Gondwana Research 27 (2015) 911–924
Contents lists available at ScienceDirect
Gondwana Research
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / g r
Vegetation history across the Permian–Triassic boundary in Pakistan
(Amb section, Salt Range)☆
Elke Schneebeli-Hermann a,e,⁎, Wolfram M. Kürschner b, Hans Kerp c, Benjamin Bomfleur d, Peter A. Hochuli e,
Hugo Bucher e, David Ware e, Ghazala Roohi f
a Palaeoecology, Institute of Environmental Biology, Faculty of Science, Laboratory of Palaeobotany and Palynology, Utrecht University, Budapestlaan 4, 3584 CD Utrecht, The Netherlands
b Department of Geosciences, University of Oslo, P.O. Box 1047, Blindern, N-0316 Oslo, Norway
c Forschungsstelle für Paläobotanik am Institut für Geologie und Paläontologie, Westfälische Wilhelms-Universität Münster, Hindenburgplatz 57, 48143 Münster, Germany
d Department of Palaeobotany, Swedish Museum of Natural History, P.O. Box 50007, SE-104 05 Stockholm, Sweden
e Institute and Museum of Palaeontology, University of Zurich, Karl Schmid-Str. 4, CH-8006 Zurich, Switzerland
f Pakistan Museum of Natural History, Garden Avenue, Islamabad 44000, Pakistan
a r t i c l e
i n f o
a b s t r a c t
Article history:
Received 27 June 2013
Received in revised form 1 November 2013
Accepted 7 November 2013
Available online 15 December 2013
Hypotheses about the Permian–Triassic floral turnover range from a catastrophic extinction of terrestrial plant
communities to a gradual change in floral composition punctuated by intervals indicating dramatic changes in
the plant communities. The shallow marine Permian–Triassic succession in the Amb Valley, Salt Range,
Pakistan, yields palynological suites together with well-preserved cuticle fragments in a stratigraphically well-
constrained succession across the Permian–Triassic boundary. Palynology and cuticle analysis indicate a mixed
Glossopteris–Dicroidium flora in the Late Permian. For the first time Dicroidium cuticles are documented from
age-constrained Upper Permian deposits on the Indian subcontinent. Close to the Permian–Triassic boundary,
several sporomorph taxa disappear. However, more than half of these taxa reappear in the overlying Smithian
to Spathian succession. The major floral change occurs towards the Dienerian. From the Permian–Triassic bound-
ary up to the middle Dienerian a gradual increase of lycopod spore abundance and a decrease in pteridosperms
and conifers are evident. Synchronously, the generic richness of sporomorphs decreases. The middle Dienerian
assemblages resemble the previously described spore spikes observed at the end-Permian (Norway) and in
the middle Smithian (Pakistan) and might reflect a similar ecological crisis.
© 2013 International Association for Gondwana Research. Published by Elsevier B.V. All rights reserved.
Keywords:
Dicroidium
Glossopteris
Pakistan
Permian–Triassic
Vegetation turn-over
1. Introduction
Whereas in marine environments the extent and chronological
course of events of the end-Permian mass extinction has been studied
in great detail over recent decades (e.g. Raup and Sepkoski, 1982;
Bowring et al., 1999; Jin et al., 2000; Benton and Twitchett, 2003; Haas
et al., 2004; Groves et al., 2007; J. Chen et al., 2011), the impact of the
mass extinction on continental vegetation remains poorly resolved
(e.g. Knoll, 1984; Rees et al., 2002; Bamford, 2004; Utting et al., 2004;
Hochuli et al., 2010; Xiong and Wang, 2011).
Despite the unresolved causal mechanism, terrestrial ecosystems
changed significantly across the Permian–Triassic boundary. Terrestrial
vertebrates, e.g. therapsids (“mammal-like reptiles”) were severely
affected by the Permian–Triassic extinction (Kemp, 2005). Several
lineages became extinct; others such as the dicynodonts were reduced
and recovered in the Middle Triassic (Benton et al., 2004; Ward et al.,
2005; Fröbisch, 2008). In contrast, skull morphology of the Cynodontia
☆ This article belongs to the Special Issue on Gondwanan Mesozoic biotas and bioevents.
⁎ Corresponding author: Tel.: +31 30 253 26 47; fax: +31 30 253 50 96a.
E-mail addresses: elke.schneebeli@pim.uzh.ch, ElkeSchneebeli@gmx.net
(E. Schneebeli-Hermann).
shows no significant change across the Permian–Triassic boundary,
but changes significantly only in the late Olenekian–Anisian (Abdala
and Ribeiro, 2010). A turnover in palaeosol characteristics has been
documented from Antarctica in association with the changes in terres-
trial vertebrates. Late Permian palaeosols are coal-bearing and coarse-
grained compared to the green–red mottled claystones of Early Triassic
age. This change has been interpreted to reflect a climatic shift to warmer
climates in the Early Triassic (Retallack and Krull, 1999).
Floral records of Late Permian and Early Triassic age have been used
as a proxy for migration pathways of newly evolved taxa such as
Dicroidium (Kerp et al., 2006; Abu Hamad et al., 2008). Dicroidium
apparently evolved in the Late Permian of the Palaeotropics (Jordan)
and migrated to higher southern latitudes probably in association
with climatic changes during the Triassic (Kerp et al., 2006; Abu
Hamad et al., 2008). However, fossil floral records have also been used
as indicators for other environmental signals (stress factors) during
the end-Permian biodiversity disruption. The abundant occurrence of
unseparated spore tetrads close to the Permian–Triassic boundary has
been interpreted to reflect a depletion of the ozone layer (e.g. Visscher
et al., 2004; Beerling et al., 2007). As suggested by these authors the
ozone layer depletion led to increasing UV-B radiation and mutagenesis
in spores, which lost their ability to separate. Another conspicuous
1342-937X/$ – see front matter © 2013 International Association for Gondwana Research. Published by Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.gr.2013.11.007
912
E. Schneebeli-Hermann et al. / Gondwana Research 27 (2015) 911–924
feature of fossil floral records is brief intervals in which pteridophyte
spores became proportionally very abundant. Instead of a long-lasting
loss of standing woody biomass (e.g. Looy et al., 1999), palynological
data from Norway indicate a rapid succession of spore dominance
and immediate recovery of gymnosperms (Hochuli et al., 2010).
Reviews of floral records have revealed that faunal mass extinctions
are commonly associated with instabilities of terrestrial ecosystems
(McElwain and Punyasena, 2007). Such short-term changes, so-called
spore spikes or spore peaks, have been observed at the Permian–Triassic
boundary (Stemmerik et al., 2001; Hochuli et al., 2010) and during
the middle Smithian (Hermann et al., 2011a). Repeated high spore
relative abundances close to the Triassic–Jurassic boundary have been
interpreted to reflect the vegetation’s reaction to environmental changes
induced by the Central Atlantic magmatic province. The volcanism
caused climatic gradients that led to compositional changes of the
regional vegetation (Götz et al., 2009; Bonis and Kürschner, 2012). A
distinct fern spike has also been associated with the Cretaceous–
Paleogene boundary (Vajda et al., 2001) and compared with the
Permian–Triassic floral succession (Vajda and McLoughlin, 2007). In
all mentioned instances high spore abundances are associated with
environmental change and faunal extinction events.
The end-Permian short-term floral changes reported from the
Northern Hemisphere called for a detailed study of the vegetation history
across the Permian–Triassic boundary in the Southern Hemisphere,
i.e. of a Gondwanan record. The Amb Valley section in the Salt Range
(Pakistan), which includes the Chhidru Formation and the Mianwali
Formation, offers the opportunity to evaluated palynological data
together with well-preserved cuticle fragments to describe the floral
succession in a well-constrained temporal framework across the
Permian–Triassic boundary.
2. Geological and palaeogeographic setting
The Amb Valley is located in the Salt Range (Pakistan), a low moun-
tain range SSW of Islamabad (Fig. 1B). It is one of numerous valleys that
yield fine exposures of the Permian–Triassic marine sedimentary
succession in this area. During the Late Permian and Early Triassic, the
Salt Range area was part of the southern Tethyan shelf of the Indian sub-
continent (Northern Indian Margin); (Fig. 1A) (e.g. Pakistani-Japanese
Research Group, 1985; Smith et al., 1994; Golonka and Ford, 2000).
The Amb section is located ~ 20 km SE of Nammal and ~5 km S of the
Sakesar mountain (Fig. 1C).
The uppermost part of the Chhidru Formation and the lowermost
part of the Mianwali Formation were sampled for the present study.
The uppermost part of the Chhidru Formation was informally named
the “white sandstone unit” by Kummel and Teichert (1970). At Amb,
the white sandstone unit consists of a 9 m thick succession of alternating
white to grey, medium-grained sandstone and dark grey siltstone. In
the Salt Range area, the upper part of the Chhidru Formation is of late
Changhsingian age based on conodont biostratigraphy (Wardlaw and
Mei, 1999; Mei et al., 2002; Shen et al., 2006) and chemostratigraphic
correlations of carbon isotope data with the GSSP of Meishan, South
China (Schneebeli-Hermann et al., 2013).
The contact between the Chhidru Formation and the overlying
Mianwali Formation represents an erosional unconformity interpreted
as a sequence boundary (Mertmann, 2003; Hermann et al., 2011b)
and representing a temporal hiatus between deposition of the two
formations. The extinction of marine biota has been described to
coincide with the formational boundary between the two formations
(Schindewolf, 1954). The overlying Mianwali Formation has been
subdivided into the Kathwai Member, the Mittiwali Member, and
the Narmia Member. The present study deals with the basal part
of the Mianwali Formation, including the Kathwai Member and
the basal part of the Mittiwali Member, namely the Lower Ceratite
Limestone and the lowermost
the Ceratite Marls
(e.g. Waagen, 1895; Kummel and Teichert, 1970; Guex, 1978; Hermann
et al., 2011b).
interval of
The position of the Permian–Triassic boundary, as defined by the
first occurrence of the conodont species Hindeodus parvus (Yin et al.,
2001) is ambiguous in the Salt Range area, partly because of the
diachronicity of lithological boundaries (e.g. Hermann et al., 2011b;
Brühwiler et al., 2012). At Nammal, the Pakistani-Japanese Research
Group (1985) divided the Kathwai Member into three units and placed
the Permian–Triassic boundary in the middle unit, whereas Wardlaw
and Mei (1999) documented the occurrence of H. parvus near the base
of the Kathwai Member in the Salt Range area (without mentioning
the exact locality); in some successions the lowermost unit of the
Kathwai Member is not preserved (Mertmann, 2003). At Nammal, the
negative carbon isotope spike marking the Permian–Triassic boundary
sections worldwide occurs in the lowermost part of the Kathwai
Member (Baud et al., 1996). Therefore, we use the formational boundary
between the Chhidru Formation and the Mianwali Formation as
an approximation for the Permian–Triassic boundary at Amb. The
overlying Lower Ceratite Limestone is of early Dienerian to earliest
middle Dienerian age (Ware et al., 2010, 2011). Ammonoids recovered
from the Ceratite Marls indicate middle Dienerian to early Smithian
ages; the lowermost 2 m included in this study are of middle Dienerian
age (Ware et al., 2010, 2011).
A
71°00’
71°30’
72°00’
Narmia
Narmia
33°00’
Landu
Surghar
Range
Daud Khel
Afghanistan
ISLAMABAD
32
Pakistan
33°00’
India
b
I s l a m a
a
d
24
B
50 km
Arabian Sea
64
72
0
Nammal
Salt Range
Salt Range
Sakesar
SakSakesae r
Amb
32°30’
72°00’
C
Salt Range
s
u
d
In
Mianwali
North-West Frontier
Province
32°30’
71°00’
Punjab
71°30’
landmass
mountains
Province boundary
Railways
Main roads
Fig. 1. A: Early Triassic palaeogeographic position of the Salt Range (after Smith et al., 1994 and Golonka and Ford, 2000). B: Location of the Salt Range and Surghar Range in Pakistan.
C: Location of the Amb valley in the Salt Range.
E. Schneebeli-Hermann et al. / Gondwana Research 27 (2015) 911–924
913
Table 2
Stratigraphic distribution of identified cuticle types in the Amb Valley section.
Sample
Lepidopteris
Dicroidium spp.
Neoggerathiopsis
Glosspteris/
Gangamopteris
3. Methods
Thirty samples were collected from fine-grained siliciclastic intervals
of the white sandstone unit (Chhidru Formation) and the basal part of
the overlying Mianwali Formation in the Amb section. The samples
were crushed and weighed (5–25 g) and subsequently treated with
hydrochloric and hydrofluoric acid according to standard palynological
preparation techniques (Traverse, 2007). A brief oxidation with
nitric acid was performed before the residues were sieved using an
11 μm mesh screen. A minimum of 250 spores and pollen grains per
sample were counted from strew mounts. Spores and pollen grains
were grouped and classified according to their botanical affinities to
aid interpretation of the vegetation history (after Balme, 1995;
Lindström et al., 1997; Taylor et al., 2006; Traverse, 2007); (Table 1).
Cuticle fragments were abundant in samples AMB 34, AMB 37,
and AMB 45 and were picked and mounted on separate slides for
identification. Palynological slides were also screened for identifi-
able cuticles.
A diversity analysis was performed on the qualitative sporomorph
datasets from the Amb Valley section using PAST (Hammer et al.,
2001). The number of pollen grains and spore genera per sample was
calculated (generic richness). Additionally, the range-through diversity
was determined, in which absences between the first and last occur-
rences were treated as the presence. For comparison, the generic rich-
ness and spore/pollen ratios of the previously described palynological
record from the Narmia Valley were calculated and illustrated
(Hermann et al., 2011a, 2012). Samples are stored in the repository of
the Palaeontological Institute and Museum of the University of Zurich
(PIMUZ repository numbers A/VI 65 and A/VI 66).
4. The cuticle record
AMB 103
AMB 26
AMB 102
AMB 120
AMB 101
AMB 24
AMB 48
AMB 47
AMB 46
AMB 45
AMB 44
AMB 43
AMB 42
AMB 41
AMB 40
AMB 21
AMB 39
AMB 20
AMB 38
AMB 37
AMB 49
AMB 36
AMB 35
AMB 34
AMB 33
AMB 32
AMB 31
AMB 30
AMB 29
AMB 28
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
Cuticle fragments were picked from palynological residue samples of
AMB 34, AMB 37, and AMB 45, and palynological slides of other samples
were scanned for additional cuticle occurrences (see Table 2). The main
categories that could be distinguished are: cuticles of Glossopteris/
Gangamopteris type (Fig. 2); the cordaitalean Noeggerathiopsis (Fig. 3A,
B), cuticle fragments of the peltasperm Lepidopteris (Fig. 3C), and
cuticles of the corystosperm Dicroidium (Fig. 3D–I). Furthermore cuticles
of other indeterminate plant groups (Fig. 3K, L) were encountered.
The identified cuticle types are described below and their strati-
graphic distribution is indicated (Table 2).
Table 1
Botanical affinities of relevant sporomorph taxa (after Balme, 1995; Lindström et al., 1997; Taylor et al., 2006; Traverse, 2007).
Bryophytes & Pteridophytes undiff.
Pteridophytes
Ferns
Gymnosperms
Corystospermales, Caytoniales,
Peltaspermales
Glossopteridales
Lycopodiopsida
Equisetopsida
Gnetopsida
Cycadopsida
Pteridospermae and
probable seed ferns
Conifers
Conifers + Pteridospermae
Gymnosperms undiff
Spores undiff and spores of uncertain affinity such as: Didecitriletes
spp., Limatulasporites spp., Lunulasporites spp., Playfordiapora
spp., Punctatisporites spp., Punctatosporites spp., and Triplexisporites spp.
Acanthotriletes spp., Apiculatisporites spp., Baculatisporites spp.,
Convolutisporites spp., Dictyophyllidites spp., Grandispora spp.,
Granulatisporites spp., Horriditriletes spp., Laevigatosporites spp.,
Leiotriletes spp., Lophotriletes spp., Osmundacidites spp.,
Polypodiisporites spp., Triquitrites spp., Verrucosisporites spp.
Endosporites papillatus
Densoisporites spp.
Kraeuselisporites spp.
Lundbladispora spp.
Calamospora spp.
Ephredipites spp., Gnetaceaepollenites spp.
Cycadopites spp., Pretricolpipollenites spp.
Falcisporites spp., Vitreisporites spp., Weylandites spp.
Protohaploxypinus spp., Striatopodocarpites spp.
Bisaccates taen
Chordasporites spp., Florinites spp., Klausipollenites spp., Lueckisporites
spp., Pinuspollenites spp., Platysaccus spp., Protodiploxypinus
spp., Sulcatisporites spp.
Bisaccates undiff
Lunatisporites spp.
Alisporites spp.
Bisaccates non-taen
Cordaitina spp., Corisaccites spp., Densipollenites spp., Guttulapollenites
hannonicus, Inaperturopollenites spp., Marsupipollenites spp., undiff
monosaccate and monosulcate pollen
914
E. Schneebeli-Hermann et al. / Gondwana Research 27 (2015) 911–924
A
B
C
D
50 µm
50 µm
50 µm
50 µm
Fig. 2. Glossopteris/Gangamopteris cuticle types from the Amb section. Sample number is followed by England Finder coordinates. PIMUZ repository number A/VI 65: A: AMB 34a, Q14/3–4.
B: AMB 40a E27/1–3. C: AMB 30a, N7/0. PIMUZ repository number A/VI 66: D: AMB 34C, Z11/1–3.
4.1. Cuticle type 1 (Glossopteridales) (Fig. 2)
4.1.1. Description
Cuticle moderately thick; epidermal cell pattern differentiated into
costal and intercostal fields (Fig. 2A, C). Cells irregularly arranged and
with variable outlines in intercostal fields (Fig. 2A), tending to longitu-
dinal alignment and elongation in costal fields (Fig. 2B, upper part of
Fig. 2C, D); anticlinal cell walls curved to highly sinuous (Fig. 2A, B).
Outer cuticle surface smooth or with a characteristic microstructure
composed of evenly distributed, small, solid papillate projections with
a density of about five to more than 20 per cell (Fig. 2A–C).
4.1.2. Remarks
Cuticle and epidermal characters of the Glossopteridales have been
studied extensively (see, e.g. Zeiller, 1896; Sahni, 1923; Pant, 1958;
Chandra, 1974; Pant and Gupta, 1968; Pant and Singh, 1968; Singh
and Maheshwari, 2000); they show a remarkable variability between
genera and species, but also within individual species (e.g. Surange
and Srivastava, 1956; Chandra, 1974; Singh, 2000). It is, therefore,
difficult to accurately delimit this plant group based on cuticle and
epidermal features alone (e.g. Singh, 2000). However, some species of
including Glossopteris colpodes, Glossopteris fibrosa,
Glossopteris,
Glossopteris hispida, Glossopteris petiolata, Glossopteris tenuifolia
and Glossopteris waltonii, possess cuticles that are characterised by a
particular surface microstructure of numerous small, solid papillae,
as described above (see, e.g. Pant, 1958; Pant and Gupta, 1968;
Maheshwari and Tewari, 1992). Similar structures have also been
described to occur on the cuticle of some Gangamopteris species
(Srivastava, 1957). This particular type of ornamentation is not known
to occur in other plant groups of that time. Together with the character-
istically sinuous anticlinal walls of epidermal cells, it forms a reliable
diagnostic feature for at least certain species of Glossopteris and
Gangamopteris.
4.2. Cuticle type 2 (Noeggerathiopsis) (Fig. 3A, B)
4.2.1. Description
Cuticle thick. Epidermal cells arranged in nearly regular longitudinal
files, oriented longitudinally; epidermis differentiated into alternating,
parallel longitudinal rows of stomata-bearing and stomata-free zones.
Epidermal cells in stomata-free (costal) rows longitudinally elongate,
with straight or slightly curving anticlinal walls, and smooth anticlinal
wall cutinisations; periclinal walls smooth or with up to three diffuse
papilla-like thickenings with circular outline. Epidermal cells in
stomata-bearing (intercostal) rows with overall similar features but
smaller, less elongate, and more variably oriented. Stomata occurring
in one or several ill-defined longitudinal files per intercostal row, each
with about five to eight (usually seven) subsidiary cells that are similar,
but more heavily cutinised than surrounding regular epidermal cells;
subsidiary cells lacking papillae; stomatal aperture narrow oval or
rectangular, longitudinally oriented; guard cells superficial or only little
sunken.
4.2.2. Remarks
The material agrees well with the epidermal and cuticular structure
of some species of Noeggerathiopsis (e.g. Lele and Maithy, 1964; Pant
and Verma, 1964); the presence of only few, ill-defined papillae
and more or less superficial guard cells show close similarity to the
cuticles of Noeggerathiopsis bunburyana (Pant and Verma, 1964) and
Noeggerathiopsis hislopii of Zeiller (1896).
It is important to note, however, that other species of Noeggerathiosis
from Gondwana, including forms that are known in remarkable
detail based on anatomically preserved material from Antarctica
(McLoughlin and Drinnan, 1996), show a very different epidermal mor-
phology with trichome-lined stomatal grooves. Similar epidermal struc-
tures are characteristic features also of the rufloriaceaen Cordaitales
typical of Angara (see, e.g., Gluchova, 2009). Indeed, we encountered a
few cuticle fragments in sample AMB 34 that show such trichome-
lined depressions, possibly stomatal chambers or grooves, in the
epidermal surface (Fig. 3J). Even though these specimens are only
small fragments, they appear remarkably similar to the cuticles of
Rufloria gondwanensis Guerra Sommer from the Permian of Rio Grande
do Sul (Guerra Sommer, 1989), which together with co-occurring
Nephropsis-type bracts (Corrêa da Silva and Arrondo, 1977) form the
so far only known Gondwanan representatives of rufloriaceaen
Cordaitales. Although we cannot identify these dispersed cuticle frag-
ments as belonging to the Rufloriaceae with certainty at present, it
may be possible that the composition of cordaitalean plants in the
Permian peri-Tethyan realm was more complex than previously
thought, combining typical Gondwanan Noeggerathiopsis plants as
well as additional taxa that are usually considered characteristic
elements of the Angaran vegetation.
4.3. Cuticle type 3 (Lepidopteris) (Fig. 3C)
4.3.1. Description
Cuticles moderately thick to thick. Epidermal cells arranged
irregularly and without preferred orientation, small, of relatively
uniform size and shape, polygonal with mostly four to six sides,
with straight anticlinal walls and smooth, even anticlinal wall
cutinisation; each regular epidermal cell bearing a distinct, centrally
positioned, hemispherical, solid papillate thickening. Stomatal
complexes evenly distributed, irregularly oriented, (sub)circular in out-
line and essentially radially symmetrical, with four to seven (usually
five or six) subsidiary cells that are similarly or less cutinised than
surrounding regular epidermal cells; subsidiary-cell papillae positioned
close to and usually overarching the stomatal pit; guard cells conspicu-
ously sunken.
E. Schneebeli-Hermann et al. / Gondwana Research 27 (2015) 911–924
915
B
50 µm
50 µm
50 µm
50 µm
20 µm
C
F
I
20 µm
E
H
50 µm
20 µm
20 µm
K
L
A
A
D
G
J
50 µm
50 µm
20 µm
Fig. 3. Cuticles from the Amb section. Identified cuticle is followed by sample number and England Finder coordinates. All PIMUZ repository number A/VI 66. A: Noeggerathiopsis sp. AMB
34CO, P26/4. B: Noeggerathiopsis sp. AMB 34CO, T26/0. C: Lepidopteris sp., AMB 24 2.Präp. a, H21/4. D: Dicroidium sp., AMB 34C, S22/4. E, F, I: Dicroidium sp., AMB 45CO, X26/4.
G, H: Dicroidium sp., AMB 45CO, P23/4. J: Rufloria/Noeggerathiopsis sp., AMB 34C, M41/1. K, L: Unidentified cuticle, AMB 34C, E33/3-4.
4.3.2. Remarks
4.4. Cuticle type 4 (Dicroidium) (Fig. 3D–I)
Cuticle and epidermal features of these fragments are typical of the
peltasperm Lepidopteris/Peltaspermum in having (1) a homogeneous
epidermal cell pattern, with irregularly arranged cells of rather uniform
size and shape, (2) irregular and roughly even distribution of stomata,
(3) stomatal complexes that are randomly oriented, almost circular in
outline, and radially symmetrical, (4) usually five or six subsidiary
cells, (5) conspicuously sunken guard cells, and (6) papillae overarching
the stomatal pit (e.g. Townrow, 1960; Meyen and Migdissova, 1969;
Bose and Srivastava, 1972; Anderson and Anderson, 1989; Poort and
Kerp, 1990; Retallack, 2002; Zhang et al., 2012). With the subsidiary
cells being similarly or even less cutinised than the surrounding regular
epidermal cells, the present material is similar to Lepidopteris species
reported from the Lower Triassic of India (e.g. Bose and Srivastava,
1972; Bose and Banerji, 1976).
4.4.1. Description
Cuticles thin or only moderately thick. Regular epidermal cells with
straight or slightly curving anticlinal walls and smooth (Fig. 3I), finely
buttressed, or interrupted (Fig. 3D) cutinisation; periclinal wall surface
either smooth, or with a single circular, low, diffuse thickening or
hollow papilla (Fig. 3E), or showing ornamentation of fine longitudinal
striae (see Fig. 3H). Epidermal cells of costal fields rounded rectangular
or elongate polygonal, arranged in longitudinal rows, mostly oriented
longitudinally; cells of free lamina irregularly arranged, of variable
size, rounded polygonal, roughly isodiametric or slightly elongated
(Fig. 3D, E, G). Stomata evenly distributed across the entire epidermis,
oriented mostly either longitudinally or transversely to adjacent vein
courses (Fig. 3D, E, G); few stomata oriented obliquely. Longitudinally
916
E. Schneebeli-Hermann et al. / Gondwana Research 27 (2015) 911–924
and transversely oriented stomatal complexes usually hourglass- or
butterfly-shaped, i.e. with two to rarely four differentiated lateral sub-
sidiary cells that are small, rounded-trapezoid, without papillae, and
commonly less cutinised than surrounding regular epidermal cells
(Fig. 3D, G–I); some obliquely oriented stomata surrounded by an
incomplete to complete ring of up to seven subsidiary cells that are
similarly differentiated; encircling cells common (Fig. 3H). Stomatal
pit rectangular or spindle-shaped, commonly bordered by thickened
proximal anticlinal walls of lateral subsidiary cells (Fig. 3D–I); guard
cells only little sunken, feebly cutinised (Fig. 3F, I).
4.4.2. Remarks
Epidermal and cuticular features of the corystosperm foliage
Dicroidium have been studied in great detail (e.g. Gothan, 1912; Jacob
and Jacob, 1950; Archangelsky, 1968; Baldoni, 1980; Anderson and
Anderson, 1983; Abu Hamad et al., 2008; Bomfleur and Kerp, 2010).
Of special importance is the characteristic stomatal organisation that
distinguishes Dicroidium from other gymnosperm groups (e.g. Lele,
1962; Rao and Lele, 1963; Retallack, 1977; Anderson and Anderson,
1983; Bomfleur and Kerp, 2010). With generally thin cuticle, weakly
developed papillae, and relatively thin subsidiary-cell cuticle, the present
material most closely resembles the cuticle morphology of Dicroidium
irnensis Abu Hamad et Kerp (Abu Hamad, 2004) and Dicroidium
jordanensis Abu Hamad et Kerp (Abu Hamad, 2004), which were recently
described from the Upper Permian of Jordan (Abu Hamad et al., 2008).
5. The palynological record
The well-preserved spore–pollen assemblages from the Amb Valley
section in the Salt Range provided a palynofloral signal across the
Permian–Triassic boundary. The assemblages derive from the upper-
most Permian Chhidru Formation, the so-called white sandstone unit,
to the lowermost Triassic basal Mianwali Formation, including the
Kathwai Member, the Lower Ceratite Sandstone and the basal part
of the Ceratite Marls. Within the corresponding time interval, the
flora underwent several quantitative changes, categorized here as
floral phases (I to IV) (Fig. 4); palynological assemblages from the
uppermost Permian white sandstone unit (phase I, 24 samples) are
dominated by taeniate and non-taeniate bisaccate pollen grains such
as Protohaploxypinus spp., Falcisporites spp. and Sulcatisporites spp.
Kraeuselisporites spp. (lycophyte spores) occur consistently in low
abundances. Ornamented trilete spores and monolete spores are pres-
ent throughout. Their abundance increases slightly towards the top of
the interval. The Griesbachian assemblage (phase II) is represented by
a single sample. Gymnosperm pollen grains are still the dominant com-
ponent, but compared to the underlying Permian assemblage, the non-
taeniate bisaccate pollen proportion is reduced. Densoisporites spp.
(lycophyte spores) and Cycadopites spp. (ginkgoalean or cycadophyte
pollen) occur regularly. The Griesbachian to earliest middle Dienerian
assemblage (phase III, three samples) shows a pronounced increase in
Densoisporites spp. abundance. Taeniate and non-taeniate pollen grains
are reduced in relative abundance or even absent (e.g. Falcisporites
O
Y
R
B
+
H
P
O
RID
E
T
P
S
N
R
E
F
E. papillatus
Pteridophytes
LYCOPODIOPSIDA
Densoisporites
PTERIDOSP
Lundbladispora
Kraeuselisporites
A
SID
A
SID
A
P
SID
TO
P
O
P
E
D
TO
UIS
A
C
E
Q
Y
N
C
E
G
Glossopteridales
Bisaccate taen
Cor-Cay-Pelta
S
R
E
NIF
O
C
P
S
O
RID
E
T
P
+
N
O
C
Bisaccate non-taen
gymnosperms undiff
Lunatisporites
Alisporites
R
FLO
AL P
S
E
S
A
H
Gymnosperms
spore/pollen
ratios
IV
III
II
I
A.a.
G.f.
G.d.
S.k.
O.sp.
H.t.
H.pp.
H.p.?
n
a
i
r
e
n
e
D
i
i
n
a
h
c
a
b
s
e
i
r
G
n
o
i
t
a
m
r
o
F
i
l
a
w
n
a
M
i
M
C
L
C
L
b
m
M
i
a
w
h
t
a
K
n
o
i
t
a
m
r
o
F
u
r
d
h
h
C
i
t
i
n
u
e
n
o
t
s
n
a
s
e
t
i
h
w
i
c
s
s
a
i
r
T
i
n
a
m
r
e
P
n
a
u
d
n
I
i
i
n
a
g
n
s
h
g
n
a
h
C
[m]
4
3
2
1
0
-1
-2
-3
-4
-5
-6
-7
-8
-9
-10
samples
AMB
103
26
102
bulk organic
cuticle wood
120
101
24
48 47
46
45
44
43
42
41
40
21
39
20
38
37
49
36
35
34
33
32
31
30
29
28
sandy
limestone
sandstone
siltstone
dolomite with
terrigenous
detritus
limestone
sandy, nodular
limestone
i
s
d
o
n
o
m
m
A
s
t
n
o
d
o
n
o
C
δ13Corg [‰]
-32 -30 -28 -26 -24
-22
<5% 5-10% 10-20% 20-30% 30-40% 40-50% 50-60% 0% 20 40 60 80100 Fig. 4. Permian–Triassic lithology, biostratigraphy, C-isotopes, and floral phases. LCL = Lower Ceratite Limestone, CM = Ceratite Marls. Ammonoid and conodont biostratigraphy: A.a. Ambites atavus, G.f. Gyronites frequens, G.d. Gyronites dubius, O.sp., ?Ophiceras sp., S.k. Sweetospathodus kummeli, H.t. Hindeodus typicalis, H.pp. Hindeodus praeparvus, H.p.? ambiguous H. parvus. Organic carbon isotopes after Schneebeli-Hermann et al. (2013). Permian–Triassic vegetation succession of the Amb Valley section with floral phase (I–IV) and spore/pollen ratios. Cor = Corystospermales, Cay = Caytoniales, Pelta = Peltaspermales, CON + PTERIDOSP = Conifers and Pteridosperms undifferentiated. E. Schneebeli-Hermann et al. / Gondwana Research 27 (2015) 911–924 917 spp., Weylandites spp., and Vitreisporites spp.). In the middle Dienerian assemblages of the basal Ceratite Marls (IV) Densoisporites spp. relative abundance exceeds 60% and the total spore component reaches 95%. The trends evident in the range-through diversity and the generic richness (number of genera) across the Permian–Triassic boundary are similar (Fig. 5). The highest diversity is reached in the uppermost part of the Chhidru Formation, in an interval of 2 m below the formational boundary. Diversity markedly decreases in the Kathwai Member where 14 genera disappear between the samples AMB 48 and AMB 24. Diversity decreases further towards the Dienerian. Nine of the 14 genera that disappear at the formational boundary, have been observed in the overlying Lower Triassic strata (eight in the lower Smithian and one in the Spathian according to Hermann et al., 2012). Considering these Smithian and Spathian occurrences, the drop in range-through diversity is less severe (grey line and dots in Fig. 5). Compared to Amb, the palynological data from Narmia are of low resolution (Fig. 6). The main differences between the two records are the high spore abundance in the uppermost Permian and the lack of appropriate Griesbachian samples. Spore abundances in the uppermost Permian (phase I) range between 50 and 60%, compared to maximum of 33% at Amb. Diversity in the Upper Permian is similar to that at Amb. However, the drop in diversity occurs only in the Dienerian, together with an increase in spore abundance (phase IV; Fig. 5). Due to the low sampling resolution, floral phases II and III are missing. 6. Discussion 6.1. ‘Mesozoic’ pteridosperms in the Upper Permian of Pakistan The pteridosperm Dicroidium (Corystospermales) has been historically considered to be restricted to the Triassic of Gondwana (e.g. Gothan, 1912). Recent reports of Dicroidium from the Upper Permian Um Irna Formation of Jordan, howeve