Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305 319

www.elsevier.com/locate/palaeo

Do red beds indicate paleoclimatic conditions?:

A Permian case study
Nathan D. Sheldon*
Department of Geology, Royal Holloway University of London, Egham, Surrey TW20 0EX, United Kingdom
Received 27 September 2004; received in revised form 25 February 2005; accepted 16 June 2005

Abstract
Terrestrial red beds have long been interpreted as desert deposits by comparison with modern red deserts. More recently red
beds have been interpreted as evidence of seasonally dry conditions and a PermoTriassic Pangean monsoon. Red beds of Cala
Viola, Sardinia are identified as paleosols and used to reconstruct Late Permian paleoclimatic conditions. Reconstruction of
paleoenvironmental conditions based on the paleosols of the Cala Viola indicates warm, humid conditions with no evidence of
dry conditions, as in a desert, or of extreme seasonality as in a monsoon. Instead, it is suggested that the red color of the
paleosols is a result of former good drainage, and that red color in general does not indicate specific paleoclimatic conditions.
D 2005 Elsevier B.V. All rights reserved.
Keywords: Permian; Sardinia; Paleosols; Paleoclimate; Monsoon

1. Introduction
A long held dogma is that widespread Triassic
terrestrial red beds indicate a global transition to warmer and drier conditions than those that characterized the
Permian (Parrish, 1995). This new climatic system has
been termed the Pangean mega-monsoon (Kutzbach
and Gallimore, 1989). The question to be addressed
herein is whether red color alone is sufficient evidence
upon which to base paleoclimatic interpretations.
Early research on modern red deserts supported the
interpretation that ancient red beds formed in hot, dry
* Tel.: +44 1784 443615.
E-mail address: [email protected].
0031-0182/$ - see front matter D 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.palaeo.2005.06.009

climates (Walker, 1976). However, modern red deserts

of Arizona and Australia are red because of sediments
recycled from paleosols of Triassic and Miocene age,
respectively, and most deserts of North and South
America, Asia, and the Middle East are grey like
their weathering source rocks. Further, many red
soils in semi-arid areas such as New Mexico derived
their red color from well-drained, warm conditions
during Pleistocene pluvials, so the red color is relict
and related to earlier paleoenvironmental conditions.
More recently, Parrish (1998, p. 192) stated that terrestrial red beds, b. . .appear to be indicative of climates that are warm and dry or seasonal with respect
to rainfall.Q A model put forth by Dubiel and Smoot
(1994) suggests that continental red bed formation is

306

N.D. Sheldon / Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305319

favored by warm climates with alternating wet-anddry seasons (monsoons) and open, desert to savanna
vegetation. However, it is not an exact analog because
a true savanna requires grasslands, which did not
evolve until the Cenozoic (Retallack, 2001a; Terry,
2001). Monsoonal tropical Pakistan and India include
both grey and red soils and paleosols (Retallack,
1991a). Wynn (2000) and Wynn and Retallack
(2001) describe reconstructed savanna ecosystems
from Cenozoic paleosols in Africa that are not significantly reddened. Furthermore, numerous examples
have been published of diagenetic reddening of nondesert paleosols by dehydration of iron oxyhydroxides
(Retallack, 1991a, 1997, 2001b; see also the review of
older literature in Blodgett et al., 1993).
Taken together, these various factors suggest that
the origin of red color may not be well-understood or
well-explained by existing models. Work presented
here on Late Permian red beds in Sardinia offers an
alternative explanation to pronounced aridity or sea-

sonality. It is suggested that the red color is indicative

of well-drained conditions and that it provides no
unequivocal information on the paleoclimatic conditions at the time of paleosol formation. Instead, paleoclimatic conditions are reconstructed on the basis of
other proxies, such as the degree of chemical weathering, nature and extent of pedogenic carbonate and
salts, and patterns of root traces and trace fossils.

2. Geologic context
Basin-and-Range topography was a result of the
CarboniferousPermian Hercynian orogeny from eastern
Europe to the southern coast of the United States
(Cortesogno et al., 1998). Collision of South Europe
with North America and Africa during the Late Devonian and Carboniferous (Condie, 1989) was followed, through Triassic time, by local rifting and
formation of continental basins in Spain, Southern

Fig. 1. Map showing the location of field sites and stratigraphic column for the Lago di Baratz section, Verrucano Sardo Formation. LB01-15 are
sample numbers, and Munsell colour of the sample is to the right of the sample number. The exact transition between the lesser developed
Mosca Pesca and Lago di Baratz paleosols and the overlying Macchia paleosols is unknown and may lie in the covered interval (shown with an
X) rather than at the first logged Macchia Rossa paleosol.

N.D. Sheldon / Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305319

France, Italy, Corsica, and Sardinia (Cassinis and

Ronchi, 1997). These basins were filled with clastic
red beds derived from the orogenic belt.
In Sardinia, those sediments are part of the Verrucano Sardo Formation exposed on the Cala Viola
(bviolet bayQ) (Fig. 1), and are divided it into four
informal units (Gasperi and Gelmini, 1979). The red
beds described in this paper are from bUnit 2,Q a 150
m package of sediments composed of sandy conglomerates, grey sandstones, and red sandstones and mudstones. The Cala Viola Nord section is capped by a
thick, quartz cobble conglomerate that is also exposed
near the base of the Cala Viola Sud section (Fig. 1).
The rocks exhibit fluvial paleochannels, tetrapod footprints, and fossil plants indicating subaerial deposition
on alluvial fans and floodplains. Gasperi and Gelmini
(1979) examined the limited available fossil assemblages and found Autunian (Early Permian) non-marine strata near the base of the sequence and Triassic
red claystones and sandstones near the top [Units 3
and 4], overlain by Middle Triassic (AnisianEarly
Ladinian) limestone (Cassinis and Ronchi, 1997; Cassinis et al., 1992). The red beds described here are
located near the top of the sequence (top of Unit 2),
and are thus Late Permian in age (Cassinis et al.,
1992).
The Lago di Baratz area (Fig. 1) is well vegetated
and exposure is generally poor. Three of the four
pedotypes are exposed in this section (Fig. 1), including the Lago di Baratz and Mosca Pesca pedotypes,
which are not preserved in either of the Cala Viola
sections. In contrast, the Cala Viola sections, with
localized gentle folding, are well-exposed in sea cliffs
and rock platforms, and are continuous and conformable with significant lateral variability difficult to
capture adequately in single stratigraphic sections
(Figs. 2 and 3). The Lago di Baratz section lies
stratigraphically below the Cala Viola sections by an
unknown thickness of mudstones and sandstones in
Unit 2 of Gasperi and Gelmini (1979). However,
given that only the middle and upper portions of
Unit 2 are red and the lower portion is primarily
grey, it is possible that the red Macchia paleosols of
the Lago di Baratz section (Fig. 1) represent the first
red beds. If this is the case, given the 7585 m
exposed in the Cala Viola sections (Figs. 2 and 3)
and a total thickness of 150 m for Unit 2 (Gasperi and
Gelmini, 1979), there can be no more than a few tens

307

Fig. 2. Stratigraphic column for the Cala Viola Nord section,

Verrucano Sardo Formation. Symbols and conventions are as in
Fig. 1. Asterisks next to sample numbers indicate samples that were
weakly reactive to dilute acid. The thicknesses of the conglomerate
that caps the Cala Viola Nord and Sud sections are variable, so the
average thickness is portrayed. Where multiple lithologies are
shown, there is significant variability along strike and additional
symbols apply to the right column, which represents the dominant
lithology.

308

N.D. Sheldon / Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305319

grain size changes (Retallack, 1997). Munsell color

was recorded as well as the qualitative degree of
calcareousness on the basis of reaction with dilute
hydrochloric acid (Retallack, 1997). Samples were
collected for petrography and geochemical analysis
from three sites north of Alghero, near Lago di Baratz
and at two localities on the Cala Viola (Fig. 1).
Geochemical data were obtained from a commercial
laboratory (Intertek of Vancouver, B.C.) using XRF,
ICP-MS, and titration (FeO) and are compiled in
Table 1. Paleosols were classified into pedotypes
(Retallack, 1997; Retallack, 2001b) on the basis of
physical and chemical characteristics (Sheldon and
Retallack, 2001; Sheldon et al., 2002), and analyzed
using the factor function approach (Jenny, 1941).
Bulk density (q) was measured by the clod method
using paraffin; analysis of 10 replicates of a single
sample gave an uncertainty of 0.09 g cm
3.

4. Evidence of pedogenesis

Fig. 3. Stratigraphic column for the Cala Viola Sud section, Verrucano Sardo Formation. Symbols and conventions are as in Fig. 1.
See Fig. 2 caption for additional information.

of meters between the top of the Lago di Baratz

section and the bottom of the Cala Viola Nord section.

3. Methods
Paleosols were recognized in the field on the basis
of ped morphology, horizonation, root traces, and

Paleosols in the Cala Viola sections fine up-profile

and are notably finer grained than the succession as a
whole (Figs. 2 and 3). This difference shows up in the
weathering profile of the sections as well (Fig. 4A)
and in contrast to the fluvially-derived sandstones
(Fig. 4B). Many of the paleosols also preserve drabhaloed root traces (Fig. 4C,F) and rarely, vertical
burrows (Fig. 4C,D). Burrows range up to one cm
in diameter and show some internal structure consistent with backfilling by an arthropod. Non-calcareous
rhizoliths are well-preserved in some of the paleosols,
both in hand specimen and thin section (Fig. 4H).
Both root traces and burrows penetrate deeply into
paleosol profiles (Fig. 4C), indicating that the paleowater table was substantially below the surface. Point
counts of thin sections (e.g., Fig. 4E,H) are consistent
with the field observation that paleosols are more finegrained than interfluve sandstones, siltstones, and
mudstones. Some of the Cala Viola paleosols have
an observed clay bulge (Fig. 5) and illuviation argillans observable in thin section consistent with subsurface accumulation of clay in a Bt horizon. A and B
horizons of paleosols have 8097% clay and phyllosilicate minerals and 320% quartz and lithics (including feldspars) with an average of less than 10%,
whereas C horizons and other fluvial sediments all

N.D. Sheldon / Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305319

309

Table 1
Geochemical data
Sample Level Reacta Horizon SiO2
(m)

TiO2 Al2O3 Feb

FeO Fe2O3 MnO MgO

LB02
LB03
LB04A
LB04B
LB04C
LB04D
LB11
LB12
LB13
LB14
LB15
CV03
CV06
CV07
CV08
CV09
CV10
CV13
CV16
CV19
CV22
CV24
CV25
CV26
CV27
CV28
CV29
CV30
CV31
CV45
CV46
CV47

3.6
4.4
5.9
6.0
6.2
6.4
13.0
13.2
13.5
13.7
13.9
1.7
4.2
5.5
5.8
6.1
6.5
10.1
12.4
14.1
17.2
19.15
19.85
20.65
22.05
23.2
24.2
24.9
26.15
46.5
48.0
49.25

0.37
0.30
0.70
0.81
0.84
0.70
0.52
0.72
0.80
0.79
0.78
0.92
0.84
0.72
0.87
0.91
0.91
0.91
0.78
0.86
0.47
0.59
0.63
0.25
0.89
0.83
0.89
0.93
0.61
0.40
0.71
0.26

0.45
0.39
0.51
0.45
0.51
0.58
0.45
0.71
0.71
0.90
0.64
0.84
1.61
0.71
1.03
0.90
0.77
0.90
1.74
1.87
1.03
1.42
0.51
0.26
0.77
0.77
0.64
0.58
0.51
0.32
0.45
0.45

Sample

Rho
(g cm
3)

LB02
LB03
LB04A
LB04B
LB04C
LB04D
LB11
LB12
LB13
LB14
LB15
CV03
CV06
CV07
CV08

2.52
2.39
2.58
2.49
2.50
2.55
2.59
2.59
2.64
2.53
2.54
2.71
2.70
2.71
2.81

N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Y
Y
N
N
Y
Y
N
N
N
N
N
N
N

A/Bwc
C
C
A/Bw
A
C
C
C
Bw
A
Bw
C
Bt
Bt
A
C
C
Bt
C
Bt
A
Bt
C
Bt
Bt
C
BC
Bw
CIA
K

98.51

97.49
98.11

97.28
97.94
90.99

86.82

77.98
78.63
66.80
70.08
69.27
67.22
75.11
66.15
61.17
62.34
61.20
58.00
56.05
48.79
54.82
56.93
57.71
56.05
55.56
56.24
30.48
56.12
51.90
14.89
56.94
53.35
59.56
57.89
55.13
58.99
58.65
83.50

Clayeynessd
0.095
0.07
0.13
0.15
0.15
0.13
0.11
0.15
0.18
0.18
0.18
0.195
0.17
0.20
0.21

12.64
9.62
14.89
17.22
17.08
15.35
14.21
16.84
18.76
18.84
18.63
19.18
16.17
16.40
19.81
20.65
20.21
19.22
14.59
16.04
9.25
13.94
11.69
4.99
19.69
15.45
19.87
20.01
11.67
6.22
12.04
8.57

1.80
4.81
7.30
1.68
1.73
6.52
2.83
5.58
6.54
5.15
7.02
6.61
6.06
6.19
6.40
6.49
6.98
6.32
6.55
3.47
3.30
4.64
4.83
2.98
7.98
5.58
6.59
8.35
4.29
1.26
3.43
1.23

P
( bases/Al)d
0.22
0.18
0.33
0.35
0.35
0.32
0.32
0.45
0.49
0.47
0.47
0.63
1.095
1.56
0.73

1.30
4.37
6.73
1.19
1.16
5.88
2.33
4.79
5.75
4.15
6.31
5.68
4.27
5.40
5.26
5.49
6.12
5.32
4.62
1.40
1.51
3.07
4.26
2.69
7.12
4.72
5.88
7.71
3.72
0.90
2.93
0.73

Salin.d

Gleyd

0.16
0.11
0.21
0.23
0.24
0.21
0.19
0.27
0.29
0.29
0.29
0.31
0.29
0.31
0.31

0.77
0.198
0.17
0.84
0.98
0.22
0.43
0.33
0.27
0.48
0.23
0.33
0.84
0.29
0.44

0.02
0.03
0.05
0.01
0.02
0.05
0.03
0.02
0.02
0.02
0.02
0.05
0.13
0.46
0.07
0.02
0.02
0.07
0.18
0.17
0.91
0.18
0.34
0.91
0.02
0.06
0.03
0.03
0.15
0.12
0.09
0.01

CaO

0.27 0.07
0.21 0.08
0.60 0.16
0.74 0.09
0.66 0.06
0.59 0.09
0.62 0.11
1.14 0.12
1.40 0.08
1.30 0.10
1.32 0.08
1.81 0.90
2.60 3.50
4.30 5.26
2.21 1.50
1.67 0.28
1.57 0.20
2.00 1.31
3.03 4.20
3.12 4.47
10.6 15.5
3.27 5.08
4.77 7.67
15.1 22.6
1.44 0.29
1.50 1.90
1.24 0.19
1.20 0.14
1.18 6.57
6.17 9.31
4.28 5.56
0.28 0.22

MAP

MAT

1540

15.3

1509
1528

12.8
13.4

1503
1522
1328

12.0
12.0
11.6

1223

11.5

Na2O K2O P2O5 LOI

Total

n/a
n/a
0.10
0.19
0.20
0.08
0.03
0.14
0.17
0.21
0.15
0.16
0.23
0.22
0.17
0.15
0.18
0.16
0.32
0.18
0.29
0.26
0.19
0.22
0.09
0.22
0.17
0.12
0.56
0.25
0.59
0.05

99.1
99.3
99.8
100.0
99.1
99.8
100.8
100.4
99.8
100.3
99.9
100.4
101.1
100.0
100.4
100.1
100.0
99.7
101.4
101.5
100.8
100.5
99.7
98.8
99.9
90.6
99.7
99.9
94.6
99.1
98.6
99.0

Bae
132
76
351
362
377
293
208
358
423
405
398
622
1636
544
998

1.85
0.97
2.75
3.40
3.56
2.88
2.48
3.95
4.82
4.69
4.70
5.22
4.05
4.41
5.47
5.72
5.27
5.30
3.79
3.87
2.26
3.58
2.89
1.10
5.79
4.24
4.95
4.77
2.41
1.57
2.85
1.57
Sr
66
36
92
127
116
95
61
107
121
123
128
117
134
122
105

0.07
0.12
0.15
0.05
0.05
0.11
0.07
0.08
0.06
0.06
0.08
0.06
0.07
0.09
0.12
0.12
0.08
0.08
0.04
0.09
0.06
0.05
0.07
0.07
0.14
0.09
0.12
0.11
0.05
0.03
0.04
0.03

4.03
4.16
5.79
5.31
5.08
5.66
4.29
4.99
5.27
5.93
5.30
6.67
9.78
12.5
7.92
6.27
6.12
7.39
10.7
11.1
26.7
11.4
14.2
35.4
5.82
6.61
5.43
5.75
11.5
14.5
9.89
2.75

Nb

Zr

Rb

25
18
31
42
37
29
23
36
34
33
38
37
34
34
30

28
18
21
28
22
22
22
28
21
28
23
23
21
16
22

164
141
257
312
281
260
229
280
215
229
225
171
284
144
114

112
72
137
194
180
151
159
240
292
293
276
266
195
210
262

(continued on next page)

310

N.D. Sheldon / Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305319

Table 1 (continued)
Sample

Rho
(g cm
3)

CIA
K

Clayeynessd

P
( bases/Al)d

Salin.d

Gleyd

MAP

MAT

Bae

Sr

Nb

Zr

Rb

CV09
CV10
CV13
CV16
CV19
CV22
CV24
CV25
CV26
CV27
CV28
CV29
CV30
CV31
CV45
CV46
CV47

2.77
2.74
2.74
2.64
2.65
2.76
2.67
2.58
2.82
2.74
2.54
2.70
2.72
2.60
2.69
2.75
2.60

96.47
96.84

0.21
0.21
0.20
0.15
0.17
0.18
0.15
0.13
0.20
0.21
0.18
0.20
0.20
0.13
0.06
0.12
0.07

0.54
0.51
0.70
1.37
1.28
6.26
1.56
2.52
16.2
0.54
0.79
0.46
0.43
1.58
5.57
2.08
0.37

0.31
0.30
0.31
0.32
0.28
0.32
0.31
0.29
0.31
0.33
0.32
0.28
0.27
0.30
0.34
0.34
0.21

0.36
0.28
0.38
0.84
2.97
1.52
1.03
0.27
0.22
0.24
0.36
0.24
0.17
0.31
0.79
0.34
1.37

1479
1490

11.5
11.8

1485

11.3

1493
1518

12.0
12.3

1428

13.4

449
418
1081
951
959
537
2439
2434
9654
507
51,452
3292
367
34,767
275
303
144

98
94
125
92
113
87
136
142
272
158
1017
262
179
688
73
107
78

33
30
27
36
42
43
33
41
29
33
22
32
47
30
34
32
17

24
21
25
19
19
10
16
12
nd
24
nd
22
21
9
20
22
23

120
153
147
237
301
93
195
306
31
125
151
159
173
269
202
390
133

280
270
260
184
196
94
163
124
40
270
178
244
241
95
79
138
102

96.68
96.95
97.79

94.67

Reactive with dilute HCl.

b
Total iron as Fe2O3.
c
Refers to samples at the boundary between two
P horizons.
d
Molar ratios: clayeyness = (Al2O3 / SiO2);
bases / Al = (CaO + MgO + Na2O + K2O) / Al2O3; salinization (Na2O + K2O) / Al2O3; gleization = (FeO / Fe2O3).
e
All trace element compositions.

have 1535% quartz and lithics, with an average of

about 20%. Many of the fluvial sediments, and two of
the paleosols, have non-calcareous sandy concretions
ranging in size from mm-scale (Fig. 4E) to decimeter
scale. Typically these are grey-green, fine- to mediumgrained sand in otherwise red sediments. Grey-green
color in iron-bearing paleosols is typically associated
with unoxidized iron. These apparent micro-reducing
conditions may be attributable to the former presence
of decaying organic matter, and may have been
formed in a fashion similar to the drab-haloed root
traces.
Bulk rock geochemical properties of fluvial rocks
may also be used to identify paleosols and to separate
paleosol orders (Sheldon et al., 2002). Net gains and
losses of different elements may be calculated by
examining the mobility of the element of interest
relative to some assumed immobile element (Chad-

wick et al., 1990; e.g., Driese et al., 2000). Elements

that are typically considered as immobile during
weathering include Ti, Zr, Nb, Hf, and under some
pH conditions, Al. Ti, Zr, and Nb were considered and
Ti was selected both because it was immobile relative
to Zr and Nb and because it is the most abundant of
the three elements. The open system mass-transport
function for element j in the weathered sample (w)
is defined as follows (e.g., Chadwick et al., 1990):
sj;w



 


qw Cj;w = qp Cj;p ei;w 1
1

where q w is the density of the weathered material,

C j,w is the chemical concentration (weight percentage)
of element j in the weathered material, q p is the
density of the parent material, and C j,p is the chemical
concentration (weight percentage) of element j in the
parent material. If s j,w = 0 (i.e., element w was immo-

Fig. 4. Field and petrographic photos. A) Outcrop photo of the Cala Viola Nord section; paleosols sit on top of the coarser, horizontal benches in
the section (arrow). B) Outcrop photo showing the complex fluvial character of Verrucano Sardo Formation. C) Profile of a Cala Viola paleosol;
the light colored vertical streaks are rhizoliths and drab-haloed root traces (arrows). D) Close-up of vertically oriented burrows in a paleosol A
horizon (arrow). E) mm-scale quartz concretion (sample CV-20). F) Root traces (arrow) deep in the C horizon of a paleosol, penetrating nearly
to into the A horizon of the underlying paleosol. G) Laterally discontinuous ground water gleying (arrow) features where the water table was
ephemerally closer to the surface. H) Thin section of a root trace (sample CV-12).

N.D. Sheldon / Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305319

311

312

N.D. Sheldon / Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305319

Fig. 5. Chemical degree of weathering. A) Ca and Sr loss in the type Macchia and Cala Viola paleosols assuming Ti is immobile. A tau value of

1 represents 100% loss of Ca relative to the parent material, and a tau value of 0 represents the parent material. Both paleosols lost Ca, with
the Cala Viola type profile showing greater Ca loss, consistent with a greater degree of chemical weathering. Other elements such as Sr show
more complicated changes, but are still consistent with a greater degree of chemical weathering in the type Cala Viola profile. B) Additional
geochemistry (CIA
K (Maynard, 1992) and clayeyness (molar ratio of alumina to silica; Retallack, 1997)). of the type Cala Viola profile The
significant offset between values low in the profile and high in profile is evidence of intense chemical weathering. The clayeyness index shows a
bbulgeQ consistent with the field identification of a Bt horizon.

bile), then e i,w can be solved for separately allowing

us to bypass volume (as in the classical definition of
strain) as follows (e.g., Chadwick et al., 1990):
ei;w


 

qp Cj;p = qw Cj;w
1

where e i,w is the strain on immobile element i in the

weathered sample. The parent materials for the profiles
were overbank mudstones and sandstones as appropri-

ate, with separate geochemical analyses for each of the

paleosol profiles (Table 1; lowermost C horizon analyses). Fig. 6 shows the losses or gains of Ca and Sr
(which occupy the same sites in most minerals) in the
type Macchia and Cala Viola paleosols assuming Ti
was immobile during weathering (calculated following
Chadwick et al. (1990)). Although both pedotypes have
lost much of their Ca relative to their parent material,
the Cala Viola paleosol has clearly been more weath-

N.D. Sheldon / Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305319

313

Fig. 6. Gleization for the type profiles of the Cala Viola and Macchia pedotypes. Gleization is the molar ratio of ferrous (Fe2+) to ferric
(Fe3+) iron.

ered, a finding consistent with the field classification of

the paleosols (Table 2). Paleosols generally exhibit the
highest degree of chemical weathering within their A
and B horizons, with a decrease in weathering down
profile. Fig. 5B shows the chemical index of alteration
without potash (Maynard, 1992) for the type Cala Viola
paleosol. This pattern is consistent with pedogenesis
rather than fluvial sedimentation where one would
predict a more erratic variation from bed to bed, but
where most values would cluster around a btypicalQ
value for the whole sedimentary succession.
4.1. Pedotypes
Four pedotypes (sensu Retallack, 1994) were
identified and given names in Italian from their
field localities or reconstructed similarity to a given
environment.
4.1.1. Cala Viola (bviolet bayQ)
The type Cala Viola paleosol crops out in the
northern Cala Viola section. Cala Viola paleosols are
Table 2
Cala Viola pedotypes
Pedotype

Diagnosis

Cala Viola Thick and red with clayey

subsurface (Bt) on alluvium
Grey-green silty soil with
Lago di
Baratz
some relict bedding and no
diagnostic subsurface horizons
Macchia
Variable thickness red
sometimes with scattered
drab haloed root traces and
no subsurface Bt or Bk
Mosca
Sandy, relict bedding,
Pesca
without horizonation

FAO

USDA

Luvisol

Alfisol

Fluvisol

Entisol

Cambisol Inceptisol

Fluvisol

Entisol

comparable to modern Alfisols (Soil Survey Staff,

1998) in the USDA soil classification scheme (Table
2). They are moderately developed (see Retallack
(1988) for definitions of the degree of development),
with no relict bedding, blocky peds, and subsurface Bt
or Bw horizons. Cala Viola profiles are typically A
BtC, and represent a fairly stable landscape (i.e.,
infrequently flooded; Table 3).
4.1.2. Lago di Baratz (bBaratzs lakeQ)
The type Lago di Baratz paleosol crops out in the
Lago di Baratz section. Lago di Baratz paleosols are
comparable to modern Entisols (Soil Survey Staff,
1998) in the USDA soil classification scheme (Table
2). They are very weakly developed, with some relict
bedding and no diagnostic subsurface horizons. Lago
di Baratz profiles are AC and represent a frequently
disturbed landscape (i.e., flooded; Table 3).
4.1.3. Macchia (bunderbrushQ)
The type Macchia paleosol crops out in the Lago di
Baratz section, and Macchia paleosols are found in
both field areas. Macchia paleosols are comparable to
modern Inceptisols (Soil Survey Staff, 1998) in the
USDA soil classification scheme (Table 2). They are
weakly developed with little relict bedding or ped
structure. Macchia profiles are A(Bw)C and represent a fairly stable landscape (Table 3).
4.1.4. Mosca Pesca (bfly fishingQ)
The type Mosca Pesca paleosol outcrops in the
Lago di Baratz section. Mosca Pesca paleosols are
comparable to modern Entisols (Soil Survey Staff,
1998) in the USDA soil classification scheme (Table
2). They are very weakly developed, preserve relict
bedding, lack ped structure, and lack diagnostic sub-

314

N.D. Sheldon / Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305319

Table 3
Paleoenvironmental interpretation
Pedotype
Cala Viola

Paleoclimate

Humid (13001500 mm/yr)

temperate
Lago di Baratz Insufficiently developed to
determine, but probably
humid
Macchia
Humid (13001500 mm/yr)
temperate
Mosca Pesca

Paleotopography

Eutrophic forest

Negligible, but
Alluvial sandstones, 100010 000 years
well-drained siltstones and mud-stones
Negligible, but poorly Coarse sandstone
1005000 years
drained

Stream-side early
successional woody
and herbaceous vegetation
Eutrophic forest
Negligible,
but moderately
to well-drained
Insufficiently developed to Stream-side early
Negligible
determine
successional herbaceous
vegetation

Parent material

Timea

Former vegetation

Alluvial sandstones, 5005000 years

siltstones, and
mud-stones
Coarse sandstone
b100 years

Estimated semi-quantitatively after Retallack (1997) and references therein.

surface horizons. Mosca Pesca profiles are ACC and

represent a frequently disturbed landscape (Table 3).

5. Diagenesis
Paleosols typically undergo a number of diagenetic
changes including loss of organic matter, burial reddening due to dehydration of oxyhydroxides (e.g.,
conversion of goethite [Fe(OH)3] to hematite
[Fe2O3]), and compaction due to burial beneath an
overburden (Retallack, 1991b). Paleosols of the Cala
Viola preserve essentially no organic matter (b 1%
by volume in thin section point counts). Studies of
Quaternary (Stevenson, 1969) and older (Retallack,
2001b) paleosols have shown that buried paleosols
lose up to an order of magnitude of organic carbon
soon after burial in well-drained soils, whereas waterlogged (hydromorphic) or peaty paleosols show significantly less to no loss of organic matter (Stevenson,
1969). Given their red color and low ferrous to ferric
ratios (see Fig. 6), much of the iron in these paleosols
has been oxidized, indicating at least a moderate
degree of aeration post-burial, and oxygen promotes
the breakdown of organic matter. This likely accounts
for the dearth of detectable organic matter.
Sheldon and Retallack (2001) showed that the
degree of compactibility varies according to the initial
physical properties of the soil. Regional stratigraphic
relationships indicate a burial depth of 24 km, so the
paleosols have been compacted to between 61.2% and
87.8% of their original thickness depending on burial
depth and soil order (see Sheldon and Retallack,

2001). Given that all of these paleosols are developed

on alluvium, an estimate based on inorganic floodplain silts and muds (see Sheldon and Retallack,
2001) of 78.686.4% of the original thickness is a
good first order generalization for the sedimentary
succession as a whole.

6. Paleoclimatic reconstruction
A number of means have been devised to reconstruct paleoclimate from paleosols. Retallack (1994)
has suggested that the depth to the Bk horizon can be
related to mean annual precipitation (see Royer (1999,
2000) and Retallack (2000) for discussion of this
approach). Although a couple of the Sardinian paleosols effervesce slightly when hydrochloric acid is
applied, there is nothing that would qualify Bk horizons (Soil Survey Staff, 1998). Royer (1999) suggested that soil carbonate is absent in regions
receiving precipitation N760 mm per year, although
this value varies with seasonality and local evapotranspiration (Retallack, 2000; Royer, 2000). This
value for the western US may be applicable to the
Sardinian paleosols given their formation within a
continental interior montane basin.
A more quantitative approach is to compare the
precipitation regimes of modern soils with indices of
chemical weathering (Sheldon et al., 2002; Sheldon,
2003). Climatic transfer functions applied to a set of
paleosols spanning the EoceneOligocene boundary
produced results that were consistent with independent estimates of mean annual precipitation and mean

N.D. Sheldon / Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305319

annual temperature (Retallack et al., 2000; Sheldon et

al., 2002; Sheldon and Retallack, 2004). Although the
paleosols in this study are much older, bulk rock
geochemical data for the Sardinian paleosols can
also be used to reconstruct the paleoenvironmental
conditions under with they formed, because the presence of root traces, rhizoliths, and burrows indicates a
formerly vegetated landscape for which modern analogues can be identified. Mean annual precipitation
can be related to the chemical index of alteration
without potash as follows (Sheldon et al., 2002):
MAP in mm 221:12e:0197CIA
K

with an R 2 = 0.72 where CIA
K is 100 times the

molar ratio of aluminum to aluminum, calcium, and
sodium (Maynard, 1992). Mean annual temperature
can be related to salinization (Retallack, 1997) where
MAT 8C
18:5S 17:3
2

with a somewhat low R = 0.37 (Sheldon et al.,

2002). As shown in Fig. 7, mean annual precipitation
increased slightly from 1300 mm/year to about 1500
mm/year, while mean annual temperature increased
slightly but held fairly steady at 1114 8C. That
result is consistent with the general lack of soil
carbonate. Two analyses (CV22 and CV25 on
Table 1) of Bt horizons are excluded from the analysis because of extremely low oxide totals owing to
high volatile contents (LOI on Table 1; 26.7 and
14.2%, respectively).

315

At the present time, soils forming under conditions

of N1200 mm/year mean annual precipitation and 11
14 8C mean annual temperature are found in Mexico
on the eastern side of the Gulf of California, in the
United States on the eastern side of the Appalachians,
northern India, Greece, and southern Italy (FAO,
19711981). Given the proximity of the Sardinian
paleosols to the Hercynian chain and their low paleolatitude (10 F 5 degrees), northern India is probably
the best modern analogue. Such comparisons are imperfect modern analogues because PermoTriassic
CO2 levels far exceeded present levels (Berner and
Kothavala, 2001; Retallack, 2001c). Nevertheless, it is
clear that these soils did not form in desert conditions.
Could they have instead formed in a monsoonal
paleoenvironment?
Modern monsoonal environments are characterized
by extreme seasonal variation, with a pronounced dry
season or seasons, and a short, very wet season or
seasons. There are two main varieties, namely, wet
monsoons as in Southeast Asia, Indonesia, northeastern Australia, and some of India, and dry monsoons as
in central Asia, parts of India, northwestern Australia,
the Arabian peninsula, and the southwestern United
States, however, there is a spectrum of conditions
between the main end-members. Soils forming under
dry monsoonal conditions are most often Vertisols,
Aridisols, and rarely, Mollisols (FAO, 19711981).
Soils forming under wet monsoonal conditions are
most often Ultisols or Vertisols (FAO, 19711981).

Fig. 7. Paleoprecipitation and paleotemperature estimates using transfer functions from regression of climatic data against chemical composition
of Quaternary soils. The standard error on the precipitation estimate is F 182 mm and the standard error on the temperature estimate is F4.4 8C.

316

N.D. Sheldon / Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305319

Table 4
Paleoclimates compared
Features

Desert

Wet
Dry
Cala viola
monsoon monsoon

Salts
Yes
CaCO3 nodules
Yes
MAP (13001500 mm) No
MAT (11148)
No
Soil Types
Aridisols

No
No
Yes
Maybe
Ultisols
Vertisols
Yes

No
Yes
Maybe
Maybe
Vertisols
Aridisols
Mollisols
Yes

No
No
Yes
Yes
Alfisols
Inceptisols
Entisols
No

Layered
Fe(OH)3CaCO3
Mukkara/gilgai
gilgai

No
No

Yes

Yes

No

Only the Cala Viola and Macchia pedotypes are

sufficiently developed to use in comparisons with
modern environments. Both pedotypes developed
on areas of little or no topographic relief, had similar
parent material, and indicate similar paleoclimatic
conditions (Table 3). The primary differences were
formation time and vegetative covering (Table 3),
though it could be argued that Cala Viola pedotypes
represent later stage succession of Macchia pedotypes, however there is not sufficient evidence to
conclude this robustly.
A dry monsoon is considered first. In Vertisols,
large precipitation differences or seasonal soil moisture deficits between wet and dry seasons change the
physical volume of smectite clay minerals in soils.
During the wet season, clays swell with the additional water. In the dry season, the clays lose the
water that they have gained and the parting between
layers shrinks. These shrink-swell cycles lead to
deep cracks in the soil. The behavior of the clay
minerals and colloids also gives rise to mukkara
structure and gilgai microrelief that readily distinguishes Vertisols (Coulombe et al., 1996; Coulombe,
1997; Retallack, 1997; Driese et al., 2000, 2003).
None of these features (deep vertical to sub-vertical
cracks, mukkara structure, gilgai microrelief) are
present in any of the Sardinian paleosols (Table 4).
Nor do the Sardinian paleosols have pedogenic calcrete and salts of Aridisols, or the abundant crumb
peds, organic matter, and fine root traces of Mollisols (Table 4).
Wet monsoons are characterized by Vertisols and
Ultisols. Ultisols are similar to Alfisols; the primary

difference is in base saturation. Alfisols are base-rich

soils that typically have forest vegetation, while Ultisols are base-poor forest soils. Because of this difference, modern Alfisols and Ultisols are distinguished
on the basis of their base status (e.g., percentage base
saturation), which is not always recorded (or measurable) in paleosols. A statistically significant method of
differentiating Alfisols from Ultisols has been derived
for paleosols. The B horizons of Alfisols have molar
ratio of bases (CaO, Na2O, MgO, K2O) to alumina
(Al2O3) greater that 0.5, whereas the B horizons of
Ultisols have base / alumina ratios less than 0.5, typically much less (Sheldon et al., 2002). Fig. 8 shows
the base / alumina ratios of paleosols in the Cala Viola
section; most have base / alumina ratios greater than
0.5, thereby confirming the field diagnosis of these
paleosols as Inceptisol-like (Macchia) and Alfisol-like
(Cala Viola) rather than Ultisol-like. Only one of Cala
Viola paleosols plots within the Ultisols field, though
others are bnear-Ultic,Q perhaps indicating some weak
monsoonal influence (Fig. 8).
Retallack (1991a) studied Miocene monsoonal
paleosols and soils of Pakistan and found that they
typically have concretions (rather than nodules) of
hematite, calcite, or interlayered calcite and hematite,
and diffuse carbonate in small nodules throughout the
profile, including the A horizon. There is essentially
no carbonate in the Sardinian paleosols and the rare

Fig. 8. Molar ratio of total bases to alumina for Cala Viola paleosols.
Only one Cala Viola paleosol (at 24.9 m) plots within the Ultisol
field, though others are bnear-Ultic.Q

N.D. Sheldon / Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305319

observed nodules are unlayered, and silica- or ironoxyhydroxide cemented. There are no nodules or
concretions consistent with a monsoonal paleoclimate.
Taken together, these various lines of evidence suggest that the Sardinian paleosols were not subject to a
strongly monsoonal paleoclimate, either wet or dry
(Table 4).
Although the Lago di Baratz and Mosca Pesca
paleosols of the Lago di Baratz section show little
significant reddening, the Macchia and Cala Viola
paleosols of both the Lago di Baratz and Cala Viola
sections are both characterized by very red color.
Intensity of color and degree of clay remobilization
are two-fold indicators of development, and can be
supported with chemical and petrographic data. Degree
of drainage also plays a role in soil color and can be
inferred from the degree of chemical gleization (molar
ratio of Fe2+ / Fe3+), soil redoximorphic features (e.g.,
reduction spots, grey/green paleosols with red mottles), and trace fossils of organisms requiring oxygen
(animal burrows and root traces). The red paleosol
types (Macchia and Cala Viola) are characterized by
low gleization ratios (e.g., Fig. 6) and nearly uniform
red color, with the exception of rare drab-haloed root
traces. Drab-haloed root traces are commonly created
by micro-reducing conditions, which occur around
decaying organic matter shortly after burial (Retallack, 1991b), and as such, would be unrelated to the
past water table depth. There are no other soil redoximorphic features (iron-manganese nodules, ferric
nodules) and the deeply penetrating root traces and
burrows (Fig. 4CD) indicate good drainage as does
the degree of chemical weathering and clay illuviation
into subsurface horizons. The root traces are drab
from the inside out, as in surface water gley, rather
than groundwater gley, yet there is no high density or
impermeable layer within the paleosols that would
perch the water table.
The sequence, as a whole, goes from weakly developed grey paleosols to more strongly developed
red paleosols, which is consistent with a dropping
base level or increased distance from a stream
(Kraus, 1999), and has no evidence significant paleotopography (Table 3). The Cala Viola Nord section of
red paleosols is capped by a thick, areally extensive
conglomerate with centimeter-sized, well-rounded
cobbles that may represent a sequence boundary because the overlying paleosols in the Cala Viola Sud

317

section are again weakly developed (Retallack, 1998;

Kraus, 1999). This suggests a long-term cyclicity in
the alluvial delivery system that could be related to
either tectonics or minor climate change, but not to
desertification or monsoonal conditions. The red color
of these paleosols appears to be primarily related to
the hydrological conditions in which they formed.

7. Conclusions
Continental red beds should be studied outcrop by
outcrop as they can form in a variety of settings, rather
than generalized to a single genetic model. Red paleosols form in environments ranging from tropical forests to deserts. Red color, in and of itself, is not
diagnostic. Paleoclimatic reconstruction of Late Permian paleosols on the basis of the degree of chemical
weathering and pedological features indicates humid,
temperate conditions with no evidence of either desert
conditions or precipitation seasonality pronounced
enough to be called a monsoon. The change from
grey paleosols to red paleosols is attributable to
changes in hydrological drainage, rather than desertification or increased seasonality. Low latitude Late
Permian paleoclimate of Sardinia was warm, temperate and perhaps mildly seasonal, but certainly not
desertic or strongly monsoonal.

Acknowledgements
The author would like to acknowledge financial
support from a Geological Society of America student
research grant to him for this project while he was a
Ph.D. student. An earlier version of the manuscript
benefited from reviews by Steve Driese and Greg
Retallack, and this version has benefited from two
anonymous reviews and a review by Lee Nordt.

References
Berner, R.A., Kothavala, X., 2001. GEOCARBIII: a revised model
of atmospheric CO2 over Phanerozoic time. American Journal
of Science 301, 182 204.
Blodgett, R.H., Crabaugh, J., McBride, E.F., 1993. The color of red
beds; a geologic perspective. Soil Science Society of America,
Special Publication 31, 127 159.

318

N.D. Sheldon / Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305319

Cassinis, G., Ronchi, A., 1997. Upper carboniferous to Lower

Permian continental deposits in Sardinia (Italy). Geodiversitas
19, 217 220.
Cassinis, G., Toutin-Morin, N., Virgili, C., 1992. Permian and
Triassic events in the continental domains of Mediterranean
Europe. In: Sweet, W.C., Yang, Z., Dickins, J.M., Yin, H.
(Eds.), PermoTriassic Events in the Eastern Tethys: Stratigraphy, Classification, and Relations with the Western Tethys,
IGCP, volume 203. Cambridge University Press, Cambridge,
pp. 60 77.
Chadwick, O.A., Brimhall, G.H., Hendricks, D.M., 1990. From a
black to a gray box a mass balance interpretation of pedogenesis. Geomorphology 3, 369 390.
Condie, K.C., 1989. Plate Tectonics and Crustal Evolution, 3rd ed.
Pergamon Press, New York.
Coulombe, C., 1997. Surface properties of Vertisols of Texas and
Mexico under different land use systems. Texas A and M
University (PhD dissertation). 293 pp.
Coulombe, C., Dixon, J.B., Wilding, L.P., 1996. Mineralogy
and chemistry of Vertisols. Developments in Soil Science 24,
115 200.
Cortesogno, L., Cassinis, G., Dallagiovanna, G., Gaggero, L.,
Oggiano, G., Ronchi, A., Seno, S., Vanossi, M., Liegeois,
J.-P., 1998. The Variscan post-collisional volcanism in
Late CarboniferousPermian sequences of Ligurian Alps,
Southern Alps and Sardinia (Italy): a synthesis. Lithos
45, 305 328.
Driese, S.G., Mora, C.I., Stiles, C.A., Joeckel, R.M., Nordt, L.C.,
2000. Mass-balance reconstruction of a modern Vertisol: implications for interpreting the geochemistry and burial alteration of
paleo-Vertisols. Geoderma 95, 179 204.
Driese, S.G., Jacobs, J.R., Nordt, L.C., 2003. Comparison of modern and ancient Vertisols developed on limestone in terms of
their geochemistry and parent material. Sedimentary Geology
157, 49 69.
Dubiel, R.F., Smoot, J.P., 1994. Criteria for interpreting paleoclimate from red beds a tool for Pangean reconstructions. In:
Embry, A.F., Beauchamp, B., Glass, B.J. (Eds.), Pangea: Global
Environments and Resources, Canadian Society of Petroleum
Geologists, Memoir, vol. 17. Canadian Society of Petroleum
Geologists, Calgary, pp. 295 310.
FAO, 19711981. Soil Map of the World IX, UNESCO.
Gasperi, G., Gelmini, R., 1979. Ricerche sul Verrocano 4. Verrocano del Nurra (Sardegna nord-occidentale). Memorie della
Societa Geologica Italiana 20, 215 231.
Jenny, H.J., 1941. Factors of Soil Formation. McGraw-Hill,
New York.
Kraus, M.J., 1999. Paleosols in clastic sedimentary rocks: their
geologic applications. Earth Science Reviews 47, 41 70.
Kutzbach, J.E., Gallimore, R.G., 1989. Pangean climates: megamonsoons of the megacontinent. Journal of Geophysical Research 94, 3341 3357.
Maynard, J.B., 1992. Chemistry of modern soils as a guide to
interpreting Precambrian paleosols. Journal of Geology 100,
279 289.
Parrish, J.T., 1995. Geologic evidence of Permian climate.
In: Scholle, P.A., Tadeusz, M.P., Ulmer-Scholle, D.S. (Eds.),

The Permian of Northern Pangea. Springer-Verlag, London,

pp. 53 61.
Parrish, J.T., 1998. Interpreting Pre-Quaternary Climate from the
Geologic Record. Columbia University Press, New York.
Retallack, G.J., 1988. Field recognition of paleosols. In: Reinhardt,
J., Sigleo, W.R. (Eds.), Paleosols and Weathering Through
Geologic Time: Principles and Applications, Geol. Soc. Am.
Special Paper, vol. 216. Geological Society of America, Denver,
pp. 1 20.
Retallack, G.J., 1991a. Miocene Paleosols and Ape Habitats of
Pakistan and Kenya, Oxford Monographs on Geology and
Geophysics, vol. 19. Oxford University Press, Oxford.
Retallack, G.J., 1991b. Untangling the effects of burial alteration
and ancient soil formation. Annual Reviews of Earth and Planetary Science 19, 183 206.
Retallack, G.J., 1994. The environmental factor approach to the
interpretation of paleosols. Soil Science Society of America,
Special Publication 33, 31 64.
Retallack, G.J., 1997. A Colour Guide to Paleosols. John Wiley and
Sons, Chichester.
Retallack, G.J., 1998. Fossil soils and the completeness of the rock
and fossil records. In: Donovan, S.K., Paul, C.R.C. (Eds.), The
Adequacy of the Fossil Record. John Wiley and Sons, New
York, pp. 133 163.
Retallack, G.J., 2000. Comment on bDepth to pedogenic
carbonate as a paleoprecipitation indicator?Q. Geology 28,
572.
Retallack, G.J., 2001a. Cenozoic expansion of grasslands and climatic cooling. Journal of Geology 109, 407 426.
Retallack, G.J., 2001b. Soils of the Past, 2nd ed. Blackwell Science,
London.
Retallack, G.J., 2001c. A 300-million-year record of atmospheric carbon dioxide from fossil plant cuticles. Nature 411,
287 290.
Retallack, G.J., Bestland, E.A., Fremd, T.J., 2000. Eocene and
Oligocene Paleosols of Central Oregon. Geological Society of
America Special Paper, vol. 344. Geological Society of America, Denver.
Royer, D.L., 1999. Depth to pedogenic carbonate horizon as a
paleoprecipitation indicator? Geology 27, 1123 1127.
Royer, D.L., 2000. Reply to Retallacks (2000) comment on bDepth
to pedogenic carbonate horizon as a paleoprecipitation
indicator?Q. Geology 28, 572 573.
Sheldon, N.D., 2003. Pedogenesis and geochemical alteration of the
picture gorge subgroup, Columbia river basalt, Oregon. Geological Society of America Bulletin 115, 1377 1387.
Sheldon, N.D., Retallack, G.J., 2001. Equation for compaction of
paleosols due to burial. Geology 29, 247 250.
Sheldon, N.D., Retallack, G.J., 2004. Regional paleoprecipitation
records from the Late Eocene and Oligocene of North America.
Journal of Geology 112, 487 494.
Sheldon, N.D., Retallack, G.J., Tanaka, S., 2002. Geochemical
climofunctions from North American soils and application to
paleosols across the EoceneOligocene boundary in Oregon.
Journal of Geology 110, 687 696.
Soil Survey Staff, 1998. Keys to Soil Taxonomy. Pocahontas Press,
Blacksburg, VA.

N.D. Sheldon / Palaeogeography, Palaeoclimatology, Palaeoecology 228 (2005) 305319

Stevenson, F.J., 1969. Pedohumus: accumulation and diagenesis
during the Quaternary. Soil Science 197, 470 479.
Terry Jr., D.O., 2001. Paleopedology of the Chadron Formation of
northwestern Nebraska: implications for paleoclimatic conditions in the North American mid-continent across the Eocene
Oligocene boundary. Palaeogeography, Palaeoclimatology,
Palaeoecology 168, 1 39.
Walker, T.R., 1976. Diagenetic origin of continental red beds.
In: Faulke, H. (Ed.), The Continental Permian in Central,

319

West, and South Europe. Riedel Publishing, Netherlands,

pp. 240 282.
Wynn, J.G., 2000. Paleosols, stable carbon isotopes, and paleoenvironmental interpretation of Kanapoi, Northern Kenya. Journal
of Human Evolution 39, 411 432.
Wynn, J.G., Retallack, G.J., 2001. Paleoenvironmental reconstruction of Middle Miocene paleosols bearing Kenyapithecus and
Victoriapithecus, Nyakach Formation, southwestern Kenya.
Journal of Human Evolution 40, 263 288.