close

Вход

Забыли?

вход по аккаунту

?

Associations between dermatoglyphic variation topography and climate in Kenya.

код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 68:395-408(1985)
Associations Between Dermatoglyphic Variation, Topography,
and Climate in Kenya
PETER ROSA
University of Stirling, Stirling, Scotland
KEY WORDS
Dermatoglyphics,Altitude, Rainfall, Kenya, Africa
ABSTRACT
Correlations between a number of dermatoglyphic finger and
palmar taxonomic traits relating to 57 male and 53 female Kenyan populations, and altitude and mean annual rainfall are significant, not only in terms
of the full range of samples, but also when samples are divided into independent smaller groups. These results are discussed and contrasted with those of
other studies which have found no relationships between dermatoglyphic variation and climate in sub-Saharan Africa.
While analysing a dermatoglyphic survey
of Kenyan peoples (Rosa, 1981; 1983&), preliminary results (Rosa, 1981: Chapter 9) indicated significant associations between
climateltopography and dermatoglyphic variation. This was surprising given the results
of Hiernaux (1968) and Hiernaux and Froment (1976), who reported that correlations
between basic dermatoglyphic finger patterns (arches, loops, and whorls) and climate
and altitude were very low and not significant in sub-Saharan Africa. These authors
concluded that “climate does not influence
the phenotypic expression of finger prints,
and that, whatever the selective forces if any,
that may act on these features, they are not
of a climatic nature, nor are they associated
with climate” (Hiernaux and Froment
1976:765).
This paper reports the results of an expanded study of associations between dermatoglyphic variation and climate/topography in Kenya, using a wide range of dermatoglyphic palmar and digital traits, and
discusses the implications of the results in
the light of their contrasting nature to those
of Hiernaux and Froment (1976) above.
BACKGROUND, MATERIALS AND METHODS
The data were collected in 1973 and 1975,
and comprise palm and finger prints from
3,552 male and 2,268 female Kenyan schoolchildren drawn from some 60 ethnic populations. After omitting those with very small
sample numbers, 57 male and 53 female populations are included in this study (Table 1).
0 1985 ALAN R. LISS, INC.
All major ethnic groups in Kenya are represented except for Kalenjin and Cushitic
speaking hunting groups. Descriptions of
populations, sampling and data collection are
contained in Rosa, 1981 and 1983&.
The dermatoglyphic traits examined in the
main survey comprise a wide range of digital
and palmar traits (for details again see Rosa,
1981 and 1983&). Those included here are
radial, ulnar, maximal unilateral and absolute ridge counts, finger pattern intensities,
finger patterns (arches, loops, whorls with
sub-patterns ulnar loops, radial loops, concentric whorls, and twinned loops), palmar
ridge counts (ab, bc, cd), palmar patterns, and
the number of interdigital triradii on the
palm. Traditional methods (Cummins and
Midlo, 1943) were used to read all traits except palmar patterns which were read according to the topological methodology of
Penrose and Loesch (1970).
The environmental variables used in this
study are altitude and mean annual rainfall,
the only ones to meet the condition that records should exist or be accessible for most
areas of Kenya sampled. Precise figures for
altitude were obtained from a series of topographical maps (scale 1:250,000: Survey of
Kenya), accurate to 500 feet, while figures
for mean annual rainfall were extracted from
a number of sources (Morgan and Shaffer,
1966); where available they were extracted
from more detailed rainfall maps (scale
Received October 22, 1984; revised May 9, 1985; revision accepted May 29,1985.
396
P.ROSA
TABLE 1. Details of Kenyan population samples in this study
No.
Pouulation
0
P
Rain"
Somali
224
103
10
500
Rendille
153
46
10
2,000
Gabbra
27
63
10
2,000
Boran
128
60
10
1,000
Burji
64
57
30
4,500
Samburu
87
46
15
2,000
Kadjiado Maasai
63
150
20
5,500
Narok Maasai
37
-
25
6,000
Ngwesi
Mukogodo
Kony
41
38
37
23
30
30
55
6,000
6,000
6,500
Nandi
52
50
55
6,000
Kipsigis
57
43
50
6,500
1.08
102
45
6,000
Key0
50
50
60
7,000
Marakwet
50
56
60
7,000
Pokot
45
54
35
4,500
Iteso
52
47
50
4,500
Turkana
66
43
10
2,000
170
107
45
4,000
Gusii
50
51
60
6,000
Tiriki
Maragoli
Bunyore
Idakho
Marama
Wanga
Marach
Bukhayo
Samia
Butsotso
Bunyala
Kabras
Bukusu
Kiambu Kikuyu
26
26
30
39
25
25
25
23
24
20
25
26
61
65
25
24
20
15
25
24
24
26
26
27
25
26
29
39
70
70
60
70
70
70
60
55
50
60
60
60
55
60
5,000
5,000
5,000
5,000
4,500
4,500
4,000
4,000
4,000
4,500
4,500
5,000
6,000
7,000
Muranga Kikuyu
56
46
60
5,000
Nyeri Kikuyu
49
51
60
6,500
Ndia
56
66
55
6,000
Kamba
78
42
45
5,000
Embu
57
72
55
4,500
Mbere
64
66
30
2,500
Tugen
Luo
Altitude**
Linguistic group
Eastern Cushitic:
Somali
Eastern Cushitic:
Archaic
Eastern Cushitic:
Galla
Eastern Cushitic:
Galla
Eastern Cushitic:
Galla
Plains Nilotic:
Maasai
Plains Nilotic:
Maasai
Plains Nilotic:
Maasai
Laikipiak Dorobo
Laikipiak Dorobo
Highland Nilotic:
Kalenjin
Highland Nilotic:
Kalenjin
Highland Nilotic:
Kalenjin
Highland Nilotic:
Kalenjin
Highland Nilotic:
Kalenjin
Highland Nilotic:
Kalenjin
Highland Nilotic:
Kalenjin
Plains Nilotic:
Karimojong
Plains Nilotic:
Karimojong
River Lake
Nilotic: Lwoo
Western Highland
Bantu
Luyia Bantu
Luyia Bantu
Luyia Bantu
Luyia Bantu
Luyia Bantu
Luyia Bantu
Luyia Bantu
Luyia Bantu
Luyia Bantu
Luyia Bantu
Luyia Bantu
Luyia Bantu
Luyia Bantu
Eastern Highland
Bantu
Eastern Highland
Bantu
Eastern Highland
Bantu
Eastern Highland
Bantu
Eastern Highland
Bantu
Eastern Highland
Bantu
Eastern Highland
Bantu
DERMATOGLYPHIC VARIATION AND CLIMATE IN KENYA
397
TABLE 1. Details of Kenyan population samples in this study (continued)
No.
Altitude* *
Linguistic group
50
4,700
83
50
7,000
62
43
50
7,000
Igoji
78
95
50
4,700
Imenti
61
39
55
5,000
Tigania
64
36
55
5,500
Igembe
66
33
30
3,000
Tharaka
54
46
20
2,500
Pokomo
114
90
10
500
Malakote
41
-
10
500
Giriama
64
38
30
200
Chonyi
79
14
40
100
Mid Mji Kenda
32
27
40
100
Dig0
116
100
40
100
Taita
55
45
40
5,000
Taveta
59
48
30
2,500
Eastern Highland
Bantu: Meru
Eastern Highland
Bantu: Meru
Eastern Highland
Bantu: Meru
Eastern Highland
Bantu: Meru
Eastern Highland
Bantu: Meru
Eastern Highland
Bantu: Meru
Eastern Highland
Bantu: Meru
Eastern Highland
Bantu: Meru
Coastal Bantu:
Pokomo
Coastal Bantu:
Pokomo
Coastal Bantu:
Mji Kenda
Coastal Bantu:
Mji Kenda
Coastal Bantu:
Mji Kenda
Coastal Bantu:
Mji Kenda
Coastal Bantu:
Taita
Coastal Bantu:
Taita
Population
0
Q
Chuka
75
61
Muthambe
76
Mwimbi
Rain*
*Inches.
**Feet.
1:250,000)which only cover a limited area of
Kenya.
Each population was given a figure for altitude and mean annual rainfall at the main
sampling location. A problem arose in the
case of a few populations [i.e., Turkana (all
but 50), Luo, Bukusu] who were sampled outside their traditional territory of origin. This
was resolved by allocating to them rainfall
and altitude figures for a location within the
territory of origin where the population is a t
its densest, on the assumption that genetic
constitution reflects area of origin rather
than residence. As most people who did not
live in their area of origin were either the
children of first or second generation migrants; or children attending secondary
school outside the traditional tribal boundaries, the assumption is reasonable.
The dermatoglyphic traits adopted here
(outlined earlier) are taxonomic traits, consisting of population means or frequencies.
For convenience they are divided here into
two groups. The first consists of %omposite”
traits, such as digital traits pooled over all
fingers by hand (e.g., right total loops, left
total ridge count)-which
provide a n overall
summary of any trends in dermatoglyphic
variation. In the second group are “non composite” traits, such as digital traits on individual fingers. Non composite traits are
complex to analyse, as they are numerous
and highly intercorrelated within major morphological areas of the hand. To allow for this
and to reduce the number of variables, principal components analyses (PCAs) were performed on three groups of non composite
traits. Firstly, quantitative digital traits [radial, ulnar, unilateral, and absolute ridge
counts and finger pattern intensities on individual digits-50 traits in all, as in Appendix I(2)]. Secondly, qualitative digital
taxonomic traits on individual fingers (arches
loops and whorls) with major sub-patterns
[ulnar, radial loops, concentric whorls,
twinned loops-70 traits in all, as in Appen-
398
P.ROSA
dix 1fl)J.
Thirdly, palmar ridge counts and
palmar patterns and triradii [38 traits in all,
as in Appendix I(3)].
A problem in adopting a PCA approach
was whether to perform each analysis separately by sex, hence duplicating the amount
of background description, or to pool the sexes
in some way. Seperate PCA analyses by sex
showed that while some differences did exist
in some components between the sexes, the
similarities far outweighed the differences.
The PCA analyses, therefore, were performed on data files pooled by sex (that is to
say composed of taxonomic traits where each
male and female sample was treated as a
separate case). Factor scores were extracted
for each unrotated and varimax rotated factor with a n eigenvalue 2 1.0, and these factors were then correlated separately by sex
with altitude and rainfall.
Spearman’s rho was adopted for basic correlations with altitude and rainfall. As rainfall is orographically induced in Kenya,
however, there is a significant correlation
between mean annual rainfall and altitude
(0.5 in this study). To adjust for this partial
correlations are also given, with rainfall controlling for altitude.
hands. In males significant correlations exist
on the right hand only, and only exceed -.3
for total absolute and ulnar ridge counts.
Total finger pattern intensity shows a consistently moderate relationship with both altitude and rainfall, with correlations higher
on the right hand in females. Significance
with altitude is confined to the right hand
for females, and although correlations with
rainfall are significant on both hands and in
both sexes, the partial correlations are not.
Total palmar ridge counts shows a significant relationship with altitude and rainfall
in both sexes, though correlations with rainfall cease to be significant in the partial correlations. The correlations, unlike those for
finger ridge counts, are positive. Total palmar pattern intensity is very significant in
males on the right hand only with altitude.
Correlations with rainfall are low and not
significant, but become so when the altitude
effect is controlled for, especially on the left
hand. The relationship is highly significant
and positive for males on the left hand.
Non composite traits
One hundred fifty non composite traits are
considered in this analysis. The correlations
of these individual traits with altitude and
RESULTS
rainfall are presented in Appendix I. For reaComposite traits
sons outlined earlier the results are analysed
Correlations between dermatoglyphic com- and discussed in terms of principal compoposite traits and altitudelmean annual rain- nents analyses (PCAs).
fall are listed in Table 2.
The results for finger patterns show that
Background
while arches show no relationship with altitude or rainfall, total loops and total whorls
The background results of the three PCAs
do. Correlations with altitude are significant are presented in Appendix 11. To save space
for the right hand only, with total loops cor- only factors which appear in Table 3 are acrelating positively and total whorls nega- tually described. In the PCA analysis of 70
tively. Correlations with rainfall are sig- finger pattern taxonomic traits, 17 factors
nificant on both hands, but when the altitude were extracted, of which U1 accounts for 23%
effect is controlled for, only correlations with of the variance and U2 11%.The others each
the left hand are significant.
account for less than 10% of the variance.
Of the pattern sub-types ulnar loops show Univariate correlations reveal at least modsimilar results to total loops above, and con- erate intercorrelations between all finger
centric whorls with total whorls, in contrast pattern taxonomic traits. It is not surprising,
to radial loops and twinned loops who exhibit therefore, to find that unrotated factor 1(Ul)
noticeably different types of trends than is a noticeable “size” factor. In general the
those for total loops and whorls, respectively. complex formed by loopdulnar loops, whorls/
Finger ridge counts show highly signifi- concentric whorls in digits 2 to 5 are differcant negative correlations with both altitude entiated from a) these same pattern types on
and rainfall. The relationship with rainfall, digit 1, b) twinned loops, c) arches, and d)
however, disappears when the altitude effect radial loops. These divisions are depicted
is controlled for. Correlations with altitude more distinctively in the varimax rotated soare strongest in females, being over -.3 for lution, where the traits just outlined tend to
all ridge count traits considered on both load together on the same component (for
399
DERMATOGLYPHIC VARIATION AND CLIMATE IN KENYA
TABLE 2. Correlations of dermatoglyphic composite traits with altitude and mean annual rainfallt
Taxonomic
trait
Altitude
w rs
Q rs
Mean annual rainfall
o rs
Q rs
A) Digital
i) Finger uatterns
- .07
- .02
-.11
RTARCH
LTARCH
- .01
.07
- .01
RTL
.32**
.42***
.27*
.16
.I7
.28*
LTL
.34**
.42***
.24*
RTUL
LTUL
.15
.19
.15
RTRL
- .03
- .05
.15
LTRL
.08
.06
.34**
RTW
-.29*
-.36**
-.26*
LTW
-.16
-.19
- .29*
RTCW
- .24*
- .20
-.35**
LTCW
.06
-.lo
- .22*
RTTWNL
-.16
-.14
- .08
LTTWNL
- .22*
-.21
- .08
ii) Finger ridge counts (unilateral, absolute, radial, ulnar)
RTRC
-.22*
-.38**
- .07
LTRC
-.12
-.38**
-.12
RTARC
-.31**
-.36**
-.22*
LTARC
-.15
-.35**
- .20
RTRRC
-.19
-.32**
-.11
LTRRC
-.I1
-.32**
-.14
-.38**
- .40***
-.14
RTURC
LTURC
-.21
-.33**
- .23*
iii) Finger pattern intensity index
RTFPII
- .20
-.28*
-.23*
LTFPII
-.11
-.17
-.24*
B) Palmar traits
i) Palmar ridge counts
RTPRC
.31**
.18
.44***
LTPRC
.29*
.27*
.30*
ii) Total palmar pattern intensity index
RTPPII
.33**
-.01
- .07
LTPPII
-.19
.04
.19
~
Partial'
wr
Q r
.07
.11
.40***
.30*
.40***
.25*
.04
.23*
-.43***
-.31*
-.30*
-.29*
- .03
- .08
- .04
.03
.I6
.23*
.10
.13
.30*
.37**
-.14
- .24
-.30*
-.35**
.02
.18
.18
.03
.04
.I7
.03
.I2
.12
.29*
-.12
-.17
- .22
-.27*
.25*
.13
-.30*
-.38**
-.35**
-.38**
-.31*
-.36**
-.36**
-.30*
- .02
-.lo
-.15
-.18
-.11
-.18
-.15
- .22
- .05
- .07
-.37**
-.24*
-.I1
- .08
- .22
- .08
-.18
- .06
-.16
- .21
p.17
p.17
.14
.31*
.18
.10
.09
.07
.16
.18
27 :i;
.30'
.45***
.17
tFifty-seven male and 53 female populations.
'Partial correlation rainfall controlling for altitude
:*P4 .05 > .01.
*:*p< .O1 > ,001.
-***p4 ,001.
example see R1 for digit 1, R9 for twinned
loops).
In the case of the 50 ridge counts and finger
pattern intensities, eight factors were extracted of which the first unrotated factor
(Ul) accounts for 58% of the variance, with
all other factors accounting for less than 10%.
Unrotated factor 1, therefore, is most important, emphasising that ridge counts and finger pattern intensities are strongly intercorrelated (much more so than finger patterns where U1 accounted for only 24% of
the variance). Within the total complex of
highly related traits, some relative differences occur. These are reflected in the varimax rotated solution, where radialhnilateral
ridge counts as a category tend to be differentiated from that formed by ulnar ridge
countslfinger pattern intensities. Absolute
ridge counts correlate highly with both these
groups of traits. As in the case of patterns,
moreover, traits on digit 1 again tend to cluster separately from those on other digits.
Correlations between qualitative and
quantitative digital taxonomic traits not reported here (see Rosa, 1981: Chapter 5) show
that ulnar ridge counts and finger pattern
intensities are highly related to loops (-vely)
and whorls (+vely) in general and more specifically to ulnar loops and concentric whorls.
In PCA analyses factors that show high Ioadings for ulnar ridge counts also show similar
high loadings for the above pattern types. In
contrast, radial and unilateral ridge counts,
when related t o patterns, only appear t o correlate relatively highly with arches. Factors
400
P. ROSA
TABLE 3. Correlations between factors derived from PCA analyses ofdermatoglyqhic non-composite tracts (see
Appendix II), with altitude and mean annual rainfall
Altitude
Factor
0 rs
0 rs
Mean annual rainfall
rs
Q rs
CI
Partial rainfall/altitude2
CY rs
Q rs
PCA (Finger patterns)
u1
-.26*
-.15
p.30"
U2
- .24*
-.17
- .09
.09
u11
.09
-.12
U17
- .04
-.11
-.34**
R1
-,35**
-.14
- .24*
R2
.ll
.03
- .01
R6
-.19
-.23*
-.16
R9
.02
- .13
.20
.14
.38**
Rl1
.12
R17
.15
.14
.10
PCA (Finger ridge counts arid finger Dattern intensities)
:,36*'
u1
-2 0
-.19
U2
.
. 32'"
- .25*
.05
U3
:29*
- .24*
.30*
-.27*
u7
.04
- .02
R1
- ,3 1 **
- 41"""
.06
-.23:*
R3
.07
.05
R4
-,35""
-.17
-.25"
R7
.04
- .25*
.02
PCA (Palmar traits)
u1
-,32**
-.12
.22*
.07
,45***
u2
,251
- .05
P.21
U6
.20
R1
14
- .00
.16
R2
27.12
.42***
R3
36 *
.03
.15
.05
-28%
R5
- 28'
-
~
-
-
-
-
-
-.39**
.04
-.31*
- .06
-.14
- .22
.13
-.15
.12
.35**
-.23*
.17
.15
- 31**
.02
-.18
- .03
.36""
.36**
.02
-.18
- ,41***
-.13
-.01
- 50***
-.3g***
.30*
-.18
.06
.OO
-.15
- .10
.06
-.25*
.09
-.17
-
-
.13
.34**
-,36~*
.24*
.28*
.03
-,23*
.33**
-,32**
.oo
.02
-.35**
.36**
- .02
.05
.24*
.02
-.38**
-.50***
-.14
.19
- 47 :6 Y. *
.ll
- .27*
,54*""
32"*
- .20
,33*:3
.22
.34**
.28*
- ,45** *
.31*
.15
,41""*
-20
-
-,
43 * P 8
-.lo
36'6%
'Only factors which correlate with a significance of a t least P < .01 a t one of t h e combinations of sex or altitude/rainfall are
included. Fifty-seven male arid fifty-three female populations. U = unrotated factor; R = variniax rutated factor.
'Partial col-i~elatir~ns
with ramfall controlling for altitude.
'P G .05 > .01
*WP < 01 > ,001
..**P< ,001.
derived from the first PCA analysis, therefore, are not independent of those derived
from the second.
Finally ten factors were extracted by the
PCA analysis of palmar traits, of which the
first accounts for 24% of the variance. The
first size component loads heavily on traits
related to interdigital triradii and interdigital patterns. Palmar ridge counts load independently together in both unrotated and
rotated solutions. Other factors separate hypothenar and thenar traits, and interdigital
patterns whose variation is not fully explained by the first factor.
Correlations with altitude and rainfall
The results of correlating the unrotated and
rotated factors from the three PCA analyses
with altitude and rainfall by sex are summarised in Table 3 . To reduce the amount of
background description in Appendix I1 and
to focus on the strongest relationships, a factor has to correlate with a significance of a t
least G.01 with any one of the various combinations of altitudelrainfall and sex, in order to qualify for inclusion in Table 3.
Altitude. Looking firstly a t finger patterns
and altitude, females show the strongest association with altitude, correlating significantly at -.3 on the most important factor,
U1. At a univariate level females correlate
significantly at p = G.05 with seven out of
eleven of the traits that are listed as loading
at .7 or above on this factor in Appendix I1
(Analysis 1).Males, however, show a nonsignificant correlation of -.15. In the univariate analyses none of the variables loading
at .7 or above on U1 are significant.
Males in contrast to females correlate significantly with U2 and R1. These two factors
are broadly equivalent to each other, emphasising loadings with loops, whorls, ulnar
loops, concentric whorls, and twinned loops
on digit 1. Females, in contrast, show more
moderate and nonsignificant correlations between altitude and these two components.
DERMATOGLYPHIC VARIATION AND CLIMATE IN KENYA
Females also show a significant relationship
with R6, (more moderate and nonsignificant
in males), which loads on loops and whorls
on digit 3.
Similar trends to those just observed for
finger patterns and altitude are exhibited by
ridge counts and pattern intensities, with females but not males as before showing significant correlations with altitude for the
powerful U1 factor, and males (but not females) showing significant correlations with
U2 and R4, factors which load heavily on
digit 1 for ulnar ridge counts, finger pattern
intensities and absolute ridge counts. In addition both males and females show significant correlations with altitude for R1, which
is a modified size component loading heavily
on unilateral and absolute ridge counts on
digits 4 and 5.
Significant correlations with altitude on
the one side and factors based on palmar
traits on the other are exclusively confined
to males. The two strongest correlations are
for size factor U1, which loads heavily on
interdigital patterns and triradii, and R3,
which loads on I1 loops and the frequency of
6 interdigital triradii (TD6). Correlations
with palmar ridge counts (U2,R2) are also
significant, as is R5 which loads on hypothenar central loops.
Rainfall. The relationship between dermatoglyphic variation and rainfall is best reviewed through the partial correlations,
which adjust for the altitude effect, and it is
on these that the following description concentrates. In the case of finger patterns, both
males and females correlate significantly
with the powerful U1 component, but when
altitude is controlled for, only the correlation
in males remains significant. On the whole
there is a distinct lack of agreement between
the sexes, with partial correlations significant with U2, U11, R2 and R6 in females but
not males, and U17, R9 and R11 in males,
but not in females.
Correlations with rainfall and the large size
factor U1 for ridge countslpattern intensities, as in the case of finger patterns, is no
longer significant in females once the altitude effect is controlled for. Other components, however, show more consistent results
between the sexes than finger patterns did.
Most noteable is the partial correlation between rainfall and U3 (radial, ulnar and finger pattern intensity on digits 4 and 5), which
is very high at - .5 in both sexes. Also consistently significant are correlations with R3
401
(Radial and unilateral ridge counts on digits
1 and 2), and with R7 (ulnar ridge countsl
pattern intensity on digit 5).
Finally correlations between rainfall and
factors based on palmar traits are much more
consistent between the sexes than was the
case with altitude. Size factor U1 partially
correlates very highly at .54 in males, and
also significantly at .34 in females. Also consistent by sex are significant correlations for
U2 (palmar ridge counts) R1 (frequencies of
4 and 5 interdigital triradii). R3 correlates
significantly in males but not in females, and
R5 in females, but not in males.
DISCUSSION
This paper has demonstrated significant
statistical relationships between a number of
dermatoglyphic taxonomic traits, and/or altitude and rainfall. In general whorl frequencies, mean finger ridge counts, finger pattern
intensity, and interdigital and palmar intensities correlate negatively with altitude
orfand rainfall, while loop frequencies and
palmar ridge count means correlate positively. The results on the whole are relatively consistent in terms of sex and
asymmetry, given that sex and symmetrical
differences are marked and nonrandom in
the Kenya data series, showing considerable
fluctuations in magnitude from one macropopulation to another in some traits (Rosa,
1981: Chapter 7). Having noted the nonrandom nature of sex and asymmetrical inconsistencies, I do not propose to analyse them
further here, owing to their considerable
complexity.
Associations with climate and altitude are
often linked with types of biological variation that result from adaptation through the
direct or indirect action of selective forces.
This kind of interpretation cannot be seriously entertained at this stage, however, as
there are other factors that could explain the
apparent associations with altitude and rainfall just reported.
The sampling strategy adopted in the main
Kenyan study was an ethno-linguistic one,
not a geographical one-i.e., a sample was
taken from each minimal ethnic unit, not
from each of a system of geographical grid
squares. This could lead to bias insofar as a
majority of ethnic units are clustered in the
climatically more favoured western and central areas of Kenya, and away from the extremes of altitude and rainfall. The arid semidesert areas with very low mean annual
402
P. ROSA
rainfall, for instance, are represented mainly
by Cushitic samples. Should the Cushitic
peoples be dermatoglyphically differentiated
from other groups because of factors associated with differentiation along ethnic or linguistic lines rather than climate, then this
could lead to spurious correlations with
climate.
The “ethno-linguistic” factor is also relevant in another context. As peoples who are
closely related linguistically in Kenya also
tend to occupy the same type of topography
and climate, we could argue that the observed significant correlations with altitude/
rainfall are spurious because they merely reflect covariation with the ethno-linguistic
factor which is more likely to be (as a result
of previous dermatoglyphic studies which
have found no relationship between dermatoglyphic variation and climate) the “real”
influence on dermatoglyphic variation.
One way of testing whether the reported
correlations with altitude and rainfall are
likely to be spurious as a result of the potential influence of the ethno-linguistic factor is
to see if patterns of significant correlations
with altitude and rainfall, observed when all
samples are included, persist within linguistic groups. This exercise is problematic in the
present study in that the samples available
from most linguistic groups are either too
homogeneous regarding altitudelrainfall, or
so few that correlations would have to be
extremely high to be significant. Of the linguistic groups in Kenya sampled in this
study, the Highland Bantu are the most acceptable for such a “within group” analysis,
numbering 16 samples and displaying
marked variation in both altitude (2-8,000
ft) and rainfall (10-65 inches).
Given the unsatisfactory nature of performing analysis within the other (non Highland Bantu) linguistic groups alone, another
way of testing the validity of the correlations
with altitude and rainfall is to see whether
they show significant and similar results in
independent groups of samples drawn from
wider linguistic divisions. To this end I chose
the non-Luyia Bantu samples as one division, and the combined Nilotic and Cushitic
samples as the other. The results are presented in Table 4.To save space, only correlations with altitude are presented. Those
with rainfall do not contradict the general
conclusions outlined later.
Looking first at the results for the Highland Bantu by themselves, the only signifi-
cant correlations with altitude are for
concentric whorls in females, and twinned
loops in males. Nevertheless many correlations, though not significant, exceed k0.2,
and in the case of males compare favourably
in magnitude with those listed for the total
sample in Table 2.
The results comparing the Highland
Coastal Bantu on the one hand with those of
the CushitesDJilotes on the other are more
powerful. Significant correlations with altitude are common within both divisions and
in both sexes. The pattern of results for males
is similar in both divisions, and compares
very favourably with that observed in Table
2 for the total sample. Females, however,
present a less consistent picture. In the case
of finger patterns, the results are relatively
similar between the two linguistic divisions,
differing mainly in the fact that the Highland/Coastal Bantu correlate significantly in
the case of twinned loops (while the Cushitesl
Nilotes do not. Conversely the Highland/
Coastal Bantu do not correlate significantly
with concentric whorls, while the Cushitesl
Nilotes do. In the case of ridge counts, however, the Nilotic females correlate very
highly indeed with altitude, but show no such
trend in the HighlandCoastal Bantu. In the
case of palmar traits, significant correlations
exist in both divisions in both sexes, though
detailed inconsistencies are also apparent (for
instance TPPII is significant in both divisions in females, but not in males).
The results in Table 4 therefore reveal that
correlations over .2 are common within the
linguistic group examined, and that four of
these are significant. In two independent
pooled linguistic divisions, correlations with
altitude are significant in both, and show
consistent trends in the results for finger patterns (males and females), finger ridge counts
(males), palmar ridge counts (males), and
TPPII (females). The results for rainfall,
while not reported, show similar kinds of
trends. I would conclude, therefore, that the
underlying direct or indirect climaticienvironmental effect implied by altitude and
rainfall may not in fact be a spurious reflection of an underlying ethno-linguistic factor
as discussed. Rather we could argue that
many of the apparent ethno-linguistic associations with dermatoglyphic variation detectable in analyses of the Kenya data (Rosa,
1981: Chapter 8; 1983b) are as likely to be a
spurious manifestation of a climaticlenvironmental effect.
403
DERMATOGLYPHIC VARIATION AND CLIMATE IN KENYA
TABLE 4. Correlations of dermatoglyphic composite traits with altitude within linguistic or combined linguistic
divisions
Highland
Bantu
RTARCH
LTARCH
RTL
LTL
RTUL
LTUL
-~
RTRL
LTRL
RTW
LTW
RTCW
LTCW
RSWNL
LTTWNL
RTRC
LTRC
RTARC
LTARC
RTRRC
LTRRC
RTURC
LTURC
RTFPII
LTFPII
RTPRC
LTPRC
RTPPII
LTPPII
~~~
~~~
~
w rs
(n= 16)
0 rs
(n=16)
- .09
.07
23
.17
.32
.14
.03
.08
- .28
- .28
- .35
- 20
- .38*
- .41*
-.18
- .29
-.25
- .25
-.19
-.22
-.12
- .03
.09
- .05
-28
-.26
- .32
-.12
-.16
.34
.36
-.28
-.17
.oo
.31
.17
.09
- .02
.48*
.57**
-.18
- .31
- .07
-.11
.17
- .06
-.22
-.lo
.24
.03
.20
.06
.23
.18
.13
- .01
Highland and Coastal
Bantu
Q rs
w rs
(n=23)
(n=22)
-.12
- .03
.50**
.20
.56**
.18
.03
.17
-.49**
- .22
-.36*
.01
- .44*
- .35*
-.40*
- .24
-.49**
-.22
- .39*
-.17
-.57**
- .38*
- .38*
-.19
.44*
.40*
-.51**
- .46*
-.38*
- .07
.54**
.15
.50**
.20
- .08
- .05
- .27
- .08
.18
.29
-.39*
- .39*
-.19
-25
-.12
-.19
- .05
-.13
-.25
- .17
- .03
.03
.ll
.30
-.43*
- .38*
Cushites and Nilotes
w rs
(n= 18)
.15
.21
.49*
34
.46*
.28
.11
.20
- .49*
- .39
-.53*
-.11
- .21
- .50*
- .36
- .34
-.51*
-.50*
- .38
-.42*
- .54*
-.37
-.41*
- .35
.61**
.44*
- .ll
- .07
Q rs
(n= 17)
.42*
~~
27
.51*
.24*
.52*
.18
- .30
.09
-.66**
-.45*
-.75***
-.45*
-.16
-.14
-.82***
-.87***
-.78***
-.76***
-.8l***
-.81***
-.79***
-.53*
-.65**
- .38
.50*
.28
.50*
.43
*P 6 .05 > .01.
**P 6 .01 > ,001.
***P 4 ,001.
A basic question arising from the results
just reported is why, if dermatoglyphic vari-ation is significantly related to climate and
altitude in Kenya, did Hiernaux (1968) and
Hiernaux and Froment (1976) find no such
evidence in sub-Saharan Africa as a whole.
Several explanations are possible.
Firstly, the number of samples available to
Hiernaux and Froment (19761, though widely
distributed throughout Africa, are still very
thinly scattered in proportion to the total
geographical area encompassed by their
study. Trends of relationship with climate
would have to be very powerful to be detectable under such a sampling framework. The
tribal samples in this study, however, are
concentrated in one country and are nearly
as numerous by sex as the total number of
samples available to Hiernaux and Froment.
Secondly, the ability to detect anything but
the most crude relationships is hampered in
sub-Saharan African dermatoglyphic studies
by the methodological inconsistencies between dermatoglyphic workers, and by the
differing quality of the data available not
only from the point of view of dermatoglyphics but also from that of climatic measurements. Although Hiernaux and Froment
(1976) included only samples from the most
reliable studies, it is still possible that inconsistencies in print reading might still be serious enough from one study to another to
obscure any subtle relationships that might
exist between dermatoglyphic variation and
extraneous factors such as climate. In contrast the dermatoglyphic methodology in this
study is internally consistent.
Thirdly, Hiernaux and Froment may be
right that no relationship-direct or indirect-exists between dermatoglyphic variation and climateltopography in sub-Saharan
Africa. In that case the relationships reported here may either eventually be proved
to be spurious when covariation with other
404
P. ROSA
factors, as yet unknown or unexplored are
considered, or, the relationships, while real,
are related to factors that are localised in
their effects.
There is, however, some indirect evidence
that could imply a relationship between dermatoglyphic variation and climate in the
broader context of sub-Saharan Africa. Geographical clines of dermatoglyphic finger
patterns in sub-Saharan Africa are well documented (Lestrange, 1953; Gessain, 1957,
Chamla, 1962; Sunderland and Coope, 19731,
and recently Vecchi (1981); while Jantz and
Hawkinson (1979) for (some) ridge counts
demonstrated significant correlations with
both latitude and longitude. Given that climate in sub-Saharan Africa varies systematically with latitude and longitude in some
parts of the sub-continent (in this study, for
instance, mean annual rainfall correlates at
-.67 with longitude within Kenya), a relationship with climate is at least possible, and
we could speculate that sophisticated analyses might find it.
Should significant statistical relationships
be proven to exist between dermatoglyphic
variation and climate, not only within
Kenya, but elsewhere, the question would
then arise as to whether some aspect of climate could have a direct influence on dermatoglyphic variation. An open mind should
be kept regarding this issue, for recent articles by Loesch and Martin (1984) have begun
to question the general belief amongst human biologists that dermatoglyphic traits are
“neutral” with respect to adaptation. At the
very least this study has demonstrated that
the a search for dermatoglyphic relationships with environmental factors such as climate should not be neglected just because
they appear a priori to be unpromising areas
of investigation.
ACKNOWLEDGMENTS
Thanks are due to the Government of
Kenya for permitting the research, to the
headmasters, teachers, and schoolchildren for
their cooperation, and to Professor E. Sunderland for his support and advice. The fieldwork was supported by the Sigma XI
Foundation, Connecticut, and the Horniman
Fund, Royal Anthropological Institute.
LITERATURE CITED
Chamla, MC (1962) La repartition geographique des
cr&tespapillaires digitales dans le monde. Nouvel essaie de synthese. L’Anthropologie 66:526-541.
Cummins, H, and Midlo, C (1943) Finger Prints, Palms
and Soles. 1961 ed. New York Dover Publications.
Gessain, M (1962)Les dermatoglyphes digitaux de Noirs
d’Afrique. L’Anthropologie 61~239-267.
Hiernaux, J (1968) La Diversite Humaine en Afrique
sub-Saharienne. Recherches Biologiques Bruxelles: Insititut de Sociologie de L‘Universite Libre de Bruxelles.
Hiernaux, J, and Froment, A (1976) The correlations
between anthropobiological and climatic variables in
sub-Saharan Africa: revised estimates. Human Biology 48357-767.
Jantz, RL, and Hawkinson, CH (1979)Finger ridge-count
variability in sub-Saharan Africa. Annals of Human
Biology 6~113-124.
Lestrange, M de (1953)Les cr6tes papillaires digitales de
1491 Noirs d’Afrique occidentale. Bull. Inst. Franc.
Afr. Noire 151278-1295.
Loesch, DZ, and Martin, NG (1984) Finger ridge patterns
and tactile sensitivity. Annals of Human Biology
11~113-128.
Morgan, WT, and Shaffer, NM, (1966) Population of
Kenya, Density and Distribution. Nairobi: Oxford University Press.
Penrose, LS, and Loesch, D (1970) Topological classification of palmar dermatoglyphics. J. of Ment. Defic. Res.
24:111-128.
Rosa, PJ (1981) An Investigation of Dermatoglyphic
Variation among Ethnic Populations in Kenya. PhD
Thesis. University of Durham.
Rosa, PJ (1983a)Descriptive report on a dermatoglyphic
survey of 6,235 schoolchildren from Kenya. University
of Durham: Department of Anthropology Report,pp. 1174.
Rosa, PJ (198313) A dermatoglyphic survey of Kenyan
schoolchildren. Annals of Human Biology 6~579-584.
Sunderland, E, and Coope, E (1973) The tribes of south
and central Ghana: A dermatoglyphic investigation.
Man 8228-265.
Vecchi, F (1981) Geographical variation of digital dermatoglyphics in Africa. Am. J. Phys. Anthropol.
54:565-580.
Finger patterns
Arches
- $02
ARCH1
ARCH2
-.12
ARCH3
-.12
ARCH4
.02
ARCH5
.09
Undifferentiated loops
L1
.39***
L2
.22
L3
.19
L4
.17
L5
.oo
Ulnar loops
uL1
+39***
uL2
.22 *
uL3
.21
uL4
.17
uL5
- .01
Radial loops
-.14
RL1
RL2
- .03
RL3
- .08
RL4
-.lo
RL5
-.05
Undifferentiated whorls
cw1
- .41***
-.12
w2
w3
-.15
w4
-.18
-.01
w5
Concentric whorls
cw1
-.31**
cw2
- .06
cw3
-.19
cw4
- .07
cw5
- .07
Twinned loops
TWNLl
-.25*
TWNL2
.01
TWNL3
.oo
TWNL4
-.17
TWNL5
.11
Ridge counts
Unilateral maximal ridge counts
F1
-.14
F2
- ,055
~~
n
Altitude
- .07
-.14
.05
- .04
-.16
-.23*
-.17
-.07
-.30*
-.23*
.02
-.15
- .09
-.13
- .27*
-.lo
-.24*
.05
- .19
.16
- .03
.19
- .04
-.12
-.12
-.25*
-.12
-.16
- .15
-.15
-.22
- .22
-.34**
-.27*
- .07
-.14
- .21
-.41***
-.19
.24*
- .21
-.21
- .05
- .28*
- .29*
-.19
-.17
- .08
-.14
- .06
-.17
- .08
-.34**
- .04
- .26*
-.38**
.oo
.02
.oo
-.18
-.38**
.07
-.01
-.29*
- .10
- .09
- .09
- .30*
- .07
- .05
- .09
-.31**
-.19
-.17
- .05
-.14
-.19
- .08
- .05
- .18
-.25*
-.11
.02
.07
-.17
-.30*
- .05
-.16
-.35**
- .07
.05
.02
.21
.07
-.lo
.09
- .05
-.42***
-.29*
-.17
-.30*
- .18
-.15
- .29*
- .21
- .04
- .09
-.35**
- .07
- .01
-.41***
(Continued)
~~
.11
-.12
-.21
- .07
-.25*
-.27*
-.36**
- .23*
- .07
-.15
-.I2
-.32**
-.29*
- .22
-.27*
- .05
.05
.19
.07
.20
.26*
.07
.29*
.15
.23
.29*
.19
.04
- .09
- .02
-.15
0
.2 1
.03
- .02
-.12
Rainfall
.40***
.12
-.19
-.14
-.30*
-.12
.08
.02
-.12
- .07
.03
-.lo
.09
.21
.12
.14
.21
.01
.09
.07
.20
.21
.20
.11
.03
.06
.08
.29*
.14
- .05
.05
-.lo
- .05
.18
.22
.18
.27*
.11
.33**
-.25*
- .02
-.11
.02
0
.16
.15
.16
.05
.12
.05
.01
-.lo
.10
.08
0
Left hand
.28*
.07
-.19
.11
-.01
.11
0.
Altitude
- .20
- .20
.10
- .20
- .06
- .07
- .20
.20
.09
.12
.41***
.33**
.20
.10
.25*
.15
.10
.21
.26*
.39**
.23*
.08
-.12
- .03
-.11
.23*
-.13
.19
.17
.12
.41***
.34**
.05
.09
.05
- .22
.20
0
.20
.24*
.22
.14
.11
Rainfall
.20
.26*
.36**
.26*
.08
.06
-.la
.11
- .21
.16
0.
- .07
0
Right hand
APPENDIX I. Correlations of dermatoglyphic non-composite taxonomic traits with altitude and mean annual rainfall
z
3
5
?!
r;
m
k
B
u
2
b
0
2,
4
0
3
%
2
0
r;
E
U
M
I
F3
- .08
F4
-.35**
F5
-.30*
Absolute ridge counts
A1
-.38**
A2
-.12
A3
-.14
A4
-.28*
A5
-.27*
Radial ridee counts
FR1
- .07
FR2
.02
- .07
FR3
FR4
-.37**
FR5
-.30*
Ulnar ridge counts
FU1
-.46***
FU2
-.23*
FU3
-.17
FU4
-.13
FU5
-.12
Finger pattern intensity index
- .40** *
D1
D2
- .05
-.lo
D3
D4
-.18
D5
- .05
Palmar patterns
Palmar ridge counts
AB
.32**
BC
.10
CD
.32**
Palmar patterns and triradii
I1
- .23*
I11
- .02
IIIT
.14
Iv
-.31**
H
.01
H
- .22
EF
- .04
D
-.32**
Frequency of interdigital triradii
TD3
.23*
TD4
.28*
TD5
- .20
TD6
-.33**
a
Altitude
~
-.40**
- .29*
-.33**
-.31*
-.18
- .04
.15
- .26*
.10
.17
.02
.01
-.16
.05
-.19
-.13
.07
.12
-.21
.14
.04
- .24*
.31**
.05
.02
- .04
.06
-.27*
-.13
.13
.08
- .04
.28*
.43***
.42***
-25*
- .06
- .07
-.14
- .21
-.25*
-.15
- .23*
-.27*
- .03
.22
-.12
.20
-.13
-.15
- .40* **
- .24*
- .04
-.17
-.28*
.07
- .08
-.14
~
-.15
.17
-.19
.20
.02
.22
-.14
-.01
- .09
-.13
.ll
.03
.19
.25*
.19
-.23*
-.39**
- .22
-.31*
- .20
.24*
-.lo
-.41***
- .03
-.35**
-.01
.02
-.19
.09
-.35**
- .03
-.30*
.25*
.oo
.30*
- .08
-.39***
.03
- .04
.09
-.08
.01
.04
-.08
- .32**
-.lo
-.24*
-.29*
-.43***
-.34**
-.09
- .05
-.16
- .28*
-.19
-.25*
-.43***
- .29*
-.11
- .22
-.12
-.22*
-.12
-.31**
- .07
- .20
- .25
-.15
-.16
-.25*
-.21
-.11
- .22
v
- .26*
Q
- .07
Rainfall
- .07
a
-.35**
-.37**
-.26*
-.14
- .18
-.21
- .48***
-.39**
- .17
- .09
-.12
-.32**
-.33**
-.50***
-.34**
-.24*
- .47 * * *
-.38**
Q
Right hand
Altitude
-.38**
-.11
-.26*
-.12
-.01
- .06
.01
-.24*
.24*
.08
.08
.26*
-.17
.ll
-.22
.02
-.35**
.22
.2 1
.32**
.36**
.07
.03
.04
-.06
.08
- .02
.oo
.05
.01
- .03
-.05
- .01
.19
-.lo
.22
-.13
- .03
- .20
- .08
-.17
-.06
-.25*
-.26*
- .08
- .01
- .06
-.32**
-.25*
-.16
- .08
-.07
-.17
- .42***
-.15
-.15
- .09
-.27*
-.17
- .08
-.11
a
- .29*
-.30*
-.27*
- .42** *
- .52***
-.31*
- .07
-.19
-.34**
-.35**
-.31*
- .24*
-.38**
_,48***
- .29*
Q
Left hand
Rainfall
APPENDIX I. Correlations of dermatoglyphic non-composite taxonomic traits with altitude and mean annual rainfall (continued)
** *
9
.16
-.27*
.24*
.12
.04
-.13
.09
-.14
.05
-.24*
.21
.12
.30*
.32**
.15
-.15
- 20
-.12
-.24*
-.21
-.30*
- .20
-.26*
- .20
-.14
-.37**
-.42***
-.32**
-.27*
- .22
-.19
-.41***
-.37*+
-.30*
-.32**
-.30*
-.27*
- .43
a
a,
rp
0
407
DERMATOGLYPHIC VARIATION AND CLIMATE IN KENYA
APPENDIX II. Details of PCA unrotated and uarimax rotated factors from PCA analyses of dermatoglyphic non
composite traits’
Analysis 1: Finger patterns (70 traits being RARCHl to LTWNL5 in Appendix I)
u1
*16.6
RW4
RCW4
RL4
RW3
RUL4
RW2
LW3
LW2
RCWB
LCW3
23.8%(1)
.79
.78
-.77
.77
-.76
.75
.74
.74
.73
.70
u11
1.6
LCW5
LARCH4
LTWNLB
RARCH2
RARCH5
R9
LTWNL4
RTWNL4
LTWNL2
LTWNL3
RTWNL2
RTWNL3
LTWNL5
2.3%
- .41
.38
.35
-.32
.31
.70
.67
.66
.65
.63
.47
.43
U 1 (cont.)
RW5
.70
RL2
-.67
RL3
- .67
LW4
.66
- .66
RUL3
RL5
- .64
- .64
RUL5
RUL2
- .62
LL4
- .62
RCWB
.60
U17
1.0
1.5%
LRL4
- .40
RARCH4
.32
R11
LARCH1
RARCHl
.80
.71
U 1 (cont.)
RW1
.60
RCWl
59
LL2
- .59
LUL4
- .58
LCW2
.58
RCW2
.58
LW5
.56
LL3
- .56
LUL3
- .52
LCW4
.51
R1
LW1
LLl
LULl
RULl
RL1
RW1
LTWNLl
RTWNLl
RCWl
LCWl
R17
LRL2
RTWNL4
RW4
.92
- .92
- .91
- .87
- .86
.82
.76
56
.56
.45
u2
7.7
11.0%
LULl
- .75
LL1
- .72
RULl
- .64
RL1
-.63
LTWNLl
.60
1w1
.59
1cw4
- .51
RARCH2
.51
1u12
- .51
RARCH3
51
R2
RL5
ru15
RW5
RTWNL5
RW4
RL4
ru14
RCW4
RCW5
RCWB
- .86
- .86
.85
.68
.54
- .53
- .53
.49
.47
.46
U2 (cont.)
LARCH2
RTWNLl
LARCH3
RARCHl
50
.49
.49
.46
R6
ru13
r13
RW3
RCW3
RTWNL3
.79
.61
52
- .86
- .84
.45
-.42
- .41
Analysis 2: Ridge counts and finger pattern intensity (50 traits being RF1 to LD5 in Appendix I)
u1
29.3
58.5%
RA4
LA3
LF3
RA3
RA2
RA5
LFR3
LA4
LA5
LF4
U1 (cont.)
.92
.91
.90
.90
.89
.89
.89
.88
.88
.87
LF5
LFR5
LFR4
RF2
LF2
RF3
RFU4
LA2
RF5
RFRB
7.5%
.77
.77
.67
.65
.64
.54
2.9
LFU5
RFR4
LD5
RF4
LD4
LFU4
RFU5
RD5
RFRB
RF5
LFR2
RFR3
RFU2
RF1
RA1
RD4
RF4
LF1
RD3
RFR4
5.8
.55
-.51
50
-49
.48
.46
.38
.37
- .34
-.33
1.3
LD4
LFU4
RFU5
LA4
u3
u2
3.7
LFUl
LD1
RFUl
LA1
RD1
RA1
U 1 (cont.)
.87
.87
.86
.86
.86
.85
.84
.84
.84
.83
U1 (cant.)
.82
.82
.80
.78
.78
.78
.77
.77
.76
.76
RFRl
LFRl
RD2
RFU3
LD3
LD2
LA1
RFR2
LFU3
RFU5
U 1 (cont.)
.76
.75
.74
.73
.70
.69
.69
.69
.66
.65
RD5
LFU4
LD4
RD1
u7
2.0%
- .46
- .44
.40
-.32
APPENDIX 11. (Continued)
.64
.63
.62
.60
408
P. ROSA
APPENDIX 11. (cont.)
R1
RFR5
RF5
RFR4
RF4
LFR4
LF4
LFR5
LF5
RA5
.86
36
.85
.84
.82
.80
.78
.76
.74
R1 (cont.)
LA5
LF3
LFR3
LA4
RF3
RFK3
RA3
LA3
R3
.66
.60
.60
59
58
57
.56
53
LFRl
RFRl
LF1
RF1
RFR2
LFR2
RF2
RAl
LA1
R4
.81
.76
.73
.72
.63
56
53
51
LFUl
LD1
RFUl
RD1
LA1
RA1
R7
.92
.90
30
.78
.77
.65
50
Analysis 3: Palmar traits (30 traits being RAB to LTD6 in Appendix I)
u1
7.1
TD
TDL
TDR4
PIIR
TDL4
.89
.85
-.84
.81
-30
.76
.75
.71
.71
.65
.64
PIE
TDR6
TDL5
TDR5
TDLG
PIIJX
u2
3.8
12.8%
RBC
.84
~~.
LAB
.80
RAB
.79
LBC
.78
RCD
.72
LCD
.69
R1
TDR5
TDR4
TDR
TDL5
TDL4
U6
24.1%
1.6
CHL
CHR
TDR3
PIIITR
~
R2
.93
-36
.76
.63
-57
RAB
LAB
RBC
LBC
RCD
LCD
*Eigenvalue and percentage variance
'Factors included in Table 2 only.
5.8%
.56
..
~~
.47
- .43
.40
R3
.85
.84
.82
.72
.72
.69
TDLG
PIIL
TDL
PIIR
R5
.89
.81
.68
.55
CHL
CHR
.87
234
LFU5
LD5
LA5
RFU5
33
31
.49
.44
Документ
Категория
Без категории
Просмотров
1
Размер файла
923 Кб
Теги
climate, associations, variation, kenya, dermatoglyphic, topography
1/--страниц
Пожаловаться на содержимое документа