0 0E
Article history:
Received 12 December 2011
Received in revised form 30 January 2012
Accepted 31 January 2012
Available online 12 February 2012
Keywords:
Calamites
pith cast
fossil forest
plant taphonomy
fossilization
Late Paleozoic
Review of Palaeobotany and Palynology 173 (2012) 1?14
Contents lists available at SciVerse ScienceDirect
Review of Palaeobota
l se1. Introduction
Calamitaleans, relatives of modern Equisetum (the ?horse-tail? or
?scouring rush?), are one of the most commonly encountered groups
of fossil plants of Pennsylvanian and Permian age (Cleal and Thomas,
1994). Comprising typically disconnected organs, including foliage,
stems, reproductive structures, and roots, they may be preserved
as adpressions, sediment-casts, and petrifactions. One of the most
(T0 assemblages), and buried in a geological instant (e.g., Dawson,
1851; Scott, 1978; Gradzinski and Doktor, 1995; Calder et al., 2006;
Falcon-Lang, 2006; DiMichele and Falcon-Lang, 2011; Fig. 1). These
upright stems are almost always preserved as sediment-casts of
some sort, retaining three-dimensional attributes and showing on
their outer surfaces the distinctive nodes and ribbed internodes
characteristic of the sphenopsid clade, looking super?cially like
giant Equisetum. Such sediment-casts also commonly are found incompelling types of calamitalean fossil ? a
the imagination of many paleontologists, pro
is the preservation of upright stems evide
? Corresponding author.
E-mail addresses: dimichel@si.edu (W.A. DiMichele)
(H.J. Falcon-Lang).
0034-6667/$ ? see front matter ? 2012 Elsevier B.V. All
doi:10.1016/j.revpalbo.2012.01.011tations to disturbed and undisturbed environments, respectively.
? 2012 Elsevier B.V. All rights reserved.a b s t r a c t
Sediment-cast calamitalean axes in growth position are one of the most common fossils in the Pennsylvanian
coal measures. In this paper, we challenge the long accepted position that these fossils represent ?pith casts?.
If correct, the hypothesis would require the sediment-cast pith to have been surrounded by a cylinder of sec-
ondary xylem during life, which later decayed away. However, sedimentary layers and structures developed
around upright calamitaleans indicate that ?uid ?ow was interacting directly with the preserved surface of
the stem, not a hypothetical woody cylinder that lay external to it. Furthermore, stem diameter?density
data for calamitalean stands already lies at the self-thinning threshold, and if actual stem diameters were sig-
ni?cantly greater than preserved diameters, this threshold would be signi?cantly exceeded. We also note
that measured diameters for upright sediment-cast calamitaleans are more similar to stem diameter data
for anatomically preserved calamitalean axes than for pith diameter data from the same axes. Our ?ndings
indicate upright calamitaleans are in fact stem casts and their preservation involved a two-stage process.
First, stems were buried in ?ood-deposited sediments, creating a mold of the external surface of the plant.
Second, following near-total decay of the axis, which may have occurred in a matter of weeks under tropical
conditions, further sedimentation in?lled the mold, forming a cast. As such, the preservation of upright cala-
mitaleans was identical to that for arborescent lycopsids, which are commonly found in the same beds. That
said, we stress that some transported sediment-cast calamitalean axes preserved in ?uvial channel facies are
certainly ?pith casts? in the traditional sense, however, their morphologies differ from those specimens tra-
ditionally called ?pith casts?. In this context, axes were buried in a single phase of sedimentation under ener-
getic ?ow conditions, resulting in the pith becoming sediment-?lled. However, intriguingly, a review of such
genuine pith casts shows that only a tiny proportion preserves large woody cylinders surrounding the pith.
This is not a taphonomic feature, but re?ects our contention that the great majority of ?oodplain-based cala-
mitaleans were reed-like plants with relatively small amounts of secondary xylem. Woody calamitaleans, in-
cluding large tree forms, are documented almost exclusively from petrifactions, and thus from peat-forming
environments (coal balls) and, more rarely, ?oodplain settings under exceptional conditions of preservation
(volcanigenic deposits, for example). These may be dif?cult to recognize in adpression preservation due to
the masking, by wood development, of node-internode features. The differing architectures may re?ect adap-a r t i c l e i n f oResearch papers
Calamitalean ?pith casts? reconsidered
William A. DiMichele a, Howard J. Falcon-Lang b,?
a Department of Paleobiology, NMNH Smithsonian Institution, Washington, DC 20560, USA
b Department of Earth Sciences, Royal Holloway, University of London, Egham, Surrey TW2
j ourna l homepage: www.end one that has caught
fessional and amateur ?
ntly in growth position
, h.falcon-lang@es.rhul.ac.uk
rights reserved.X, UK
ny and Palynology
v ie r .com/ locate / revpa lbohorizontal position, sometimes clearly ?attened in place and buried
during ?oods, but not infrequently showing some transport from
their site of growth.
Whether in growth position or transported, the three-dimensional
casts are generally referred to as ?pith casts?, re?ecting a belief that
they are casts of the internal hollow pith area that is a key character-
istic feature of the equisetalean clade. The origin of this taphonomic
most cases at the nodes. This is re?ected in the anatomy of the
plant. The stem has a central pith that is lined along its external mar-
gin by parenchyma but is hollow in the center through the internodal
regions. At the nodes, the pith becomes solid, consisting of drum-
head-like plates of parenchyma, termed nodal diaphragms (Fig. 2A).
Surrounding this pith region is a ring of primary vascular bundles,
each of which contains a distinctive carinal canal, surrounded by pri-
mary xylem (Fig. 2B). From each primary xylary bundle, a wedge of
secondary xylem develops, the extent of which is variable. In speci-
mens with thick wood, the individual bundles fuse laterally with ad-
jacent bundles after some distance to form a solid ring of wood. This
wood generally retains wide parenchymatous rays separating the
wedges developing from each primary xylem bundle.
In their landmark textbook, Taylor et al. (2009) have described
how this anatomical structure might result in the distinctive ?pith
casts? preserved in the fossil record. We quote them here because
their exposition spells out the taphonomic implication of the pith
cast hypothesis very clearly. ?Pith casts of calamites were formed by
sediments that ?lled the hollow central canal [i.e. the pith] and solid-
i?ed before the more resistant tissues of the stem, such as the primary
and secondary xylem, were broken down by various biological agents.
Following the decay of the remainder of the axis, additional sediment
?lled in around the cast, thus resulting in a mold-cast preservationFig. 1. A ?eld sketch of a stand of sediment-cast calamitaleans, preserved in growth po-
2 W.A. DiMichele, H.J. Falcon-Lang / Review of Palaeobotany and Palynology 173 (2012) 1?14model may be traced to a collaboration between E.W. Binney and J.D.
Hooker, in the course of their pioneering work on coal balls (Hooker
and Binney, 1854). In a letter dated August 25, 1854, Hooker de-
scribed the concept to his close friend, Charles Darwin (Letter 1581,
www.darwinproject.ac.uk), as follows: ?I spent a day at Manchester
with Binney at Fossil plants, a study I hate & despise & am always
sneaking after all the same. I think we have proof positive now that
all Calamites are mere casts of piths! I am glad of it for the Survey
people used always to laugh at us for maintaining that we did not
know that Calamites was an identi?able vegetable form ? The fact
is that the stri? are the impressions of the interspaces between the
medullary rays, & the scars are points at which bundles go from a
pseudo-medullary sheath through the woody wedges to the bark.?
sition in the Middle Pennsylvanian Pictou Group of Nova Scotia, Canada (reproduced
from Dawson, 1851).The calamitalean stem, like that of all members of the Equisetalean
clade, is organized on a body plan of distinctly separate nodes and in-
ternodes, with all leaves and lateral appendages borne in whorls, in
Fig. 2. General anatomy of calamitaleans. A) Longitudinal section of a calamitalean axis sh
Lancashire, England (courtesy of Hans Steur). B) Cross section of a Lower Pennsylvanian c
hollow pith (p) with marginal parenchyma (mp), surrounded by a ring of vascular bundles
Geological Survey, photograph number: P685958).type. The structural organization of a calamite stem includes wedges
of primary xylem that extended into the central canal [pith], with
broad channels of parenchyma representing the vascular rays be-
tween. On the surface of a pith cast, these appear as a series of ribs
and furrows ? furrows mark the former position of the primary
xylem wedges, whereas ribs correspond to the vascular rays between
them? (Taylor et al., 2009, p. 350).
To be absolutely clear, in order to form such ?pith casts? of upright,
autochthonous stems, sediment is envisioned to have surrounded
and, sometime during or after burial, in?lled the partially hollow cen-
tral portion of a stem. Following decay of the wood, the space created
is inferred to have been in?lled by further sedimentation. The result-
ing outer surface of the cast is thus supposed to show the inner sur-
face of the pith. As Taylor et al. (2009) imply some of the support
for this viewpoint comes from the well-known form of most anatom-
ically preserved calamitaleans, which have a circumferential ring of
secondary xylem, outside the pith and primary vascular system.
Such wood buries the mostly hollow pith region and would have pro-
vided a means for the stem to stand erect and resist collapse as it was
owing a solid pith diaphragm (pd), from a specimen from the Lower Pennsylvanian,
alamitalean axis from Yorkshire, England, showing the characteristic structure of the
(vb) containing carinal canals and minimal extraxylary tissue (et) (Copyright, British
buried by the weight of ?ood-borne sediment. Numerous examples of
such woody calamitalean stems are known from coal-ball perminera-
lizations collected directly from coal beds, silici?ed nodules, or trees
preserved as petrifactions in ash beds (e.g., Fig. 3; Williamson and
Scott, 1895; Anderson, 1954; Andrews and Agashe, 1965; R??ler
and Noll, 2006).
However, the occurrence of a woody cylinder also raises a prob-
lem for the traditional taphonomic model because, if correct, the pre-
served pith cast ought be surrounded by a thick covering of wood and
bark, which is not seen in the many examples of upright calamita-
leans that we have observed. In the traditional model, summarized
by Taylor et al. (2009), this problem is solved by the later decay of
the wood and a secondary phase of sedimentation in?lling the gap.
Seward (1898) also seems to have been well aware of this problem,
but attempted to resolve it in a different way, by assuming that the
wood had been compressed to a negligible thickness. In this regard,
he argued that ?a zone of wood 27 mm in thickness is reduced in
the process of carbonisation to a layer 1 mm thick? (p. 366). However,
it is uncertain whether burial would result in the kind of lateral com-
been referred to as ?pith casts?. This is not a new idea, but one implic-
itly accepted by the some of the earliest investigators of calamitalean
forests (Dawson, 1851, 1868, 1892).
Like the lycopsids, the sediment-cast calamitalean stems and, im-
portantly for our argument, subterranean rhizomes, can show varying
degrees of external decay, corrosion or abrasion. Perhaps best illus-
trated by Weiss (1884), there are clear examples showing that the
layers of bark and perhaps secondary tissues found external to the
typically ribbed pith cast were very thin, at most, and easily removed
or decayed (Fig. 4). In other instances, these tissues appear to be var-
iously compressed against the ribs such that the ribbing is clearly vis-
ible through the bark, again indicating that they were veritable
?skins?, thin and ephemeral (Fig. 5).
These two plant groups, the calamitaleans and the arborescent
lycopsids, in contrast to the many other kinds of plants common in
Carboniferous and Permian deposits, appear indeed to have had hol-
low cavities in the stems, especially in the lower parts, during life,
as a result of their fundamental architecture and an anatomy that
permitted inside-out decay. In addition, they populated the wetter,
low-lying parts of landscapes, including channel margins and intra-
channel bars, likely to attract sediment-laden ?oods (DiMichele and
Falcon-Lang, 2011). Furthermore, they were exceptionally well an-
chored in the substrate, unlike contemporary tree ferns or pteridosperms,
either by extensive stigmarian rhizomorphs, or by development from
subterranean rhizomes (Fig. 6), and thus were capable of remaining
erect when inundated by ?ood-borne sediments. And ?nally, they pos-
sess unusually diagnostic external morphologies, rendering them easily
and immediately recognizable.
These characteristics facilitated the peculiar mold-cast taphonomy
that has so often preserved them in growth position. However, rather
3W.A. DiMichele, H.J. Falcon-Lang / Review of Palaeobotany and Palynology 173 (2012) 1?14pressive force to achieve this effect, and certainly not without signif-
icantly deforming the sediment-cast as well.
The presumed existence of a layer of wood and bark is similarly
problematic for interpreting transported ?pith casts?. Where these
horizontally disposed stems occur in parautochthonous associations
in ?ood-basin shales, they may show varying degrees of in?lling by
mud, and may lie atop one another in thick mats. In most instances,
the stems are surrounded by a thin coaly rind with the external fea-
tures of ribbing and nodes variably, but usually fairly clearly, visible.
Where horizontally disposed stems occur in channel sandstone bod-
ies, they are usually more dispersed, but also can comprise more con-
centrated assemblages. In both these instances, it is perplexing why
the presumed ?pith cast? is visible and not deeply buried in the sec-
ondary xylem layer, unless the fossils were somehow fractured
along the contact zone between the pith and wood. Yet even if this
had fortuitously (and repeatedly) occurred, one would also expect
to see a section through the wood, lying either side of the pith. And
it seems unreasonable to assume that horizontally disposed calamita-
lean stem remains were transported as sediment-?lled ?pith casts?,
which would have been too fragile to survive transport as bed load
in moving water.
In this paper, we argue that the great majority of sediment-cast
calamitaleans, both autochthonous and allochthnous, do not repre-
sent internal ?pith casts? at all but, in fact, external stem casts very
similar to the stem casts of arborescent lycopsids, with which they
commonly co-occur (DiMichele and Falcon-Lang, 2011; Gastaldo,
1986); we know of no cases where these lycopsid tree casts have
Fig. 3. A Lower Pennsylvanian calamitalean axis from Lancashire, England showing de-
velopment of a relatively thick (~30 mm) woody cylinder (Copyright, British Geological
Survey, photograph number: P685953).Fig. 4. A Calamites from the Stephanian C of Wettin near Halle/Saale, Germany showing the
exteriormarkings of pith (nodes and internodes) covered by a thin ?skin? of surrounding tis-
sue. Reproduced from Weiss (1884, plate I, ?g. 1). Scale: 77 mm diameter at node labeled a.
4 W.A. DiMichele, H.J. Falcon-Lang / Review of Palaeobotany and Palynology 173 (2012) 1?14than the pith being generally ?lled ?rst, resulting in an internal pith
cast, we argue here that the stems were initially buried by sediment.
The buried stem then decayed in the moist tropical climate, leaving a
mold of the external surface, which was subsequently cast by a later
phase of sedimentation. Sediment-cast calamitaleans are, therefore,
stem casts and not generally pith casts.
2. Sediment-cast calamitaleans in growth position
As already discussed, according to conventional thinking based on
permineralized calamitalean axes (Anderson, 1954; Andrews, 1952;
Cichan and Taylor, 1983), upright stems in growth position were
woody in life, to some degree, and represent only sediment-cast
pith regions, their surfaces re?ecting the outer margin of the pith
area. This model raises three initial questions, which allow the ?pith
cast hypothesis? to be critically tested. (1) How do measurements of
pith and stem diameter in anatomically preserved axes compare
with measurements of the presumed ?pith casts?? (2) Given the
Fig. 5. A specimen of Calamitina paucirama Weiss from near Jedlina Zdroj (formerly
Charlottenbrunn), Lower Silesia, Poland. An upright stem showing a thin layer of exter-
nal tissue still visible, strongly adpressed against ribs beneath. Small ovals at nodes are
interpreted as leaf scars and large ovals as branch scars. The lines beneath branch scars
appear to be deformations on the surface of the specimen, possibly taphonomic in or-
igin. Reproduced from Weiss (1844, plate XI, ?g. 1). Scale: 40 mm diameter axis.typical close proximity of standing casts, just how thick could the
woody cylinder have been if it were present during life? (3) If there
was a woody cylinder of considerable thickness surrounding the pri-
mary body and hollow pith, which later decayed away, to what extent
is this supported by sedimentological evidence?
2.1. Comparison of sediment-cast dimensions with anatomically
preserved calamitaleans
As an initial test of the ?pith cast hypothesis?, we collated data on
the diameter of sediment-cast calamitalean upright stems, and the di-
ameter of the piths and stems of anatomically preserved calamita-
leans (Fig. 7). Sediment-cast diameters were obtained from two
?eld studies undertaken over a 15 year period by one of us (HFL):
the two intervals investigated comprised the Early Pennsylvanian
Joggins Formation of Nova Scotia, Canada, a predominantly wetland
coastal plain succession (n=65; Falcon-Lang et al., 2006) and the
Early Pennsylvanian Tynemouth Creek Formation of New Brunswick,
Canada, a mostly dryland alluvial plain succession (n=143; Falcon-
Lang et al., 2010). Only sediment-casts that were nearly three-
dimensional were measured; however, as most of these axes showed
Fig. 6. A specimen of Calamites (Stylocalamites) suckowii Brongniart from the Orzesze
(formely Orzesche) Mine in Upper Silesia, Poland, showing a horizontal rhizome and
vertical side shoots. Secondary rootlets can be seen in various states of preservation at-
tached at the nodes on both the rhizome and the upright shoots. Reproduced from
Weiss (1884, plate IV, 1). Scale: horizontal axis is 69 mm diameter at widest point.some compression, both the long and short diameters were mea-
sured, and the median value obtained.
For permineralized calamitaleans, data were collated from publi-
cations (e.g., Anderson, 1954; Andrews, 1952; Cichan and Taylor,
1983; Reed, 1952; R??ler and Noll, 2006, 2007; Scott et al., 1997;
Wang et al., 2006), online museum collections with reliable scale
bars (e.g. http://geoscenic.bgs.ac.uk/asset-bank/action/viewHome),
and online fossil shops, using the same measurement protocols. How-
ever, in this case, both pith diameter and stem diameter were mea-
sured, and Pennsylvanian (n=14) and Permian (n=22) datasets
were kept separate. In most cases, stem diameter was taken as the
diameter of the secondary xylem, as extraxylary tissues commonly
are not preserved. No attempt was made to control the position in
the tree; however, in order to partly exclude lateral branches and
the upper parts of trunks, only axes with a diameter >20 mm were
selected for study.
The two Canadian ?eld studies show that calamitalean sediment-
casts have diameters in the range of 20?160 mm. However, over 80%
of records we examined are in the range of 40?70 mm diameter with
a mean for all datasets of 59 mm (n=208). In addition, we have
noted, on various ?eld excursions, prostrate axes up to 200 mm
(e.g., DiMichele et al., 2010), although ?attening may have enlarged
some of these. Furthermore, older published literature suggests a
We are cognizant of potential dif?culties in comparing measure-
ments of sediment-cast specimens with anatomically preserved spec-
imens. The former mostly come from occurrences in growth position
and represent the basal part of trunks, whereas the latter may include
material from the basal or upper trunk, and even large lateral
branches (>20 mm diameter). Nonetheless, some conclusions may
be drawn. Clearly the modal size of sedimentcast axes (40?70 mm)
is much more closely similar to the mean size of Pennsylvanian and
Permian stem diameters (52 mm and 81 mm, respectively) than it is
to mean pith diameters (29 mm and 28 mm diameter, respectively).
Furthermore, pith diameters above 40 mm have been reported only
rarely (5 out of the 36 specimens in our datasets) and dimensions
never exceed 82 mm. Yet, sediment-casts are dominantly in this larg-
er size range and may be up to double the diameter of the largest
documented pith dimensions. This initial analysis suggests that up-
right sediment-cast calamitaleans are stem casts not pith casts.
5W.A. DiMichele, H.J. Falcon-Lang / Review of Palaeobotany and Palynology 173 (2012) 1?14very small number of larger (up to 650 mm) calamitalean axes (e.g.,
Renault, 1893?1896). Overall, our?ndings are fairly similar to an earlier
compilation by Kidston and Jongmans (1917), who concluded thatmost
calamitalean casts were in the range of 50?200 mm diameter.
For anatomically preserved axes, pith diameter ranges from 9 to
75 mm (mean 29 mm) and stem diameter from 23 to 90 mm (mean
52 mm) for Pennsylvanian data (n=14). In contrast, while Permian
axes (n=22) have similar-sized piths (9?82 mm; mean 28 mm),
stem diameter is rather larger (20?160 mm; mean 81 mm). Further-
more, the ratio between stem: pith diameter ratio and is 2.53 and
2.96 for Pennsylvanian and Permian datasets, respectively. In prepar-
ing these statistics, we excluded three very large stems (600 mm,
240 mm and approximately 250 mm) with relatively small piths
(15 mmor less) recently described from the Early Permian of Chemnitz,
Germany (R??ler and Noll, 2006) and the Late Permian of southwest
China (Wang et al., 2006), whose inclusion would have signi?cantly
skewed the results.
2.2. Tree proximity, stand density and the self-thinning threshold
Fig. 7. Calamitalean diameter data. A) Diameter of pith casts in growth position in the
Lower Pennsylvanian Joggins and Tynemouth Creek formations of Canada; B) Stem di-
ameter in anatomically preserved calamitaleans, subdivided by Pennsylvanian occur-
rences (blue) and Permian occurrences (red); C) Pith diameter in anatomically
preserved calamitaleans, subdivided by Pennsylvanian occurrences (blue) and Permian
occurrences (red). See text for explanation.A secondapproach to testing the ?pith cast hypothesis? is to consider
the density of calamitalean forests. Whereas upright stem casts of cala-
mitaleans can occur singly, they more frequently occur in quite dense
stands. One of the best sites for observing upright calamitalean forests
in plan view is the Gardner Creek section of the Early Pennsylvanian
Tynemouth Creek Formation of New Brunswick, Canada (Fig. 8;
Falcon-Lang, 2006). Measurements of seven separate fossil horizons in-
dicate densities ranging from4 to 9 stems perm2 and amean sediment-
cast diameter of 62 mm (Falcon-Lang, 2006, and unpublished data). In
the Early Pennsylvanian Joggins Formation of Nova Scotia, Canada
(Falcon-Lang et al., 2006), similar bedding plane exposures indicate
calamitalean densities on the order of 2 to 13 stems per m2 with a
mean stem diameter of 53 mm (Falcon-Lang, 1999).
A universal relationship between plant metabolism and the spatial
distribution of trees in three dimensions has been posited for modern
multi-species forests composed of individuals of different ages
(Enquist and Niklas, 2001; Enquist et al., 2009; West et al., 2009). An
even-aged forest composed primarily or entirely of a single species is
a special case of this more general relationship, which presumably ap-
plies to ancient forests as well. One expectation, termed ?self-thinning?
is that tree density declines as tree diameter, and hence biomass, in-
creases, a function of the fact that trees compete for limited resources
(sunlight, water, soil nutrients), and as they mature, uncompetitive
neighboring trees die off (e.g., Silvertown and Doust, 1993). Conse-
quently, when diameter?density data for modern forests are plotted,
it is possible to ascertain the upper threshold for stem diameter at a
given density (e.g., Cao et al., 2000). This relationship has been ques-
tioned for some clonal plants but studies show that it does apply to
Fig. 8. Example of a bedding surface in the Lower Pennsylvanian Tynemouth Creek
Formation, New Brunswick, Canada, for which calamitalean density data were
obtained. Diameter of upright trees is ~50 mm, for scale.
woody, tropical species with perennial life-histories, while at the same
time annual plants may not follow the self-thinning trajectory (e.g., de
Kroon and Kalliola, 1995; Peterson and Jones, 1997). Thus, given the
woody and presumed perennial nature of calamitalean clones (recog-
nizing that not all calamitalean species were clonal), we believe we
can safely assume that the self-thinning laws apply. Our Pennsylvanian
calamitalean forests from Canada already lie more or less on the self-
thinning threshold, if not above it, representing trees with the maxi-
mumpossible stemdiameter for the extremely high densitiesmeasured.
However, if these sediment-cast axes represent only part of the
tree diameter in life, how might this affect our calculations? As
noted above, the ratio between pith diameter and stem diameter is
2.53 and 2.96 for Pennsylvanian and Permian calamitaleans showing
anatomical preservation. As such, if we add wood to a typical
60 mm diameter ?pith cast? we might infer that it must have repre-
sented, in life, a tree with a diameter in the range of 152?178 mm.
Applying these ?corrected? diameters to our high density stands
would push these forests signi?cantly above the self-thinning thresh-
old, in other words being impossibly dense stands for the inferred
mean diameter (Fig. 9). It is possible, though not likely, that extinct
trees with unusual architectures, like the calamitaleans, may have
had quite different self-thinning thresholds than the conifers and di-
cots on which the modern analysis is based. Greater uncertainly may
lie in the realm of how such clonal trees would compete for resources
with their non-clonal neighbors ? as individual trees? Allowing for
these important caveats, this density: diameter analysis further chal-
lenges the concept that upright calamitaleans are pith casts of trees
that were much more substantial in life.
A more straightforward piece of evidence that leads to exactly the
?bush? (Fig. 10), a pattern reported relatively commonly in the liter-
ature (e.g. Dawson, 1851; DiMichele et al., 2009; Fig. 1). The only
credible interpretation of this phenomenon is that the preserved
axis diameter observed today is the same, or at least nearly the
same, as in life.
2.3. Relationship of upright calamitaleans to sedimentary layers
A third way to test the ?pith cast hypothesis? is to examine the re-
lationship of upright sediment-cast axes to entombing sedimentary
layers and structures. If we maintain that such calamitaleans are
casts of the ?pith?, then the following seem to be the rarely stated as-
sumptions: (1) The axis would need to have been open to the pene-
tration of sediment, basically through its entire preserved length,
meaning that the nodal diaphragms (which partitioned the pith in
life) either would have to have been gone prior to, or were breached
during, sediment ?lling. (2) The fully buried axis would have needed
support for an open central cavity to become ?lled by sediment,
meaning either by a rind of mechanically strong tissue (secondary
xylem) or by burial in sediment deposited before or during central
6 W.A. DiMichele, H.J. Falcon-Lang / Review of Palaeobotany and Palynology 173 (2012) 1?14same conclusion is the observation of adjacent sediment-cast calami-
taleans that are in very close proximity to one another. For example,
in the Tynemouth Creek Formation, there are several examples of ad-
jacent axes in growth position that, in plan view, either touch one an-
other or are spaced only a few millimeters apart. In related examples
seen in vertical section from the same site, calamitalean axes are com-
monly observed budding off from a parent rhizome, forming a dense
Fig. 9. Log-transformed relationship between tree density and tree diameter based on
modern forestry observation showing the self-thinning threshold, and data for three
forests of varying initial density: medium-density stand A (1000 trees per hectare,
tph); high density stand B (4000 tph); extremely high density stand C (10,000 tph).
Figure and dataset modi?ed after Cao et al. (2000). Data for calamitalean stands in
the Lower Pennsylvanian Tynemouth Creek (TCF) and Joggins (JF) formations are plot-
ted, together with ?adjusted? values assuming that real tree diameter is 2.5 times larger
than observed diameter. Unadjusted calamitaleans lie on or near the self-thinning
threshold, but adjusted stands signi?cantly exceed it.cavity ?lling. Thus, the stem most likely would have been either par-
tially or fully buried by sediment prior to, or contemporaneously
with, the in?lling of the pith area. (3) The rind would (a) have had
to undergo decay subsequent to burial, in order to allow surrounding
sediment to lie in direct contact with the pith cast, or (b) need to have
been highly compacted against the pith cast after burial, remaining
only as a thin layer (the contention of Seward, 1898).
If the organic material sometimes found adherent to the ?pith? cast
were the result of compaction of the rind (wood and parenchyma)
with only minimal decay, then the degree of compaction would need
to have been considerable. As noted above, some compression speci-
mens (e.g., Figs. 5?6) show that tissues outside the distinctly ribbed
portions of the stem were very thin. In woody permineralized speci-
mens, secondary xylem of typically 10?30 mm radius (Anderson,
1954; Cichan and Taylor, 1983; Wang et al., 2006) has been reported,
requiring compaction of perhaps 20:1 or more to account for the thin-
ness of the rinds observed on most calamitalean casts. Were this to
have been the case, the great differences in compaction ratios between
organic material and siliciclastic sediment must be considered (Nadon,
1998). Plant material compacts between 10:1 and 80:1, by some
estimates (e.g., Winston, 1986), whereas the compaction ratios of silici-
clastic sediment are approximately 2: 1 for siltstone and close to zero
for sandstone, both of which commonly entomb and cast calamitalean
stems. Thus, whether the rind decayed or whether it was compacted,
the expectation is that one should see a distinct zone of sediment
Fig. 10. Two calamitalean axes budding off a ?parent? axis in the Lower Pennsylvanian
Tynemouth Creek Formation, New Brunswick, Canada. This is evidence that these plants
were of the same diameter in life, as preserved today. Hammer for scale is 0.4 m long.
between the outer edge of the ?pith? cast and the inferred original outer
edge of the stem.
If the entombing sediment were stiff (had ?set up?) before the hy-
pothetical decay or compression of the rind began, there ought to be
evidence for a cavity remaining after decay of the rind tissue. Thus, as
Taylor et al. (2009) have stressed, an upright calamitalean ?pith cast?
should show clear evidence for a double-?lling event: the ?rst stage
involving ?lling of the hollow pith area, and the second the ?lling of
the hollow left by decay or partial decay of the rind tissue. This should
be detectable on outcrop as one or more concentric ?llings around the
?pith cast?, in total radius equal to the thickness of the original organ-
ic rind. Such ?lling sediment possibly would be distinct from the gen-
eral sediment of the deposit, in terms of color, texture, or bedding. On
the other hand, if the sediment entombing the exterior of the full
stem were still in a ?uid or semi-?uid state, some deformation of
that sediment around the central ?pith? cast might be expected as
the organic rind either decayed or compacted. At least a few examples
of these hypothetical relationships should have been noted in the lit-
erature, given the observational powers of most ?eld geologists;
however, to our knowledge, none have, though there are many exam-
ples where undeformed laminae truncate directly against upright
calamitalean stems (Fig. 11).
2.4. Vegetation-induced sedimentary structures
A fourth test of the ?pith cast hypothesis?, and one closely related
to that discussed above, is the occurrence of what has been termed
vegetation-induced sedimentary structures (VISS) by Rygel et al.
(2004). These comprise sedimentary structures generated by the
7W.A. DiMichele, H.J. Falcon-Lang / Review of Palaeobotany and Palynology 173 (2012) 1?14Fig. 11. A Lower Pennsylvanian calamitalean tree buried in tidalites in Alabama, USA
(reproduced from Gastaldo et al., 2004) with arrows highlighting rhythmites (scale
in cm). Note that the laminations are upturned and truncate directly at the edge of
the upright stem.?uid ?ow around an obstacle such as, in this case, an upright calami-
talean tree (Fig. 12). Such structures might include laminations that
dip towards the trunk (centroclinal cross-strati?cation; Leeder et al.,
1984), formed as a result of the acceleration of water around the
tree, which generates scours that are subsequently draped when the
?ow wanes. Another style of VISS is the occurrence of laminations
that are upturned towards the tree and record mounding of sediment
against the obstacle at times of relative low ?ow velocity. Of key im-
portance is that close examination of these features demonstrates
that (1) laminations and structures truncate directly against the
edge of the preserved sediment-cast tree and (2) they indicate that
?uid ?ow was interacting directly with the preserved surface of the
stem not a hypothetical woody cylinder that lay external to it.
Thus, all these discussions lead us to infer that sediment-cast cala-
mitaleans in growth position are either (1) external stem casts or
(2) pith casts of an architectural variety that had only a very small
amount of secondary xylem and whose outer surface more or less
corresponded to the preserved surface of the sediment-cast. Our liter-
ature review indicates that anatomically preserved calamitaleans
typically had a variably thick cylinder of secondary xylem, generally
favoring the former hypothesis. However, an exceptional 75 mm
diameter stem described by Anderson (1954), which showed a
65 mm diameter pith and only ~5 mm radius of secondary xylem de-
velopment, also offers some support for the latter hypothesis. It is
possible that such a ?reed-like? calamitalean could hypothetically
generate pith casts with minimal sediment disturbance due to com-
pression and decay of outer layers. Nonetheless, it should be possible
to distinguish between external stem casts and internal pith casts
based on the structures preserved.
2.5. Could upright calamitaleans be ?pith casts? of reed-like plants?
If sediment-cast calamitaleans are pith casts of reed-like plants,
then the cast is a mold of the outer surface of the pith. However, in
contrast, if sediment-cast calamitaleans are stem casts then the fea-
tures preserved re?ect the structure of the outer surface of the plant
stem or some of the layers immediately beneath that outer surface
(analogous to the ?decortication? stages of arborescent lycopsid
stems). Yet, here lies the problem because, as Taylor et al. (2009,
p. 350) note, ?the impression of the outer surface of the xylem cylin-
der and the pith cast look almost identical, even in the alternation of
ribs and furrows at the node?. Nonetheless, close examination of the
carbonaceous rind that commonly surrounds sediment-cast calamita-
leans should allow internal and external casts to be distinguished.
Although these rinds are usually quite thin (b1 mm), they more
commonly than not show diagnostic features suggestive of the
outer surface of the stem (ribs, nodes, branch scars, even attached lat-
eral branches in whorls, locally with attached stroboli or leaves;
Figs. 13?15), although such features may be dampened on some spec-
imens (Fig. 5), due to the exposure of underlying tissues, such as a
thin layer of wood. However, coaly organic matter frequently adheres
to the indentations between the ribs, indicating that the organic man-
tle of the stem re?ects such architecture from its inner to its outer
surface, and that such stems had limited wood development.
A comparison with stem casts of arborescent lycopsids is useful at
this point to illustrate how a decayed external surface of a trunk
might show composite features from different internal tissues. Most
commonly, the outer surface of the stem preserves the characteristic
leaf cushions on the external side of the coaly rind, so there is no
question that the outer surface of the stem is in view, even if the car-
bon has been lost from the immediate surface of the leaf cushions.
However, in contrast, in some Sigillaria specimens, the diagnostic
paired to fused parichnos aerating strands may instead appear on
the outer margins of sediment-?lled casts (Fig. 16), re?ecting the
deeper internal surfaces of the preserved bark, and suggesting a con-
siderable amount of external tissue decay prior to burial of the stem
externally and ?lling of the interior hollow area with sediment. Thus,
the occasional presence of features such as infranodal canals on the
surface of sediment-cast calamitaleans does not falsify the hypothesis
that these are external stem casts. Even if these plants were reed-like,
with relatively little wood development, they are still preserved as
external stem casts.
2.6. Casts of rhizomes
developed), or have been compressed to a tissue-paper thin layer
through which the ribs could show. In addition, the roots are attached
to the specimen directly at the nodes, both on the horizontal axis and
on the attached upright axes, suggesting little or no tissue beyond the
surface of the modern-day cast (the interpretation we prefer). Alter-
natively, one might presume either that a now-missing woody rind
decayed while leaving intact the roots, around which wood would
have developed as cambial activity proceeded, or that the wood was
compressed to a thin sheet while, at the same time, the roots that
passed through it were uncompressed or minimally compressed.
2.7. Positive proof that sediment-cast calamitaleans are not pith casts
Positive proof that sediment-cast calamitaleans are not pith casts
comes from a remarkable fossil preserved in seasonally dry red bed
Fig. 12. Vegetation-induced sedimentary structures (VISS) surrounding an upright calamitalean tree in the Lower Pennsylvanian Joggins Formation, Nova Scotia, Canada (reproduced
from Rygel et al., 2004, ?g. 10).
8 W.A. DiMichele, H.J. Falcon-Lang / Review of Palaeobotany and Palynology 173 (2012) 1?14Among the most signi?cant challenges to the ?pith cast hypothe-
sis? are horizontal, likely subterranean rhizomes. Weiss (1884, his
plate IV, 1) illustrates a nice example of such a specimen (reproduced
here as Fig. 6). Although slightly ?attened, it is cast with sediment,
has roots at the nodes on both sides, indicating that it was subterra-
nean at the time of death and preservation, and has carbonized re-
mains of tissue adherent to the ribbed surface, through which the
ribs are clearly visible. In addition, it bears two upright axes, each of
which also bears roots at the nodes, indicating that they, too, were
embedded in the sediment at the time of casting.
It stretches credulity to imagine that this, or other specimens like
it, are casts of pith regions of an originally much larger, woody axis.
The presumed woody rind would have had to decay below ground
without leaving a noticeable cavity in the original sediment (nor
would there be an obvious way to ?ll such a cavity were it to haveFig. 13. A specimen of Calamites multiramis Weiss from Manebach, Germany showing
lateral branches bearing strobili, accessioned in the Museum f?r Naturkunde der
Humboldt-Universit?t zu Berlin (courtesy of Manfred Barthel, Berlin, Germany). Scale
in mm (small divisions).Fig. 14. A specimen of Calamites ramosus Artis from the Westphalian of Nowa Ruda
(formerly Neurode), Lower Silesia, Poland, showing lateral branches and Annularia
foliage, some of which may be still attached. This specimen was reproduced in Weiss
(1884, plate V) and is in the collection of the Museum f?r Naturkunde der
Humboldt-Universit?t zu Berlin (photo Hans Kerp, M?nster, Germany).
deposits in the Tynemouth Creek Formation of New Brunswick. Here,
an upright calamitalean exposed in plan view appears to show a
?wound? that allowed sediment to enter the pith at the time the
stem was buried, forming a poorly preserved pith cast, but positioned
inside the stem cast (Fig. 17). Presumably, this fossil formed by a
three-stage process. First, the tree must have been wounded, creating
a breach in its secondary xylem cylinder, i.e., an opening connecting
the pith area to the outside (cf. Stopes, 1907). Second, the tree was
buried in ?ne-grained sand, forming an external mold in the usual
way, but with some sediment entering the pith through the opening
afforded by the wound, creating an internal cast as well. It is possible
that wounding and burial happened in the same ?ood event. Third,
after the sediment-cast pith had hardened in the dry climate, the
woody cylinder surrounding the pith decayed, and the hole it left be-
hind itself became in?lled with sand of a different character than that
in which the stem was buried. Consequently, this extraordinary fossil
Fig. 15. A specimen of Calamites goeppertii Ettingshausen from the mid-Stephanian of
Reisbach, Saar, Germany, showing branch scars at the nodes (axis 60 mm diameter).
This specimen is in the collection of the Laboratory of Palaeobotany and Palynology,
Utrecht, Netherlands (photo, Hans Kerp, M?nster, Germany).
Fig. 16. A Middle Pennsylvanian sediment-cast sigillarian lycopsid tree (Syringodendron
sp.) buried in growth position at Jenlin strip mine, Indiana, USA. Note the paired
parichnos scars in parallel bands indicating that the external leaf cushions had been
sloughed off or decayed away prior to burial (DiMichele et al., 2009). Scale: upper
trunk, 0.3 m diameter.
9W.A. DiMichele, H.J. Falcon-Lang / Review of Palaeobotany and Palynology 173 (2012) 1?14Fig. 17. The exception that proves the rule: a rare example of an external stem cast con-
taining an internal pit cast from the Lower Pennsylvanian Tynemouth Creek Formation,
New Brunswick, Canada. A) Two stem casts in growth position, observed in plan view,
scale: coin 21 mm diameter. B) Annotated photo of the upper stem cast, scale: coin
21 mm diameter. This tree was apparently wounded, allowing sediment to ?ll the
pith during the initial burial event. Later the secondary xylem decayed and the space
was in?lled by a second phase of sedimentation. Ribbed internodes are visible both
on the outer surface of the pith cast, and also on the outer surface of the stem cast.
3. Horizontally disposed stems and stem casts
Horizontally disposed, adpressed remains of calamitalean stems
are far more common than autochthonous remains found in upright
posture. These vary from (par)autochthonous, where they were
knocked over and buried nearly in place by ?oods, to allochthonous,
transported some distance from their site of growth. In all instances,
the stem remains may be accompanied by some degree of decay
and disarticulation. And not all specimens have a sediment-cast pith
region preserved. The degree of ?attening, therefore, can be consider-
able, resulting in an effectively two-dimensional fossil. Most com-
monly, such adpressed, prostrate remains have an organic rind
similar to that found on upright stems. Thus, even though these re-
mains have a different taphonomic history from upright cast stems,
many of the same problems are faced regarding interpretation of
the organic rind and sediment-cast central area, if one exists.
3.1. Critique of current thinking about prostrate calamitalean stems
In the literature, horizontally disposed calamitalean stems are not
differentiated from upright, autochthonous stem remains, both re-
ferred to as ?pith casts.? This interpretation raises some serious mat-
10 W.A. DiMichele, H.J. Falcon-Lang / Review of Palaeobotany and Palynology 173 (2012) 1?14is a stem cast of the external surface but also contains a smaller inter-
nal cast of the pith cavity! Both external and internal surfaces pre-
serve internodal ribbing.
The thickness of the sedimentary ?ll of what we infer is the
decayed woody cylinder is only on the order of 7?10 mm radius in
these specimens. This, in turn, supports the idea that sediment-cast
calamitaleans were generally rather ?reed-like? plants with a large
pith cavity and relatively small amount of secondary xylem develop-
ment. This reed-like architecture contrasts with the calamitaleans of
undisturbed, Pennsylvanian peat-substrate environments (anatomi-
cally preserved in coal balls), which were mostly substantially more
woody (typically >10?30 mm radius of wood; Anderson, 1954;
Cichan and Taylor, 1983), with the exception of Anderson's unusual
reed-like form noted above. It also differs from certain Permian clades
that developed gigantic woody trunks (radii up to 320 mm) and ap-
pear to have been adapted to more water-stressed settings (Barthel
and R??ler, 1996; Kerp, 1984; Naugolnykh, 2005; R??ler and Noll,
2002).
2.8. A model for the formation of upright sediment-cast calamitaleans
Without having ?rst-hand observations of the process of preserva-
tion of sediment-cast calamitaleans in growth position, we offer the
following model as a likely explanation accounting for most of the ob-
servations seen on outcrop:
(1) Those calamitaleans most commonly preserved as sediment-
casts are associated with facies representing disturbed envi-
ronments, along river banks or on ?oodplains. Most of these spe-
cies were small and slender, much like over-sized Equisetum in
terms of their growth architecture. They were not extensively
woody, being supported by relatively small amounts of second-
ary xylem, or none at all, and therefore probably were of fairly
limited stature.
(2) The habitats of the Pennsylvanian sediment-cast calamitaleans
were predominantly ?uvial and prone to ?ooding. These ?oods
rapidly buried whole stands of trees. In some instances the
plants were able to recover from the ?oods and re-grow, as
in the example presented by Gastaldo (1992). In other cases,
the plants were killed, the stems broken open or sheared off,
and then ?lled with sediment from the surrounding environ-
ment, especially if it were still ?ooded. In some instances,
they were cast with similar sediment brought in by a later,
but nearly contemporaneous ?ood. Because calamitaleans,
like modern equisetaleans, had plates of parenchymatous
tissue, like drum heads, spanning the nodal regions (Fig. 2A),
these must have been decayed prior to in?lling, either in life
or after death, or destroyed mechanically during sediment
entry.
(3) Most upright calamitalean stems demonstrate some degree of
departure from vertical (e.g., Appleton et al., 2011, ?g. 10), as
if they were pushed by currents bringing in the sediment in
which they were buried, or by the weight of the sediment it-
self, were it to have come in preferentially on one side of the
stem (Fig. 18). This implies that these plants were not able to
stand upright in the face of ?oodwaters to the same degree
as the basal portions of lycopsid trees. The latter were ?rmly
anchored to the substrate by their stigmarian rooting systems
and appear to have been stif?y vertical in their construction,
held so by the well documented periderm/bark that provided
the main support structure of the stem (e.g., Philips and
DiMichele, 1992). The ?exibility of the calamites further sup-
ports the idea that they were reed-like with minimal support-
ing woody tissue.
(4) Preservation at an angle also suggests that ?lling of the stem
cavity, in most instances, would have happened after theupright stems were stabilized by a surrounding mass of sedi-
ment, with exposed parts broken off or decayed suf?ciently
to permit entry of sediment to the hollow central areas. This
calls either for two sedimentation events, not unlikely if the
environments of preservation were being inundated actively
(e.g., Gastaldo, 1992, for example), or for the stems to be bro-
ken off during active burial so that sediment could both sur-
round them and enter the hollow central area (with due
consideration given to the presence of nodal septa, mentioned
above).
(5) The thin outer rind of tissues (primary vasculature and a thin
woody layer) was then coali?ed and compressed. However,
due to its initial thinness this compaction did not lead to signif-
icant distortion of sediment against the outer surface of the
stem. The outer surfaces of these stems preserved the nodal
morphology, internodal ribbing and appendicular scars at the
nodes typical of this clade, though these may have been accen-
tuated as the organic rind decayed revealing the internal layer,
or as it compacted and vitri?ed in its compression against the
sediment-cast.
Fig. 18. A Middle Pennsylvanian sediment-cast calamitalean axis preserved in growth
position in Indiana, USA. The specimen is inferred to have been tilted at a low angle
by ?oodwaters during burial (DiMichele et al., 2009). Rock hammer for scale, 22 mm
in width at blunt end.ters with regard to a model of formation. The clearest articulation of a
enriched in foliage and axes of pteridosperms and ferns, axes of
lycopsids, and organic debris of various kinds and sizes. Why, one
might ask, should the organic exteriors of calamitaleans be preferen-
tially decayed under such circumstances while the organic matter de-
Fig. 19. Adpressed, and partially sediment-cast calamitalean axes arranged in dense,
randomly orientated prostrate accumulations; Baker Coal, Middle Pennsylvanian,
Indiana, USA. Ruler for scale divided into one-inch intervals (courtesy of Scott Elrick,
Illinois State Geological Survey).
Fig. 20. Mat of calamitalean axes without sediment-cast central areas. USNM Specimen
528424, USNM locality number 40005, late Virgilian (Stephanian C), Late Pennsylvanian,
11W.A. DiMichele, H.J. Falcon-Lang / Review of Palaeobotany and Palynology 173 (2012) 1?14sedimentary model is that of Gastaldo et al. (1989), who considered
adpressed stem remains to have two distinct components. The ?rst
was the in?lling of the hollow central cavity by sediment, deposited
either from suspension or bed load. This is necessarily followed by a
separate, temporally later event, the decay and/or coali?cation of
the tissues external to the pith cast.
The ?rst of these assumptions seems essentially irrefutable.
Adpressed calamitalean remains frequently have a core of sediment,
of variable thickness, generally identical in grain size, color, and bed-
ding/lamination to the sediment that surrounds and entombs the
stems. In this situation, the stems were certainly hollow at the time
of burial and the sediment surrounding them, carried in the ambient
aqueous environment, mostly likely was deposited contemporane-
ously with that within their hollow centers (whether that took
place in one or more depositional events). Consider also the following
observations that characterize most of these specimens: (1) the stems
are generally fully buried in that sediment, (2) their hollow core ?ll-
ings are usually only wafers and not complete round pith region casts,
or there is no pith cast at all, and (3) there is no reported evidence of
infaunal activity within the central sediment if present (as noted by
Gastaldo et al., 1989, for the in?lling of a Holocene woody log).
These factors combine to suggest that complete burial of such speci-
mens took place during the transport and burial event, in most in-
stances. In other words, it was only the rare, exceptional specimen
that lay open on the surface or at the bottom of a stream, fully or par-
tially exposed, accumulating sediment through multiple ?oods over
an extended period of time. In most cases such pith-area ?lling as
did occur (zero to some) was contemporaneous with burial of the
stem remains.
Far more problematic is the second part of the model, which, al-
though articulated clearly by Gastaldo et al. (1989), is an implicit in-
ference in all studies of these remains, from the earliest conception of
the pith cast interpretation. Its fundamental assumption is that a
woody rind surrounded the pith area, that this rind decayed or was
compressed, and that the external morphology of the specimen re-
veals the internal characteristics of the outer surface of the pith. As
discussed for upright stem remains, this interpretation requires that
the ribs of the fossil stem, which are inferred to represent the mani-
festation of the carinal canals and any surrounding sclerenchymatous
or primary xylem tissue, somehow retain rigidity while woody sec-
ondary tissues outside of them decay and collapse. Where there is a
coali?ed layer outside of the sediment-cast, such tissues external to
the pith must have completely collapsed against a rigid frame, such
that they take on its form. It would have to be true, then, if the form
of these coali?ed compressions were to re?ect the inner part of the
pith cavity, that the form of the pith was created by the strength of
the tissues immediately surrounding the hollow pith area, because
the classic node?internode form and internodal ribs can be seen
even in specimens where there is no sedimentary ?lling of the pith
region. In such simple compressions, supposed external surface
form of the ?pith? remains visible on the outside of the axis, in the or-
ganic compression.
Using this model, it becomes essentially impossible to explain the
ribbed, nodal architecture seen on mats of calamitalean stems
(Fig. 19), deposited in the same sedimentary event, especially exam-
ples where none have sediment-cast ?piths? (Fig. 20). Such mats re-
quire that the external tissues of each stem decay back to the pith
margin, the stems then de?ating upon one another. It requires that
smaller axes with attached foliage and branches, on which ribs and
nodes are clearly visible (Fig. 21), somehow undergo differential
decay of (woody) tissues external to the pith, but not of the foliage
and branches (Figs. 13?15). And it is doubly dif?cult to explain how
such axes could form in those instances where there is no evidence
of a sedimentary wafer in the hollow, central, pith region.
Finally, calamitalean stem compressions are common components
of adpressed mega?oras in which the remainder of the assemblage isrived from virtually every other plant group is not decayed (Fig. 22)?
This is especially highlighted by the case of sediment-?lled lycopsid
stems and stigmarian roots (again, never referred to as ?pith casts?),
but which are often preserved with partial sediment in?lls, around
which the stem has collapsed under sediment pressure following
burial.
3.2. A model for the formation of prostrate, adpressed calamitalean axes
In keeping with the model proposed by Gastaldo et al. (1989), we
interpret prostrate calamitalean stems to have formed by burial in
sediment, much like other parautochthonous to allochthonous organ-
ic plant remains. (1) This formation may have been effected simply byMarkley Formation, North-central Texas, USA. Scale in mm (small divisions).
Fig. 21. Prostrate calamitalean axes showing demonstrably external features from the Lowe
t, Br
of t
).
12 W.A. DiMichele, H.J. Falcon-Lang / Review of Palaeobotany and Palynology 173 (2012) 1?14the burial and compression of the stem, without any in?lling of its
pith by sediment. (2) On the other hand, sediment in?lling of a hol-
low central area may have occurred to varying degrees, measureable
by the degree of ?attening of the stem and the thickness of the sedi-
ment wafer that ?lls its central area. (3) Such internal sediment,
when present, is most often identical to the sediment that entombs
the total the fossil assemblage, including the calamitalean stems.
This strongly suggests that the sediment entered a broken or frag-
mentary calamitalean stem at the time of burial. Not ruled out is the
possibility that some prostrate calamitalean stems could have had
their hollow central areas fully or partially ?lled by multiple ?ood/
branch scars (arrow) at nodes, from Royston, Yorkshire, England; scale: 8 mm (Copyrigh
scars on a specimen from Adeer Pit, Ayrshire, Scotland. That this is the external surface
brate, scale: 8 mm (Copyright, British Geological Survey, photograph number: P687965sedimentary events, were they to have lain exposed or partially bur-
ied on a surface or subaqueously, but such instances appear to be rare,
given the lack of reports of sedimentary indicators of this kind of ?ll-
ing. (4) The buried stems, including their relatively thin external
Fig. 22. Calamitalean axes (two, in center of photograph), closely adpressed upon one
another, showing clearly jointed and ribbed external morphology beneath thin, exteri-
or layers of carbonized tissue. There is no evidence of a central ?lling by sediment. Note
the co-occurrence of pteridosperm axes and Neuropteris foliage (middle and bottom)
and marattialean tree fern stem (top right). Immediate roof shales of the Spring?eld
(No. 5) coal bed, late Desmoinesian age, Middle Pennsylvanian, Carbondale Formation,
Indiana, USA. Rock hammer for scale, 22 mm in width at blunt end.tissue layers, are compressed in the same way that other compression
fossils are in the assemblage of which they are a part. (5) The external
features of these stems do not result from decay of tissues external to
the pith. As with upright stems, they are re?ective of the external
morphology of the stem during life. Their external surfaces do not re-
?ect the morphology of the outer margin of the pith, brought-out or
enhanced by decay of exterior tissues and the processes attendant
fossilization and compression.
We should note, independently of this model, that sediment-?lled,
prostrate ?pith casts? in adpression preservation, must have formed
in situ. It would be very dif?cult, indeed, to imagine that these plant
r Pennsylvanian of Yorkshire, England. A) Calamitalean showing leaf scars and whorled
itish Geological Survey, photograph number: P685992); B) Close-up of whorled branch
he plant is demonstrated by the occurrence of attached spirorbids, an aquatic inverte-fossils were ?rst cast by sediment and then transported in water to
their site of deposition. Issues of both ?otation and survival of a
sediment-cast in traction, in the bed-load fraction of a stream,
would appear to rule out any such considerations.
In this model, we refer back to the comments of R??ler and Noll
(2006) regarding the much greater-than-presently-appreciated di-
versity of the calamitaleans. As already discussed, it appears that the
calamitaleans of highly-disturbed riparian environments may have
been morphologically distinct from those of more stable peat-
substrate habitats or long-persistent, low sediment ?ux, swampy
sites. Repeated disturbance and burial (e.g. Gastaldo, 1986) favored
those plants capable of recovering rapidly from such events.
3.3. Do genuine calamitalean pith casts exist? ? Yes
All that said, there do appear to be rare examples of genuine pith
casts preserved in the fossil record. In one case that we have documen-
ted above in the Lower Pennsylvanian Tynemouth Creek Formation, an
autochthonous, upright stem has a sediment core that is different litho-
logically from the sediment between that core and the outer edge of
the stem (Fig. 17). This appears to have formed by an initial in?lling of
a hollow central area, followed by decay of a woody zone with subse-
quent sediment in?lling, ful?lling the expectations of the model of
Taylor et al. (2009). The specimen still retains the outermost morphol-
ogy of a calamitalean stem, however, reinforcing our argument that
most stems of this size are stem, rather than pith, casts. Furthermore,
the pith cast preserved in this example is somewhat unlike normal
sediment-cast calamitaleans, being narrower with poorly developed
nodes and internodes.
In addition we have observed several examples of prostrate stems
that are demonstrably pith casts. These occur in the Lower Pennsylvanian
Joggins Formation (Falcon-Lang et al., 2006) and above the Herrin Coal
in Illinois (DiMichele et al., 2007). They comprise a sediment-?lled cen-
tral pith area surrounded by an external layer of woody tissue, the
latter preserved in a ?attened, adpressed state (Fig. 23). Therefore,
rather than requiring a process of double-?lling, there was but a single
in?lling event, followed by compaction of surrounding woody tissue
into a thin sheet.
4. Discussion
The concept of the calamitalean pith cast is venerable and of long
standing. As already noted, it originated with Joseph Hooker at a time
when he was fast rising to become one of the superstars of mid-
Nineteenth Century science. With these credentials, it is unsurprising
that the concept obtained mainstream acceptance. The pith cast
hypothesis was, however, not universally accepted. William Dawson,
one of the best-known Carboniferous palaeontologists of the Nine-
13W.A. DiMichele, H.J. Falcon-Lang / Review of Palaeobotany and Palynology 173 (2012) 1?14teenth Century, held a more nuanced view, writing that, ?sometimes
what we call a Calamite is a mere cast of its pith showing longitudinal
striae and constrictions at the nodes. Sometimes we have the form of
the outer surface of the woody cylinder, showing longitudinal ribs,
nodes, and marks of the emission of the branchlets. Sometimes we have
the outer surface of the plant covered with a smooth bark showing ?at
ribs, or almost smooth, and having at the nodes regular articulations
with the bases of the verticillate branchlets, or on the lower part of the
stem the marks of the attachment of the roots (Dawson, 1892, p.
122?123; many of the features that Dawson discusses are illustrated
in Figs. 4?6, 13, 15, 20, 21). Dawson's view is broadly consistent
with our own position, although we believe that pith casts comprise
the minority of such fossils.
However, while some Nineteenth Century scientists were aware
that calamitalean axes could be preserved as both as stem casts and
pith casts, since the dawn of the Twentieth Century that nuance
seems to have been lost. It is important to stress that we are not ques-
tioning the basic idea that the (partially) hollow central pith area of
calamitalean stems could, and frequently did, become cast by an in-
?ux of sediment. Our challenge is to the notion that such fossils rep-
resent only the central areas of once much larger diameter stems in
which the pith region was surrounded by a thick mantle of various
Fig. 23. A rare example of a genuine calamitalean pith cast, surrounded by its adpressed
outer woody cylinder, from the Lower Pennsylvanian Joggins Formation of Nova Scotia,
Canada (courtesy of Melissa Grey, Joggins Fossil Institute, specimen number
NSM008GF031.151). Scale bar in cm (numbered intervals).tissues, mostly wood. Rather, we have argued that the common or-
ganic compressions and casts with an organic rind represent in?lled
stems, the external surfaces of which represent closely the external
appearance of the plant in life. In this way, the calamitalean stem
cast parallels, but with much less interior decay (because of less orig-
inal tissue mass along the inside of the hollow central area), the stem
cast of a giant lycopsid tree, which no one would ever seriously con-
sider calling a ?pith cast?. In this sense, a pith cast is only part of the
larger stem cast. The stem fossil, if cast, includes the coaly rind, part
of original outer tissue, and the sediment-?lled hollow central cavity
cast, which may include both the original hollow pith area, and any
area opened to in?ll by decay of the tissues lining the pith cavity,
which must have been considerable.
As with most conventional wisdoms, when situations or conditions
are encountered that do not ?t themodel, somehead-scratching ensues
generally followed by a digression explaining why the data are some-
how ?awed, thus preserving the truth as established by authority. An
example of this is provided by Arber (1918), who puzzled about the ab-
sence of ?infranodal canals? below the nodes in some calamitalean
specimens, traditionally thought to be pith casts. Arber concluded
that, in fact, such specimens are not pith casts in the strictest sense,
but ?incrustations of surfaces that lay external to the pith? (Arber,
1918, p. 213), which he called ?sub-medullary casts.? These sub-
medullary casts supposedly resulted from ?considerable decay? of the
outer margin of the pith cavity prior to casting. Arber credits his insight
to observation of a petri?ed stem in the collection of Dr. D.H. Scott,
re?ecting the unspoken certitude, well established by that time, that
calamitalean stems were universally woody. So, here is another exam-
ple calling for decay of the tissues external to the pith, from the inside
outward, yet still leaving the basic form of the pith, the ribs and
nodes, visible. This explanation, on the surface of it, fails the test of
?total evidence,? in that it explains one particular observation, the ab-
sence of small openings below the nodal plate in some calamitalean
stem fossils, while failing to account for many other factors, such as
the widespread nature of this phenomenon, the need for partial decay
of the most resistant parts of the stem, leaving the tissues most likely
to decay behind (in tens of thousands of examples), or the occurrence
of such morphologies in specimens that do not have cast central areas.
Our contention that most calamitalean stem fossils represent ex-
ternal stem casts, in turn, implies that many of these ?ood-basin
plants had a somewhat ?reed-like? growth habit quite similar to mod-
ern Equisetum, with little development of secondary xylem. In con-
trast there are the large diameter stems of Calamites gigas (Kerp,
1984; Taylor et al., 2009, ?g. 10.35), which have been interpreted as
succulents, with specialized mechanisms for water storage, permit-
ting them to live in dry environments (Barthel and R??ler, 1994;
R??ler and Noll, 2002; see Naugolnykh, 2005 for an alternative inter-
pretation), something quite different from the generally accepted
conceptions of Carboniferous calamitaleans. Such growth morphol-
ogies are markedly different from the woody calamitaleans, which
seem to have preferred swampy, stable substrates, both on peat and
clastic soils, but were themselves diverse in size and form. Thus, a sec-
ondary insight, in agreement with R??ler and Noll (2006), is that the
calamitaleans were much more diverse in terms of their growth ar-
chitectures and phylogenetic diversity than is generally appreciated.
Using due caution, we recognizing that not all calamitalean axes rep-
resent stem casts or were reed-like in habitat, and that genuine pith
casts do occur, the most unequivocal examples of which also show evi-
dence for wood development. All such things considered, however, the
adpression fossil record is dominated by the stem casts of reed-like
calamitaleans.
Acknowledgments
HFL acknowledges a NERC Advanced Fellowship held at Royal
Holloway, University of London. We are extremely grateful to Hans
Gastaldo, R.A., 1992. Regenerative growth in fossil horsetails (Calamites) following
14 W.A. DiMichele, H.J. Falcon-Lang / Review of Palaeobotany and Palynology 173 (2012) 1?14Steur (Ellecom, Netherlands), Bob Gastaldo (Colby College, USA), Mike
Rygel (SUNYNewYork, USA), Manfred Barthel (Berlin, Germany), Hans
Kerp (M?nster, Germany), Scott Elrick (Illinois State Geological Survey,
USA) and Melissa Grey (Joggins Fossil Institute, Canada) who kindly
provided the photographs used in Figs. 2A, 11, 12, 13, 14?15, 19 and
23, respectively. The British Geological Survey kindly provided the
images used in Figs. 2B, 3 and 19 via its Geoscenic Service. The
comments of Ronny R??ler (Chemnitz, Germany) and an anonymous
reviewer greatly improved this manuscript. We are also extremely
grateful to the editor, Hans Kerp (M?nster, Germany), who directed
us to some important historic literature and illustrations (i.e., Weiss,
1884), which we have used in Figs. 4, 5, and 6.
References
Anderson, B.R., 1954. A study of American petri?ed calamites. Annals of the Missouri
Botanical Garden 41, 395?418.
Andrews, H.N., 1952. Some American petri?ed calamitean stems. Annals of the Missouri
Botanic Garden 39, 189?206.
Andrews, H.N., Agashe, S.N., 1965. Some exceptionally large calamite stems. Phytomor-
phology 15, 103?108.
Appleton, P., Malpas, J., Thomas, B.A., Cleal, C.J., 2011. The Brymbo fossil forest. Geology
Today 27, 107?113.
Arber, E.A.N., 1918. A note on submedullary casts of coal-measure calamites. Geological
Magazine 20, 212?214.
Barthel, M., R??ler, R., 1994. Calamiten aus dem Oberrotliegend des Th?ringer Waldes ?
Was ist ?Walchia imbricata?. Abhanlungen Naturhistorisches Museums Schleusingen
9, 69?80.
Barthel, M., R??ler, R., 1996. Pal?ontologische Fundschichten im Rotliegend von
Manebach (Th?r. Wald) mit Calamites gigas (Sphenophyta). Ver?ffentlichungen
des Naturhistorischen Museums Schleusingen 11, 3?21.
Calder, J.H., Gibling, M.R., Scott, A.C., Davies, S.J., Hebert, B.L., 2006. A fossil lycopsid forest
succession in the classic Joggins section of Nova Scotia: paleoecology of a
disturbance-prone Pennsylvanian wetland. In: Greb, S., DiMichele, W.A. (Eds.), Wet-
lands through Time: Geological Society of America, Special Paper, 399, pp. 169?196.
Cao, Q.V., Dean, T.J., Baldwin, C., 2000. Modelling the size?density relationship in
direct-seeded slash pine stands. Forest Science 46, 317?321.
Cichan, M.A., Taylor, T.N., 1983. A systematic and developmental analysis of Arthropitys
deltoides sp. nov. Botanical Gazette 144, 285?294.
Cleal, C.J., Thomas, B.A., 1994. Plant Fossils of the British Coal Measures. Palaeontological
Association, London. 222 pp.
Dawson, J.W., 1851. Notice on the occurrence of upright Calamites near Pictou, Nova
Scotia. Quarterly Journal of the Geological Society of London 7, 194?196.
Dawson, J.W., 1868. Acadian Geology. Macmillan and Company, London. 694 pp.
Dawson, J.W., 1892. The Geological History of Plants. D. Appleton and Company, New
York. 290 pp.
de Kroon, H., Kalliola, R., 1995. Shoot dynamics of the giant grass Gynerium sagittatum
in Peruvian Amazon ?oodplains, a clonal plant that does show self-thinning.
Oecologia 101, 124?131.
DiMichele, W.A., Falcon-Lang, H.J., 2011. Fossil forests in growth position (T0 assemblages):
origin, taphonomic biases and palaeoecological signi?cance. Journal of the Geological
Society 168, 585?605.
DiMichele, W.A., Falcon-Lang, H.J., Nelson, W.J., Elrick, S., Ames, P.R., 2007. Ecological
gradients within a Pennsylvanian mire forest. Geology 35, 415?418.
DiMichele, W.A., Nelson, W.J., Elrick, S., Ames, P.R., 2009. Catastrophically buried
Middle Pennsylvanian Sigillaria and calamitean sphenopsids from Indiana, USA:
what kind of vegetation was this? Palaios 24, 159?166.
DiMichele, W.A., Chaney, D.S., Kerp, H., Lucas, S.G., 2010. Late Pennsylvanian ?oras in
western equatorial Pangea, Ca?on del Cobre, New Mexico. New Mexico Museum
of Natural History and Science Bulletin 49, 75?113.
Enquist, B.J., Niklas, K.J., 2001. Invariant scaling relations across tree-dominated com-
munities. Nature 410, 655?660.
Enquist, B.J., West, G.B., Brown, J.H., 2009. Extensions and evaluation of a general quan-
titative theory of forest structure and dynamics. Proceedings of the National Acad-
emy of Sciences of the USA 106, 7046?7051.
Falcon-Lang, H.J., 1999. Fire ecology of a Late Carboniferous ?oodplain, Joggins, Nova
Scotia. Journal of the Geological Society of London 156, 137?148.
Falcon-Lang, H.J., 2006. Vegetation ecology of Early Pennsylvanian alluvial fan and
piedmont environments in southern New Brunswick, Canada. Palaeogeography,
Palaeoclimatology, Palaeoecology 233, 34?50.
Falcon-Lang, H.J., Benton, M.J., Braddy, S.J., Davies, S.J., 2006. The Pennsylvanian tropical
biome reconstructed from the Joggins Formation of Canada. Journal of the Geolog-
ical Society of London 163, 561?576.
Falcon-Lang, H.J., Gibling, M.R., Benton, M.J., Miller, R.F., Bashforth, A.R., 2010. Diverse
tetrapod trackways in the Lower Pennsylvanian Tynemouth Creek Formation,
southern New Brunswick, Canada. Palaeogeography, Palaeoclimatology, Palaeoe-
cology 296, 1?13.
Gastaldo, R.A., 1986. Implications on the paleoecology of autochthonous Carboniferous
lycopods in clastic sedimentary environments. Palaeogeography, Palaeoclimatology,
Palaeoecology 53, 191?212.burial by alluvium. Historical Biology 6, 203?220.
Gastaldo, R.A., Demko, T.M., Liu, Y., Keefer, W.D., Abston, S.L., 1989. Biostratinomic pro-
cesses for the development of mud-cast logs in Carboniferous and Holocene
swamps. Palaios 4, 356?365.
Gastaldo, R.A., Stevanovi?-Walls, I.M., Ware, W.N., 2004. Erect forests are evidence for
coseismic base-level changes in Pennsylvanian cyclothems of the Black Warrior
Basin, U.S.A. In: Pashin, J.C., Gastaldo, R.A. (Eds.), Sequence Stratigraphy, Paleoclimate,
and Tectonics of Coal-bearing Strata. AAPG Studies in Geology, 51, 219?238.
Gradzinski, R., Doktor, M., 1995. Upright stems and their burial conditions in the coal-
bearing Mudstone Series (Upper Carboniferous), Upper Silesia Coal Basin, Poland.
Studia Geologica Polonica 108, 129?147.
Hooker, J.D., Binney, E.W., 1854. On the structure of certain limestone nodules
enclosed in seams of bituminous coal, with a description of some Trigonocarpons
contained in them. Philosophical Transactions of the Royal Society of London
145, 149?156.
Kerp, J.H.F., 1984. Aspects of Permian Palaeobotany and Palynology. V. On the nature of
Asterophyllites dumasii Zeiller, its correlation with Calamites gigas Brongniart and
the problem concerning its sterile foliage. Review of Palaeobotany and Palynology
41, 301?317.
Kidston, R., Jongmans, W.J., 1917. A monograph of the Calamites of Western Europe.
Medeelingen van de Rijkopsporing van Delfstoffen, 7. 207 pp.
Leeder, M.R., Nami, M., Scott, A.C., Gardiner, A., 1984. Studies on the sedimentology,
palaeontology and palaeoecology of the Middle Coal Measures, (Westphalian B,
Upper Carboniferous) at Swillington, Yorkshire. II. Centroclinal cross-strata pre-
served around an in situ vertical tree trunk. Transactions of the Leeds Geological
Association 10, 17?21.
Nadon, G.C., 1998. Magnitude and timing of peat-to-coal compaction. Geology 26,
727?730.
Naugolnykh, S.V., 2005. Permian Calamites gigas Brongniart, 1828: the morphological
concept, paleoecology, and implications for paleophytogeography and paleoclima-
tology. Paleontological Journal 39, 321?332.
Peterson, C.J., Jones, R.H., 1997. Clonality in woody plants: a review and comparison
with clonal herbs. In: de Kroon, H., van Groenendael, J. (Eds.), The Ecology and Evo-
lution of Clonal Plants. Backhuys Publishers, Leiden, pp. 263?289.
Philips, T.L., DiMichele, W.A., 1992. Comparative ecology and life history biology of ar-
borescent lycopods in Late Carboniferous swamps of Euramerica. Annals of the
Missouri Botanical Garden 79, 560?588.
Reed, F.D., 1952. Arthroxylon, a rede?ned genus of Calamite. Annals of the Missouri
Botanic Garden 39, 173?187.
Renault, B., 1893?1896. Bassin houiller et Permian d'Autun et d'Epinac. Flore Fossile, 2e
partie. Gites Min?raux de la France. 578 pp.
R??ler, R., Noll, R., 2002. Der permische versteinerte Wald von Araguaina/Brasilien ?
Geologie, Taphonomie und Fossilf?hrung. Ver?ffentlichungenMuseum f?rNaturkunde,
Chemnitz 25, 5?44.
R??ler, R., Noll, R., 2006. Sphenopsids of the Permian (I): the largest known anatomi-
cally preserved calamite, an exceptional ?nd from the petri?ed forest of Chemnitz,
Germany. Review of Palaeobotany and Palynology 140, 145?162.
R??ler, R., Noll, R., 2007. Calamitea Cotta, the correct name for calamitean sphenopsids
currently classi?ed as Calamodendron Brongniart. Review of Palaeobotany and
Palynology 144, 157?180.
Rygel, M.C., Gibling, M.R., Calder, J.H., 2004. Vegetation-induced sedimentary struc-
tures from fossil forests in the Pennsylvanian Joggins Formation, Nova Scotia.
Sedimentology 51, 531?552.
Seward, A.C., 1898. Fossil Plants, Volume 1. Cambridge University Press, Cambridge.
452 pp.
Scott, A.C., 1978. Sedimentological and ecological control of Westphalian B plant
assemblages from West Yorkshire. Proceedings of the Yorkshire Geological Society
41, 461?508.
Scott, A.C., Galtier, J., Mapes, R.H., Mapes, G., 1997. Palaeoecological and evolutionary
signi?cance of anatomically preserved terrestrial plants in Upper Carboniferous
marine goniatite bullions. Journal of the Geological Society of London 154,
61?68.
Silvertown, J.W., Doust, J.L., 1993. Introduction to Plant Population Biology, 3rd Edition.
Blackwell Scienti?c, Oxford. 210 pp.
Stopes, M.C., 1907. Note on a wounded Calamite. Annals of Botany 21, 277?280.
Taylor, T.N., Taylor, E.L., Krings, M., 2009. Palaeobotany: the biology and evolution of
fossil plants. Academic Press. 1230 pp.
Wang, S.J., Hilton, J., Galtier, J., Tian, B., 2006. A large anatomically preserved calamitean
stem from the Upper Permian of southwest China and its implications for calami-
tean development and functional anatomy. Plant Systematics and Evolution 261,
229?244.
Weiss, C.E., 1884. Beitr?ge zur fossilen Flora. III. Steinkohlen?Calamarien. Abhandlungen
zur geologischen Specialkarte von Preussen. Das Band 5 (2) 28 plates.
West, G.B., Enquist, B.J., Brown, J.H., 2009. A general quantitative theory of forest struc-
ture and dynamics. Proceedings of the National Academy of Sciences of the USA
106, 7040?7045.
Williamson, W.C., Scott, D.H., 1895. Further observations on the organization of the fos-
sil plants of the coal measures. Pt. I. Calamites, Calamostachya, and Sphenophyllum:
Philosophical Transactions of the Royal Society, London, Series B, 185, pp. 863?959.
Winston, R.B., 1986. Characteristics, features and compaction of plant tissues traced
from permineralized peat to coal in Pennsylvanian coals (Desmoinesian) from
the Illinois basin. International Journal of Coal Geology 6, 21?41.