Underground chamber systems excavated by cenozoic ground sloths in the state of Rio Grande do Sul, Brazil

Abstract

A regional survey of the caves of the state of Rio Grande do Sul, Brazil, listed hundreds of caves with varied origins. Among these, we identified two large caves with special characteristics, one located in the city of Santa Cruz do Sul and the other in the municipality of Vale Real. Both evolve in an approximately horizontal plane, extend tens of meters into the interior of the respective elevations, and do not show any signs of present or past underground drainage. Concave surfaces on the walls stand out in their morphologies, which suggest that the caves originally constituted a system of ellipsoidal chambers, each chamber with original height of approximately 1.5 m and diameter between 3.5 and 7 m. The individual chambers were connected by relatively short tunnels. These chambers are best preserved in the cave of Santa Cruz do Sul; the cave of Vale Real appears to have been much disfigured, both because of inorganic processes and anthropogenic action. The general characteristics of these two caves enable us to propose that they originated from the activity of digging ground sloths of the South American Megafauna. Their dimensions suggest that they were excavated and inhabited not by individual animals, but by groups of sloths. The original polished walls suggest that the caves were used for extended periods, possibly in the order of centuries. Their morphology is so different from other tunnel systems that were excavated by the Megafauna and found in southern Brazil, that we suggest that this morphology of chamber systems is associated with a specific species of ground sloth, whose identification will be very difficult.

Full text

273
Rev. bras. paleontol. 18(2):273-284, Maio/Agosto 2015
© 2015 by the Sociedade Brasileira de Paleontologia
doi: 10.4072/rbp.2015.2.08
UNDERGROUND CHAMBER SYSTEMS EXCAVATED BY CENOZOIC
GROUND SLOTHS IN THE STATE OF RIO GRANDE DO SUL, BRAZIL
HEINRICH THEODOR FRANK, CAMILA ELIZA ALTHAUS, ERIK MARTINS DARIO,
FERNANDO RUBBO TRAMONTINA, RAFAEL MARQUEZAM ADRIANO,
MARIANA DE LIMA ALMEIDA
Instituto de Geociências, Universidade Federal do Rio Grande do Sul, Cx. P. 15001, Av. Bento Gonçalves, 9500, 92501-970,
Porto Alegre, RS, Brasil. heinrich.frank@ufrgs.br, camilaalthaus@gmail.com, erikdario@hotmail.com,
tramontinarubbo@hotmail.com, rafa2005adriano@hotmail.com,
mary_lima_almeida@hotmail.com
GABRIELA FEITEN FERREIRA, RAFAELA NOGUEIRA & ROGÉRIO BREIER
Universidade do Vale do Rio dos Sinos, Av. Unisinos, 950, 93022-000, São Leopoldo, RS, Brasil.
gabriele.paleotocas@gmail.com, rafaela.paleotocas@gmail.com, rbreier@via-rs.net
ABSTRACT – A regional survey of the caves of the state of Rio Grande do Sul, Brazil, listed hundreds of caves with varied
origins. Among these, we identifi ed two large caves with special characteristics, one located in the city of Santa Cruz do
Sul and the other in the municipality of Vale Real. Both evolve in an approximately horizontal plane, extend tens of meters
into the interior of the respective elevations, and do not show any signs of present or past underground drainage. Concave
surfaces on the walls stand out in their morphologies, which suggest that the caves originally constituted a system of ellipsoidal
chambers, each chamber with original height of approximately 1.5 m and diameter between 3.5 and 7 m. The individual
chambers were connected by relatively short tunnels.These chambers are best preserved in the cave of Santa Cruz do Sul;
the cave of Vale Real appears to have been much disfi gured, both because of inorganic processes and anthropogenic action.
The general characteristics of these two caves enable us to propose that they originated from the activity of digging ground
sloths of the South American Megafauna. Their dimensions suggest that they were excavated and inhabited not by individual
animals, but by groups of sloths. The original polished walls suggest that the caves were used for extended periods, possibly
in the order of centuries. Their morphology is so different from other tunnel systems that were excavated by the Megafauna
and found in southern Brazil, that we suggest that this morphology of chamber systems is associated with a specifi c species
of ground sloth, whose identifi cation will be very diffi cult.
Key words: caves, megafauna, ground sloths, tunnels, chambers, Brazil.
RESUMO – Um levantamento regional das cavernas existentes no estado do Rio Grande do Sul, Brasil, listou centenas
de cavernas com variadas origens. Dentre estas, foram identifi cadas duas cavernas de grande porte com características
diferenciadas, uma localizada na cidade de Santa Cruz do Sul e a outra no município de Vale Real. Ambas desenvolvem-se
aproximadamente em um plano horizontal, estendem-se dezenas de metros para o interior das respectivas elevações e não
apresentam sinais de drenagens subterrâneas presentes ou passadas. Nas suas morfologias destacam-se superfícies côncavas
nas paredes que sugerem que as cavernas se formaram a partir de um sistema de câmaras elípticas, cada câmara com altura
original aproximada de 1,5 m e diâmetro original entre 3,5 e 7 m. As câmaras individuais eram unidas através de túneis
relativamente curtos. Estas câmaras estão mais bem preservadas na caverna de Santa Cruz do Sul; a caverna de Vale Real
apresenta-se muito desfi gurada, tanto por processos inorgânicos como por ação antropogênica. As características gerais das
duas cavernas permitem propor que se originaram pela ação de escavação de preguiças gigantes da Megafauna sul-americana.
Suas dimensões sugerem que foram escavadas e habitadas não por animais isolados, mas por grupos de preguiças. As paredes
originais polidas sugerem que as câmaras foram usadas por longos períodos (séculos). Sua morfologia é tão contrastante com
àquela dos sistemas de túneis encontrados na Região Sul do Brasil, também escavados pela Megafauna, que propomos que esta
morfologia de abrigo subterrâneo se refere a uma espécie específi ca de preguiça-gigante, cuja identifi cação será muito difícil.
Palavras-chave: cavernas, megafauna, preguiças-gigantes, túneis, câmaras, Brasil.

274 REVISTA BRASILEIRA DE PALEONTOLOGIA, 18(2), 2015
INTRODUCTION
In southern Brazil, spanning the states of Rio Grande
do Sul and Santa Catarina, hundreds of tunnels that were
probably dug by fossorial vertebrates of the Cenozoic
Megafauna, such as giant armadillos (Dasypodidae) and
ground sloths (Xenarthra) have been found (Bergqvist &
Maciel, 1994; Buchmann et al., 2009; Frank et al., 2012a,b,
2013). The widths and heights of the original tunnels range
between 0.7-4.0 and 0.6-2.0 m, respectively. Individual
tunnels may reach lengths of more than 50 m. Over 80% of
the tunnels are found completely fi lled with sediments. Larger
outcrops, like anthropogenic cuts with lengths between some
tens of meters to a few hundred meters, often show several
tunnels side by side; such tunnel clusters may be formed by
more than 30 tunnels side by side (Frank et al., 2012a, p.
146). Investigation of occurrences with open tunnels shows
that the tunnels form complex systems that can have an added
tunnel length exceeding 300 m (Ruchkys et al., 2014). The
regional density of tunnel systems may be of more than a
tunnel system per square kilometer, some regions present
one or more tunnel systems in every hill. The usual pattern
of Cenozoic paleovertebrate tunnels in southern Brazil,
therefore, is that the tunnel systems are formed without the
presence of chambers. This paper presents the features of two
large caves that most probably arose from a different kind
of underground structure produced by fossorial Cenozoic
paleovertebrates. This other structure is a huge underground
shelter composed of a system of large ellipsoidal chambers
linked by relatively short and low tunnels.
METHODS
Several of the huge tunnels dug by fossorial vertebrates of
the Cenozoic Megafauna in southern Brazil still remain open
today and are well known by the landowners, who visit them
sporadically driven by curiosity. Images and descriptions
of some of these tunnels can be found even on the internet.
Acknowledging this fact, an extensive digital prospecting
was performed, looking for images related to terms
commonly used for cavities, such as “caves”, “grottoes”,
“holes” and others, focusing the search on occurrences in the
state of Rio Grande do Sul. This survey resulted in a list of
over 200 identifi ed caves in dozens of municipalities. Most
caves are shallow rock shelters, overhangs or other kinds
of cavities located on steep rock walls or behind waterfalls,
others were formed by rocky landslides or by creeks or
rivers through hydraulic action – all these are unrelated to
fossorial paleovertebrates. Using the available information,
two caves with different characteristics were selected. One is
located in an urban park in the city of Santa Cruz do Sul and
the other in the municipality of Vale Real (Figure 1). Both
seemed to extend deep into hills constituted by sandstones,
were not located on more or less vertical cliffs and are not
associated to running water (behind waterfalls, besides a
creek or river, etc.).
To begin, both caves were inspected and photographed,
discussing its characteristics with the team. Once the caves
were confi rmed as potentially of fossorial paleovertebrate
origin, detailed photographic surveys of the morphology of
the caves and of the features on the walls were performed,
Figure 1. Location of the caves and geological outline of the region. A, Vale Real cave; B, Santa Cruz do Sul cave.

275
FRANK ET AL. – UNDERGROUND CHAMBER SYSTEMS EXCAVATED BY CENOZOIC GROUND SLOTHS
focusing the search on still existing original characteristics.
Using a theodolite and related topographic equipment, we made
planimetric surveys of the caves. Details about the appearances
of these two caves over more than half a century ago were
obtained through interviews with older locals who had visited
the caves for a long time. For both caves, petrographic thin
sections were made from the caves’ host rock.
GEOLOGY
Most land in the three southernmost Brazilian states is
located in the Upper Ordovician to Cretaceous Paraná Basin
(Figure 1). This intracratonic basin comprises an area of more
than 1.106 km2 (Zalán et al., 1990; Milani et al., 1998) and is
fi lled with a lower sequence of sedimentary rocks and an upper
sequence of volcanic rocks. The volcanic rocks constitute the
Lower Cretaceous Serra Geral Formation, part of the Paraná-
Etendeka Continental Flood Basalt Province (Peate, 1997).
The youngest sedimentary formation, immediately beneath
the volcanic rocks, is the Late Jurassic – Early Cretaceous
Botucatu Formation (Scherer, 2000), whose most prominent
lithotype is a fi ne- to coarse-grained reddish sandstone with
large scale aeolian cross-bedding, a relict of a >1.5 million
km2 arid continental area that extended beyond the limits of
the basin. Both caves described in this paper were excavated
in this sandstone which, because of its generally medium
lithifi cation degree and high stability, was a preferred host rock
for paleovertebrate tunnels in this region. The petrographic
data of sandstone samples from both caves is listed in Table 1
and the similarities between the caves are shown in Table 2.
DESCRIPTION
Vale Real Cave
The Vale Real cave is located in the rural area of the
municipality of Vale Real (29º20´24.40´´S, 51º13´29.70´´W)
(Figure 1), in a private estate. It is known as the “Cave of the
Indians”. Inscriptions on the walls show that the cave has
been visited by European settlers since at least 1878. The
description of the cave, presented below, addresses (i) its size
and general characteristics, (ii) the internal morphology and
(iii) the characteristics of the lateral walls and of the ceiling.
The cave has an arc-shaped entrance with a width and
height of 8 and 2.6 m, respectively (Figure 2A). In the front
of the cave entrance there is a fl at area with a length of 9.5
m perpendicular to the entrance. According to reports of
old residents in the region, this area was produced by the
deposition of sediments taken from inside the cave by treasure
hunters, who excavated some parts of the fl oor and some walls
between 2007 and 2009. The length of the cave, measured
perpendicular to the slope of the hill, is of 36 m, but its linear
development is almost 55 m with a very irregular perimeter
Table 1. Petrographic data of the sandstone that hosts both caves.
Vale Real Santa Cruz do Sul
Sand size fine (average 0,1-0,2) fine (average 0,1)
Selection medium medium
Composition quartz (>80%), feldspar quartz (>90%), feldspar
Roundness good good
Sphericity average to good average to good
Grain contacts long short
Porosity 15-20 % 20-30 %
Iron oxide films >60% of the grains >90% of the grains
Table 2. Similarities between the caves of Santa Cruz do Sul and Vale Real.
Santa Cruz do Sul Vale Real
Host rock Botucatu sandstone Botucatu sandstone
Distance to the nearest stream <100 m ~200 m
Position of the cave almost horizontal almost horizontal
Approximate area (m2) 332 480
Approximate volume (m3) 600-650 1000
Depth perpendicular to hillside 28,5 m 36 m
Approximate total length ~50 m ~55 m
Defined ellipsoidal spaces >6 >6
Features of groundwater flow almost none almost none
Collapsing features on the roof abundant abundant
Original surfaces on the walls >6 >4
Original surfaces on the roof 1 1
Evidence of buried portions several several

276 REVISTA BRASILEIRA DE PALEONTOLOGIA, 18(2), 2015
Figure 2. Floorplan of the Vale Real cave. A, entrance of the cave
(observer = 1.84 m); B, view from c (see plan) towards the entrance
(on the left, around the pillar, the excavations of the treasure hunters
can be seen); C, view from E towards d. Lowercase corresponds to
concave surfaces.
(Figures 2B,C). The cave has an area of approximately 480 m2
and a volume of around 1,000.00 m3. The fl oor is developed
on three levels: an initial part (Figure 2: from the entrance
to the dashed line) is almost horizontal at the same level as
the fl at area in front of the cave entrance. A middle portion
(Figure 2: between “D” and “E”) is slightly ascending and
a last level (Figure 2: from “E” until the end of the cave) is
slightly descending. The maximum gap between the entrance
of the cave and the highest point of the fl oor (Figure 2: near
to “E”) is approximately 1 m.
Today, the heights of the cave vary between zero and
nearly 3.5 m. The area between the pillar and points “E” and
“F” (Figure 2) has had its fl oor and some parts of the wall
intensely excavated by treasure hunters. Longtime residents
of the regions all agree that there was a “well” (a hole with
water) inside the cave (at the position shown in Figure 2), with
a diameter of 60-70 cm and of unknown depth. This well was
fi lled with sand by the treasure hunters. On 2 spots (Figure 2:
“a” and “d”) there are cracks and cavities of different sizes
(maximum axis of less then 40 cm) at the upper part of the
wall, forming an alignment that coincides with the plane
of the cross-bedded strata of sandstone that hosts the cave
(Figure 3C). In the last chamber of the cave (Figures 2C, 4D),
the lower part of the wall is covered with a black coating from
the fl oor up to a height of approximately 50 cm. The roof of
the cave and some of the upper portions of the lateral walls
also exhibit dark colors, contrasting with the lighter colored
portions of the walls where collapsing has occurred, which
exposed the original red and pink color of the sandstone.
Approximately six spots have white millimeter-thick coatings,
which may be considered speleothems (Figure 4a). Together,
these spots cover less than 20 m2, mainly at the upper limit
of the lateral walls and sometimes on the roof.
Despite the anthropogenic activity and several collapsing
features (described below), it is possible to recognize,
on the lateral walls, several concave surfaces formed by
smooth and continuous vertical and horizontal arcs. As a
whole, these concave surfaces tend to constitute ellipsoidal
spaces, with the two major axes disposed horizontally and
the smaller axis vertically. In the rear part of the ellipsoidal
spaces, there is no opening or continuation of the cave; they
always end up as dead ends (Figure 4). The concave surfaces
may be simple or compound. The simple surfaces have a
side wall formed by a continuous arc whose curvature may
be sharper or gentler. On Figure 2, these concave surfaces
are in lowercase and all surfaces except “b” and “c” are
simple surfaces. Despite the fact that surface “a” is partially
buried, it defi nes an ellipsoidal space precisely (Figure 4a).
The composite surfaces have a lateral wall formed by a
succession of several very smooth concave surfaces (Figures
2b,c). Each one of these spaces clearly shows three successive
concave surfaces which form a larger arc.
The lateral walls and the roof are composed of (i) collapsed
surfaces, and (ii) smooth surfaces. The collapsed surfaces
are those who show signs of detachment of platy sandstone
blocks, exposing the pink original color of the sandstone.
Usually, the collapsed surfaces are linked by more or less
vertical fracture surfaces (Figure 3A). Much of the lateral
walls and virtually the entire cave roof are formed by such
surfaces. The platy shape of the blocks was constrained by the
large-scale cross-stratifi cation of the sandstone, whose most
common attitude is N20°W20°NE. The size of the collapsed
blocks, considering the largest continuous fl at surfaces on
the roof, reaches 1.5 m. The smooth surfaces, on the other
hand, show a very uniform, smooth and slightly wavy surface.
They cover an area of approximately 6 m2 on the roof of the
cave (Figure 2: dotted area between “D” and “g”; Figure
3D) and the entire surface of the ellipsoidal spaces indicated
“a” and “g”. In these spaces, fractured portions show that
these smooth surfaces are formed by a layer of sandstone
with a more pronounced diagenesis, with a lighter color and
a thickness ranging from 1 to 20 mm (Figure 4h). The west
side of the pillar is also formed by a smooth surface with
inscriptions dated from 1897 (“Jacob Schmitt”) and from 1898
(in indecipherable gothic German). In front of this side of the
pillar, on the wall of concavity “c”, the oldest inscriptions
are dated from 1904 (“Fredolino Zimmer”) and 1907. These
inscriptions testify that these walls are original.
A
B
C

277
FRANK ET AL. – UNDERGROUND CHAMBER SYSTEMS EXCAVATED BY CENOZOIC GROUND SLOTHS
Santa Cruz do Sul Cave
This cave, also known as “Cave of the Indians”
(portuguese: “Gruta dos Índios”), is located in the “Park of
the Cave” (“Parque da Gruta”) in the city limits of Santa
Cruz do Sul (29º42´39´´S, 52º24´30´´W) (Figure 1). At the
entrance of the cave, which is frequently visited when the Park
is open, there is an iron gate with padlock; inside there are
several fi xtures (Figures 5A,B). The cave extends for 28.5 m
perpendicular to the slope of the hill, comprising an area of
approximately 332 m2 and a volume between 600 and 650 m3.
Through a morphological analysis the cave can be
subdivided into two spaces of approximately equal length, but
with markedly different states of conservation regarding the
original morphology and the surface structures of the walls:
the front space and the distal space, divided by the dashed
line C-C´ in Figure 5.
The front space forms an approximately linear, NE-SW
oriented cavity whose widths range between 5 and 8 m. Its
height varies from virtually zero at its NE edges to more than
3 m near the cave entrance. The arrangement of the walls is
very irregular and does not reveal any recognizable pattern
(Figure 5B). The walls show many fractures and collapsing
surfaces, which can be roughly perpendicular or parallel to
the large-scale cross-bedding of the sandstone who hosts
the cave. Close to the cave entrance (Figure 5; point “D”),
the roof is much higher, like a dome, following an almost
vertical fracture in the sandstone. Similar collapsing features,
but on a smaller scale, can be seen on the entire length of
the walls and the roof. According to information provided
by older citizens, decades ago the government undertook
extensive excavations in the cave in order to facilitate the
access to the public. During this work, the large fallen blocks
were removed from the cave, resulting in the current height
of the fl oor and the roof. These citizens also reported that
some archeological excavations were conducted in the cave
in search of Indian artifacts, but nothing was found. Due to
these intense anthropogenic interventions, neither the current
dimensions, nor the surface of the walls and of the fl oor of
this front space can be considered original, and the current
situation does not allow any conclusion about its original
morphology. Moreover, all the walls and the entire roof, with
the exception of the highest and most inaccessible portions,
have been scratched or painted with names and dates. The
oldest date back to more than half a century ago.
The distal space shows a well-defi ned morphology. All of
its lateral walls are formed by successive concave surfaces
with a remarkable uniformity in the degree of concavity
(Figures 6A,B). These surfaces are not only vertically
concave, but also horizontally; concave surfaces on opposite
walls create several ellipsoidal spaces whose longer axes are
always horizontal, but of different lengths. The vertical axis
of the ellipsoidal spaces, considering the vertical arch formed
by several of the analyzed concave surfaces, has a length of
around 1.5 m. The best-preserved ellipsoidal space is located
at the rear end of the cave (Figures 5F, 6A). Ten individual
concave surfaces were identifi ed (Figure 5: lowercase), all
of them with their base at approximately the same horizontal
Figure 3. Aspects of the Vale Real cave. A, image of the roof with
collapsed surfaces, following the stratification of the sandstone;
B, pillar, seen form W to E. On the floor around the pillar are
anthropogenic excavations. On this side of the pillar there is an
inscription of 1897 “Jacob Schmitt”; C, opening that coincides with
the stratification of the sandstone at the upper part of the ellipsoidal
space 1; D, completely smooth roof, corresponding to the dotted area
on Figure 2. View from S to N.
A
B
C
D

278 REVISTA BRASILEIRA DE PALEONTOLOGIA, 18(2), 2015
plane. The roof of the cave in the distal space shows several
collapsing features, but much less pronounced than those in
the front space. At a single point on the roof (Figure 5: dotted
area; Figure 6E) there is a continuous, very smooth, slightly
concave surface similar to those at the lateral walls, without
any collapsing features. The superposition of two concave
surfaces (Figures 5g-h), also occurs there, with the lower one
buried with sediments by more than 60%. When the lower
concave surface (Figure 5g) is combined with the opposite
one (Figures 5j), also buried by at least 50%, it is possible
to defi ne an ellipsoidal space located at a level some 50 cm
below those of the concave surfaces from “a” to “f”. The wall
surfaces and most of the roof also are covered by scratchings
left by the visitors of the cave.
The surfaces of the walls are either fractured or smooth.
The fractured surfaces may be concordant or discordant with
the large-scale cross-bedding of the sandstone and expose
the sandstone with its original pinkish color. They are found
mostly on the roof of the cave and are much less abundant
than in the front space of the cave.
The smooth surfaces are those of the concave surfaces
and have a whitish-brown color. They are extremely uniform
and smooth, almost polished, and have only some irregular
grooves. In a single point there are some broad grooves,
roughly parallel, near the roof of the cave (Figure 6C). These
smooth surfaces show a brown crust with a slightly irregular
thickness between 1 and 5 mm, sometimes very well exposed
at fracture surfaces (Figure 6D). Petrographical analysis of the
crust revealed a composition of a cryptocrystalline brownish-
reddish material who shows many orange internal refl ections
with crossed polarizers on reflected light microscopy,
suggesting the presence of goethite (hydrated iron oxide
phases). Under this crust there is a black fi lm with a thickness
of less than 1 millimeter covering the reddish sandstone.
DISCUSSION
The nature of the discussion about the origin of these two
caves is much more speleological than palaeontological. What
evidences or traces do these caves display that point to the
processes, factors or characters that generated them?
The most common origins of caves can be quickly
excluded for this case study. The two caves were not formed
by the dissolution of the host rock as solutional caves in
Figure 4. Ellipsoidal spaces in the Vale Real cave; the lowercase corresponds to those of Figure 2 and the text (observer = 1.75 m). Image “h”
shows in detail the whitish centimetric crust of more lithified sandstone that forms the smooth surface of the ellipsoidal space “g”. Pen cap = 5.6 cm.
a
b
c
gh
f
e
d

279
FRANK ET AL. – UNDERGROUND CHAMBER SYSTEMS EXCAVATED BY CENOZOIC GROUND SLOTHS
Figure 5. Floor plan of the Santa Cruz do Sul cave. A, entrance of the cave (observer = 1.85 m); B, view from D (see floor plan) towards E
(observer = 1.8m). The C-C´ line divides the front space from the distal space. Lowercase from “a” to “j” are concave surfaces; F indicates the
position of the best-preserved ellipsoidal space (chamber) and the dotted area indicates the position of the still original roof (see text).
Figure 6. Aspects of the Santa Cruz do Sul cave. A, Concave surface “e” (Figure 5), defining the ellipsoidal space F at the rear end of the
cave; B, succession of concave and smooth surfaces on opposite walls of the cave, defining several ellipsoidal spaces. On the roof there are
many fracture surfaces. Height of the luminaire = 80 cm; C, large parallel grooves near the roof of the cave, on an original concave surface;
D, irregular, dark, centimetric crust (arrow) of a concave surface on a fracture located on the intersection of concave surface “a” (Figure 5) with
line C-C´. The smooth surface is on the right; E: concave, continuous, smooth and whitish roof (dotted area on Figure 5) (observer = 1.75 m).
A
AB
C
D
E
B

280 REVISTA BRASILEIRA DE PALEONTOLOGIA, 18(2), 2015
carbonate rock like limestone and dolostone (e.g. Palmer,
1991). Likewise, they were not formed by the action of waves
along beaches (abrasion caves or sea caves) (e.g. Waterstrat
et al., 2010) nor they are spaces between fallen blocks (talus
caves or boulder caves) (e.g. Sjöberg, 1987), cavities inside
fault zones (crevice or structural caves) (e.g. Margielewski
& Urban, 2003) or caves related to volcanic activity (lava
tubes or lava caves) (e.g. Peterson et al., 1994). Processes
like the ones responsible for “granite karst” (e.g. Osborne
et al., 2013) and “kamenitza” or “gnamma” (e.g. Goudie,
2013, p. 69) must also be excluded. Caves of anthropogenic
origin (e.g. Kostof, 1989; Wimmer, 2000) generally display
a unique set of features: usually the walls are vertical, the
roofs horizontal, the corridors have a rectangular section and
the chambers (rooms) are square or rectangular. Since they
were excavated in relatively friable rocks, one can easily
recognize on the walls the marks of the tools used during the
excavation, such as hatchets, pickaxes and hoes. In the same
way, the irregular subterranean spaces left by mining activity
also display numerous features and artifacts related to the
extraction of the ore. Furthermore, it should be noted that the
indigenous people who populated the Brazilian territory in
pre-colonial times did not know metallic artifacts, using only
stones, pottery and utensils made from wood (Prous, 1992).
With these instruments, the excavation of sandstones with
an advanced degree of lithifi cation, such as the sandstones
of the Botucatu Formation, is virtually impossible. As such,
the origin of the two herein presented caves as excavations by
pre-colonial people can be discarded. Recent (post-colonial)
antrophogenic action may be destructive or not. Destructive
action usually is due to treasure hunters, who dig the ground,
work the walls with tools and leave tool marks everywhere
(Figure 7A). Caves adapted for religious use (most often as
grottos of Our Lady of Lourdes) generally have their fl oor
leveled, fallen blocks have been removed and the entrance
was cleared (Figure 7B).
Worldwide, caves in quartz sandstones are very common
and so they are in the quartz sandstones of the Botucatu
Formation, in which small caves with a volume of up to a few
cubic meters are easily found. Most are located on more or less
vertical cliffs that are currently exposed to sun, rain and wind
and some of them, in past times, maybe to surfi cial drainage.
Dozens of such caves have been spotted by our team during
the search for paleomammal tunnels. These small cavities
have an irregular morphology (roughly spherical, cylindrical
or conical), extend at maximum two or three meters inside
the hills and are not related to fl owing underground waters
or to paleovertebrates. Usually it is very diffi cult to identify
traces that allow conclusive statements about their origin,
but most probably the above-cited factors must have been
responsible. These small caves and other hollows in sandstone
with obvious different genesis (talus and crevice caves, large
overhangs, cavities behind waterfalls, etc.) need not to be
discussed here.
To pinpoint the differences between large caves of
inorganic origin in quartz sandstone and the ones related to
paleovertebrates, we will discuss fi rst the inorganic cave-
forming processes and the characteristics of the caves thus
formed and, in a second moment, highlight the evidences that
indicate that the two above-described caves were produced by
paleomammals. Features of both types of caves are compared
in Table 3.
Big-sized caves in quartz sandstones are formed by a
set of processes recognized only in recent decades. These
processes closely resemble the cave-forming ones operating
in carbonate rock (e.g. Wray, 2009; Sauro, 2014). To form a
cave in quartz sandstone, the fi rst step is the dissolution of
quartz by waters that infi ltrate along intergranular boundaries,
bedding planes, joints, microfi ssures and other weaknesses
of the rock (Jennings, 1983; Wray, 1995, 1997). The same
process acts in quartzites (e.g. Briceño & Schubert, 1990).
Due to this essential precursor cave-forming process, such
caves are termed “solutional caves”. Not only is the siliceous
cement removed from the rock, but also silica from the sand
grains, etching the surface of the grains and loosening them,
turning the rock poorly cemented (e.g. Wray, 2009; Figure 8).
This process was termed “arenisation” by Martini (1979).
Once arenised, the now friable sandstone breaks down easily
and can be mechanically removed by fl owing underground
waters. After opening the fi rst passages, the water fl ow widens
these pathways forming roughly circular sub-horizontal tubes
or conduits (“karst tubes”), whose diameters range from a few
Figure 7. Recent (post-colonial) anthropogenic action in caves. A, excavations (to the left of the person) on the floor and lateral walls made
by treasure hunters in the Vale Real cave (seen from “E” to “F” of Figure 2) (observer = 1.72 m); B, adaptations made to fit a cave in volcanic
rock for religious use (Grotto of Our Lady of Lourdes in the municipality of Santa Tereza, Rio Grande do Sul, Brazil) (left person = 1.75 m).
AB

281
FRANK ET AL. – UNDERGROUND CHAMBER SYSTEMS EXCAVATED BY CENOZOIC GROUND SLOTHS
centimeters up to 1.5 m (Wray, 1995, 2009). Such tubes may
constitute a regional phenomenon (Wray, 2009), but in our
survey we have found only some isolated ones and a single
occurrence with well-developed tubes forming a network at
the base of a sandstone cliff (Pasqualon et al., 2013). Lateral
and vertical collapsing of the sandstone along the tubes widens
these conduits, then the fallen blocks disintegrate and the
loose sand is removed by the water, resulting in big-sized
caves that most often have an active water fl ow (“stream
caves”). Water flow also occurs through discontinuities
(joints, fi ssures, etc.) and following the primary structures of
the rock (bedding planes, etc.), opening these features and
promoting the collapsing of blocks. Therefore, this kind of
caves generally has highly asymmetric shapes that depend
on the lithological contacts, bedding planes and structural
pattern of the rocks (see Brazilian examples by Wernick et
al., 1973, 1977; Monteiro & Ribeiro, 2001; Spoladore &
Cottas, 2005; Robaina & Bazzan, 2008; Hardt et al., 2009).
If dry, they show very distinctive features of past drainages.
More or less common are features like incised channels
and tunnel-like lateral caves that constitute outlets similar
to soil pipes (Gunn, 2004). As a whole, superfi cial features
(grikes, drainage runnels, solutional basins, solution noches
and dolines) and underground features (caves, etc.) closely
resemble carbonate karst and, therefore, this is also called
“karst” (e.g. Shade, 2002; Self & Mullan, 2005). As a rule,
caves with this origin are easily recognized as unrelated to
the digging action of paleovertebrates.
The hypothesis that the caves were initially excavated by
mammals of the Cenozoic Megafauna is supported by several
features of the caves. The most important is the morphology of
the original chambers and walls. Both have to be distinguished
from a much larger number of surfaces that formed later due to
the collapsing of the walls and roofs, anthropogenic diggings,
the entry of subsurface waters, scratchings of cave visitors,
and other destruction features. Original features are rare,
especially in the Vale Real cave. In this cave the existence of
ellipsoidal spaces defi ned by concave surfaces (“chambers”) is
an important clue, combined with the absence of any features
related to a different genetic process. Original walls have been
found there only on a few of the concave surfaces and on a
single spot on the roof. These original features are represented
by smooth walls. Such smooth, almost polished walls have
been found in several paleomammal tunnels excavated in
sandstones, and are most likely the result of the rubbing of
the hairy back of the ground sloths against the walls and roof
during the long-term (centuries or more) usage of the tunnels
by many generations of animals (see Frank et al., 2012b, fi gs.
3C, E; Frank et al., 2013, fi g. 9C). This explains the absence
of digging scratches, which were erased by this rubbing. Only
sometimes, some scratches stay preserved on a few concave
spots of the walls (Frank et al., 2012b, fi g. 12A). The original
smooth surfaces in the Vale Real cave are usually of a much
darker color than the surfaces formed by collapsing features,
and, when ruptured, these surfaces reveal that they are formed
by a distinctive centimetric layer of sand with a different
color and lithifi cation degree (Figure 4h). When combined,
the concave surfaces with smooth walls constitute ellipsoidal
spaces that were buried by up to 50% in sand and fallen blocks
in this cave. Similar considerations apply to the Santa Cruz
do Sul cave. The front half of this cave was so thoroughly
destroyed to suit the cave for visitors that no original features
were left. On a single spot (to the right of the letter “E” in
Figure 5) there seems to remain a segment of a tunnel, but
it is almost completely fi lled with sand and rock fragments
and it was not excavated by our team. The rear half of the
cave, on the other hand, shows very well preserved concave
surfaces that outline several defi ned ellipsoidal spaces. The
Table 3. Characteristics of large caves in quartz-sandstones produced by underground waters compared to caves dug by fossorial mammals
of the Cenozoic Megafauna in South America.
Caves of inorganic origin Paleomammal burrows
Network of subsurface drainage conduits or tubes (Ø = 2-150 cm) Very common Absent or very discrete
Seeping waters entering the cave through roof and lateral walls Very common to absent Very common to absent
Speleothems (many different forms, tens of different minerals) Absent to common Absent to very discrete
Entering flowing waters at the rear end and along the lateral walls of the cave Common to absent Absent
Leaving flowing waters at the front end of the cave Sometimes absent, usually
present in large amounts Absent to present in
small amounts
Features of current or past water drainages, like incised channels Always present Absent to discrete
Declivity of the cave as a whole Slight to strong None to slight
Influence of the structural pattern of the rock on cave morphology Strong Absent
Influence of bedding planes and lithological contacts of the rock on cave morphology Strong Absent
Waters flowing from the cave into current drainages (creeks, etc.) Very common to absent Rare to absent
Smooth, almost polished roof and lateral walls Absent Common to absent
Concave lateral walls Rare to absent Common
Digging scratches on the roof and the lateral walls Absent Common to absent
Vertical parallel grooves on the lateral walls due to seeping waters from the walls Common to absent Common to absent
Collapsing features, especially on the roof Widespread Generally less developed
Associated surficial typical karst forms Absent to widespread Absent

282 REVISTA BRASILEIRA DE PALEONTOLOGIA, 18(2), 2015
uniformity of sizes of these spaces, with original heights of
approximately 1.5 m and the testimony of older people about
the former existence of low tunnels connecting the spaces, are
a strong indication of an origin related to fossorial mammals.
The few scratches found on the walls (Figure 6C) are not well
preserved, as expected, and cannot reinforce the hypothesis.
In this cave there is also a roof segment that is smooth and
dark, being original (Figure 6E).
Other clues regarding the origin of both caves as
paleomammal diggings arise from the analysis of other
aspects, the host rock being an important one. Several
hundreds of typical fossorial paleomammal tunnels have
been found excavated in the sandstones of the Botucatu
Formation, showing that the animals were able to dig into this
rock despite its toughness (Frank et. al., 2012a,b). The big
dimensions of the caves, including the size of the individual
ellipsoidal spaces, indicates that the digging animals were not
giant armadillos (Dasypodidae), which had a maximum body
weight of less than 300 kg and a body width of no more than
80 cm. The ellipsoidal spaces, with heights of around 1.5 m
or more and horizontal axes exceeding 7 m, are suggestive
of fossorial ground sloths (Xenarthra), whose body mass
exceeded 800 kg (Fariña & Vizcaíno, 1995). The location of
both caves, on hillsides close to waterways, follows the pattern
of many of these paleomammal tunnels, which always had
an entrance directed to a nearby water source (Frank et al.,
2012b, p. 153), sometimes being right besides of a waterfall.
In Santa Cruz do Sul, for example, the stream that passes in
front of the cave is so important regionally that it was used by
the fi rst settlers almost a century ago to build the fi rst water
reservoir for the emerging city.
In regard of infi ltrating (running) ground water in both
caves, it is present but in very small amounts, generating
small features on only a few spots. Near the roof of concave
surfaces “a” and “d” of the Vale Real cave; there are two rows
of small voids whose alignment follow the stratifi cation of the
sandstone (Figure 4c). These small cavities were most likely
formed by infi ltrating groundwater, but in such small amounts
that the water fl ow cannot be considered responsible for the
genesis of the cave. Similarly, in the Santa Cruz do Sul cave
there is an almost vertical fracture in the sandstone near the
entry of the cave, through which small amounts of water enter
the cave in rainy periods (Figure 5, point D). Another feature
of most of the caves produced by the action of groundwater
is a signifi cant vertical gap between the deepest portion of
the cave and the entrance, because of the downward fl ow of
the water. In both caves presented in this paper, however, the
cave fl oor is almost horizontal if we consider the probably
original concave surfaces and disregard the anthropogenic
interventions.
It is possible that pre-existing smaller caves of inorganic
origin were adapted by the fossorial paleomammals to fi t
their needs. One such situation was described by Carmo et
al. (2011). But the present state of both caves does not offer
any feature that indicates such an evolution.
As far as this was possible, the reconstruction of the
original morphology of both caves suggests a very different
pattern if compared to the common burrows of fossorial
mammals of the South American megafauna. In Argentina,
several contributions analyze burrows from the Cenozoic
Megafauna (Imbellone & Teruggi, 1988; Imbellone et
al., 1990; Quintana, 1992; Zárate et al., 1998; Vizcaíno et
al., 2001; Isla & Dondas, 2001), always emphasizing its
morphology as shorter or longer tunnels. In Brazil, among
hundreds of entirely clogged tunnels and many short open
tunnel remnants, some better-preserved tunnel-based
Figure 8. Tunnel-based systems and chamber-based systems. A, schematic hypothetic floorplan of a tunnel-based system, with some longer
tunnels connected to shorter ones, several dead-end tunnels and several entrances. Image, tunnel in the city of Erechim; B, schematic hypothetic
floorplan of a chamber-based system, with several ellipsoidal spaces connected, when necessary, by short tunnels. The system has only one
entrance. Image, last chamber in the Santa Cruz do Sul cave. Scale applies to both systems.
AB

283
FRANK ET AL. – UNDERGROUND CHAMBER SYSTEMS EXCAVATED BY CENOZOIC GROUND SLOTHS
systems were found, composed of longer (> 20 m) tunnels
with a width of around 1.4 m that may be connected to
each other by tunnels with similar or smaller diameters, but
usually without forming chambers (Frank et al., 2012b, p.
146). Many of these tunnels fi nish as dead-ends. A single
occurrence of these tunnel-based systems has shown a small
chamber, with a diameter of 1.6 m and a height of 0.7 m
(Frank & Buchmann, 2009). Using the better-preserved
distal space of the Santa Cruz do Sul cave as model, we can
reconstruct an excavation with a very different morphology:
instead of interconnected long tunnels with occasionally
small chambers (Figure 8A), we obtain a system of large
ellipsoidal chambers connected, when necessary, by short
and low tunnels (Figure 8B). Converging chambers in
these chamber-based systems may constitute huge caves
whose walls are all formed by concave surfaces in several
orientations. Since fossorial animals always produce a
defi ned burrow pattern, the strong contrast between the
tunnel-based systems and the chamber-based systems
suggests that each one applies to a specifi c fossorial species,
most probably to two species of ground sloths because of
the huge size of the tunnels and chambers.
When we consider the frequency of both systems, another
conclusion seems possible. Regarding the tunnel-based
systems, our team has found approximately 1,500 tunnels
in the state of Rio Grande do Sul until now. In some regions
with favorable rock and relief, each hill seems to host one
or more tunnel complexes. The contrast between these very
abundant tunnel-based systems and the only two chamber-
based systems most probably means that the sloth species
responsible for the chambers was much less abundant than the
sloth species that excavated the tunnels. Certainly there are
tens of thousands of paleomammal excavations still hidden
in the hills and mountains, but this contrast in abundance
between the two systems found until now will probably be
maintained even after more fi ndings in the future.
CONCLUSIONS
The caves of Santa Cruz do Sul and of Vale Real, which
are both large and hosted in sandstone, do not show any
evidence of an origin related to underground fl owing waters,
any other inorganic process or human pre-colonial action. The
still original characteristics in both caves had to be identifi ed
among a lot of features produced by infi ltrating rainwater,
collapsing walls and human action, and this search revealed
only some smooth walls and smooth parts on the roofs, in
addition to a few grooves on the lateral walls that are probably
excavation marks. Combining these characteristics with the
general original morphology of the caves, especially with the
ellipsoidal spaces located at the walls, it is highly probable that
both caves were originally underground shelters excavated by
fossorial mammals of the Cenozoic megafauna, most probably
ground sloths due to the size of the caves.
After more than a century of public access to the caves,
with hundreds of visitors every year, it is very unlikely that
any direct evidence of the excavators can be found, such as
bones, hair or claw fragments. The assignment of the caves
to fossorial mammals is also supported by existence of
several dozens of better preserved big sized tunnels in both
southernmost states of Brazil (Rio Grande do Sul and Santa
Catarina) and a great number of smaller tunnels, which show
that the fossorial habits of some of the megafauna mammals
were commonplace in Southern Brazil during the Cenozoic.
ACKNOWLEDGMENTS
We are indebted to the team of the city hall of Santa Cruz
do Sul for their license to work inside the “Indian Cave” and
to L.A. Massochini, whose property hosts the Vale Real cave,
for his support during our fi eldwork there. Many thanks to V.
Krapovickas and to two anonymous referees for their comments
that greatly improved the manuscript.
REFERENCES
Bergqvist, L.P. & Maciel, L. 1994. Icnofósseis de Mamíferos
(Crotovinas) na Planície Costeira do Rio Grande do Sul, Brasil.
Anais da Academia Brasileira de Ciências, 66:189-197.
Briceño, H.O. & Schubert, C. 1990. Geomorphology of the Gran
Sabana, Guyana Shield, southeastern Venezuela. Geomorphology,
3:125-141. doi:10.1016/0169-555x(90)90041-n
Buchmann, F.S.C.; Lopes, R.P. & Caron, F. 2009. Icnofósseis
(paleotocas e crotovinas) atribuídos a Mamíferos extintos no
Sudeste e Sul do Brasil. Revista Brasileira de Paleontologia,
12:247-256. doi:10.4072/rbp.2009.3.07
Carmo, F.F.; Carmo, F.F.; Buchmann, F.S.C.; Frank, H.T. & Jacobi,
C.M. 2011. Primeiros registros de paleotocas desenvolvidas em
formações ferríferas, Minas Gerais, Brasil. In: CONGRESSO
BRASILEIRO DE ESPELEOLOGIA, 31, 2011. Resumos, Ponta
Grossa, SBE, p. 531-540.
Fariña, R.A. & Vizcaíno, S.F. 1995. Hace solo diez mil años. 6th ed.
Montevideo, Editorial Fin de Siglo, 123 p.
Frank, H.T. & Buchmann, F.S.C. 2009. A resting chamber in a
large palaeovertebrate underground shelter. In: ENCONTRO
BRASILEIRO DE ESTUDOS DO CARSTE 2009. Anais, São
Carlos, CD-ROM.
Frank, H.T.; Buchmann, F.S.C.; Lima, L.G.; Fornari, M.; Caron, F.
& Lopes, R.P. 2012a. Cenozoic vertebrate tunnels in Southern
Brazil. In: R.G. Netto; N.B. Carmona & F.M.W. Tognoli
(eds.) Ichnology of Latin America - selected papers, Porto
Alegre, Sociedade Brasileira de Paleontologia, p. 141-157
(Monografi as 2).
Frank, H.T.; Lima, L.G.; Gerhard, N.P.; Caron, F.; Buchmann, F.S.C.;
Fornari, M. & Lopes, R.P. 2013. Description and interpretation
of Cenozoic vertebrate ichnofossils in Rio Grande do Sul
State, Brazil. Revista Brasileira de Paleontologia, 16:83-96.
doi:10.4072/rbp.2013.1.0
Frank, H.T.; Oliveira, L.D.; Vicroski, F.N.; Breier, R.; Pasqualon,
N.G.; Araújo, T.; Buchmann, F.S.C.; Fornari, M.; Lima, L.G.;
Lopes, R.P. & Caron, F. 2012b. The complex history of a
sandstone-hosted cave in the State of Santa Catarina, Brazil.
Espeleo-Tema, 23:87-101.
Goudie, A.S. 2013. Arid and semi-arid geomorphology. 1st ed.
Cambridge, Cambridge University Press, 461 p. doi:10.1017/
CBO9780511794261
Gunn, J. 2004. Encyclopedia of caves and karst science. New York,
Taylor and Francis Group, 1940 p.

284 REVISTA BRASILEIRA DE PALEONTOLOGIA, 18(2), 2015
Hardt, R.; Rodet, J.; Pinto, S.A.F. & Willems, L. 2009. Exemplos
brasileiros de carste em arenito: Chapada dos Guimarães (MT)
e Serra de Itaqueri (SP). Espeleo-Tema, 20:7-23.
Imbellone, P. & Teruggi, M. 1988. Sedimentación crotovinica
en secuencias cuaternarias bonaerenses. In: REUNIÓN
ARGENTINA DE SEDIMENTOLOGIA, 2, 1988. Anais, Salta,
ARS, p. 125-129.
Imbellone, P.; Teruggi, M. & Mormeneo, L. 1990. Crotovinas
en sedimentos cuaternarios del partido de La Plata.
In: INTERNATIONAL SYMPOSIUM ON LOESS -
CADINQUA, 1, 1990. Anais, Mar del Plata, p. 166-172.
Isla, F.I. & Dondas, A. 2001. Facies fl uviales del Pleistoceno de
Mar Del Plata, Argentina. Revista de la Asociación Geológica
Argentina, 56:259-267.
Jennings, J.N. 1983. Sandstone pseudokarst or karst? In: R.W.
Young & G.C. Nanson (eds.) Aspects of Australian sandstone
landscapes. Bundanoon, Australian and New Zealand
Geomorphology Group, p. 21-30 (Special Publication 1).
Kostof, S. 1989. Caves of god: Cappadocia and its churches. Oxford,
Oxford University Press, 350 p.
Margielewski, W. & Urban, J. 2003. Crevice-type caves as
initial forms of rock landslide development in the Flysch
Carphatians. Geomorphology, 54:325-338. doi:10.1016/S0169-
555X(02)00375-6.
Martini, J.E.J. 1979. Karst in black reef quartzite near Kaapsehoop,
Easters Transvaal. Annales of the South African Geological
Survey, 13:115-128.
Milani, E.J.; Faccini, U.F.; Scherer, C.M.S.; Araújo, L.M. &
Cupertino, J.A. 1998. Sequences and stratigraphic hierarchy of
the Paraná Basin (Ordovician to Cretaceous), Southern Brazil.
Boletim IG-USP, Série Científi ca, 29:125-173.
Monteiro, R.C. & Ribeiro, L.F.B. 2001. Espeleogênese em cavernas
areníticas: algumas considerações aplicada à Província
Espeleológica da Serra de Itaqueri, Estado de São Paulo, Brasil.
In: BRAZILIAN CONGRESS OF SPELEOLOGY, 26, 2001.
Anais, SBE, Brasília, p. 138.
Osborne, R.A.L.; Weliange, W.S.; Jayasingha, P.; Dandeniya, A.S.;
Algiriya, A.K.P. P. & Pogson, R.E. 2013. Caves and karst-like
features in Proterozoic gneiss and Cambrian granite, Southern
and Central Sri Lanka: an introduction. Acta Carsologica, 42:25-
48. doi:10.3986/ac.v42i1.630
Palmer, A.N. 1991. Origin and morphology of limestone
caves. Geological Society of America Bulletin, 103:1-21.
doi:10.1130/0016-7606(1991)103<0001:OAMOLC>2.3.CO;2
Pasqualon, N.G.; Frank, H.T.; Silva, M.Y.L.; Althaus, C.E.;
Almeida, M.L.; Breier, R.; Oliveira, D.M. & Bischoff, L. 2013.
Pseudopaleotocas em arenitos da Formação Botucatu (Jsup-
Kinf): estudo de caso. In: Congresso Brasileiro de Paleontologia,
23, 2013. Boletim de Resumos, Gramado, SBP, p. 311.
Peate, D.W. 1997. The Paraná-Etendeka province. In: J.J. Mahoney &
M. Coffi n (eds.) Large igneous provinces. Washington, American
Geophysical Union, p. 217-254 (Geophysical Monographs 100).
Peterson, D.W.; Holcomb, R.T.; Tilling, R.I. & Christiansen, R.L.
1994. Development of lava tubes in the light of observations at
Mauna Ulu, Kilauea Volcano, Hawaii. Bulletin of Volcanology,
56:343-360. doi:10.1007/bf00326461
Prous, A. 1992. Arqueologia Brasileira. Brasília, Editora da UnB,
607 p.
Quintana, C.A. 1992. Estructura interna de una paleocueva,
posiblemente de un Dasypodidae (Mammalia, Edentata) del
Pleistoceno de Mar del Plata (Provincia de Buenos Aires,
Argentina). Ameghiniana, 29:87-91.
Robaina, L.E.S. & Bazzan, T. 2008. Feições cársticas em rochas
siliciclásticas no Oeste do Estado do Rio Grande do Sul. Revista
Brasileira de Geomorfologia, 9:53-64.
Ruchkys, U.A.; Bittencourt, J.S. & Buchmann, F.S.C. 2014.
A paleotoca da Serra do Gandarela e seu potencial como geossítio
do Geoparque Quadrilátero Ferrífero, Minas Gerais. Caderno
de Geografi a, 24:249-263.
Sauro, F. 2014. Structural and lithological guidance on
speleogenesis in quartz-sandstone: evidence of the arenisation
process. Geomorphology, 226:106-123. doi:10.1016/j.
geomorph.2014.07.033
Scherer, C.M.S. 2000. Aeolian dunes of the Botucatu Formation
(Cretaceous) in southernmost Brazil: morphology and origin.
Sedimentary Geology, 137:63-84. doi:10.1016/S0037-
0738(00)00135-4
Self, C.A. & Mullan, G.J. 2005. Rapid karst development in an
English quartzitic sandstone. Acta Carsiologica, 34:415-424.
Shade, B.L. 2002. The genesis and hydrogeology of a sandstone
karst in Pine County, Minnesota. University of Minnesota,
Ph.D. thesis, 171 p.
Sjöberg, R. 1987. Caves as indicators of neotectonics in Sweden.
Zeitschrift für Geomorphologie, 63:141-148.
Spoladore, A. & Cottas, L.R. 2005. A Gruta do Portão de Cima e a
Gruta do Portão de Baixo: duas cavernas areníticas no município
de Sengés - PR. Geografi a, 14:71-83.
Vizcaíno, S.F.; Zárate, M.; Bargo, S.M. & Dondas, A. 2001.
Pleistocene burrows in the Mar Del Plata area (Argentina)
and their probable builders. Acta Palaeontologica Polonica,
46:289-301.
Waterstrat, W.J.; Mylroie, J.E.; Owen, A.M. & Mylroie, J.R. 2010.
Coastal caves in Bahamian eolian calcarenites: differentiating
between sea caves and fl ank margin caves using quantitative
morphology. Journal of Cave and Karst Studies, 72:61-74.
doi:10.4311/jcks2009es0086.
Wernick, E.; Pastore, E.L. & Pires Neto, A. 1973. Cavernas em
arenito. Notícia Geomorfológica, 13:55-67.
Wernick, E.; Pastore, E.L. & Pires Neto, A. 1977. Cuevas en
areniscas, Rio Claro, Brasil. Boletín de la Sociedad Venezolana
de Espeleología, 8:1699-1707.
Wimmer, H. 2000. Die Regional-typisierung der Erdställe. Der
Erdstall, 26:54-56.
Wray, R.A.L. 1995. Solutional landforms in quartz sandstones of the
Sydney Basin. School of Geosciences (Geography), University
of Wollongong, Ph.D. thesis, 381 p.
Wray, R.A.L. 1997. A global review of solutional weathering forms
on quartz sandstones. Earth-Science Reviews, 42:137-160.
doi:10.1016/s0012-8252(96)00056-6
Wray, R.A.L. 2009. Phreatic drainage conduits within quartz
sandstone: evidence from the Jurassic Precipice Sandstone,
Carnarvon Range, Queensland, Australia. Geomorphology,
110:203-211. doi:10.1016/j.geomorph.2009.04.007
Zalán, P.V.; Wolff, S.; Conceição, J.C.J.; Marques, A.; Astolfi , M.A.M.;
Vieira, I.S.; Appi, V.T. & Zanotto, O.A. 1990. Bacia do Paraná.
In: G.P. Raja Gabaglia; E.J. Milani (eds.) Origem e evolução de
bacias sedimentares, Petrobrás/Serec/Cen-Sud, p. 135-168.
Zárate, M.A.; Bargo, M.S.; Vizcaíno, S.F.; Dondas, A. & Scaglia,
O. 1998. Estructuras biogênicas en el Cenozoico tardio de Mar
del Plata (Argentina) atribuibles a grandes mamíferos. Revista
da Asociación Argentina de Sedimentologia, 5:95-103.
Received in October, 2014; accepted in May, 2015.