frequently in 
the Rotuma series. Incisor shoveling, for example, was 
used because some teeth are present with, and have 
limited wear of, the trait that permits confident scoring, 



56 ROTUMA DENTITION

T
A

B
L

E
 1

. D
en

ta
l t

ra
it

 fr
eq

u
en

ci
es

 in
 R

ot
u

m
a 

an
d 

co
m

pa
ra

ti
ve

 s
er

ie
s

 
 

B
re

ak
 

 
. 

 
E

as
te

r 
 

N
ew

 
 

 
 

 
N

W
To

ot
h

 
Tr

ai
t 

P
oi

n
t 

R
ot

u
m

a 
M

ar
q

u
es

. 
Ta

h
it

i 
Is

la
n

d
 

G
u

am
 

B
ri

ta
in

 
A

u
st

ra
li

a 
T

h
ai

la
n

d
 

In
d

on
es

ia
 

P
er

u
 

C
oa

st

U
I1

 
W

in
gi

n
g 

1/
1-

4 
50

.0
 (

2)
 

16
.0

 (
50

) 
—

 (
0)

 
—

 (
0)

 
18

.2
 (

22
) 

25
.0

 (
11

2)
 

13
.3

 (
30

) 
31

.0
 (

42
) 

14
.8

 (
61

) 
83

.7
 (

49
) 

33
.3

 (
66

)
U

I2
 

Sh
ov

el
 

2-
7/

0-
7 

66
.7

 (
6)

 
41

.5
 (

70
) 

10
0.

0 
(1

) 
0.

0 
(1

) 
50

.0
 (

34
) 

21
.6

 (
13

0)
 

45
.8

 (
24

) 
60

.4
 (

48
) 

67
.3

 (
58

) 
97

.7
 (

43
) 

96
.8

 (
63

)
U

I1
 

D
u

b-
sh

ov
. 

2-
6/

0-
6 

50
.0

 (
4)

 
9.

4 
(6

4)
 

0.
0 

(1
) 

—
 (

0)
 

5.
6 

(3
6)

 
3.

7 
(1

07
) 

7.
4 

(2
7)

 
12

.5
 (

48
) 

30
.5

 (
59

) 
93

.8
 (

16
) 

51
.0

 (
49

)
U

I2
 

Tu
b.

 d
en

ta
l 

+
/

+
, 0

 
66

.7
 (

6)
 

60
.9

 (
46

) 
0.

0 
(1

) 
0.

0 
(1

) 
83

.3
 (

6)
 

45
.8

 (
12

0)
 

54
.5

 (
22

) 
47

.2
 (

36
) 

63
.8

 (
47

) 
52

.9
 (

34
) 

77
.0

 (
61

)
U

C
 

B
u

sh
m

an
 

1-
3/

0-
3 

0.
0 

(8
) 

1.
4 

(7
4)

 
0.

0 
(2

) 
0.

0 
(1

) 
2.

8 
(3

6)
 

2.
2 

(1
37

) 
8.

3 
(2

4)
 

2.
4 

(4
2)

 
2.

2 
(8

9)
 

0.
0 

(6
0)

 
0.

0 
(8

7)
U

C
 

D
A

R
 

2-
5/

0-
5 

83
.3

 (
6)

 
69

.2
 (

39
) 

—
 (

0)
 

—
 (

0)
 

79
.0

 (
19

) 
36

.6
 (

60
) 

80
.0

 (
5)

 
72

.2
 (

18
) 

77
.8

 (
54

) 
81

.0
 (

21
) 

85
.3

 (
34

)
U

P
1 

U
to

-A
zt

ec
 

+
/

+
, 0

 
0.

0 
(6

) 
—

 (
0)

 
—

 (
0)

 
—

 (
0)

 
0.

0 
(8

) 
—

 (
0)

 
0.

0 
(9

) 
0.

0 
(3

1)
 

0.
0 

(4
) 

0.
0 

(1
18

) 
0.

0 
(3

4)
U

M
3 

M
et

ac
on

e 
1-

5/
0-

5 
10

0.
0 

(5
) 

—
 (

0)
 

—
 (

0)
 

—
 (

0)
 

10
0.

0 
(6

) 
—

 (
0)

 
10

0.
0 

(3
) 

10
0.

0 
(3

1)
 

10
0.

0 
(1

) 
98

.1
 (

10
7)

 
10

0.
0 

(1
3)

U
M

2 
H

yp
oc

on
e 

1-
5/

0-
5 

10
0.

0 
(4

) 
98

.3
 (

12
1)

 
10

0.
0 

(6
) 

10
0.

0 
(9

) 
96

.4
 (

56
) 

97
.2

 (
17

9)
 

10
0.

0 
(4

5)
 

97
.1

 (
70

) 
93

.8
 (

19
5)

 
94

.7
 (

24
4)

 
93

.2
 (

14
8)

U
M

1 
C

u
sp

 5
 

1-
5/

0-
5 

25
.0

 (
8)

 
61

.5
 (

13
0)

 
10

0.
0 

(5
) 

66
.7

 (
6)

 
47

.7
 (

44
) 

68
.4

 (
15

5)
 

79
.2

 (
24

) 
34

.9
 (

63
) 

33
.7

 (
18

1)
 

15
.4

 (
16

2)
 

24
.0

 (
12

1)
U

M
1 

C
ar

ab
el

li
 

2-
8/

0-
8 

22
.2

 (
9)

 
37

.4
 (

15
5)

 
16

.7
 (

6)
 

10
.0

 (
10

) 
36

.2
 (

47
) 

41
.9

 (
17

7)
 

45
.8

 (
24

) 
39

.4
 (

61
) 

38
.7

 (
20

9)
 

38
.6

 (
16

3)
 

23
.4

 (
12

4)
U

M
3 

P
ar

as
ty

le
 

1-
5/

0-
5 

0.
0 

(5
) 

5.
4 

(5
6)

 
0.

0 
(2

) 
0.

0 
(1

) 
0.

0 
(6

) 
5.

7 
(1

23
) 

7.
7 

(5
2)

 
0.

0 
(3

4)
 

2.
3 

(8
6)

 
4.

7 
(1

28
) 

3.
9 

(1
29

)
U

M
1 

E
n

am
el

 e
xt

 
1-

3/
0-

3 
44

.4
 (

9)
 

45
.5

 (
14

5)
 

55
.6

 (
9)

 
46

.2
 (

13
) 

31
.4

 (
51

) 
20

.7
 (

15
0)

 
50

.7
 (

73
) 

47
.5

 (
80

) 
66

.5
 (

22
1)

 
66

.5
 (

31
0)

 
71

.6
 (

22
2)

U
L

P
12

 
O

d
on

to
m

e 
+

/
+

, 0
 

0.
0 

(1
0)

 
1.

5 
(1

35
) 

0.
0 

(8
) 

0.
0 

(1
4)

 
20

.0
 (

5)
 

2.
1 

(1
46

) 
0.

0 
(2

7)
 

0.
0 

(1
1)

 
1.

7 
(1

15
) 

3.
2 

(1
54

) 
5.

2 
(9

7)
U

P
1 

1-
ro

ot
 

1/
1-

3 
87

.5
 (

8)
 

60
.0

 (
13

5)
 

80
.0

 (
15

) 
87

.0
 (

23
) 

49
.1

 (
57

) 
48

.8
 (

12
5)

 
59

.3
 (

81
) 

52
.2

 (
67

) 
53

.9
 (

26
7)

 
89

.3
 (

28
0)

 
90

.9
 (

23
1)

U
M

2 
3-

ro
ot

 
3-

5/
1-

5 
10

0.
0 

(3
) 

61
.7

 (
12

0)
 

36
.4

 (
11

) 
40

.0
 (

20
) 

82
.0

 (
50

) 
86

.3
 (

13
9)

 
76

.7
 (

60
) 

85
.3

 (
34

) 
78

.0
 (

22
3)

 
44

.8
 (

23
9)

 
43

.4
 (

19
1)

U
M

3 
R

ed
/

p
eg

/
ca

 
+

/
+

, 0
 

0.
0 

(5
) 

40
.0

 (
95

) 
25

.0
 (

8)
 

46
.2

 (
13

) 
81

.5
 (

27
) 

16
.7

 (
15

6)
 

10
.1

 (
69

) 
29

.5
 (

44
) 

41
.7

 (
17

5)
 

20
.9

 (
30

2)
 

8.
1 

(1
86

)
L

P
2 

>
1 

L
 c

u
sp

 
2-

3/
0-

3 
10

0.
0 

(7
) 

90
.0

 (
90

) 
10

0.
0 

(4
) 

87
.5

 (
16

) 
93

.9
 (

33
) 

91
.5

 (
15

3)
 

91
.3

 (
23

) 
87

.5
 (

56
) 

81
.4

 (
70

) 
24

.4
 (

82
) 

40
.5

 (
79

)
L

M
1 

A
n

te
ri

or
 f

ov
ea

 
1-

5/
0-

5 
10

0.
0 

(6
) 

—
 (

0)
 

—
 (

0)
 

—
 (

0)
 

10
0.

0 
(4

) 
—

 (
0)

 
—

 (
0)

 
66

.7
 (

3)
 

—
 (

0)
 

88
.1

 (
59

) 
85

.7
 (

7)
L

M
2 

Y
 g

ro
ov

e 
Y

/
Y,

 +
, X

 
22

.2
 (

9)
 

16
.4

 (
11

6)
 

33
.3

 (
9)

 
12

.5
 (

16
) 

18
.5

 (
65

) 
25

.4
 (

18
1)

 
19

.1
 (

47
) 

20
.5

 (
73

) 
15

.8
 (

10
1)

 
7.

9 
(1

52
) 

12
.4

 (
13

7)
L

M
1 

C
u

sp
 6

 
6/

4-
6 

71
.4

 (
7)

 
55

.0
 (

10
9)

 
71

.4
 (

7)
 

37
.5

 (
8)

 
53

.7
 (

54
) 

49
.0

 (
15

7)
 

54
.5

 (
22

) 
22

.6
 (

62
) 

34
.6

 (
81

) 
63

.7
 (

91
) 

40
.7

 (
91

)
L

M
2 

4-
cu

sp
s 

3-
4/

3-
6 

28
.6

 (
7)

 
39

.6
 (

10
6)

 
33

.3
 (

9)
 

53
.8

 (
13

) 
13

.6
 (

59
) 

51
.8

 (
17

0)
 

12
.5

 (
32

) 
44

.1
 (

68
) 

37
.1

 (
97

) 
9.

4 
(1

59
) 

2.
9 

(1
36

)
L

M
1 

D
efl

ec
ti

n
g 

w
r. 

1-
3/

0-
3 

85
.7

 (
7)

 
56

.8
 (

74
) 

10
0.

0 
(3

) 
—

 (
0)

 
74

.4
 (

43
) 

59
.8

 (
10

7)
 

40
.0

 (
10

) 
73

.1
 (

26
) 

54
.5

 (
44

) 
65

.5
 (

58
) 

51
.1

 (
45

)
L

M
1 

Tr
ig

on
id

 c
re

st
 

+
/

+
, 0

 
12

.5
 (

8)
 

1.
4 

(7
2)

 
0.

0 
(6

) 
0.

0 
(8

) 
0.

0 
(6

) 
0.

0 
(1

13
) 

5.
6 

(1
8)

 
0.

0 
(1

6)
 

7.
8 

(5
1)

 
3.

3 
(1

21
) 

11
.3

 (
71

)
L

M
1 

P
ro

to
st

yl
id

 
1-

8/
0-

8 
62

.5
 (

8)
 

13
.9

 (
12

2)
 

14
.3

 (
7)

 
30

.0
 (

10
) 

30
.0

 (
60

) 
21

.2
 (

18
9)

 
6.

7 
(3

0)
 

40
.8

 (
76

) 
35

.6
 (

90
) 

26
.9

 (
16

0)
 

37
.2

 (
11

3)
L

M
1 

C
u

sp
 7

 
1-

5/
0-

5 
0.

0 
(9

) 
7.

2 
(1

25
) 

12
.5

 (
8)

 
0.

0 
(1

3)
 

11
.8

 (
68

) 
13

.0
 (

19
2)

 
6.

2 
(3

2)
 

11
.5

 (
87

) 
10

.9
 (

10
1)

 
10

.8
 (

16
7)

 
7.

9 
(1

27
)

L
C

 
2-

ro
ot

s 
2/

1-
2 

0.
0 

(4
) 

0.
0 

(1
00

) 
0.

0 
(1

5)
 

0.
0 

(2
7)

 
0.

0 
(5

5)
 

0.
0 

(9
7)

 
0.

0 
(6

0)
 

0.
0 

(3
6)

 
0.

0 
(9

2)
 

1.
5 

(1
99

) 
0.

0 
(1

56
)

L
M

1 
3-

ro
ot

s 
3-

4/
1-

4 
10

.0
 (

10
) 

2.
1 

(1
41

) 
23

.1
 (

13
) 

7.
4 

(2
7)

 
0.

0 
(7

5)
 

4.
5 

(1
55

) 
6.

9 
(7

2)
 

8.
9 

(9
0)

 
12

.2
 (

13
1)

 
5.

1 
(2

17
) 

16
.4

 (
21

4)
L

M
2 

1-
ro

ot
 

1/
1-

4 
44

.4
 (

9)
 

24
.2

 (
12

0)
 

20
.0

 (
15

) 
26

.9
 (

26
) 

9.
5 

(6
3)

 
2.

0 
(1

49
) 

6.
8 

(7
3)

 
11

.9
 (

67
) 

27
.3

 (
12

1)
 

35
.2

 (
18

2)
 

36
.3

 (
20

1)
L

P
1 

To
m

es
 r

oo
t 

4-
5/

0-
5 

0.
0 

(2
) 

0.
0 

(8
5)

 
20

.0
 (

15
) 

3.
6 

(2
8)

 
0.

0 
(7

) 
10

.6
 (

94
) 

14
.8

 (
54

) 
7.

3 
(4

1)
 

11
.1

 (
90

) 
2.

3 
(8

6)
 

11
.5

 (
16

5)



57

whereas other traits are either absent (missing data) or 
their amount of occlusal wear exceeds the maximum 
for confident scoring (see various comments about 
wear in Turner et al., 1991). Most of these crown traits, 
but not the root traits, have been previously subjected 
to hereditary study and are believed to have a strong 
genetic component in their occurrence and expression 
(Scott, 1973; Harris, 1977; Nichol, 1990). Table 1 shows 
the various trait frequencies for Rotuma and the 
comparative assemblages. Counts are by individuals, 
sexes are pooled, and dichotomizing frequency break 
points are identified in Table 1, which are necessary for 
the computation of both chi-square and the multivariate 
Mean Measure of Divergence statistic (Berry and 
Berry, 1967; Sjøvold, 1973). This multivariate statistic is 
preferred over others because of its relative simplicity 
and because it readily handles the problem of missing 
data.

RESULTS

Univariate comparisons using chi-square (1 df, 
Yates corrected when any expected cell is less than 5, 
P significant at 0.05, all observed cells had to be greater 
than 0) between Rotuma and the ten comparative 
Oceanic and circum-Pacific samples in Table 1 gave 
the following percentages of significant trait frequency 
differences: Rotuma and Easter, 0.0% (0 out of 23 possible  
comparisons); Indonesia, 0.0% (0/29); Marquesas, 3.7% 
(1/27); Tahiti, 4.0% (l/25); Thailand, 6.7% (2/30); Guam, 
6.7% (2/30); Peru, 6.7% (2/30);  Northwest Coast of 
Alaska and western Canada, 10.0% (3/30); Australia, 
10.3% (3/29); New Britain, 14.8% (4/27). In terms of 
culture area and linguistic family classifications, Easter, 
Marquesas, and Tahiti are Polynesian; Indonesia and 
Thailand are Southeast Asian; Guam is Micronesian; 
Australia and New Britain are Australmelanesian; and 
Northwest Coast and Peru (the two areas that Hyerdahl 
suggested Polynesians might have come from) are 
Native American-Amerind.

The Rotuma dental traits that showed significant 
inter-group frequency differences were: Shoveling 
(Rotuma vs. New Britain, 2 = 4.3; Peru, 4.5; Northwest 
Coast, 4.7); double-shoveling (New Britain, 8.6); upper 
molar cusp 5 (Tahiti, 4.3; New Britain, 5.3; Australia, 
6.7); peg-reduced-congenitally absent upper third 
molars (Guam, 10.1); >1 lingual cusp of lower second 
premolar (Northwest Coast, 7.1; Peru, 14.3); lower 
molar cusp 6 (Thailand, 6.0); 4-cusped lower second 
molar (Northwest Coast, 5.9); protostylid (New Britain, 
5.8; Marquesas, 9.4; Australia, 10.9); 1-rooted lower 
second molar (Thailand, 4.7; Guam, 6.0; Australia, 8.8; 
New Britain, 27.7). Several nearly significant frequency 
differences possibly would have been significant had the 
Rotuma series been larger, and these differences likely 
would have enhanced the differences between Rotuma 
and the Australmelanesian and American dental series.

Since five percent significant differences can be 
expected on the basis of chance alone, these univariate 
comparisons suggest that this Rotuma dental sample is 
statistically indistinguishable from those originating in 
the Marquesas, Tahiti, Easter, and Indonesia locations 
(which includes teeth from younger levels at Niah 
Cave, Malay near Singapore, other Malays, Philippines, 
Bangkok, and the Atayal of Taiwan), and only barely 
distinguishable from teeth from Guam, Thailand 
(archaeological Don Klang, Ban Tong, Non Nok Tha, 
and Ban Chiang pooled), and Peru. The Rotuma dental 
sample is easily distinguished from those originating in 
New Britain, north and south Australia, and Northwest 
Coast of North America.

Univariate comparisons show Rotuma to be 
indistinguishable from the known Polynesian samples. 
Compared with Australmelanesians, Rotuma and 
the Polynesians possess relatively high frequencies 
of incisor shoveling, deflecting wrinkle, protostylid, 
3-rooted lower first molar, and 1-rooted lower second 
molars. They have low frequencies of the pronounced 
mesial-ridged upper canine (Bushman canine), 
Carabelli’s cusp, and the parastyle. These frequencies 
are characteristic of the Southeast Asian dental pattern 
I have called Sundadonty in contrast to the Northeast 
Asian and New World pattern termed Sinodonty 
(Turner, 1979, and elsewhere).

Multivariate comparisons were made using 23 to 30 
traits available in the comparative samples (Table 2). 
Because the Rotuma series is small, few of the computed 
Mean Measures of Divergence (MMD) are significant. 
This coupled with the fact that the number of traits 
compared differed slightly between comparative pairs, 
indicates that more reliability should be placed on 
the univariate findings and inferences. Definitely, no 
strictly “literal” interpretation should be made of the 
MMD values, however, relatively, they generally follow 
what was inferred from the univariate comparisons. 
Given that the number of trait pairs was not identical 
in all inter-group comparisons, I perhaps should have 
attempted to “standardize” the MMD values. It is a 
happy coincidence that I did not, because following 
the submission of this article to Dental Anthropology, I 
have read the important article by Harris and Sjøvold 
(2004) that, among other MMD considerations, 
convincingly demonstrates the inappropriateness of 
MMD standardization.

As with the univariate comparisons, the MMD 
values of Table 2 show that Rotuma is more like most 
Polynesians than like New Britain. Australia occupies 
an intermediate position, both in relation to Rotuma 
(MMD = 0.061) and New Britain (MMD = 0.057). 
Rotuma has no measurable MMD dissimilarity to 
Tahiti, Thailand and Indonesia, and effectively no 
divergence from Easter. This odontological association 
of Rotuma with Polynesians and Southeast Asians is 

C.G. TURNER II



58

quite suggestive of proximate (when the Rotuma people 
were alive) and close ties with known Polynesians, and 
close ultimate links with Southeast Asian Sundadonts. 
Insofar as sample size permits, Rotuma cannot be 
multivariately distinguished from Polynesians, whereas 
it can be when compared with Melanesians. Since 
Polynesian and Melanesian populations are the most 
realistic geographic sources for this Rotuma sample, the 
former are a better bet than the latter for having been 
close relatives. The Micronesian people of Guam belong 
to the Sundadont dental class, so it is not unexpected 
that inter-group similarities and differences parallel 
those of Rotuma. Given the very great oceanic distance 
separating Guam and Rotuma, their relative similarity 
is best attributed to their shared ultimate Sundadont 
ancestry in Southeast Asia. A similar inference was made 
earlier by Harris et al. (1975:231) regarding the stronger 
dental relationships between the Yaps of Micronesia 
and Polynesians, in contrast to the much weaker 
relationship between Yaps and Australmelanesians. 
In large-scale comparisons, both Pietrusewsky (1990), 
using craniometric observations, and the author 
(Turner, 1990) using dental morphology, found Guam 
skulls and teeth to be much more like those of Southeast 
Asians and Polynesians than like various Australian, 
Melanesian, and Tasmanian samples. However, when 
Turner pooled Rotuma and Fiji dental samples, because 
of their relative Oceanic closeness, this combined group 
was most like samples from Early Malay Archipelago, 
and from Melanesian-Polynesian border islands. 
Fiji has a history of both Polynesian and Melanesian 
occupation. The Fiji dental sample was considered to 
be Polynesian, but it would appear this was incorrect 

because the Fiji-Rotuma combination clustered with 
Australmelanesians instead of Sundadont Southeast 
Asians, Polynesians, and Guam Micronesians. Yet, 
the study by Weets (1996) on the dentition of Vanuatu 
islanders, near Fiji, in eastern Melanesia, found that 
these people were more like Polynesians and east Asians 
than like Melanesians. Hence, large-scale boundaries in 
Oceania defined culturally and linguistically generally 
have high correspondence with dentally-defined 
communities. Rotuma alone classifies as Polynesian, but 
when combined with nearby Fiji its affiliation becomes 
ambiguous. 

DISCUSSION

Both univariate and multivariate statistical 
comparisons of the small Rotuma dental sample 
indicate a closer relationship with Polynesians than 
with Melanesians or American Indians. In addition, 
the Rotuma teeth are very similar to those of Southeast 
Asians. Since numerous other assessments of affinity 
based on these same dental traits have produced 
expected results when evaluated with independent 
archaeological, linguistic, or ethnographic information 
(Scott and Turner, 1997), there is good reason to 
hypothesize a strong Rotuma-Polynesian linkage, 
depending, of course, on how one feels about the size of 
the Rotuma sample. Although no cultural remains were 
found with the Rotuma bones and teeth, Shutler and 
Evrard (1991) argued that the Rotuma oral traditions 
strongly indicated a Polynesian cultural affiliation.

Differences between human groups are due to 
evolutionary processes, with genetic drift or founder’s 
effect figuring prominently in small groups, especially 

TABLE 2. Mean Measures of Divergence for Rotuma and comparative dental samples1

 ROT TAHI THAI INDO EAST NWC AUST GUAM MARQ PERU NEW B

ROT  25 30 29 23 30 29 30 27 30 27
TAHI 0.000  25 25 23 25 25 25 25 25 25
THAI 0.000 0.041  29 23 30 29 30 27 30 27
INDO 0.000 0.071 0.000  23 29 29 29 27 29 27
EAST 0.007 0.000 0.015 0.045  23 23 23 23 23 23
NWC 0.057 0.223 0.223 0.135 0.303  29 30 27 30 27
AUST 0.061 0.000 0.067 0.050 0.033 0.262  29 27 29 27
GUAM 0.074 0.130 0.065 0.050 0.100 0.313 0.097  27 30 27
MARQ 0.090 0.021 0.058 0.073 0.000 0.324 0.039 0.060  27 27
PERU 0.138 0.307 0.347 0.275 0.270 0.093 0.452 0.450 0.452  27
NEW B 0.160 0.087 0.074 0.182 0.028 0.528 0.057 0.152 0.075 0.665 

1Whole numbers are the number of pairs of traits used to calculate the inter-group MMD values. Thus, there were 
25 traits involved in the MMD comparison between Rotuma and Tahiti, and 30 used for the Rotuma-Thailand 
MMD. MMD values are shown as fractions with small values representing greater similarity than larger values. 
Thus, Rotuma is more similar to Easter (0.007) than it is with New Britian (0.160). MMD values were calculated 
according to C.A.B. Smith (Berry and Berry, 1967) with the modifications suggested by Sjøvold (1973) and Green 
and Suchey (1976).

ROTUMA DENTITION



59

for traits of little or no identifiable adaptive value, such 
as enamel extensions or occlusal surface characters 
that wear off early in life. Such traits, especially if their 
mode of inheritance is relatively simple, should show 
increased inter-group frequency differences, with 
increased amounts of temporal separation. For example, 
the MMD between American Indians and Northeast 
Asians is 0.154 (Turner, 1986). Most archaeological 
evidence suggests that these two geographic groups 
have been physically separated on the order of 12,000 to 
15,000 years (Fiedel, 2004; several others). The averaged 
MMD between Rotuma and the Polynesians, compared 
with Thailand-Indonesians is 0.038, about four times 
less that the Indian-Northeast Asian MMD value. As 
time and MMD values between separated groups has 
been suggested as roughly proportional (Turner 1986), 
then an MMD separation estimate between Polynesians 
and Southeast Asians would be about 3,000 years. Such 
an estimate corresponds fairly well with radiocarbon 
dates of about 1,000 B.C. from early Tonga (Shutler 
and Shutler, 1975; Bellwood, 1979) and several other 
excavated Polynesian sites (Green, 1994).

Pooling the same Rotuma and Polynesian samples 
and comparing their MMD values with New Britain 
gives an averaged MMD of 0.087. This is two to three 
times greater that the Rotuma-Polynesian/Southeast 
Asian comparison and produces an estimated 8,000 to 
9,000 years of separation. Such a date vastly exceeds 
any dated archaeological site in Polynesia. So, it would 
seem that archaeological chronometrics when linked to 
mega-regional dental MMD values, also leans towards a 
Polynesian identification of this Rotuma dental sample. 
Finally, nothing more needs to be said regarding a New 
World origin for Polynesians, other than there is no 
dental evidence in support of this hypothesis. This is as 
true today as it was more than 25 years ago when the 
author and G. Richard Scott (1977) set out to describe 
and assess the affinity of living Easter Islanders based 
on dental morphology. Then, as now, Easter and all 
other Polynesian dental evidence points to Southeast 
Asia as the Polynesian homeland, not the Americas nor 
Melanesia.

Finally, a word or two needs to be said about the 
population history of the ultimate ancestral homeland 
of the Rotuma and other Polynesian islanders. This 
ancestral homeland is usually considered to be in 
Southeast Asia, which is referred to as Sundaland 
when in ice age Pleistocene times sea levels were lower 
and all of island and mainland Southeast Asia were 
connected by dry land. The prehistoric and recent 
teeth of the people of Sundaland possess the dental 
pattern previously referred to as Sundadonty. Recently, 
Matsumura and Hudson (2005) have challenged the 
local evolution hypothesis used to explain the origin 
of Sundadonty, returning instead to the older idea of 
“southern Mongoloids” being the result of Neolithic 

migrants from China mixing with Southeast Asian 
Australmelanesians. There are several reasons why the 
old migrant-mixture scenario is flawed, not the least 
of which is that hybridized populations sometimes do 
not breed true (Turner, n.d.). There can be resulting 
offspring that exhibit the original characteristics of 
the parental stocks instead of the hybrid intermediacy 
condition. None of the samples of Polynesians that I 
have examined exhibit a dental pattern that could be 
considered Australmelanesian or Chinese (Sinodonty). 
Despite the absence of archaeological evidence that the 
Rotuma dental sample should be considered Polynesian, 
oral tradition, cemetery location, island location, dating, 
and dental characteristics strongly suggest that it is 
Polynesian. Hence, it provides yet another Polynesian 
isolate that supports the local evolution hypothesis for 
the origin of Sundadonty.

CONCLUSION

A small but geographically rare sample of 
archaeologically-derived teeth from Rotuma 
Island shares more crown and root morphological 
resemblances with teeth from Polynesian and Southeast 
Asian dental samples than it does with teeth from 
the Melanesian island of New Britain. On the basis of 
these comparisons, it is concluded that this Rotuma 
dental sample originated from a population that had 
a greater epigenetic relationship with Polynesians than 
with Melanesians. Being relatively near the border 
zone between Melanesia and Polynesia, Rotuma Island 
may have had chronologically or geographically both 
Melanesian and Polynesian occupants; however, the 
sample discussed herein can easily be hypothesized as 
having been Polynesian. Dental morphology, linguistic 
classification,  and oral traditions independently favor a 
Polynesian affiliation for these Rotuma human remains. 
The Rotuma teeth also help reconfirm the local evolution 
hypothesis for the origin of Sundadonty.

ACKNOWLEDGMENTS

Data and earlier analyses on the comparative 
series was gathered with grants from the National 
Geographic Society, The Wenner-Gren Foundation 
for Anthropological Research, and Arizona State 
University Faculty Grant Program. The following 
institutions housed the comparative series when 
studied: American Museum of Natural History, New 
York; Bishop Museum, Honolulu; National Museum of 
Natural History, Smithsonian Institution, Washington, 
D.C.; Burke Museum, University of Washington, 
Seattle; Archaeological Survey of Canada, Ottawa; 
University of British Columbia, Vancouver; Peabody 
Museum, Harvard University, Cambridge; University 
of Arkansas, Fayetteville; University of Pennsylvania 
Museum, Philadelphia; University of Nevada, Las 
Vegas; University of Hawaii, Honolulu, Field Museum 

C.G. TURNER II



60

of Natural History, Chicago; Academia Sinica, Taipei; 
San Diego Museum of Man, San Diego; Simon Fraser 
University, Burnaby. Data processing and entry were 
aided by Linda Nuss (Watson). Richard Shutler kindly 
made the dental sample available for study. This is 
contribution number 33 in my Peopling of the Pacific 
Basin and Adjoining Areas series.

LITERATURE CITED

Bellwood P. 1979. Man’s conquest of the Pacific: the 
prehistory of Southeast Asia and Oceania. New 
York: Oxford University Press.

Berry AC, Berry RJ. 1967. Epigenetic variation in the 
human cranium. J Anatomy 101:361-379.

Fiedel S. 2004. The Kennewick follies. J Anthropol 
Research 60:75-110.

Green R. 1994. Changes over time: recent advances in 
dating human colonisation of the Pacific Basin area. 
In: Sutton DG, editor. The origins of the first New 
Zealanders. Auckland, p 1-33.

Green R, Suchey J. 1976. The use of the inverse sine 
transformation in the analysis of non-metrical 
cranial data. Am J Phys Anthropol 45:61-68.

Harris EF. 1977. Anthropologic and genetic aspects of the 
dental morphology of Solomon islanders, Melanesia. 
Ph.D. dissertation, Arizona State University, Tempe.

Harris EF, Turner CG II, Underwood JH. 1975. Dental 
morphology of living Yap islanders, Micronesia. 
Archaeol & Phys Anthropol in Oceania 10:218-234.

Harris EF, Sjøvold T. 2004. Calculation of Smith’s Mean 
Measure of Divergence for intergroup comparisons 
using nonmetric data. Dental Anthropolology 17:
83-93.

Heyerdahl T. 1952. American Indians in the Pacific. 
Chicago: Rand McNally.

Howells WW. 1973. The Pacific islanders. New York: C. 
Scribner’s Sons.

Matsumura H, Hudson MJ. 2005. Dental perspectives 
on the population history of Southeast Asia. Am J 
Phys Anthropol 127:182-209.

Nichol CR. 1990. Dental genetics and biological 
relationships of the Pima Indians of Arizona. Ph.D. 
dissertation, Arizona State University, Tempe.

Pietrusewsky M. 1990. Craniometric variation in 
Micronesia and the Pacific: a multivariate study. 
In: Hunter-Anderson RL, editor. Recent advances 
in Micronesian archaeology. Supplement 2, 
Micronesica. Mangilao: University of Guam Press, 
p 373-402.

Scott GR. 1973. Dental morphology: a genetic study 
of American White families and variation in living 
Southwest Indians. Ph.D. dissertation, Arizona State 
University, Tempe.

Scott GR and Turner CG II. 1997. The anthropology of 
modern human teeth: dental morphology and its 
variation in recent human populations. Cambridge: 
University of Cambridge Press.

Shutler R Jr. 1998. Rotuma in historic and prehistoric 
perspective. Asian Profile 26:251-254.

Shutler R Jr., Evrard JS. 1991. Rotuma: a case of 
archaeology documenting the Rotuman oral 
tradition of the first Tongan landing. Man and 
Culture in Oceania 7:133-137.

Shutler R Jr., Shutler ME. 1975. Oceanic prehistory. 
Menlo Park: Cummings Publishing Company.

Sjøvold T. 1977. Non-metrical divergence between 
skeletal populations. Ossa 4:1-133.

Turner CG II. 1979. Sinodonty and Sundadonty: a dental 
anthropological  view of Mongoloid microevolution, 
origin, and dispersal into the Pacific basin, Siberia, 
and the Americas. In: Vasilievsky RS, editor. Late 
Pleistocene and early Holocene cultural connections 
of Asia and America. Novosibirsk: USSR Acad Sci, 
Siberian Branch, p.762-76 [Russian].

Turner CG II. 1986. Dentochronoloigcal separation 
estimates for Pacific rim populations. Science 232:
1140-1142.

Turner CG II. 1990. Origin and affinity of the people 
of Guam: a dental anthropological assessment.  
In: Hunter-Anderson RL, editor. Recent advances 
in Micronesian archaeology. Supplement 2, 
Micronesica. Mangilao: University of Guam Press, 
p 403-416.

Turner CG II. n.d. Dental morphology and the 
population history of the Pacific Rim and Basin: 
commentary on Hirofumi Matsumura and Mark J. 
Hudson. Am J Phys Anthropol [Under Review.]

Turner CG II, Scott GR. 1977. Dentition of Easter 
islanders. In: Dahlberg AA, Graber TM, editors. 
Orofacial growth and development. The Hague: 
Mouton Publishers, p 229-249.

Turner CG II, Nichol CR, Scott GR. 1991. Scoring 
procedures for key morphological traits of the 
permanent dentition: the Arizona State University 
dental anthropology system. In: Kelley MA, Larsen 
CS, editors. Advances in dental anthropology. New 
York: Wiley-Liss, p 13-31.

Weets JD. 1996. The dental anthropology of Vanuatu, 
eastern Melanesia. M.A. thesis, Anthropology, 
Arizona State University, Tempe.

ROTUMA DENTITION