See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/319993662

Paleosol and ichnofossil evidence for significant Neotropical habitat variation

during the late middle Miocene (Serravallian)

Article in Palaeogeography Palaeoclimatology Palaeoecology · September 2017

DOI: 10.1016/j.palaeo.2017.09.024

CITATIONS READS

19 884

5 authors, including:

Angeline Catena Daniel I Hembree

Diablo Valley Colllege University of Tennessee
16 PUBLICATIONS 111 CITATIONS 98 PUBLICATIONS 1,225 CITATIONS

SEE PROFILE SEE PROFILE

Federico Anaya Darin A. Croft

Universidad Autónoma "Tomás Frías" Case Western Reserve University
40 PUBLICATIONS 1,207 CITATIONS 175 PUBLICATIONS 4,372 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Angeline Catena on 26 October 2017.

The user has requested enhancement of the downloaded file.

 Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology

journal homepage: www.elsevier.com/locate/palaeo

Paleosol and ichnofossil evidence for significant Neotropical habitat MARK

variation during the late middle Miocene (Serravallian)
Angeline M. Catenaa,⁎, Daniel I. Hembreeb, Beverly Z. Saylorc, Federico Anayad, Darin A. Crofta
a
Department of Anatomy, Case Western Reserve University, 10900 Euclid Ave., Cleveland, OH 44106-4930, United States
b
Department of Geological Sciences, 316 Clippinger Laboratories, Athens, OH 45701-2979, United States
c
Department of Earth, Environmental and Planetary Sciences, Case Western Reserve University, 10900 Euclid Ave., Cleveland, OH 44106, United States
d
Departamento de Ingeniería Geológica, Universidad Autónoma Tomás Frías, Potosí, Bolivia

A R T I C L E I N F O A B S T R A C T

Keywords: We use paleopedology and ichnology to elucidate the habitat of the late middle Miocene fossil site of Quebrada
Continental Honda, southern Bolivia. The paleosols represent three pedotypes, Type 1 and Type 2 paleosols are interpreted as
Paleoenvironment Inceptisols (Eutrudepts) and Entisols (Udifluvents), respectively, which formed on proximal and distal flood-
Neogene plains in a seasonal, sub-humid to humid wooded grassland-like vegetative community. Type 3 paleosols are
Mammals
interpreted as Inceptisols (Calciustepts) that formed in more densely vegetated wooded grassland-like vegetative
Alluvial
communities on distal floodplains in a strongly seasonal sub-humid to semi-arid environment. The ichnofossil
Provinciality
assemblage of Quebrada Honda includes Celliforma, Coprinisphaera, Taenidium, Fictovichnus, Planolites, Skolithos,
Katarrhedrites, and root traces and represents heterogeneous communities dominated by soil arthropods and
plants. The physical and geochemical properties of the paleosols, including low maturity, poor development of
horizons, and the overall moderate estimates of mean annual precipitation, indicate changes in soil moisture due
to seasonal precipitation and flooding and low but varying degrees of temporal stability. The diverse ichnofossil
assemblage of Quebrada Honda reflects environments with greater primary productivity and temporal stability
than those of nearby Cerdas, Bolivia, which are several million years older. Quebrada Honda's inferred pa-
leoenvironments differ markedly from those of La Venta, Colombia, indicating that dissimilar habitats may
partly or principally account for the vastly different faunas of these two well-sampled and contemporaneous
fossil localities.

1. Introduction Antoine et al., 2016b; Croft et al., 2016; Kerber et al., 2016). Therefore,
applying new analyses to known Neotropical sites is just as important as
The modern tropics of Central and South America (Neotropics) have the search for new fossil sites.
been the focus of myriad ecological studies that have provided insights The fossil site of Quebrada Honda is among the richer Neotropical
into the biodiversity and functioning of modern tropical floras, faunas, mammal sites presently known (MacFadden and Wolff, 1981; Marshall
and ecosystems (Sarmiento, 1984; Hoorn et al., 2010a; Hughes et al., and Sempéré, 1991; Croft, 2007, 2016; Engelman et al., 2016). > 30
2013; Rull, 2014; Smith et al., 2014). By contrast, comparatively little is species of well-preserved mammals have been described there in ad-
known about the structure and functioning of ancient Neotropical dition to several types of non-mammalian vertebrates (Croft et al.,
ecosystems. This is largely the result of a sparse terrestrial fossil record 2013; Cadena et al., 2015). Among Neotropical sites of Miocene age,
(Flynn and Wyss, 1998; MacFadden, 2006). Only a handful of relatively Quebrada Honda is second only to La Venta, Colombia (Fig. 1A), in the
well-sampled paleontological sites older than a few million years are total number of well-documented mammal species that have been re-
known from tropical latitudes of South America (Croft, 2007; Carrillo corded, but the ancient depositional environments and habitats pre-
et al., 2015; Antoine et al., 2016a). Of these, only a small proportion served there have never been studied in detail. As a result, little is
derive from a thick sedimentary sequence that not only preserves a known about this ecosystem as a whole and the factors potentially re-
diverse fossil fauna, but also is suitable for associated geochronologic, sponsible for similarities and differences between the mammals of
paleoenvironmental, and paleoecological studies (e.g., Kay and Quebrada Honda and those of other Neotropical localities.
Madden, 1997a; Cozzuol, 2006; Croft, 2007; Tejada-Lara et al., 2015; The principal aim of the present study is to use paleosols and

⁎
Corresponding author.
E-mail address: [email protected] (A.M. Catena).

http://dx.doi.org/10.1016/j.palaeo.2017.09.024
Received 1 May 2017; Received in revised form 20 September 2017; Accepted 20 September 2017
Available online 22 September 2017
0031-0182/ © 2017 Elsevier B.V. All rights reserved.
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

Fig. 1. A) Locations of Quebrada Honda (star) and La Venta

(black dot) in South America; B) Inset box showing
Quebrada Honda (star, 21° 57‘ S, 65° 25‘ W) and Cerdas (C,
20° 52′ S, 66° 19′ W) in Bolivia (modified from Croft, 2007);
C) Google Earth image showing outcrops of the Quebrada
Honda local area (QH, green dot) and the Río Rosario local
area (RR, blue dot) (Image © 2017 CNES/Astrium, © 2017
Google). (For interpretation of the references to color in
this figure legend, the reader is referred to the web version
of this article.)

ichnofossils as independent lines of evidence to reconstruct the middle study. The Quebrada Honda LA (abbreviated throughout as QH) in-
Miocene paleoenvironment of Quebrada Honda. Quebrada Honda is cludes badlands with approximately 300 m of stratigraphic section lo-
highly suitable for such a study because the thick sedimentary sequence cated 1–2 km southeast of the town of the same name. The Río Rosario
includes abundant paleosols that, in turn, preserve abundant ichno- LA (abbreviated throughout as RR) includes approximately 300 m of
fossils and vertebrate fossil remains (Croft et al., 2013; Catena et al., section located 5–6 km north of the town of Quebrada Honda. Although
2015). We also compare paleopedological and paleoichnological data stratigraphic units cannot be laterally traced between QH and RR, the
from Quebrada Honda to similar data from both La Venta and the two local areas have been correlated with one another based on li-
slightly older locality of Cerdas, Bolivia to test previous hypotheses thology and local paleomagnetic sections (MacFadden et al., 1990).
about the paleoecologies of Quebrada Honda mammals and possible The entire QH section has been constrained to approximately
reasons for their taxonomic distinctiveness relative to La Venta and 13.0–11.9 Ma (Serravallian Age; Polarity Chrons C5AA and C5A of the
Cerdas. Geomagnetic Polarity Time Scale or GPTS (Ogg, 2012)) by the local
paleomagnetic sections combined with two 40K/40Ar dates, one on sa-
nidine (12.83 ± 0.07 Ma at approximately 85 m) and another from
2. Geographic and geologic setting biotite (11.96 ± 0.11 Ma at approximately 260 m) derived from vol-
canic tuffs (MacFadden et al., 1990). The most fossiliferous zones of QH
The fossil site known as Quebrada Honda (ca. 21°57′S, 65°25′W) is are near the base of the section, below the 12.83 ± 0.07 Ma dated
located in the Eastern Cordillera of southern Bolivia, 65 km southwest “Unit 9” tuff of MacFadden et al. (1990). These zones are within the
of the city of Tarija (Fig. 1A, B). The Neogene sedimentary rocks at local magnetic polarity zone N1 which correlates to Polarity Chron
Quebrada Honda are mapped as part of the Honda Group by the Ser- C5AAn (MacFadden et al., 1990) and spans 13.18–13.03 Ma based on
vicio Geológico de Bolivia and are composed of tan to reddish brown Ogg (2012).
claystones, siltstones, and fine-grained sandstones; interbedded tuffac- The RR section has an estimated age of 13.0–11.9 Ma and has been
eous horizons, and local conglomerates and coarser-grained sandstones tentatively correlated to the lower and middle portions of Polarity
(MacFadden and Wolff, 1981; MacFadden et al., 1990; Auerbach et al., Chron C5A via the QH section (MacFadden et al., 1990). Its most fos-
2015). Modern elevation is approximately 3500 m, but it was probably siliferous portion is the lowest 100 m (MacFadden et al., 1990; personal
much lower during the middle Miocene (see below). observation), which is primarily of reversed polarity (MacFadden et al.,
Four broad areas of badland exposures (local areas or LAs) have 1990) and correlates to zone C5Ar.3r of the GPTS (13.03–12.89 Ma;
been recognized near the town of Quebrada Honda: Quebrada Honda Ogg, 2012). The mammals of QH and RR are considered to represent a
LA, Río Rosario LA, Papachacra LA, and Huayllajara LA. Collectively, single, relatively contemporaneous fauna (Croft and Anaya, 2006;
these comprise the “Quebrada Honda” site, and their local faunas Croft, 2007; Croft et al., 2011).
comprise the Quebrada Honda Fauna (Croft, 2007; Croft et al., 2011; The lower 100 m of both the QH and RR sections are composed of
Engelman et al., 2015, 2016; Brandoni et al., 2017). Most geological siltstones, sandy siltstones, sandstone, conglomerates, and laterally
and paleontological studies have focused on the Quebrada Honda and continuous volcanic tuffs, some of which have been reworked into
Río Rosario LAs (e.g., MacFadden et al., 1990; Sánchez-Villagra et al., biotite sandstones (Fig. 2). The biotite sands and tuffs can be traced
2000; Goin et al., 2003; Engelman and Croft, 2014; Garzione et al., laterally for hundreds of meters and are useful for stratigraphic
2014) (Fig. 1C); therefore, we examined the same areas for in this

382
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

Fig. 2. Stratigraphy of the lower QH and RR

beds. A) Composite stratigraphic section of
northern and southern exposures of QH and lo-
cations of QHS1 (blue vertical bar) and QHS2
(red vertical bar). Solid lines indicate strati-
graphic correlations between a biotite sand and a
tuff in the northern and southern exposures, da-
shed line indicates distance between exposures;
B) Composite stratigraphic section of RR with the
location of RRS1 (green vertical bar). Paleosols
approximately colored to match outcrops. (For
interpretation of the references to color in this
figure legend, the reader is referred to the web
version of this article.)

correlations among topographic features within each local area but are 3. Methodology
not physically distinctive enough to permit correlation between local
areas (e.g., QH and RR). The sedimentary units of the QH section have The lower (most fossiliferous) portions of the QH and RR sections
been interpreted as laterally-limited fluvial channel and floodplain were chosen for study. The major lithologic units at RR and the
deposits with average sediment accumulation rates of 15–50 cm/kyr northern and southern sides of the main ridge of exposures at QH were
(Auerbach et al., 2015). Paleosols within both the QH and RR sections measured and described to produce composite stratigraphic sections
have been noted within fine-grained sedimentary units interpreted as (Fig. 2). We identified paleosols within the fine-grained units of the QH
overbank deposits (Auerbach et al., 2015); paleopedogenic features and RR sections. Two areas within QH were selected for detailed pa-
include mottling, pedogenic carbonate, and abundant root traces leopedogenic and ichnologic analysis; QHS1, the northern exposure,
(Garzione et al., 2014; Auerbach et al., 2015). was sampled in the 0–18 m stratigraphic interval (Figs. 2A, 3), and
Garzione et al. (2014) used clumped isotope paleothermometry of QHS2, the southern exposure, was sampled in the 22–25 m strati-
pedogenic carbonates (a soil temperature proxy) to estimate the late graphic interval (Figs. 2A, 4). The sampled areas are located within
middle Miocene mean annual air temperature (MAAT) at QH at 4–13 °C 300 m of each other, and their stratigraphic positions were correlated
and the associated elevation at 2.6 ± 0.6 km. However, paleo- with a biotite sand unit (Fig. 2A). In RR, we used five sampling areas, all
temperature estimates derived from the physiognomy of fossil leaves at located within 100 m of each other, to form an 18.5 m composite sec-
other Bolivian localities suggest Central Andean paleoelevations < tion (RRS1; Fig. 5) within the 22–41 m stratigraphic interval (Fig. 2B).
1.5 km during this same interval (Gregory-Wodzicki, 2000; Garzione This stratigraphic position for the RR composite section was chosen
et al., 2008). Recently, Cadena et al. (2015) estimated QH elevation because it includes the highest density of vertebrate specimens and was
at < 1000 m based on likely physiological limitations of a giant tortoise accessible.
(Testudinidae: Chelonoidis) and a freshwater turtle (Chelidae: Acantho- At each sampling area, a 1 m-wide trench was excavated to a depth
chelys) from the site. These discrepancies may be due to the fact that of 20–50 cm to describe trace fossils and paleosol profiles (including
Garzione et al. (2014) sampled pedogenic carbonates at greater depths lithology, color, structure, nodules, and redox depletions, concentra-
within paleosol profiles at QH than at other localities (due to a lack of tions, etc.). Where possible, specimens of representative trace fossils
well-defined argillic horizons at QH), which would result in under- were collected for additional description. Paleosol classification follows
estimates of MAAT and overestimates of paleoelevation (Garzione et al., the descriptive system of Mack et al. (1993); paleosols are also assigned
2014). a modern soil classification that follows the system of Soil Survey Staff
(2010). Bulk paleosol samples were collected from each identified pa-
leosol horizon for analyses of thin sections and geochemistry.

383
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

Fig. 3. Stratigraphic column of QHS1 with Type 1 (blue labels) and Type 2 (red labels) paleosols. Each profile is marked by brackets and interpreted horizons are labeled; column is
continuous from left to right. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Sixty-five thin sections (QH: n = 30; RR: n = 35), mounted on X-ray fluorescence (XRF). Samples (0.85 g each) were dried, pulverized,
2.5 × 5.0 cm slides, were prepared by Texas Petrographic Services, sieved to 180 μm, and analyzed using lithium borate fusion XRF. Bulk
Inc., Houston TX, USA. The thin sections were studied and photo- geochemical data are reported in oxide percents (Supplementary
graphed under a polarizing microscope (Motic BA300, Moticam 10) to Table 2); weight percents were normalized to their molecular weights
identify micromorphological features including grain and plasmic mi- and used to calculate six molecular weathering ratios (MWRs; base loss,
crofabrics (Supplementary Table 1), peds, nodules, illuviated clays, calcification, leaching, lessivage, oxidation, salinization), and chemical
lithic fragments, and small trace fossils. Micromorphological descrip- index of alteration without potash (CIA-K), a weathering index
tions of thin sections follow the nomenclature of Brewer (1976) and (Retallack, 2001; Sheldon and Tabor, 2009) (Supplementary Tables 3,
Fitzpatrick (1993). 4). Mean annual precipitation (MAP) was calculated from CIA-K
Fifty-five samples (QH: n = 23; RR: n = 32) were analyzed by ALS (Sheldon and Tabor, 2009). Following Catena et al. (2016), values for
Minerals (Reno, Nevada) for bulk geochemistry of major elements using the six MRW from 0.0–0.50 were considered low, from 0.51–1.00 were

384
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

4. Results

4.1. Sedimentology

The measured Honda Group beds of QH primarily consist of massive

tan, brown, and reddish brown siltstones, sandstones, vitric biotite tuffs
and alluvial conglomerates. The siltstones are laterally continuous
tabular sheets with gradational upper and lower contacts and are the
most common lithologic units at both QH and RR. Siltstones in both
local areas commonly contain mottles, clay clasts and lithic clasts, and
those of RR commonly contain pedogenic carbonate nodules. The bio-
tite sandstones are laterally continuous tabular sheets with sharp ero-
sional lower contacts and gradational upper contacts. Biotite sandstones
with areas of planar lamination and vitric biotite tuffs are present at
both QH and RR. Sandstones are fine to medium-grained and commonly
contain euhedral biotite crystals and subangular to subrounded pebble-
sized siliciclastic and volcanic-derived lithics. These units are laterally
continuous and traceable for distances of > 1 km. The biotite sand-
stones represent volcanic deposits that were reworked by alluvial re-
distribution and later modified by pedogenic processes. Thin con-
glomerates with sharp lower contacts and gradational upper contacts
that contain angular to subrounded, primarily quartzite, clasts are
present at multiple levels at QH but are rare at RR.

4.2. Continental ichnology

4.2.1. Rhizoliths
Rhizohaloes: Within the rhizosphere, organic compounds that are
produced by roots and microbes preferentially reduce and mobilize
pedogenic iron and manganese (Retallack, 2001; Kraus and Hasiotis,
2006). Reduction and mobilization processes result in elongate, drab-
colored mottles that extend into the paleosol matrix from the rhizo-
sphere (Kraus and Hasiotis, 2006). QH and RR paleosols contain elon-
gate, vertical, downward tapering, and horizontally branching light
greenish grey (G1 7/2) to light grey (5Y 7/1) rhizohaloes (Fig. 6A–C).
QH paleosols also contain rare yellowish red (5YR 5/6) rhizohaloes
with a light greenish grey lining. The contacts between the lined rhi-
zohaloes and the surrounding matrix are sharp, whereas the contacts
between the unlined rhizohaloes and the surrounding matrix vary from
sharp to diffuse. The rhizohaloes are circular in cross-section, with
diameters ranging from 0.3–1.8 cm. Observed lengths of the rhizoha-
loes range from 0.7–5.0 cm. Observed horizontal branch lengths range
from 0.2–0.5 cm. The rhizohaloes are similar in lithologic composition
to the surrounding matrix.
Root casts: Root casts are formed when root molds or tubules are
filled with sediment or cement (Klappa, 1980). The sediment within the
mold or tubule preserves the architecture of the root (Klappa, 1980). RR
paleosols contain abundant, primarily vertical, horizontally branching
and downward tapering root casts. The root casts are circular in cross-
section; observed total lengths range from 0.5–5.5 cm, and observed
widths range from 0.2–1.3 cm. An inverted “Y” shaped, horizontal
branching pattern is common, with branch lengths ranging from
0.3–2.1 cm. Root casts are filled with tan to brown mudstone that is
finer grained than the surrounding matrix and shows illuviation fea-
tures in thin section (Fig. 6D–F); the centers of the root casts commonly
Fig. 4. Stratigraphic column of QHS2 with Type 3 paleosols (green labels). Each profile is contain calcite (Fig. 6F) and rarely contain fine-grained siltstone.
marked by brackets and interpreted horizons are labeled. (For interpretation of the re- Rhizotubules: Rhizotubules form when calcium carbonate pre-
ferences to color in this figure legend, the reader is referred to the web version of this cipitated from evaporating soil water fills tubular voids left in a soil
article.)
after a root decays (Klappa, 1980; Buol et al., 2011). Carbonate within
the mold preserves the morphology of the root. Primarily vertical,
considered moderate, and > 1.00 were considered high. The CIA-K downward tapering and horizontally branching calcareous rhizotubules
values from 0 to 50%, 51–75%, and 76–100% were considered low, are rare at QH but are abundant at RR. The rhizotubules are circular to
moderate, and high, respectively. MAP estimates from 0 to 800 mm slightly ellipsoidal in cross-section; observed lengths and widths range
(arid to semi-arid), 801–1600 mm (temperate sub-humid to humid), from 0.2–3.1 cm and 0.05–0.7 cm, respectively (Fig. 6G–H).
and 1600+ mm (wet tropical) were considered low, moderate, and Rhizoconcretions: Rhizoconcretions are calcareous, tubular con-
high, respectively (modified from Kottek et al., 2006). cretions of calcium carbonate that accumulate around root structures

385
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

Fig. 5. Composite stratigraphic column of RRS1 with Type 1 (blue labels), Type 2 (red labels) and Type 3 (green labels) paleosols. Each profile is marked by brackets and interpreted
horizons are labeled; column is continuous from left to right. Dashed lines indicate the borders between the five subsections used to create the composite. (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of this article.)

rather than filling open voids left by roots (Klappa, 1980; Retallack, 4.2.2. Fossil burrows
2001; Kraus and Hasiotis, 2006). Carbonate can accumulate around 4.2.2.1. Skolithos. Vertical to subvertical, unlined, non-branching
living plant roots through repeated wetting and drying cycles and also shafts with sharp boundaries between the burrow and the
via preferential water flow through soil voids that remains after root surrounding matrix are common in all QH and RR paleosols
death (Klappa, 1980; Birkeland, 1999; Retallack, 2001). Vertical, (Fig. 7A–B). The shafts are linear and circular to ellipsoidal in cross-
branching and non-branching rhizoconcretions are present in QH and section. The burrows have observed lengths of 1.2–4.1 cm, observed
RR paleosols (Fig. 6I–K). The rhizoconcretions are circular in cross- diameters of 0.2–0.5 cm, and are filled with massive fine- to medium-
section, have an observed length of up to 10.2 cm, and have widths grained sandstone that is coarser grained and commonly lighter in color
ranging from 1.0–2.9 cm. When preserved, the terminal points of the than the surrounding matrix. These burrows are interpreted as
rhizoconcretions are tapered. temporary dwelling structures of small soil arthropods (e.g., Hasiotis,

386
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

Fig. 6. Rhizoliths. A–C Rhizohaloes: A) Lateral view of lined rhizohalo with a sharp contact with the surrounding matrix (inside of dashed line, from QHS1, T1P P2); B) Cross sections of
several pale-green rhizohaloes (at arrows) with sharp contacts (from QHS1 T1P P1); C) Lateral view of lined rhizohalo with horizontally-oriented branches (inside of dashed line, from
QHS1, T1P P1. D–F Root casts: D) Lateral view of an elongate, vertically-oriented root cast (at arrows) preserved in two halves of the same sample (from RRS1 T3P P13); E) Oblique view
of vertically- oriented, horizontally-branching root casts (at arrows) with calcite-filled centers (from RRS1 T3P P11); F) Lateral view and cross section of a calcite-filled root casts
surrounded by illuviated clay in a thin section (at arrows, under polarized light) (from RRS1 T3P P2). G–H Rhizotubules: G) Lateral view of small, branching rhizotubules in a thin section
under cross-polarized light (from RRS1 T3P P5); H) Lateral view of a large, elongate, horizontally-oriented rhizotubules (at arrow) (from RRS1 T3P P4). I – K Rhizoconcretions: I) Lateral
view of a downward branching rhizoconcretions (outlined) (from RRS1 T3P P15); J) Isolated rhizoconcretion with tapered termination point (at left; from QHS2 T3P P5); K) Elongate,
rounded rhizoconcretion (from QHS2 T3P P5). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

2002; Bowen and Hembree, 2014). grained sandstone that is the same color as the surrounding matrix but
coarser-grained. The burrows are interpreted as dwelling and
4.2.2.2. Planolites. Horizontal, unlined, non-branching tunnels with reproduction structures of small arthropods such as beetles,
sharp boundaries between the burrow and the surrounding matrix are millipedes, and spiders (e.g., Hasiotis, 2002; Hembree, 2009; Mikuś
common in all QH and RR paleosols (Fig. 7C–D). The tunnels are and Uchman, 2013; Bowen and Hembree, 2014; Hils and Hembree,
circular to concave ellipsoidal in cross-section with observed lengths of 2015).
2.2–5.4 cm and diameters of 0.2–0.8 cm. The burrows are filled with
massive fine- to medium-grained sandstone that is coarser and 4.2.2.4. Taenidium. Unlined, unbranched, subvertical to vertical shafts
commonly lighter in color than the surrounding matrix. These with meniscate fill are found in both the QH and RR paleosols but are
burrows are interpreted as deposit feeding, locomotion, and more common throughout the QH section (Fig. 8). The burrows are
temporary dwelling structures of small soil arthropods (e.g., Hasiotis, circular in cross-section and 0.4–1.4 cm wide; observed lengths are
2002; Hembree, 2009; Bowen and Hembree, 2014). 1.5–3.5 cm, but burrow openings are not preserved. The shafts are filled
with medium- to dark brown claystone or very fine siltstone that is
4.2.2.3. Macanopsis. Unlined, 2.1–3.2 cm long, vertical, 0.3–0.5 cm distinct from the surrounding matrix. A meniscate structure is present
diameter shafts with sharp walls ending in terminal chambers of in several of the burrow fills; the sediment packages are well preserved
0.5–1.7 cm diameter are rare in QH and RR paleosols (Fig. 7E–F). and 0.2–0.3 cm in length. The burrows have been interpreted as
The shafts and chambers are filled with a massive, fine- to medium- deposit-feeding structures of small soil arthropods such as larval and

387
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

Fig. 7. Passively filled burrows, highlighted in B–F with dashed lines placed slightly beyond their actual boundaries. A) Thin section with cross section of a vertically-oriented granotubule
assigned to Skolithos (under polarized light, from RRS1 T1P P1); B) Cross-sections of passively filled burrows assigned to Skolithos (from QHS1 T1P P13; C) Lateral view of an elongate,
passively filled tunnel assigned to Planolites (From QHS1 T1P P13); D) Lateral view of a horizontal, passively filled tunnel assigned to Planolites (From QHS1 T2P P1); E) Lateral view of a
passively-filled, vertically-oriented tunnel with a terminal chamber assigned to Macanopsis (From QHS1 T1P P3); F) Lateral view of a passively-filled burrow with a terminal chamber
assigned to Macanopsis (From RRS1 T3P P6).

adult beetles (e.g., D'Alessandro and Bromley, 1987; Keighley and

Pickerill, 1994; Savrda et al., 2000).

4.2.2.5. Celliforma. Flask- to capsule-shaped cell molds and casts are

common in QH paleosols and abundant in RR paleosols (Fig. 9A–C).
The cell casts and molds are 0.5–1.0 cm in diameter, with exposed cast
lengths ranging from 0.7–2.1 cm. The length to width ratios of the cells
range from 1:2.0–2.6. The cells are smooth-walled; in cross-section the
cells are lined by a thin, 0.05–0.1 cm layer of mudstone. The fill of the
cells is commonly a tan to light brown, fine siltstone that is similar in
color and texture to the surrounding matrix; in RR paleosols, the fill of
several cells is composed of calcium carbonate. The cells are primarily
vertical, but sub-vertical and horizontal cells are also present in both
QH and RR paleosols. Some cells are isolated (Fig. 9D), but the cell casts
and molds are more commonly found in clusters or rows (Fig. 9C). The
cells are interpreted as the internal casts of terminal chambers of brood
burrows produced by ground-nesting bees (Hymenoptera) (Brown,
1934, 1935; Retallack, 1984; Cane, 1991; Genise, 2000, 2016;
Hasiotis, 2002).

4.2.2.6. Fictovichnus Form 1. Elongate, smooth-walled, ovoid cells are

rare in QH and RR paleosols and typically occur in clusters of two to
three (Fig. 9E), but some cells are isolated (Fig. 9F). The cells are
0.8–1.0 cm in diameter and 1.5–1.0 cm in length, with length to width
ratios of approximately 1:1.8. In cross-section, the cells have thin,
0.05 cm thick, brown mudstone linings and are filled with a fine, tan- to
brown siltstone that is similar in color and texture to the surrounding
matrix. One sample from a QH paleosol includes two cross-cutting cells
extending from 0.3–0.5 cm vertical- to sub-vertical shafts (Fig. 9E). The
shafts have an exposed length of 1.4–1.6 cm. The cells are interpreted as
casts of insect cocoons, potentially produced by solitary wasps (e.g.,
Rosenheim, 1987; Genise et al., 2000; Hasiotis, 2002; Genise and
Cladera, 2004; Genise, 2016).
Fig. 8. Backfilled burrows. A) Lateral view of a subvertical shaft with a meniscate fill
(sediment packages separated by dashed lines) assigned to Taenidium (from RRS1 T2P
4.2.2.7. Fictovichnus Form 2. A single, egg-shaped, mudstone-covered
P1); B) Oblique view of a vertical shaft with a meniscate fill (outlined) assigned to
Taenidium (from RRS1 T1P P2). cast is present in one QH paleosol (Fig. 9G). The cast is 3.0 cm in length
and 2.0 cm in diameter with a length to width ratio of 1:1.5. The cast is
preserved in 3D relief and, as such, the inner morphology of the

388
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

Fig. 9. Actively filled brood structures, highlighted with outlines in A, C, E–F. A–D Celliforma: A) Lateral view of two subvertically-oriented capsule-shaped Celliforma (from QHS1 T1P
P2); B) Lateral view of three clustered, vertically-oriented, capsule-shaped Celliforma (at arrows) (from QHS2 T3P P1); C) Cross-sections of molds of six Celliforma arranged in two rows of
three (from QHS1 T1P P10); D) Isolated, vertically-oriented, smooth-walled Celliforma (from RRS1 P11). E–G Fictovichnus: E) Lateral view of two clay-lined, cross cutting, sub-vertically
oriented cocoons assigned to Form 1 (from QHS1 T1P P3); F) Oblique view of an isolated, vertically-oriented cocoon assigned to Form 1 (from RRS1 T3P P1). G) Lateral view of clay-lined,
egg-shaped chamber assigned to Form 2 (from QHS1 T2P P8). H–I Coprinisphaera: H) Top view of a round, mudstone-filled chamber (from QHS1 T2P P9); I) Thin section of round,
mudstone-filled chamber; area to left of dashed line is the mudstone-filled center of the chamber (under polarized light).

structure is not discernable. The exterior of the cast is finer grained than
the surrounding matrix. The cast is interpreted as a larval beetle cocoon
(Johnston et al., 1996).

4.2.2.8. Coprinisphaera. Isolated, circular to ellipsoid balls are found in

situ in QH paleosols and in float below RR paleosols (Fig. 9H–I). The
balls range in width and height from 2.5–3.5 cm and 1.8–2.7 cm,
respectively, with width to height ratios ranging from 1:1.3–1:1.4.
The balls have a smooth outer lining composed of brown- to reddish-
brown claystone that is 0.05–0.07 cm thick. The fill of the balls is a tan-
to-brown mudstone (Fig. 9I) that is finer grained than the surrounding
matrix. The balls are interpreted as brood chambers of dung beetles
(Scarabaeoidea) that were formerly filled by dung (e.g., Sauer, 1955;
Retallack, 1984; Cambefort and Hanski, 1991; Hasiotis, 2002; Philips,
2011; Simmons and Ridsdill-Smith, 2011; Genise, 2016).
Large diameter tunnels and chambers (cf. Katarrhedrites): Several
Fig. 10. Specimen assigned to Katarrhedrites (from QHS2 T3P P5). Superior view of
elongate, horizontal tunnel fragments and a single expanded chamber Katarrhedrites (center image), and the left and right edges (left and right images, re-
are present in one QH paleosol. The tunnels are straight with walls that spectively). Arrows indicate the top side of the chamber.
lack distinct bioglyphs and have exposed lengths of up to 6.5 cm. The
tunnels are elliptical in cross-section with widths ranging from
are assigned to cf. Katarrhedrites (Hembree and Hasiotis, 2008). Given
3.0–4.0 cm, heights ranging from 3.0–3.5 cm, and length to width ratios
the simple nature of the tunnels and chambers, possible tracemakers
of 1:1.2–1.3. The expanded chamber (Fig. 10) has a total exposed
include small mammals known from body fossils at the site (the octo-
length of 10.1 cm and irregular widths and heights ranging from
dontoid rodents Acarechimys and Quebradahondomys, the small pa-
2.9–6.6 cm and 2.7–3.2 cm, respectively. The bottom of the chamber is
leothentid marsupials Chimeralestes and Palaeothentes), another
flat, whereas the top is convex. The chambers and associated tunnels

389
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

cutting relationships are also observed. These relationships, combined

with grain size differences relative to the surrounding matrix, indicate
that the mottles result from organism-substrate interactions rather than
abiotic factors. Irregular mottles can be produced by sediment filling in
open shafts and by intense mixing of sediment layers of differing grain
sizes through bioturbation (Gingras et al., 1999). These structures were
likely produced during locomotion and deposit-feeding behavior, but it
is not possible to identify the trace maker(s) due to their irregular
morphology.

4.3. Quebrada Honda and Río Rosario Paleosols

Paleosols were identified and described within the three sampled

sections. Three pedotypes were distinguished based on macro- and
micromorphological pedogenic properties, ichnofossil assemblages, and
bulk geochemistry. These are Type 1 and Type 2 paleosols in QHS1 and
RRS1 and Type 3 paleosols in QHS2 and RRS1.

4.3.1. Type 1 paleosol (T1P)

T1Ps are present in QHS1 (n = 15) and RRS1 (n = 3) (Figs. 3, 5).
T1Ps typically occur as multiple (2–4), 15–118 cm thick, stacked,
overlapping paleosol profiles or rarely as individual profiles (P12 in
QHS1, P3 in RRS1) (Figs. 3, 5). Individual T1P profiles are character-
ized by poorly- to moderately-defined paleosol horizons; from the top
down, the horizons are typically composed of brown (7.5YR 4/3) to
light yellowish brown (7.5YR 6/3) massive silty mudstones, thinly la-
minated (0.1 cm-scale) sandy siltstones, and massive, medium-grained
sandstones. Conglomerates are also found in the lower portions of some
T1P profiles. Mudstones and siltstones have angular blocky to sub-
angular blocky structures. Contacts between the different lithologies
are typically gradational. The colors of the sandstones, siltstones and
mudstones are generally strongest (solid brown) at their base and be-
come more variegated (light yellowish brown) toward the top. In-
dividual T1Ps commonly contain iron nodules, fine-to medium- pebble-
sized lithic clasts, and reddish brown (2.5YR 4/4) claystone intraclasts
averaging 0.5 cm in height and 1.0–1.5 cm in width dispersed
throughout the profiles. Dark brown argillans (7.5 YR 3/4) and slick-
ensides averaging 3.5 cm in length and 2.2 cm in width are rare and
occur near the top of the profiles. Small 0.7–1.2 cm diameter carbonate
nodules are also rare and occur in the middle of the profiles.
The ichnofossil assemblage associated with the sandy siltstones and
the silty mudstones of T1Ps is dominated by common light brown and
rare pale green irregular mottles as well as primarily vertical, pale
green, greenish grey and yellowish red rhizohaloes with two size dis-
tributions relative to length (0.7–1.1 cm, average = 0.8 cm; 1.2–5.0,
average = 3.6 cm) as well as branching rhizotubules (average
length = 0.3 cm). The rhizotubules are larger, up to 1.1 cm in length,
Fig. 11. Irregular mottles. A) Oblique view of brown mottles with sharp boundaries and more abundant in the RRS1 T1Ps than in those of QHS1. Within a
(indicated by dashed line) within a lighter brown matrix (from QHS1 P13); B) Lateral single T1P profile, Planolites, Skolithos, and Celliforma are common and
view of tan mottles with sharp boundaries within a brown matrix (from QHS1 T1P P15); often found together. Macanopsis and Taenidium are less common;
C) Oblique view of light brown mottles with diffuse boundaries within a brown matrix Macanopsis is found in P3 and P12 of QHS1, and Taenidium is found in
(from RRS1 T3P P9). (For interpretation of the references to color in this figure legend,
P3 of QHS1 and P1 of RRS1. Coprinisphaera is present in P1 and P5 of
the reader is referred to the web version of this article.)
QHS1. Fictovichnus Form 1 is also rare, only found within P3 of QHS1.
The micromorphology of T1Ps (Fig. 12A–D) is characterized by an
vertebrate not represented by osteological remains (e.g., small lizard), intertextic to agglomeroplasmic fabric with some rarer areas of por-
or by a large invertebrate such as a spider, scorpion, or crayfish phyroskelic fabric. The matrix microfabric is primarily silasepic to in-
(Hembree et al., 2012; Hembree, 2014; Hils and Hembree, 2015; sepic; samples with argillasepic and mosepic microfabrics are present
Hasiotis and Mitchell, 2008). but rare. The framework grains are primarily quartz, plagioclase feld-
spar, and biotite that generally have thin weathering rinds. Subangular
4.2.2.9. Mottles. All QH and RR paleosols contain common, sharp- blocky peds are distinct at a horizon that occurs 10 cm below the up-
walled, irregular mottles (Fig. 11). Mottle fill is unstructured and permost surface of P2 in RRS1 (Fig. 13A). Small pedogenic iron nodules
composed of tan-to-brown, fine- to medium-grained sandstone that is are also common within the samples. Biological features visible in thin
coarser grained and typically lighter in color than the surrounding section include circular to irregular areas of coarse fill, elongate rhi-
matrix. The mottles have irregularly exposed widths but are typically zotubules, and granotubules (small pedotubules filled with clastic
elliptical and 0.8–3.3 cm in cross-section. The mottles are dispersed grains; following Brewer, 1976).
throughout the paleosols but commonly occur in distinct zones; cross- Sixteen T1P samples from QHS1 (n = 10) and RRS1 (n = 6) were

390
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

Fig. 12. Microfabrics (under polarized light). A–D) Matrix microfabrics of the Type 1 paleosols: A) Agglomeroplasmic grain fabric within a matrix of silasepic plasmic microfabric; B)
Agglomeroplasmic grain fabric within an insepic plasmic microfabric, localized areas of high birefringence clay (at arrows); C) Intertextic grain fabric within a matrix of silasepic plasma;
D) Localized areas of granular grain fabric (at arrows); grain fabric is primarily agglomeroplasmic within an insepic plasmic microfabric. E – H) Matrix microfabrics of Type 2 paleosols: E)
Agglomeroplasmic grain fabric within an insepic plasmic microfabric with patchy, localized areas of high birefringence clay (at arrows); F) Localized area of insepic clay surrounded by a
granular grain fabric; G) Intertextic grain fabric within a matrix of insepic plasmic microfabric; H) Argillasepic microfabric of a laminated clay clast. I – L) Matrix microfabrics of Type 3
paleosols: I) A granotubule (at arrow) within an intertextic grain fabric and insepic plasmic microfabric; J) Localized area with a clinobimasepic microfabric, highly birefrigent streaks (at
arrows); K) A rhizotubule (at arrow) surrounded by small iron nodules within a porphyroskelic grain fabric and an insepic plasmic microfabric; L) An isolated pellet (at arrow) within a
intertextic grain fabric and an insepic plasmic microfabric.

analyzed for bulk geochemistry (Supplementary Tables 2, 3). Within profiles are composed of solid brown (7.5YR 4/3) to dark brown (7.5YR
the QHS1 profiles: 1) oxidation and lessivage values are low and show 3/3) silty mudstones that locally contain primary sedimentary struc-
little variability; 2) salinization is low to moderate; 3) base loss is tures in the form of 0.1 cm-scale planar laminations and massive, silty
moderate in P15 but otherwise high; 4) calcification is low in P10 but sandstones that commonly contain medium-to coarse pebble-sized
otherwise moderate; 5) leaching is high throughout; and 6) CIA-K is lithics and rarely contain primary sedimentary structures. The T2P
moderate in P2, P4, and P15 but otherwise high. Within the RRS1 mudstones and sandstones have angular to subangular blocky struc-
profile: 1) oxidation and lessivage values are low and show little tures, and the contacts between them are sharp. Small iron nodules and
variability; 2) salinization is low to moderate; 3) base loss is moderate claystone intraclasts that average 0.7 cm in thickness and 1.5–2.0 cm in
but high in P1; 4) calcification is moderate to high with a peak in P2; 5) width commonly occur within the upper 25 cm of the profiles. Slick-
leaching is high; and 6) CIA-K is moderate. ensides averaging 2.0 cm in length and 2.5 cm in width (Fig. 13B) as
well as dark brown (7.5 YR 3/3) argillans are rare and found in the
4.3.2. Type 2 paleosol (T2P) upper 20 cm of the profiles (Fig. 13C).
T2Ps are present in QHS1 (n = 9) and RRS1 (n = 1) (Figs. 3, 5). The The typical ichnofossil assemblage associated with the silty mud-
T2Ps are 35–90 cm thick and occur either as individual profiles or stones of the T2Ps includes vertical, downward branching rhizotubules
multiple (2–4), stacked profiles. Individual T2Ps are characterized by (average length = 0.3 cm), irregular light brown mottles, and
poorly-defined paleosol horizons. From the top down, individual T2P Celliforma. Pale green rhizohaloes, Skolithos, Planolites, Taenidium, and

391
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

Coprinisphaera are less common and found in the T2Ps of QHS1;

Fictovichnus Form 2 is rare and found only in P8 of QHS1 T2Ps.
The micromorphology of the T2Ps (Fig. 12E–H) is characterized by
intertextic to agglomeroplasmic fabrics with some rarer areas of gran-
ular fabric. The matrix microfabric is primarily silasepic; insepic mi-
crofabric is present, but rare. The framework grains are primarily
quartz, plagioclase feldspar and biotite that rarely have weathering
rinds. Small pedogenic iron nodules are common. Biogenic features
visible in thin section include rhizotubules and irregular mottles.
Eight samples of T2Ps from QHS1 (n = 7) and RRS1 (n = 1) were
analyzed for bulk geochemistry (Supplementary Tables 2, 3). Within
the profiles: 1) oxidation and lessivage values are low and show little
variability; 2) calcification is low to moderate; 3) salinization is mod-
erate; 4) base loss and leaching are high; 5) CIA-K is moderate in P2 and
P7 of QHS1 but otherwise high.

4.3.3. Type 3 paleosol (T3P)

T3Ps are present in QHS2 (n = 5) and RRS1 (n = 15) (Fig. 12I–L).
The T3Ps occur as multiple (3–6), 50–172 cm, stacked, overlapping
paleosol profiles. From the top down, individual T3Ps are characterized
by moderately- to well-defined paleosol horizons composed of solid to
variegated reddish brown (5YR 4/4), brown (7.5YR 4/3), to yellowish
brown (10YR 5/4) massive silty mudstones, sandy siltstones that rarely
contain primary sedimentary structures in the form of millimeter-scale
planar laminations, and medium-grained massive sandstones. The
mudstones, siltstones, and sandstones have an angular blocky to sub-
angular blocky structures, and the contacts between them are grada-
tional. Within a profile, colors are generally strongest toward the base
and more variegated toward the top. Carbonate nodules ranging in
diameter from 0.5–3.5 cm are abundant and dispersed but more
common in the middle of the profiles. Iron nodules and medium- to
coarse-grained pebble-sized lithic clasts are commonly dispersed
throughout T3P profiles; claystone intraclasts averaging 0.5 cm in
height and 1.0–1.5 cm in width are less common but also dispersed
throughout the profiles. Argillans are rare and found within the upper
30 cm of the profiles within RRS1; 0.1 cm-scale planar laminations are
also rare and only found within the lower 15 cm of the T3P profiles in
QHS2.
The typical ichnofossil assemblage associated with the sandy silt-
stone and silty mudstone of the T3Ps is characterized by an abundance
of high density vertical and horizontal branching rhizotubules (average
length = 1.5 cm) and common, but less abundant, tan root casts
(average length = 3.2 cm) with calcareous or silty centers, calcareous
rhizoconcretions (average length = 6.1 cm; average width = 2.0 cm),
Celliforma, and irregular mottles. Planolites and Skolithos are common
and often found together. Macanopsis occurs within both the QHS2 and
RRS1 sections, but is rare. Coprinisphaera and Katarrhedrites only occur
in P5 of QHS2, and Fictovichnus Form 1 only occurs in P1 of RRS1.
The micromorphology of the T3Ps is characterized by agglomer-
oplasmic fabric with some intertextic fabric and rarer areas of granular
and porphyroskelic fabric. The matrix microfabric is primarily insepic
with some silasepic in both the QHS2 and RRS1 sections; rare mosepic
and clinobimasepic microfabrics are present in P7 of RRS1. The fra-
mework grains are primarily quartz, plagioclase feldspar, and biotite
that generally have thin weathering rinds. Small pedogenic iron no-
dules and carbonate cement are common in the T3Ps of both sections. A
single distinct horizon of rounded blocky peds is present in 10 cm below
the surface of P11 of RRS1 (Fig. 13D). Biological features visible in thin
Fig. 13. Pedogenic features. A) Subangular blocky peds from a Type 1 paleosol (under
polarized light, RRS1 T1P P2); B) Dark brown clay argillans (below dashed line; from section include circular to irregular areas of coarse-fill, elongate rhi-
QHS1 T2P P2); C) Small slickensides (at arrows; from QHS1 T2P P3); D) Rounded blocky zotubules, root casts, granotubules, and pellets.
peds from a Type 3 paleosol, borders highlighted by dashes (under polarized light, RRS1 Thirty samples of T3Ps from QHS2 (n = 4) and RRS1 (n = 26) were
T3P P11). (For interpretation of the references to color in this figure legend, the reader is analyzed for bulk geochemistry (Supplementary Tables 2, 3). Within
referred to the web version of this article.) the QHS2 profiles: 1) oxidation and lessivage are low; 2) salinization is
low to moderate; 3) calcification is moderate; 4) base loss is high in P1
but otherwise moderate; 5) leaching is high throughout; and 6) CIA-K is
moderate and shows little variability. Within the RRS1 profiles: 1)

392
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

oxidation and lessivage are low and show little variability; 2) salini- paleosol profile is, therefore, separated by a layer of non-pedogenically
zation is low to moderate; 3) calcification is moderate to high; 4) base modified sediment. Compound paleosol profiles commonly form in
loss is low in P1 but otherwise moderate to high; 5) leaching is high areas proximal to active channels, such as natural levees, where sedi-
throughout; and 6) CIA-K is low in P1 and P6 but otherwise moderate to mentation events are frequent and rapid, and erosion is minimal (Kraus,
high. 1999).
The T2Ps are classified as ferric Protosols based on poorly defined
5. Paleoenvironmental interpretation soil horizons and ferric nodules (Mack et al., 1993) and interpreted as
Entisols, specifically Udifluvents (Soil Survey Staff, 2010; Buol et al.,
The paleosols of Quebrada Honda record paleoenvironmental, pa- 2011). Entisols are immature, poorly developed soils that have little or
leoecological, and paleoclimatic data that allow for the reconstruction no development of pedogenic horizons (Soil Survey Staff, 2010; Buol
and comparison of terrestrial environments within the ancient et al., 2011). Udifluvents are fine-grained Entisols with very simple,
Neotropics. The studied sections of the QH and RR local areas represent commonly stratified profiles (Soil Survey Staff, 2010; Buol et al., 2011).
soils forming within dynamic alluvial systems and are useful for in- The micromorphological features of the T1Ps and T2Ps indicate
terpreting paleoenvironmental changes that are a result of changes in fluctuations in moisture conditions. The patches of high birefringence
sedimentation rates, parent material, topography, and climate. clay in thin section found within the T1P and T2P profiles were de-
posited by the downward movement of water and clay particles through
5.1. T1P and T2P the soil profile and also indicate periods of improved drainage
(Retallack, 2001); the same is true of pedogenic iron nodules, which are
Due to their similarities, T1Ps and T2Ps are discussed together. The precipitated during redox cycles (Retallack, 2001; Stiles et al., 2001;
physical structures, ichnology, and geochemistry of the T1Ps allow for Kraus and Hasiotis, 2006). Flooding events are indicated by clay clasts
the recognition of 19 composite paleosols composed of A, B, B/C and C and siliciclastic lithics distributed throughout the T1P and T2P profiles,
horizons. The A horizons are composed of heavily bioturbated silty and it is likely that the flooding events had an influence on the fluc-
mudstones that frequently include zones of Celliforma. The B horizons tuating soil moisture regimes. However, the distribution of high bi-
are composed of massive silty mudstones; B/C horizons are composed refringence clay within the soil profiles and the overall lack of surface
of laminated sandy siltstones and fine-grained sandstones that contain gley features indicates that the influence of flooding events on fluc-
mottles, clay clasts, and lithics, and the C horizons are composed of tuations in soil moisture regimes was likely minor and the fluctuations
massive, sandy siltstones, sandstones, and conglomerates. in soil moisture were instead largely a result of long term saturation of
The stacked T1P profiles overlap and are separated either by layers the soil due to seasonal variations in moisture.
of less pedogenically modified sediment or by layers with higher con- Taken together, the T1Ps and T2Ps represent soil formation on
centrations of lithics and clay clasts that partially extend into the upper floodplains in a dynamic alluvial environment. The T2Ps likely formed
surfaces of the underlying profiles. These layers form gradational zones proximal to active channels. The T2Ps are thin and have weakly de-
between the profiles and represent intervals of high sediment deposi- veloped profiles; this, combined with the compound stacking pattern, is
tion, truncations of active paleosol profiles, and pauses in pedogenesis. typical of soils that form adjacent to channels on natural levees or
Composite paleosols form when the rate of pedogenesis is not steady crevasse splay deposits (Kraus, 1999). Entisols with a compound
because it is interrupted by episodes of high deposition (Kraus, 1999). stacking pattern indicate short periods of pedogenic development and
The depositional events bury the active soil surface and, over time, the temporally unstable landscapes with rapid, short term, re-occurring
development of the new overlying soil extends into, and partially flooding and sedimentation events (Birkeland, 1999; Kraus, 1999; Buol
overprints, the buried paleosol. This results in vertically successive and et al., 2011). These flooding events and the associated temporal land-
partially overlapping profiles that have a combination of pedogenic scape instability results in the co-occurrence of primary sedimentary
features of both the overlying and underlying soils (Kraus, 1999; Mason structures and bioturbated zones (Birkeland, 1999; Retallack, 2001;
and Jacobs, 2007). Buol et al., 2011). Given their immaturity, thickness, and stacking
Due to the presence of poorly to moderately defined soil horizons, pattern, the time of formation of the T2Ps was likely on the order of
moderate to high calcification ratios, and ferric nodules, the T1Ps are 101–102 years (Birkeland, 1999; Buol et al., 2011).
classified as ferric Protosols (Mack et al., 1993) and interpreted as In- The T1Ps likely formed in a more distal area of the alluvial system,
ceptisols, specifically Eutrudepts (Soil Survey Staff, 2010; Buol et al., such as the distal floodplain. The T1Ps have overlapping, composite
2011). Inceptisols are immature soils that lack many diagnostic features profiles that are generally thicker and better developed than the T2Ps,
but are characterized by weak development and poorly to moderately which indicates lower sedimentation rates, higher rates of pedogenesis,
developed soil horizons (Mack et al., 1993; Soil Survey Staff, 2010; Buol and greater temporal stability due to longer intervals between deposi-
et al., 2011). Eutrudepts are Inceptisols that have free carbonate within tional events (Kraus, 1999). In alluvial settings, Inceptisols generally
the soil and may also have a high base status (Soil Survey Staff, 2010; form more distal to active channels than Entisols but are still affected by
Buol et al., 2011). The T1Ps do possess some features of more mature intermittent flooding and sedimentation events (Birkeland, 1999;
soil orders such as Vertisols and Alfisols, including slickensides and Collinson, 1996). Given their thickness, composite stacking pattern, and
argillans (Mack et al., 1993; Retallack, 2001), but these features are level of development, the time of formation for the T1Ps was likely on
rare and localized. T1Ps lack other diagnostic features of Vertisols in- the order of 102–103 years (Birkeland, 1999; Buol et al., 2011). The
cluding wedge-shaped peds and clay-rich horizons (Mack et al., 1993; increased level of development of the T1Ps compared to the T2Ps is
Retallack, 2001). attributed to decreased sedimentation and increased temporal stability
The sandy siltstones and silty mudstones of the T2Ps contain due to a more distal position with respect to active channels and does
properties of both C and A horizons, including massive beds, primary not likely reflect differences in climate or biota between the T1Ps and
sedimentary structures, zones of intense bioturbation, and zones of T2Ps.
brood chambers and, therefore, represent A/C transitional horizons. In The ichnofossils preserved in the T1Ps and T2Ps suggest a savanna,
the stacked T2P profiles, the features of the overlying paleosol do not or relatively open wooded grassland-like vegetative community in a
extend into the upper surfaces of the underlying paleosols, but are in- seasonal, subhumid-to-humid climate. The size distribution and two
stead separated by a sharp contact. Compound paleosols form when types of preservation (rhizohaloes and rhizotubules) of the root traces
rapid but short-term sedimentation events bury the active soil surface present in the T1Ps and T2Ps is typical of a modern wooded grassland
(Kraus, 1999); however, unlike composite profiles, compound profiles where fine, mm-scale roots are found alongside larger roots (Cole,
are buried beyond the reach of later pedogenesis (Kraus, 1999). Each 1986). The presence of both grey and red rhizohaloes and calcareous

393
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

rhizotubules suggests fluctuating moisture conditions. Grey rhizohaloes reliability of MAP estimates for the T2Ps is less certain; however, the
commonly form when a soil is water saturated and relatively depleted MAP estimates of T2Ps in both QHS1 and RRS1 are also moderate,
of oxygen, as iron is reduced adjacent to roots and other organic matter comparatively constant through the section, and average approximately
and then translocated outwards to form drab-colored depletion zones 1050 mm/year.
(Kraus and Hasiotis, 2006). The red color of some of the rhizohaloes The habitats in which the T1Ps and T2Ps formed may have re-
indicates the presence of hematite and implies oxidized and moder- sembled portions of the modern Caatinga ecoregion of northeastern
ately-well drained conditions (Kraus and Hasiotis, 2006). Calcareous Brazil with respect to MAP, seasonality, and elevation. The Caatinga is
rhizotubules are typically found in soils that undergo periodic drying generally humid to semi-arid with bi- to trimodal rainfall of approxi-
(Kraus and Hasiotis, 2006); open channels created by decayed roots are mately 500–1300 mm/year (Liebmann and Allured, 2005; Lavin, 2006;
filled with calcite precipitated from evaporating ground water (Buol Alvares et al., 2013), and its elevation ranges from < 100 m to ap-
et al., 2011). proximately 1200 m (Alvares et al., 2013). The vegetational structure
Roots and burrows that are left open to the surface (i.e., Skolithos, the Caatinga ecoregion is variable and includes areas of open and closed
Planolites, Macanopsis) are constrained to the oxygenated upper vadose shrublands, dry forests, savannas, and semi-evergreen forests (Eva
to upper phrenic zones of the soil and also indicate moderately- to well- et al., 2002), but ground-covering shrubs and small trees predominate
drained soils (Hasiotis, 2002, 2007). Backfilled, meniscate burrows (Lavin, 2006).
(i.e., Taenidium) that are not left open to the surface are generally found
in the A and upper B horizons of moist, unsaturated soils (Villani et al., 5.2. T3P
1999; Hasiotis, 2002, 2007). The burrows and brood chambers of the
T1Ps and T2Ps further indicate moderately well-drained soils. Dung The T3Ps represent composite paleosol profiles composed of A, B,
beetles are abundant in environments with seasonal changes in Bk, B/C and C horizons. The A horizons are composed of heavily bio-
moisture and generally construct their burrows during the dry season in turbated silty mudstones with zones of brood chambers. The B horizons
A or upper B horizons of well-drained soils (Cambefort and Hanski, are composed of massive silty mudstones, whereas the Bk horizons are
1991). Wasp cocoons, beetle cocoons, and ground-dwelling bee nests composed of silty mudstones with accumulations of carbonate nodules.
are also commonly found in A and upper B horizons of well-drained, B/C horizons are composed of siltstones and sandy siltstones that
cohesive soils in alluvial environments (Retallack, 1984; Cane, 1991; contain mottles, clay clasts, lithics, and preserve some primary sedi-
Hasiotis, 2002). mentary structures in the form of planar laminations, and the C hor-
The prevalence of rhizoliths and brood structures produced by izons are composed of massive sandstones and sandy siltstones.
beetles, bees, and wasps are also indicative of an environment with high Due to the presence of moderately- to well-defined soil horizons,
primary productivity. Brood chambers such as Coprinisphaera, carbonate horizons, and ferric nodules, T3Ps are classified as Calcisols
Celliforma, and Fictovichnus Forms 1 and 2 are packed with pollen, and interpreted as well-developed calcic, ferric Inceptisols, or
nectar, plant material, or mammalian herbivore dung (Retallack, 1984; Calciustepts (Mack et al., 1993; Soil Survey Staff, 2010; Buol et al.,
Cane, 1991), which adds organic matter to the soil. Deposit-feeding soil 2011). Calcisols are soils in which calcic horizons are prominent
arthropods cycle nutrients through soils through the processes of pedogenic features (Mack et al., 1993), and Calciustepts are Inceptisols
feeding and excretion (Hasiotis, 2002, 2007); as such, the presence of with calcic horizons (Soil Survey Staff, 2010; Buol et al., 2011).
deposit feeding structures such as Taenidium and Planolites also in- The mm- and cm- scale rhizoliths represent the roots of small- to
dicates soils with high organic content (Hasiotis, 2002, 2007). medium-sized plants; the abundant mm-scale root traces combined with
The geochemical features of the T1Ps and T2Ps also indicate sea- the less common cm-scale root traces are typical of relatively open but
sonal fluctuations in moisture conditions during soil development. densely vegetated wooded grassland-like soils (Retallack, 2001). The
During periods of low soil moisture, when soil profiles are at least calcareous rhizotubules and rhizoconcretions indicate drier conditions,
moderately-well drained, base cations and soluble compounds are re- as these types of rhizoliths commonly form in soils that undergo peri-
moved and translocated from weatherable minerals, which results in odic drying (Retallack, 2001; Kraus and Hasiotis, 2006; Buol et al.,
moderate-to-high values for base loss and leaching (Retallack, 2001; 2011). The clay linings of the root casts show illuviation structures in
Schaetzl and Anderson, 2009; Sheldon and Tabor, 2009). During per- thin section and also indicate well-drained soil conditions. Roots respire
iods of high soil moisture, when soil profiles are poorly drained, base aerobically and do not penetrate far below the water table; thus, the
cations are not transported through the soil profile; instead, iron is presence of both vertically and horizontally oriented rhizotubules in-
reduced and mobilized, resulting in low values for lessivage and oxi- dicates periodically high water tables (Cohen, 1982; Retallack, 2001).
dation, respectively (Retallack, 2001; Schaetzl and Anderson, 2009). Katarrhedrites, Fictovichnus, passively filled burrows, and Celliforma also
While low lessivage values can also indicate dry conditions (Retallack, indicate well-drained, stable, cohesive soils that supported several an-
2001; Schaetzl and Anderson, 2009), the patchy distribution of high nual breeding seasons (Hasiotis, 2007). High organic content in the
birefringence clay, the ferric nodules in thin section, and the MAP es- upper soil horizons is also indicated by the abundance of rhizoliths and
timates indicate low lessivage values due to periods of high soil brood chambers (Hasiotis, 2002, 2007). Overall the ichnofossils and the
moisture to be more likely. During cycles of wetting and drying, both T3Ps represent an organic rich, well-drained soil ecosystem in a rela-
salts and carbonates precipitate above and within the zone of water tively open environment that experienced strong moisture seasonality
table fluctuation, resulting in overall moderate values for salinization and regular drying.
and moderate to high values for calcification (Retallack, 2001; Schaetzl The pedogenic, geochemical, and micromorphological features of
and Anderson, 2009). the T3Ps are indicative of pronounced fluctuations in soil moisture.
The CIA-K values for the T1Ps indicate mostly moderate MAP that Argillans, the patches of high birefringence clay and the illuviated clay
ranged from approximately 650–1100 mm/year. The MAP estimates of of the root casts visible in thin section, as well as the abundance of
the T1Ps in QHS1 remain comparatively constant with a mean of ap- pedogenic carbonate in the subsurface is characteristic of well-drained
proximately 1050 mm/year and with the level of variation falling soils (Retallack, 2001). Pedogenic iron nodules indicate fluctuating
within the range of error for the MAP estimate ( ± 181 mm/yr; Sheldon moisture conditions (Retallack, 2001; Stiles et al., 2001).
and Tabor, 2009). The MAP estimates of the T1Ps in RRS1, however, The low levels of oxidation throughout the T3P profiles as well as
are more variable and have a lower mean value (approximately the low levels of base loss in P1 of RRS1 indicates periods of high soil
830 mm/year). Among the T1Ps of RRS1 there is a general decrease in moisture with poor drainage (Retallack, 2001; Schaetzl and Anderson,
MAP from the bottom of the section to the top (Supplementary Table 3). 2009; Sheldon and Tabor, 2009). The high leaching values, moderate to
Due to the low degree of pedogenic development of Entisols, the high calcification values, and the overall moderate to high base loss

394
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

values indicate drier soil conditions (Retallack, 2001). As with the T1Ps Magdalena Valley (unrelated to the Honda Group of southern Bolivia),
and T2Ps, fluctuations in moisture conditions are likely the result of which comprise an approximately 1000 m thick sequence of volcanic
seasonal variations in precipitation. litharenites, sandstones, mudstones, and coarse-pebble conglomerates
The CIA-K values for the T3Ps in QHS2 indicate moderate and divided between the lower La Victoria and upper Villavieja formations
steady MAP values with a mean of approximately 980 mm/year and (Guerrero, 1997). Sedimentary units of both formations are dominated
variation falling within the range of estimation error. In RRS1, the MAP by sandstones and mudstones interpreted as channel, sheet, levee, and
estimates of the T3Ps are low to moderate and range from approxi- crevasse splay deposits attributed to meandering rivers with depths of
mately 500–1140 mm/year, with a mean of approximately 930 mm/ up to 12.8 m for the La Victoria Formation and 3.0 m for the Villavieja
year. The MAP estimates of T3Ps in RRS1 are more variable than those Formation (Guerrero, 1997). The upper portion of the La Victoria
in QHS1, with nine shifts that are beyond the range of estimation error. Formation also includes conglomerates interpreted to have been de-
The largest inter-profile shift is from 496 mm/yr in P1 to 1081 mm/yr posited by braided streams (Guerrero, 1997).
in P3. These changes in estimated MAP within the T3Ps suggest var- The paleosols of both the La Victoria and Villavieja formations at La
iations in soil moisture on longer time scales and periods of increased Venta are clay-rich and composed of greenish grey and red-to-purple
aridity in RRS1. The shifts in estimated MAP, combined with the overall mudstones that contain drab mottles, rhizohaloes, rhizoconcretions,
abundance of pedogenic carbonate in the subsurface, indicate a sub- slickensides, desiccation cracks, iron oxides, and dispersed 1–3 cm
humid to semi-arid environment with marked seasonally distributed diameter carbonate nodules (Guerrero, 1997). They also contain
precipitation (Retallack, 2001; Schaetzl and Anderson, 2009). Like the abundant wood remains, some of which have been identified as Gou-
T1Ps and T2Ps, the reconstructed habitat of the T3Ps may also resemble pioxylon stutzeri of the staff tree family Celastraceae (Guerrero, 1997).
portions of the Caatinga ecoregion. Less mature paleosols interpreted as calcic Entisols and calcic In-
The T3Ps likely formed on relatively stable landscape surfaces distal ceptisols are common in the La Victoria Formation, whereas more
to active channels with low rates of sedimentation that were subject to mature paleosols interpreted as calcic Vertisols and Oxisols are common
occasional, rapid influxes of sediments. Sedimentation events are in- in the Villavieja Formation. Entisols, Inceptisols, and Vertisols can form
dicated by the lithics and clay intraclasts, and landscape stability is quickly (101–103 years) in environments that range from semi-arid to
indicated by the abundance of roots, brood chambers, and open bur- humid, whereas Oxisols have long formation times ranging from 104 to
rows. The pedogenic carbonates, which are classified as stage II and 106 years and typically form in humid, tropical environments
stage III accumulations (Machette, 1985; Retallack, 2001), also indicate (Retallack, 2001; Schaetzl and Anderson, 2009; Buol et al., 2011). The
landscape stability. Based on moderate level of pedogenesis, lower general differences in paleosol maturity between the two formations
moisture content, and amount of carbonate in the subsurface, the time has been attributed to geologic factors, such as basin subsidence, rather
of formation of the T3Ps is interpreted to be on the order of than climatic factors; the estimated sedimentation rate of the La Vic-
102–103 years (Birkeland, 1999; Retallack, 2001). toria Formation is approximately 95 cm/kyr, while that of the Villavieja
Formation is approximately 41 cm/kyr (Guerrero, 1997). Seasonal
6. Comparison with La Venta, Colombia; faunal provinciality in changes in soil moisture are indicated by carbonate nodules, rhizo-
South America concretions, slickensides, iron oxides, and desiccation cracks which are
produced by the shrinking and swelling of pedogenic clays.
The tropics of Central and South America (Neotropics) contain un- The paleoenvironment of La Venta has been interpreted as ex-
paralleled levels of tetrapod biodiversity, including a wide variety of tensive, moderately-stable to stable, river-associated, sub-humid to
mammals (Wilson and Reeder, 2005). This mammalian diversity is due humid tropical forests based on paleosols and trace fossils (Guerrero,
in part to faunal provinciality (regional variation) that developed in at 1997), and its MAP has been estimated at 1500–2000 mm/year based
least some regions by the early Miocene (Croft et al., 2004) and persists on fossil vertebrates and plants (Kay and Madden, 1997a, 1997b). By
today. Several factors may have contributed to the development of this contrast, we reconstruct Quebrada Honda as a seasonal, humid to semi-
biogeographic pattern. Geographic barriers present in the Miocene, arid, river-associated wooded grassland-like habitat with varying de-
such as marine incursions related to global sea-level rise and tectonic grees of temporal stability. The markedly different habitats of La Venta
loading (Hoorn, 1993, 2006; Lovejoy et al., 1998), as well as Andean and Quebrada Honda during the late middle Miocene, particularly in
uplift and the associated development of extensive wetlands and river regards to landscape stability and soil moisture, indicate that their di-
systems (Hoorn et al., 2010a, 2010b) could have led to divergence of vergent mammal faunas partly or principally reflect distinct biomes and
mammalian communities over relatively short distances by vicariance ecological communities. By extension, we conclude that the develop-
(i.e., disrupting dispersal between regions). Another possibility, not ment of mammalian faunal provinciality in South American was linked
mutually exclusive, is that observed differences in mammalian com- to intracontinental habitat divergence resulting from climatic and tec-
munities mainly reflect sampling distinct paleoenvironmental condi- tonic factors.
tions and habitats in different regions.
Explanations for Miocene mammalian faunal provinciality have 7. Comparison with Cerdas, Bolivia
been difficult to assess due to a scarcity of well-sampled Neotropical
fossil localities. Our study of Quebrada Honda paleosols and ichno- The early middle Miocene (Langhian age) fossil locality of Cerdas,
fossils presents a unique opportunity to test the habitat divergence Bolivia is located approximately 170 km northwest of Quebrada Honda
hypothesis by comparing this locality to La Venta. This latter site, lo- and has also been the subject of detailed paleosol and ichnofossil-based
cated in the Magdalena Valley of central Colombia, is roughly con- paleoenvironmental analysis (Catena et al., 2016). The fossiliferous
temporaneous with Quebrada Honda yet has only 2–3 genera of levels within the Cerdas beds have an extrapolated age spanning
mammals in common (the interatheriid notoungulate Miocochilius, the 16.5–15.3 Ma (MacFadden et al., 1995), coinciding with the latter part
octodontoid rodent Acarechimys, and possibly an astrapothere; Croft, of the middle Miocene Climatic Optimum (MMCO; Zachos et al., 2001)
2007, 2016; Croft et al., 2011; Engelman et al., 2016). In total, La Venta (Croft et al., 2016). The sediments of the Cerdas beds are similar to
preserves the remains of > 80 mammal species in addition to a di- those of Quebrada Honda and have been interpreted as laterally ex-
versity of reptiles, birds and fishes (Kay and Madden, 1997a; Croft, tensive braided alluvial deposits that include both overbank and fluvial
2016). Various geological and geochronological studies have placed La channel deposits (Croft et al., 2009; Auerbach et al., 2015). However,
Venta's paleocommunity in a comprehensive temporal and pa- the channel deposits of Cerdas are thicker, more common and have
leoenvironmental framework (Kay et al., 1997). higher average estimated sediment accumulation rates than those of
The sediments of La Venta are part of the Honda Group in the Quebrada Honda (approximately 50 cm/kyr and 15 cm/kyr,

395
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

respectively) (Auerbach et al., 2015). Together, the sediment accumu- seasonal subhumid to humid, wooded grassland-like vegetative com-
lation rates and thicknesses of the sediments suggest that the fluvial munity in a shifting alluvial environment with varying temporal sta-
systems of Quebrada Honda were smaller than those of Cerdas bility; the T3Ps are reconstructed as a subhumid to semiarid, densely
(Auerbach et al., 2015) and potentially more stable. Cerdas paleosols vegetated wooded grassland-like vegetative community in a distal
include Halusteps and Drystudepts (Inceptisols) as well as Calciustolls floodplain with periods of increased aridity and high degrees of sea-
(Mollisols) that are interpreted to represent soils that formed in sea- sonality and temporal landscape stability.
sonal, humid to semi-arid, patchy to densely vegetated shrublands and Paleosols and ichnofossil assemblages of Quebrada Honda are dis-
open environments in an alluvial setting with estimated MAPs of tinct from those of the well-studied, contemporaneous Neotropical lo-
570–1050 mm (Catena et al., 2016). The slickensides and more abun- cality of La Venta, Colombia, and reflect dissimilar environments.
dant clay argillans, amount of carbonate within the subsurface, and Whereas both La Venta and Quebrada Honda paleosols are derived
larger ranges and variances in the estimated MAP indicates that the from alluvial sediments, the alluvial systems in La Venta were larger
seasonal changes in soil moisture regimes were more pronounced at and had greater sediment accumulation rates than those of Quebrada
Quebrada Honda than Cerdas. Honda. Overall, the paleosols of La Venta are thicker and more mature
Cerdas paleosols resemble those of Quebrada Honda, but the ich- than those of Quebrada Honda and represent pedogenesis in a relatively
nofossil assemblages of the two localities are distinct, with that of temporally stable, forested, humid to sub-humid tropical ecosystem
Cerdas including only rhizohaloes, fine (mm-scale) rhizotubules, with estimated MAP as much as twice that of Quebrada Honda.
Skolithos, Planolites, Macanopsis and Parowanichnus (ant nest) (Catena The paleosols of Cerdas, Bolivia are derived from alluvial sediments
et al., 2016). Cerdas rhizoliths are primarily cm-scale rhizohaloes that and resemble those of Quebrada Honda. However, its ichnofossils are
are either clumped or relatively evenly distributed (Catena et al., 2016), less diverse and abundant than those of Quebrada Honda, and the
whereas those of Quebrada Honda are overall more abundant and ichnofossil assemblages of the two localities are distinct overall. The
densely concentrated. The greater abundance of rhizoliths in Quebrada paleosols of Cerdas reflect lower net primary productivity and temporal
Honda, as well as the presence of cocoons and brood chambers, in- stability than those of Quebrada Honda and represent pedogenesis in
dicates higher soil organic content and net productivity within Queb- areas with larger fluvial systems. The higher primary productivity of
rada Honda. The differences in ichnofossil assemblages could be the Quebrada Honda compared to Cerdas may explain why mesothere no-
result of Cerdas' generally lower primary productivity and less pro- toungulates are curiously rare at Quebrada Honda but abundant at
nounced seasonality. Cerdas and other Miocene sites in the region.
Although fewer fossils have been collected at Cerdas than at This research presents the first rigorous paleoenvironmental re-
Quebrada Honda, with the result that fewer taxa have been identified, construction of the late middle Miocene locality of Quebrada Honda
there are clear differences between the faunas of the two sites (Croft and provides strong evidence that differences in environmental condi-
et al., 2009, 2016). One of the most conspicuous differences is the tions, primarily resulting from differences in MAP, soil moisture re-
abundance of mesotheriid notoungulates, which were medium-sized gimes and vegetational structures, could account for the dissimilarities
(approximately 10–30 kg) herbivores that were likely semi-fossorial or of the soil and vertebrate faunas of the Quebrada Honda and La Venta
fossorial (Croft et al., 2004; Shockey et al., 2007; Croft, 2016). Remains localities. Future investigations should focus on how Miocene geo-
of mesotheres are the most common fossils at Cerdas (Croft et al., 2009; graphic barriers such as Andean uplift and the associated development
Townsend and Croft, 2010) but are rare at Quebrada Honda (Croft and of extensive wetlands and river systems could have contributed to en-
Anaya, 2006; Croft, 2007; personal observation). In this regard, Cerdas vironmental heterogeneity and possibly also to provinciality among
resembles most other Miocene fossil sites in the middle latitudes of Neotropical mammalian faunas through vicariance.
northern Chile (e.g., Chucal and Caragua; Flynn et al., 2002, 2005; Supplementary data to this article can be found online at https://
Croft et al., 2004, 2007; Montoya-Sanhueza et al., 2017) and Bolivia doi.org/10.1016/j.palaeo.2017.09.024.
(e.g., Nazareno, Quehua, Achiri; Villarroel, 1974; Marshall and
Sempéré, 1991; Oiso, 1991). Quebrada Honda is unusual in the ap- Acknowledgments
parent rarity of mesotheres, which almost certainly reflects some aspect
of the sampled paleohabitats. Paleosol and ichnofossil data suggest that We thank two anonymous reviewers for their comments and sug-
main habitat difference between Quebrada Honda and Cerdas is the gestions that greatly improved this paper. We thank Federico Anaya,
greater primary productivity (i.e., vegetation structure/density) of the Alfredo Carlini, Beth Carroll, Martin Ciancio, Camilla Crifo, Al Deino,
former. Thus, Miocene mesotheres may have preferred less productive Luis Gibert and Andrew McGrath for their assistance in the field,
(i.e., less heavily vegetated) habitats. This hypothesis fits with the partnership and support of this research. Thin sections were prepared
general observation that mesothere remains are common to abundant by Texas Petrographic Services, Inc. Bulk geochemical analyses were
at sites throughout Argentina during the late Neogene, an interval performed by ALS Chemex. Funding for this research was made possible
during which forested habitats were replaced by more open habitats in by the National Science Foundation (EAR 0958733, EAR 1423058 to
many parts of the Southern Cone (Pascual and Jaureguizar, 1990; DAC), and the Paleontology Society Kenneth E. & Annie Caster Award
Hynek et al., 2012; Palazzesi and Barreda, 2012). It also explains the (to AMC).
absence of mesotheres at well-sampled Miocene sites where remains of
monkeys and other tropical forest-dwelling mammals have been found References
(e.g., Santa Cruz and Collón-Cura formations of Argentina, and La
Venta, Colombia). Additional paleoecological studies at Quebrada Alvares, C.A., Stape, J.L., Sentelhas, P.C., Gonçalves de Moraes, J.L., Sparovek, G., 2013.
Honda and other geographically proximate sites should permit further Köppen's climate classification map for Brazil. Meteorol. Z. 22, 711–728.
Antoine, P.O., Abello, M.A., Adnet, S., Altamirano Sierra, A.J., Baby, P., Billet, G., Boivin,
testing of this hypothesis. M., Calderón, Y., Candela, A., Chabain, J., Corfu, F., Croft, D.A., Ganerød, M.,
Jaramillo, C., Klaus, S., Marivaux, L., Navarrete, R.E., Orliac, M., Parra, F., Pérez,
8. Conclusions M.E., Pujos, F., Rage, J., Ravel, A., Robinet, C., Roddaz, M., Tejada-Lara, J.V., Vélez-
Juarbe, J., Wesselingh, F.P., Salas-Gismondi, S., 2016a. A 60-million-year Cenozoic
history of western Amazonian ecosystems in Contamana, eastern Peru. Gondwana
Paleosols of Quebrada Honda were described from two local areas, Res. 31, 30–59.
QH and RR, and contain three pedotypes representing distinct pa- Antoine, P.O., Salas-Gismondi, R., Pujos, F., Ganerød, M., Marivaux, L., 2016b. Western
Amazonia as a hotspot of mammalian biodiversity throughout the Cenozoic. J.
leoenvironments within a shifting alluvial system. Pedotypes are dis- Mamm. Evol. 24, 5–17.
tinguished by their maturity, preserved horizons, ichnofossil assem- Auerbach, D.J., Garzione, C.N., Smith, J.J., Quade, J., MacFadden, B., Anaya, F., Croft,
blages and geochemistry. The T1Ps and T2Ps are reconstructed as a D.A., 2015. Miocene Aridification of the Southern Altiplano and Its Relation to the

396
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

Growth of the Central Andes. (manuscript). paleontology. Trends Ecol. Evol. 13, 449–454.
Birkeland, P.W., 1999. Soils and Geomorphology. Oxford University Press, New York, NY, Flynn, J.J., Novacek, M.J., Dodson, H.E., Frasinetti, D., McKenna, C., Norrell, M.A., Sears,
USA. K.E., Swisher, C., Wyss, A.R., 2002. A new fossil mammal assemblage from the
Bowen, J.J., Hembree, D.I., 2014. Neoichnology of two spirobolid millipedes: improving southern Chilean Andes: implications for geology, geochronology, and tectonics. J. S.
the understanding of the burrows of soil detritivores. Palaeontol. Electron. 17, 1–48. Am. Earth Sci. 15, 285–302.
Brandoni, D., Carlini, A.A., Anaya, F., Gans, P., Croft, D.A., 2017. New Remains of Flynn, J.J., Croft, D.A., Charrier, R., Hérail, G., Garcia, M., 2005. New Mesotheriidae
Megathericulus patagonicus Ameghino, 1904 (Xenarthra, Tardigrada) from the (Mammalia, Notoungulata, Typotheria), geochronology and tectonics of the Caragua
Serravallian (middle Miocene) of Bolivia; Chronological and biogeographical im- area, northernmost Chile. J. S. Am. Earth Sci. 19, 55–74.
plications. J. Mamm. Evol. http://dx.doi.org/10.1007/s10914-017-9384-y. Garzione, C.N., Hoke, G.D., Libarkin, J.C., Withers, S., MacFadden, B., Eiler, J., Ghosh, G.,
Brewer, R., 1976. Fabric and Mineral Analysis of Soils, Second ed. Krieger, New York, NY, Mulch, A., 2008. Rise of the Andes. Science 320, 1304–1307.
USA. Garzione, C.N., Auerbach, D.J., Smith, J.J., RosaRío, J.J., Passey, B.H., Jordan, T.E., Eiler,
Brown, R.W., 1934. Celliforma spirifer, the fossil larval chambers of mining bees. J. Wash. J.M., 2014. Clumped isotope evidence for diachronous surface cooling of the
Acad. Sci. 24, 532–539. Altiplano and pulsed surface uplift of the Central Andes. Earth Planet. Sci. Lett. 393,
Brown, R.W., 1935. Further notes on fossil larval chambers of mining bees. J. Wash. Acad. 173–181.
Sci. 25, 526–528. Genise, J.F., 2000. The ichnofamily Celliformidae for Celliforma and allied ichnogenera.
Buol, S.W., Southard, R.J., Graham, R.C., McDaniel, P.A., 2011. Soil genesis and classi- Ichnos 74, 267–282.
fication. John Wiley & Sons, Ames, IA, USA. Genise, J.F., 2016. Ichnoentomology: Insect Traces in Soils and Paleosols Topics in
Cadena, E.A., Anaya, F., Croft, D.A., 2015. Giant fossil tortoise and freshwater chelid Geobiology. 37 Springer International Publishing, Switzerland.
turtle remains from the middle Miocene, Quebrada Honda, Bolivia: Evidence for Genise, J.F., Cladera, G., 2004. Chubutolithes gaimanensis and other wasp trace fossils:
lower paleoelevations for the southern Altiplano. J. S. Am. Earth Sci. 64, 190–198. breaking through the taphonomic barrier. J. Kansas Entomol. Soc. 77, 626–638.
Cambefort, Y., Hanski, I., 1991. Dung Beetle Population Biology. In: Hanski, I., Genise, J.F., Mángano, M.G., Buatois, L.A., Laza, J.H., Verde, M., 2000. Insect trace fossil
Cambefort, Y. (Eds.), Dung Beetle Ecology. Princeton University Press, United States, associations in paleosols: the Coprinisphaera ichnofacies. PALAIOS 15, 49–64.
pp. 36–50. Gingras, M.K., Pemberton, S.G., Mendoza, C., Henk, F.H., 1999. Modeling fluid flow in
Cane, J.H., 1991. Soils of ground-nesting bees (Hymenoptera: Apoidea): texture, trace fossils; assessing the anisotropic permeability of Glossifungites surfaces. Pet.
moisture, cell depth and climate. J. Kansas Entomol. Soc. 64, 406–413. Geosci. 5, 349–357.
Carrillo, J.D., Forasiepi, A., Jaramillo, C., Sánchez-Villagra, M.R., 2015. Neotropical Goin, F., Sánchez-Villagra, M.R., Kay, R.F., Anaya-Daza, F., Takai, M., 2003. New pa-
mammal diversity and the Great American Biotic Interchange: spatial and temporal laeothentid marsupial from the middle Miocene of Bolivia. Palaeontology 46,
variation in South America's fossil record. Front. Genet. 5, 451. 307–315.
Catena, A.M., Saylor, B., Croft, D.A., 2015. Reconstructing the Middle Miocene Gregory-Wodzicki, K.M., 2000. Uplift history of the Central and Northern Andes: a re-
Paleoenvironment of Quebrada Honda, Bolivia Using Ichnology and Paleopedology. view. GSA Bull. 112, 1091–1105.
Geological Society of America Annual Meeting (Abstracts with Programs: 262782). Guerrero, J., 1997. Stratigraphy, sedimentary environments, and the Miocene uplift of the
Catena, A.M., Hembree, D.I., Saylor, B.Z., Anaya, F., Croft, D.A., 2016. Colombian Andes. In: Kay, R.F., Madden, R.H., Cifelli, R.H., Flynn, J.J. (Eds.),
Paleoenvironmental analysis of the Neotropical fossil mammal site of Cerdas, Bolivia Vertebrate Paleontology in the Neotropics: The Miocene Fauna of La Venta,
(middle Miocene) based on ichnofossils and paleopedology. Palaeogeogr. Colombia. Smithsonian Institution Press, Washington, DC, pp. 13–43.
Palaeoclimatol. Palaeoecol. 459, 423–439. Hasiotis, S.T., 2002. Continental Trace Fossils. SEPM Short Course Notes 51. SEPM
Cohen, A.S., 1982. Paleoenvironments of root casts from the Koobi Fora Formation, Society for Sedimentary, Tulsa, OK, USA.
Kenya. J. Sediment. Res. 52, 401–414. Hasiotis, S.T., 2007. Continental ichnology: fundamental processes and controls on trace
Cole, M.M., 1986. The Savannas: Biogeography and Geobotany. Academic Press, London. fossil distribution. In: Miller IIIW. (Ed.), Trace Fossils: Concepts, Problems, Prospects.
Collinson, J., 1996. Alluvial sediments. In: Reading, H.G. (Ed.), Sedimentary Elsevier, Amsterdam, pp. 268–284.
Environments: Processes, Facies and Stratigraphy, third ed. Blackwell Sciences Ltd., Hasiotis, S.T., Mitchell, C.E., 2008. A comparison of crayfish burrow morphologies:
Oxford, UK, pp. 37–82. Triassic and Holocene fossil, paleo-and neo-ichnological evidence, and the identifi-
Cozzuol, M.A., 2006. The Acre vertebrate fauna: age, diversity, and geography. J. S. Am. cation of their burrowing signatures. Ichnos 2, 291–314.
Earth Sci. 213, 185–203. Hembree, D.I., 2009. Neoichnology of burrowing millipedes: linking modern burrow
Croft, D.A., 2007. The middle Miocene (Laventan) Quebrada Honda Fauna, southern morphology, organism behavior, and sediment properties to interpret continental
Bolivia, and a description of its notoungulates. Palaeontology 50, 277–303. ichnofossils. PALAIOS 24, 425–439.
Croft, D.A., 2016. Horned Armadillos and Rafting Monkeys: the Fascinating Fossil Hembree, D.I., 2014. Large complex burrows of terrestrial invertebrates: neoichnology of
Mammals of South America. Indiana University Press, Bloomington, Indiana, USA. Pandinus imperator (Scorpiones: Scorpionidae). In: Hembree, D.I., Platt, B.F., Smith,
Croft, D.A., Anaya, F., 2006. A new middle Miocene hegetotheriid (Notoungulata: J.J. (Eds.), Experimental Approaches to Understanding Fossil Organisms: Lessons
Typotheria) and a phylogeny of the Hegetotheriidae. J. Vertebr. Paleontol. 26, From the Living. Springer, Netherlands, pp. 229–263.
387–399. Hembree, D.I., Hasiotis, S.T., 2008. Miocene vertebrate and invertebrate burrows defining
Croft, D.A., Flynn, J.J., Wyss, A.R., 2004. Notoungulata and Litopterna of the early compound paleosols in the Pawnee Creek Formation, Colorado, USA. Palaeogeogr.
Miocene Chucal Fauna, northern Chile. Fieldiana Geol. 50, 1–49. Palaeoclimatol. Palaeoecol. 270, 349–365.
Croft, D.A., Flynn, J.J., Wyss, A.R., 2007. A new basal glyptodontid and other Xenarthra Hembree, D.I., Johnson, L.M., Tenwalde, R.W., 2012. Neoichnology of the desert scorpion
of the early Miocene Chucal Fauna, northern Chile. J. Vertebr. Paleontol. 27, Hadrurus arizonensis: burrows to biogenic cross lamination. Palaeontol. Electron. 15
781–797. (10A), 34.
Croft, D.A., Anaya, F., Auerbach, D., Garzione, C., MacFadden, B.J., 2009. New data on Hils, J.M., Hembree, D.I., 2015. Neoichnology of the burrowing spiders Gorgyrella inermis
Miocene Neotropical provinciality from Cerdas, Bolivia. J. Mamm. Evol. 175–198. (Mygalomorphae: Idiopidae) and Hogna lenta (Araneomorphae: Lycosidae).
Croft, D.A., Chick, J.M.H., Anaya, F., 2011. New middle Miocene caviomorph rodents Palaeontol. Electron. 18, 1–62.
from Quebrada Honda, Bolivia. J. Mamm. Evol. 18, 245–268. Hoorn, C., 1993. Marine incursions and the influence of Andean tectonics on the Miocene
Croft, D.A., Anaya, F., Catena, A., Ciancio, M., Engelman, R., 2013. New species, local depositional history of northwestern Amazonia: results of a palynostratigraphic
faunas, and paleoenvironmental data for the middle Miocene Quebrada Honda study. Palaeogeogr. Palaeoclimatol. Palaeoecol. 105, 267–309.
Fauna, Bolivia. J. Vertebr. Paleontol (SVP Program and Abstracts Book: 109). Hoorn, C., 2006. Mangrove forests and marine incursions in Neogene Amazonia (Lower
Croft, D.A., Carlini, A.A., Ciancio, M.R., Brandoni, D., Drew, N.E., Engelman, R.K., Anaya, Apaporis River, Colombia). PALAIOS 21, 197–209.
F., 2016. New mammal faunal data from Cerdas, Bolivia, a low latitude Neotropical Hoorn, C., Wesselingh, F., Hovikoski, J., Guerrero, J., 2010a. The development of the
site that chronicles the end of the Middle Miocene Climatic Optimum in South Amazonian mega-wetland (Miocene; Brazil, Colombia, Peru, Bolivia). In: Hoorn, C.,
America. J. Vertebr. Paleontol. http://dx.doi.org/10.1080/02724634.2016.1163574. Wesselingh, F.P. (Eds.), Amazonia, Landscape and Species Evolution: A Look into the
D'Alessandro, A., Bromley, R.G., 1987. Meniscate trace fossils and the Muensteria- Past. Blackwell Publishing, West Sussex, pp. 123–142.
Taenidium problem. Palaeontology 30, 743–763. Hoorn, C., Wesselingh, F.P., ter Steege, H., Bermudez, M.A., Mora, A., Sevink, J.,
Engelman, R.K., Croft, D.A., 2014. A new species of small-bodied sparassodont Sanmartin, I., Sanchez-Meseguer, A., Anderson, C.L., Figueiredo, J.P., Jaramillo, C.,
(Mammalia: Metatheria) from the middle Miocene locality of Quebrada Honda, Riff, D., Negri, F.R., Hooghiemstra, H., Lundberg, J., Stadler, T., Särkinen, T.,
Bolivia. J. Vertebr. Paleontol. 34, 672–688. Antonelli, A., 2010b. Amazonia through time: Andean uplift, climate change, land-
Engelman, R.K., Anaya, F., Croft, D.A., 2015. New specimens of Acyon myctoderos scape evolution and biodiversity. Science 330, 927–931.
(Metatheria, Sparassodonta) from Quebrada Honda, Bolivia. Ameghiniana 52, Hughes, C.E., Pennington, R.T., Antonelli, A., 2013. Neotropical plant evolution: assem-
204–225. bling the big picture. Bot. J. Linn. Soc. 171, 1–18.
Engelman, R.K., Anaya, F., Croft, D.A., 2016. New palaeothentid marsupials Hynek, S.A., Passey, B.H., Prado, J.L., Brown, F.H., Cerling, T.E., Quade, J., 2012. Small
(Paucituberculata) from the middle Miocene of Quebrada Honda, Bolivia, and their mammal carbon isotope ecology across the Miocene–Pliocene boundary, north-
implications for the palaeoecology, decline and extinction of the Palaeothentoidea. J. western Argentina. Earth Planet. Sci. Lett. 321, 177–188.
Syst. Palaeontol. http://dx.doi.org/10.1080/14772019.2016.1240112. Johnston, P.A., Eberth, D.A., Anderson, P.K., 1996. Alleged vertebrate eggs From Upper
Eva, H.D., de Miranda, E.E., Di Bella, C.M., Gond, V., Huber, O., Sgrenzaroli, M., Jones, S., Cretaceous redbeds, Gobi Desert, are fossil insect (Coleoptera) pupal Chambers:
Coutinho, A., Dorado, A., Guimarães, Elvidge, C., Achard, F., Belward, A.S., Fictovichnus new ichnogenus: Canadian J. Earth Sci. 33, 511–525.
Bartholomé, E., Baraldi, A., De Grandi, G., Vogt, P., Fritz, S., Hartley, A., 2002. A Kay, R.F., Madden, R.H., 1997a. Paleogeography and paleoecology. In: Kay, R.F.,
vegetation map of South America. In: EUR20159 EN. Joint Research Centre of the Madden, R.H., Cifelli, R.H., Flynn, J.J. (Eds.), Vertebrate Paleontology in the
European Commission, Luxembourg. Neotropics: The Miocene Fauna of La Venta, Colombia. Smithsonian Institution Press,
Fitzpatrick, E.A., 1993. Soil Microscopy and Micromorphology. John Wiley & Sons, New Washington, DC, pp. 520–550.
York, NY, USA. Kay, R.F., Madden, R.H., 1997b. Mammals and rainfall: paleoecology of the middle
Flynn, J.J., Wyss, A.R., 1998. Recent advances in South American mammalian Miocene at La Venta (Colombia, South America). J. Hum. Evol. 32, 161–199.

397
A.M. Catena et al. Palaeogeography, Palaeoclimatology, Palaeoecology 487 (2017) 381–398

Kay, R.F., Madden, R.H., Cifelli, R.H., Flynn, J.J., 1997. Vertebrate Paleontology in the ncomms2299.
Neotropics: The Miocene Fauna of La Venta, Colombia. Smithsonian Institution Press, Pascual, R., Jaureguizar, E.O., 1990. Evolving climates and mammal faunas in Cenozoic
Washington, DC. South America. J. Hum. Evol. 19, 23–60.
Keighley, D.G., Pickerill, R.K., 1994. The ichnogenus Beaconites and its distinction from Philips, T.K., 2011. The evolutionary history and diversification of dung beetles. In:
Ancorichnus and Taenidium. Palaeontology 37, 305–338. Simmons, L.W., Ridsdill-Smith, T.J. (Eds.), Ecology and Evolution of Dung Beetles.
Kerber, L., Negri, F.R., Ribeiro, A.M., Nasif, N., Souza-Filho, J.P., Ferigolo, J., 2016. Wiley-Blackwell, pp. 21–46.
Tropical fossil caviomorph rodents from the southwestern Brazilian Amazonia in the Retallack, G.J., 1984. Trace fossils of burrowing beetles and bees in an Oligocene pa-
context of the South American faunas: systematics, biochronology, and paleobio- leosol, Badlands National Park, South Dakota. J. Paleontol. 58, 571–592.
geography. J. Mamm. Evol. 24, 57–70. Retallack, G.J., 2001. Soils of the Past: An Introduction to Paleopedology, Second ed.
Klappa, C.F., 1980. Rhizoliths in terrestrial carbonates: classification, recognition, gen- Blackwell Science Ltd., Oxford, United Kingdom.
esis, and significance. Sedimentology 26, 613–629. Rosenheim, J.A., 1987. Nesting behavior and bionomics of a solitary ground-nesting
Kottek, M., Grieser, J., Beck, C., Rudolf, B., Rubel, F., 2006. Worldmap of the Köppen- wasp, Ammophila dysmica (Hymenoptera: Sphecidae): influence of parasite pressure.
Geiger climate classification updated. Meteorol. Z. 15, 259–263. Ann. Entomol. Soc. Am. 80, 739–749.
Kraus, M.J., 1999. Paleosols in clastic sedimentary rocks: their geologic applications. Rull, V., 2014. Neotropical biodiversity: timing and potential drivers. Trends Ecol. Evol.
Earth Sci. Rev. 47, 41–70. 26, 508–513.
Kraus, M.J., Hasiotis, S.T., 2006. Significance of differentmodes of rhizolith preservation Sánchez-Villagra, M.R., Kay, R.F., Anaya-Daza, F., 2000. Cranial anatomy and palaeo-
to interpreting paleoenvironmental and paleohydrologic settings: examples from biology of the Miocene marsupial Hondalagus altiplanensis and a phylogeny of ar-
Paleogene paleosols, Bighorn Basin, Wyoming, USA. J. Sediment. Res. 76, 633–646. gyrolagids. Palaeontology 43, 287–301.
Lavin, M., 2006. Floristic and geographical stability of discontinuous seasonally dry Sarmiento, G., 1984. The Ecology of Neotropical Savannas. Harvard University Press,
tropical forests explains patterns of plant phylogeney and endemism. In: Pennington, United States.
R.T., Lewis, G.P., Ratter, J.A. (Eds.), Neotropical Savannas and Seasonally Dry Sauer, W., 1955. Coprinisphaera ecuadoriensis, un fosil singular del Pleistoceno. Bol. Inst.
Forests: Plant Diversity, Biogeography, and Conservation. The Systematics Cienc. Nat. 1, 123–132.
Association Special Volume Series. 69. pp. 43–448. Savrda, C.E., Blanton-Hooks, A.D., Collier, J.W., Drake, R.A., Graves, R.L., Hall, A.G.,
Liebmann, B., Allured, D., 2005. Daily precipitation grids for South America. Bull. Am. Nelson, A.I., Slone, J.C., Williams, D.D., Wood, H.A., 2000. Taenidium and associated
Meteorol. Soc. 86, 1567–1570. ichnofossils in fluvial deposits, Cretaceous Tuscaloosa Formation, eastern Alabama,
Lovejoy, N.R., Bermingham, E., Martin, A.P., 1998. Marine incursion into South America. southeastern USA. Ichnos 7, 227–242.
Nature 396, 421–422. Schaetzl, R.J., Anderson, S., 2009. Soils: Genesis and Geomorphology. Cambridge
MacFadden, B.J., 2006. Extinct mammalian biodiversity of the ancient New World tro- University Press, Cambridge, United Kingdom.
pics. Trends Ecol. Evol. 21, 157–165. Sheldon, N.D., Tabor, N.J., 2009. Quantitative paleoenvironmental and paleoclimatic
MacFadden, B.J., Wolff, R.G., 1981. Geological investigations of Late Cenozoic verte- reconstruction using paleosols. Earth Sci. Rev. 95, 1–52.
brate-bearing deposits in southern Bolivia. In: Anais do II Congresso Latino- Shockey, B.J., Croft, D.A., Anaya, F., 2007. Analysis of function in the absence of extant
Americano de Paleontología. 2. pp. 765–778. functional homologues: a case study using mesotheriid notoungulates (Mammalia). J.
MacFadden, B.J., Anaya, F., Perez, H., Naeser, C.W., Zeitler, P.K., Campbell Jr., K.E., Inflamm. 33 (2), 227–247.
1990. Late Cenozoic paleomagnetism and chronology of Andean basins of Bolivia: Simmons, L.W., Ridsdill-Smith, T.J., 2011. Reproductive competition and its impact on
evidence for possible oroclinal bending. J. Geol. 98, 541–555. the evolution and ecology of dung beetles. In: Simmons, L.W., Ridsdill-Smith, T.J.
MacFadden, B.J., Anaya, F., Swisher, C.C., 1995. Neogene paleomagnetism and oroclinal (Eds.), Ecology and Evolution of Dung Beetles. Wiley-Blackwell, pp. 1–20.
bending of the central Andes of Bolivia. J. Geophys. Res. 100, 8153–8167. Smith, B.T., McCormack, J.E., Cuervo, A.M., Hickerson, M.J., Aleixo, A., Cadena, C.D.,
Machette, M.N., 1985. Calcic soils of the southwestern United States. In: Weide, D.L. Pérez-Emán, J., Burney, C.W., Xie, X., Harvey, M.G., Faircloth, B.C., Glenn, T.C.,
(Ed.), Soils and Quaternary Geology of the Southwestern United States. 203. GSA Derryberry, E.P., Prejean, J., Fields, S., Brumfield, R.T., 2014. The drivers of tropical
Special Paper, pp. 10–21. speciation. Nature 515, 406–409.
Mack, G.H., James, W.C., Monger, H.C., 1993. Classification of paleosols. Geol. Soc. Am. Soil Survey Staff, 2010. Keys to Soil Taxonomy, 11th ed. USDA Natural Resources
Bull. 105, 129–136. Conservation Service, Washington, DC, USA.
Marshall, L.G., Sempéré, T., 1991. The Eocene to Pleistocene vertebrates of Bolivia and Stiles, C.A., Mora, C.I., Driese, S.G., 2001. Pedogenic iron‑manganese nodules in Vertisols:
their stratigraphic context: a review. In: Suárez-Soruco, R. (Ed.), Fósiles y Facies de a new proxy for paleoprecipitation? Geology 29, 943–946.
Bolivia–Vol I Vertebrados. Revista Técnica de Yacimientos Petrolíferos Fiscales Tejada-Lara, J.V., Salas-Gismondi, R., Pujos, F., Baby, P., Benammi, M., Brusset, S., De
Bolivianos, Santa Cruz, pp. 631–652. Franceschi, D., Espurt, N., Urbina, M., Antoine, P.O., 2015. Life in proto-Amazonia:
Mason, J.A., Jacobs, P.M., 2007. Nature of Paleosol. In: Elias, S. (Ed.), Encyclopedia of Middle Miocene mammals from the Fitzcarrald Arch (Peruvian Amazonia).
Quaternary Science. Elsevier, Amsterdam, pp. 2086–2095. Palaeontology 58, 341–378.
Mikuś, P., Uchman, A., 2013. Bettle burrows with a terminal chamber: a contribution to Townsend, B., Croft, D.A., 2010. Middle Miocene mesotheriine diversity at Cerdas,
the knowledge of the trace fossil Macanopsis in continental sediments. PALAIOS 28, Bolivia, and a reconsideration of Plesiotypotherium minus. Palaeontol. Electron. 13 (1)
403–413. (1 A:36 pp.).
Montoya-Sanhueza, G., Moreno, K., Bobe, R., Carrano, M.T., García, M., Corgne, A., 2017. Villani, M.G., Allee, L.L., Diaz, A., Robbins, P.S., 1999. Adaptive strategies of edaphic
Peltephilidae and Mesotheriidae (Mammalia) from late Miocene strata of Northern arthropods. Annu. Rev. Entomol. 44, 233–256.
Chilean Andes, Caragua. J. S. Am. Earth Sci. 75, 51–65. Villarroel, C., 1974. Un Mésothériiné nouveau (Notoungulata, Mammalia) dans le
Ogg, J.G., 2012. Geomagnetic Polarity Time Scale. In: Gradstein, F.M., Ogg, J.G., Schmitz, Miocène supérieur de Bolivie. C. R. Seances Acad. Sci. D 279, 551–554.
M., Ogg, G. (Eds.), The Geologic Time Scale 2012. Elsevier, Amsterdam, pp. 85–113. Wilson, D.E., Reeder, D.M., 2005. Mammal Species of the World: A Taxonomic and
Oiso, O., 1991. New land mammal locality of middle Miocene (Colloncuran) age from Geographic Reference, Second ed. John Hopkins University Press, Baltimore,
Nazareno, southern Bolivia. In: Fósiles y facis de Bolivia. 1. pp. 653–672. Maryland.
Palazzesi, L., Barreda, V., 2012. Fossil pollen records reveal a late rise of open-habitat Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends, rhythms, and
ecosystems in Patagonia. Nat. Commun. 3, 1294. http://dx.doi.org/10.1038/ aberrations in global climate 65 Ma to present. Science 292, 686–693.

398

View publication stats