Paleobotanists have likened the splitting open of sedimentary rocks and finding fossil plants contained within to the opening of Christmas presents. The anticipation, the awe, and the following pleasure and excitement all fuel the analogy – but it breaks down eventually, as the fossils come with neither a parts list nor the ever-important instruction manual. Paleobotanists are then left with major questions. What does the presence of this suite of fossils in the rock really mean? What inferences and conclusions can be made about the past vegetation based on the presence of these – and equally, the absence of other – plant remains? Paleobotany, as well as paleozoology, is in part driven by uniformitarianism – if even for no other reason than the fact that only in extant plants and animals can we observe and experiment with their environmental relationships. If one assumes that modern, observable, basic physiological and environmental principles can be applied to past vegetation, then it is possible to make inferences about those floras. It is into this arena that the paper by Hovenden & Van der Schoor (this issue, pp. 581–590) can be placed.
‘A fossil deposit is a wondrous thing – it can comprise layers of rock that each hold a ‘freeze-frame’ from the moving picture that is the evolution of life’
Good news for paleoecologists
By elegantly testing the age old ‘heredity vs environment’ argument in the leaf characteristics of extant Nothofagus cunninghamii (Hook) Oerst. from Tasmania (Australia), Hovenden & Van der Schoor have not only provided excellent data on the growth and interpretation of this species through time, but have helped outline numerous clear signs (both warnings and guideposts) to other researchers in the fields of foliar physiognomy and general paleobotany. The specific aim of the paper is to test the relative contribution of genotypic control and environmental plasticity as it applies to some major (and frequently used) foliar physiognomic characters, including leaf length, area, thickness and stomatal density. The authors cite past work as showing strong genotypic control of leaves in this species, but the ‘good news’ from their data (at least for paleoecologists) is that new results suggest that the environment shows a significantly greater effect in both leaf length and area, with thickness being the only character obviously more rigorously controlled by genotype. This last character is of little consequence to paleobotanical studies, in that only very rarely are fossils preserved in such a way as to allow total leaf thickness to be measured. The other important result presented in the paper (for paleobotanists again) is that this species supports the old adage elucidated by Bailey & Sinnott (1916) that ‘the more tropical the flora, the larger the leaves’ and the more recent corollary ‘the higher the altitude, the smaller the leaves (and as Hovenden & Van der Schoor reiterate, the greater the stomatal density)’.
In illustrating the great potential value this paper has to the discipline of paleobotany, there are three areas upon which to focus:
1The importance of species-specific studies, including an example of strong (and even inverse) results obtained from different species.
2The importance of multiple life-forms in such studies.
3The importance of the effect that dispersal of leaves prior to deposition may have on interpreting fossil deposits or, more specifically, the effect that leaves with an origin at higher elevation relative to their deposition may have on a physiognomic signature at that location.
Will all relatives behave the same?
There are some instances in this natural world where one can take comfort in assuming that relationships with natural (and unnatural) processes are shared between closely related organisms. For example, it can be assumed that a deadly toxin known to be lethal to frogs in one stream will penetrate the porous skin of other frog species and kill them if their habitat also becomes polluted with this toxin. Frogs are frogs in this example and the species (or genus) will likely have little consequence on the observable outcome. For the most part such obvious broadly based assumptions can be considered safe hypotheses, but there always needs to be great care taken to make sure that you aren’t giving too much ecological credit where it isn’t due.
We would like to introduce a case from another Gondwanic family, the Cunoniaceae, where assumptions regarding the responses to change in environmental conditions cannot be shared among species. Recently collected data on two members of this family have shown greatly different morphological responses related to changes in environmental conditions associated with changing elevation (P. Gordon, unpublished). Specifically, these two species (Geissios biagiana (F. Muell.) F. Muell. and Pullea stutzeri (F. Muell.) Gibbs) show opposite linear relationships in stomatal density with increase in altitude (Fig. 1). This study was conducted on Mt Lewis, on the northern edge of the Atherton Tablelands in northern Queensland, Australia.
This relates to Hovenden & Van der Schoor's rightful assertion that there is strong evidence for the degree of environmental plasticity and adaptation being species dependant. Another example can be found in the Lauraceae (cinnamon family) where it has been shown that correlations between physiognomic features and the environment are predictable for Neolitsea dealbata (R.Br.) Merr. (Greenwood et al., 2003). However, other taxa in this family (e.g. Litsea leefiana (F. Muell.) Merr.) do not show this trend (D. Christophel, unpublished).
With these studies in mind, we are compelled to make some cautionary comments directed towards paloebotanists and paleoecologists who attempt to place extant species/environment relationships on extinct taxa of assumed close relation in order to obtain proxy paleoenvironmental data. With relatively little published in the area it is our opinion that the number of examples where there are found to be species-specific responses will only continue to grow. This possible confounding factor is one that needs close and careful consideration and the Hovenden & Van der Schoor paper helps to highlight the issue.
Vines, herbs and shrubs
A fossil deposit is a wondrous thing – it can comprise layers of rock that each hold a ‘freeze-frame’ from the moving picture that is the evolution of life. What makes up this abstract picture of a past environment? The study of fossil plant deposits is a specific and important example of where this basic question needs to be asked. What is the habit of the plants from which the leaves, flowers, fruits and seeds come? It cannot be assumed that all leaves belonging to one species come from one plant, and it would be equally naive to assume that all leaves in the deposit came from only trees without any input from vines, herbs or shrubs. The study of foliar physiognomy seems to be heavily biased towards tree species and there is little available literature describing the morphological relationships of other plant forms to their surrounding environmental conditions. While the target species of Hovenden & Van der Schoor's paper is also a tree species, they mention several relatively recent studies that focus on the foliar physiognomy of other life forms. This research is invaluable in helping to understand and decipher any proxy-related estimation of paleoclimate.
If differing plant forms are shown to have dissimilar morphological responses to environmental conditions, then it is crucial to take account of the relative proportion of these forms and factor this into the interpretation of a fossil flora – or at least to be aware of the different signatures that components may yield. The reality remains that except in a few obvious cases, unless the identity of fossil taxa can be determined, its life form may remain unknown. Some exceptions do exist, however, in that the presence of regular chordate bases, tendrils, or sharply bent petioles will often suggest vines.
Using the previously mentioned tree species (G. biagiana) and two vine species that are also common in the rainforest habitat of north-eastern Queensland (Australia), Hibbertia scandens (Wild.) Gilg and Maesa dependens F. Muell., we have compared the response in leaf length to increasing elevation (Fig. 2). Here it is seen that the two vine species chosen show very similar, significant reductions in leaf length with increasing altitude, as does the selected tree species. As previously mentioned, this fits with the generalization cited by Hovenden & Van der Schoor, and clearly suggests that plant forms such as vines can and do show very similar foliar physiognomic responses to the one commonly cited for trees. This group of plants, however, could well show species-specific rather than genus or family specific responses as found in trees, and further work is required.
While Webb's (1959) classification of the Australian rain forests was based in large part on the physiognomy of canopy leaves, many of his forest types are named and qualified by the presence and/or frequency of woody vines, which implies their importance in the vegetation. The work of Greenwood (1992) in examining the physiognomic signatures of leaf litter from Webb's seminal localities, modified Webb's rainforest classification system to make it more useful to paleobotanists. Greenwood found a significant drop in average size for leaves in the litter of each of the rain forest types relative to the canopy. While he attributed much of it to taphonomic bias, we believe that some of that bias was likely caused by ‘compositional’ bias – for example, the litter being composed of more life forms. Quantification of this on a ‘per species’ basis is required to conform with the conclusions and suggestions found in the Hovenden & Van der Schoor paper concerning species-specific variations.
Spacial and altidudinal dispersal: autochthany versus allochthany
A final aspect of Hovenden & Van der Schoor's paper that holds strong relevance to the field of paleobotany is their noting of the significant variation in the foliar physiognomic response seen in N. cunninghamii over the relatively small geographic scale of 15 km. The idea of long or even short distance dispersal of leaves before their final deposition has long been of concern to paleobotanists interested in reconstruction of paleoclimate (e.g. Hill, 1981). Mt Lewis, where the two studies illustrated earlier were conducted, covers approx. 25 km across its altitudinal gradient. Leaves of one highly distinctive taxon [Stenocarpus davalliodes D. Foreman and B. Hyland (Proteaceae)] which occurs only in small stands at an altitude of 1190 m have been found along stream beds at 670 m, at least 7 km from the nearest tree (D. Christophel, personal observation). Thus paleoecological interpretation of the climate of a given flora must also consider the relative allochthany/autochthany of the plant specimens included.
The paper by Hovenden & Van der Schoor may therefore be seen as a very significant contribution, not only to the understanding of the biology of a very important Tasmanian tree species, but as an excellent example of the type of studies necessary for our ultimate understanding of the relationship between heredity and environment both in the contemporary scene and through time. The importance of this kind of research to the field of paleoecology cannot be overstated, and paleobotanists will continue to benefit from their and related future studies.