Geoscience Reference
In-Depth Information
Seed dispersion
Even although morphology often indicates the
general means of dispersal, an apparently obvious
modification for dispersal may not always predict
the actual process (Howe & Smallwood 1982).
Many studies on the ecology of seed dispersal
(Howe & Smallwood 1982; Willson & Traverset
2000) agree on the complexity of the mechanisms
involved. Propagule dispersal can only be well inter-
preted in natural environments with direct obser-
vations. Pioneering works (Ridley 1930) as well as
recent studies (Cain et al. 2000; Tackenberg et al.
2003) have however provided much information on
living plants that can be used to infer about fossils.
Even although Upper Devonian seeds encompass a
wide range of morphological variations, nearly all
observed morphologies point to wind as the major
dispersal agent. Other means cannot be excluded,
but the available information is poor.
DiMichele et al. (1989) discussed for Tournai-
sian spermatophytes a similar diversity pattern
centred on integument architecture only. In their
model, seed plants are seen as having evolved
in an 'open' selective landscape; natural selection
does not appear to have strongly influenced early
seed plant diversification. This low interspecific
competition allowed seed plants to explore various
morphologies and thus radiate. The underlying
hypothesis is that seed plants established in large
areas with reduced or no biotic competition; survi-
val in such ecosystems was controlled by abiotic
selection processes. In order to correctly understand
this abiotic selection context, the environment has
to be characterized. Geology, and more precisely
sedimentology, is the only available tool.
large specimens of Moresnetia replaced by pure
Rhacophyton layers. Since this succession occurs
within less than 1 m of sediment, we tentatively con-
clude that it represents an ecological succession; this
is a reasonable support for the Rothwell & Scheck-
ler's (1988) hypothesis.
Cressler (2006) investigated the Upper Devonian
landscape and its biotic associations at the extensive
Red Hill locality. He studied the sedimentology of
the different layers from Red Hill precisely. He
was then able to place the various fossiliferous
layers in the landscape, identify the plant fossili-
ferous layer as a floodplain pond and locate the
parautochtonous plant fossils in both successional
sequence and distance from the shore. This precise
positioning, coupled with a layer by layer statistical
study of plant remains provided the spatial organiz-
ation of these deposits. Considering all deposition
biases, Cressler (2006) was able to estimate the
original growth position of the vegetation on the
landscape. His model proposes that niche partition-
ing occurred at a high taxonomic level during the
Upper Devonian, exactly as suggested by DiMi-
chele & Bateman (1996) for the Carboniferous.
Due to a well-documented argumentation, they were
able to demonstrate ecological affinities of entire
clades at the class level. The terrestrial surface
was thus divided in sub-environments within each
class-level taxa. The Upper Devonian environment
predates this organization with Progymnosperms
(Archaeopteris) established on the well-drained
areas, ferns (Rhacophyton) growing as monotypic
populations widespread in the surrounding land-
scape, lycopsids living on the edge of ponds and
spermatophytes being opportunistic pioneer plants
in burned areas.
Taffs Well and the Baggy Beds in Great Britain
represent continental and marine deposits, respect-
ively. The Taffs Well continental deposits present
a more diverse assemblage, seed plants included.
Only one species, Xenotheca devonica (belonging
to the Moresnetia-type), is represented at Baggy
Beds (Hilton & Edwards 1999). The occurrence of
the single Moresnetia-type in the most distal depos-
its has been interpreted as an isolated event that
drifted one particular population of that plant. This
is probably not the case, as the extensive, mainly
neritic Belgian deposits gave similar results with
seed plants of
Early seed plant ecology
Rothwell & Scheckler (1988) were the first to tenta-
tively locate early seeds in the Upper Devonian
environments. They used the geological settings to
extrapolate early seed plant ecology and concluded
that the earliest gymnosperms were 'pioneer coloni-
zers of newly emerged, primary successional habi-
tats near shorelines'. They based this assumption
on a particular succession of autochotonous to
hypoautochtonous beds at Elkins.
First mats of only seed plants are observed, then
monotypic Rhacophyton (early fern) layers and,
finally, a diverse allochtonous assemblage. The suc-
cessions present in Belgium, at the Langlier quarry,
are roughly similarly organized (Thorez J. pers.
comm., 2006), and provided some of the best pre-
served specimens of Moresnetia (Fairon-Demaret
& Scheckler 1987). This succession begins with
an almost pure bed of Archaeopteris (progymno-
sperm) followedby a thick bed of profusely branched
the Moresnetia-type being the
most common.
The enormous fossil plant material collected
by Fran¸ois Stockmans, who did one of the most
complete and systematic Upper Devonian fossil
plant collections, comes from 25 quarries and out-
crops. In this collection, Moresnetia represents 60%
of the spermatophyte remains. The remaining 40%
are covered by the other seed types present in
Belgium (Condrusia-type, Dorinnotheca-type and
