ELSEVIER
Palaeogeography, Palaeoclimatology, Palaeoecology 149 (1999) 15–26
Leaf venation density as a climate and environmental proxy: a critical review and new data D. Uhl Ł , V. Mosbrugger Institut und Museum fu¨r Geologie und Pala¨ontologie der Universita¨t Tu¨bingen, Sigwartstrasse 10, D-72076 Tu¨bingen, Germany Received 27 January 1997; revised version received 23 July 1997; accepted 8 June 1998
Abstract In palaeobotany, leaf venation density is still primarily used as a taxonomic character although numerous studies on recent plants reveal that leaf venation density may be influenced by various environmental factors. To promote the use of leaf venation density as a palaeoclimate=palaeoenvironmental proxy we give a brief review of these studies and provide some additional data. Our review shows that environmental factors that increase transpiration of plants or decrease water availability also tend to increase the leaf venation density. Based on the analysis of leaves of some recent and fossil plants we found: (1) Venation density may be measured as vein length per area or as distance between veins, but the first parameter is more reliable. (2) Depending on the plant species, leaf venation density may or may not vary with leaf size. This ‘leaf size effect’ has to be taken into account when leaf venation density is to be used as a palaeoclimate= palaeoenvironmental proxy. (3) No significant effect of a changing atmospheric carbon dioxide concentration on leaf venation density was observed. (4) In Permian seed plants, intraspecific variation of leaf venation density was similar to that observed in modern angiosperms. Obviously, even in these seed plants, leaf venation density can be used as a palaeoclimate=palaeoenvironmental proxy. 1999 Elsevier Science B.V. All rights reserved. Keywords: taphonomy; palaeoclimatology; palaeoecology; leaf venation density
1. Introduction Ever since the beginning of palaeobotany as a modern science, scientists have used plants as indicators of palaeoclimates. Even pioneers like Schlotheim and Brongniart recognized the potential of fossil plants as ‘keys’ to past climates (cf. Chaloner and Creber, 1990). Since this time we have gained an increasing knowledge about fossil and recent plants, which can help us to reconstruct palaeoclimates and to understand the influence of climate change on plants. Ł Corresponding
author. Fax: C49-7071-949040; E-mail:
[email protected]
Basically, three different approaches have been used to obtain palaeoclimatic or palaeoenvironmental information from plant fossils. The first is based on a comparison of fossil with recent taxa and is known as the ‘nearest-living-relative’ method (Chaloner and Creber, 1990): it is assumed that the climatic=environmental requirements of fossil taxa are similar to those of their nearest living relatives. The most recent variation of this nearest-living-relative approach has been described by Mosbrugger and Utescher (1997). In a second type of approach the chemical signatures of fossil plants, in particular their δ13 C, is used as a climate=environmental proxy (e.g. Frielingsdorf, 1992). In contrast, the ‘constructional morphology ap-
0031-0182/99/$ – see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 1 - 0 1 8 2 ( 9 8 ) 0 0 1 8 9 - 8
16
D. Uhl, V. Mosbrugger / Palaeogeography, Palaeoclimatology, Palaeoecology 149 (1999) 15–26
proach’ is based on the assumption that certain constructional features of plants (e.g. growth rings in wood, drip tips of leaves, etc.) can be used as climate=environmental proxies and are largely independent of the systematic position of the plant groups. Well-known examples of this type of approach are the leaf physiognomy analysis (e.g. MacGinitie, 1974; Greenwood, 1992; Wolfe, 1993) to derive various climatic parameters and the use of stomatal densities for estimations of ancient atmospheric carbon dioxide concentrations (e.g. Beerling and Chaloner, 1993a,b, 1994; McElwain et al., 1995). Thus, fossil land plants can be used in various ways as climatic proxies and various methods and approaches are available. However, all these methods and approaches have their inherent shortcomings, limitations and problems; in fact, there is no optimal, universally applicable and absolutely reliable technique for deriving palaeoclimatic=palaeoenvironmental information from fossil plants (cf. Birks and Birks, 1980; Chaloner and Creber, 1990; Mosbrugger and Utescher, 1997; Wilf, 1997; Wilf et al., 1998; Wiemann et al., 1998). Hence, we need a multitude of approaches and techniques and more research is needed to learn how plants can be used as palaeoclimate=palaeoenvironmental proxies. In this paper we concentrate on the problem if the density of leaf venation can be used as a climate=environmental proxy. The dependence of leaf venation density on environmental parameters has been analysed repeatedly, in particular in the first half of this century, but the results of these studies have found almost no application in terrestrial palaeoclimatology or palaeoecology; instead, leaf venation density is still most commonly used in palaeobotany as a taxonomic character. In the following we present a critical review of previous studies and some new data to illustrate the problems and possibilities of the use of leaf venation density as a palaeoclimate=palaeoenvironmental proxy.
2. Known facts In the middle of the 19th century botanists used venation as a fixed character of leaf morphology, mainly as an aid in systematics (e.g. Ettingshausen, 1861). Even today venation is used in systematics
(e.g. Mory, 1992), but since the end of the last century scientists recognized that venation patterns and venation densities of plants are variable characters, influenced on an ontogenetic and evolutionary level by climatic and environmental parameters. Since the beginning of our century it is known, that in grasses venation density increases with increasing insertion height above the ground (Zalenski, 1904). This correlation was designated ‘Zalenski’s law’ (Maximov, 1929). It was confirmed for other grasses and herbs (e.g. Rippel, 1919; Lebedincev, 1927), for some trees (Manze, 1968) and for some temperate as well as tropical forests (Gorysina et al., 1961; Ho¨germann, 1990; Roth and Yee, 1991). In some plants, however (e.g. Populus species: Critchfield, 1960; Hedera helix, Mahonia grandiflora: Lalanne, 1890; Amygdalus nana D Prunus tenella: Barykina, 1967) length of venation per area decreases with increasing height above ground. Leaf size may also influence venation density. Gupta (1961) observed an increase of venation density with increasing leaf size in Nicotiana tabacum, whereas Schuster (1908) reports a decrease of venation density with increasing leaf size in Pisum sativum. Manze (1968) found no influence of leaf size in Acer platanoides. Only in very small adult leaves he observed that the number of veins per cm was larger than in normal-sized leaves. It is long known that sun and shade leaves differ markedly in their leaf size (sun leaves are smaller than shade leaves) and hence it is expected that they also differ in their venation density. In fact, in most of the investigated plants (including deciduous trees, evergreen trees, bushes and herbs) sun leaves have a higher venation density than shade leaves (e.g. Schuster, 1908; Schramm, 1912; Wylie, 1951; Manze, 1968). Schuster (1908) demonstrated that the observed difference in venation density between sun and shade leaves cannot entirely be explained by the difference in leaf size. Only in a few of the investigated species (e.g. Philadelphus suspensa, Symphoricarpus racemosus) there was no difference in venation density between sun and shade leaves, and in Oxalis corniculata, venation density was higher in shade leaves than in sun leaves (Schuster, 1908). In experimental studies with various herbaceous plants illumination power and growing temperature also had a positive effect on leaf venation density
D. Uhl, V. Mosbrugger / Palaeogeography, Palaeoclimatology, Palaeoecology 149 (1999) 15–26
(Heilbronn, 1926; Maximov, 1929; Hughes, 1959; Chonan, 1967; Thiraporn and Geisler, 1978). Even shortage of nutrients may increase length of veins per area, as was observed in experimental studies with some herbs and grasses (e.g. Marenkov et al., 1971; Bolton and Brown, 1980; Ka¨mpf and Lange, 1983) as well as in field studies including trees (Beiguelman, 1962) and shrub-bog plants (Philpott, 1956). Presumably the most important factor influencing venation density is water. In a study including 93 species, mostly herbs and some deciduous trees, Zalenski (1902) observed that venation density is higher in plants from dry habitats than in plants from mesic habitats. He demonstrated that venation densities of systematically unrelated plants from one locality differ less from each other than venation densities of different species of one genus or family from different localities. In other field observations the correlation between venation density and dryness of a locality was confirmed for many trees, bushes and herbs (e.g. Zeuner, 1932; Keller, 1933; Manze, 1968; Herbig and Kull, 1991).
17
Further evidence for a correlation between venation density and water availability comes from experimental studies. Length of veins per area increased with experimentally induced increasing dryness of the soil and the surrounding air in several herbs (Napp-Zinn, 1984, 1988, and citations therein). Eberhard (1903) and Lebedincev (1927), however, also observed an increase in venation density in plants grown under nearly saturated air humidity. Such an effect was also reported for tropical rainforests, where humidity is normally close to the saturation point (Pyykko¨, 1979). Table 1 summarizes the environmental factors that may influence leaf venation density. This compilation illustrates that most factors which increase transpiration rates and reduce water availability tend to increase the leaf venation density but also the stomatal density (measured as number of stomata per area or as stomatal index). This may also explain why in grasses venation density increases with increasing windspeed (Whitehead, 1964–1965; Grace and Russell, 1977). It should also be noted, however,
Table 1 Influences of different factors on leaf venation density, stomatal density and transpiration a Influence on venation density
Influence on Influence on stomatal density transpiration
Investigated in
Data from (partim)
C( )
C( )
C
grasses, herbs, trees
Zalenski, 1902; Manze, 1968 Manze, 1968;Wylie, 1951
Increasing insertion level Sun-leaves vs. shade leaves Increasing temperature
C( )
C( )
C
grasses, herbs, trees
C
N.D.
C
Increasing soil dryness
C
C
C
Tropaeolum coccineum, Zea mays grasses, herbs, trees
Decreasing air moisture Nutrient deficiency
Cb
Cb
Cb
grasses, herbs, trees
C
C
C
grasses, herbs, trees
Increase of atmospheric CO2 Leaf size
0* C=
* 0 *,c
=0
N.D.
mostly trees
Heilbronn, 1926; Thiraporn and Geisler, 1978 Napp-Zinn, 1984, 1988 and citations therein Napp-Zinn, 1984, 1988 and citations therein Bolton and Brown, 1980; Ka¨mpf and Lange, 1983 Ku¨rschner, 1996; this work
herbs, trees
Gupta, 1961; Manze, 1968
C D increase; C ( ) D decrease only in a few species observed, in the most species increase; D decrease, 0 D no influence; N.D. D not determined. a Compilation of some environmental factors influencing venation density, stomatal density and transpiration. The data are compiled from the literature cited in Section 2 and from our study (data marked with * cf. Section 4). It is obvious that factors increasing transpiration also increase venation density and stomatal density. The main problem within this list, is that most of these data only show qualitative tendencies and only a few show statistically reliable quantitative results. b In nearly saturated air moisture venation density can be very high. c Probably species specific.
18
D. Uhl, V. Mosbrugger / Palaeogeography, Palaeoclimatology, Palaeoecology 149 (1999) 15–26
that there is no universally valid correlation between leaf venation density and environmental parameters: different plant species may react in a different way. Hitherto there have been only two investigations to determine palaeoclimatic data with the aid of venation density. Zeuner (1932, 1936) studied leaves from the Upper Miocene of Oehningen. Based on leave venation density data he concluded that transpiration rates in the Upper Miocene of Oehningen were higher than today and similar to those in the recent Mediterranean area. Manze (1968) investigated venation densities of leaves from the Upper Miocene of Oehningen and Schrotzburg and from the Upper Miocene=Lower Pliocene of the Lower Rhine Embayment to derive qualitative estimates of palaeoprecipitation. Since the work of Manze (1968) only a few authors (e.g. Wolfe, 1977; Mai, 1995) mentioned that leaf venation could probably be a useful tool in terrestrial palaeoclimatology but no detailed study in this field was undertaken.
3. Problems and questions As shown above a lot of information is available about the influence of various climatic and environmental factors on venation density. Despite these numerous studies, however, there are still many open questions and problems. First of all, most investigations cited above use a qualitative approach or a statistically insufficient database. Hence, many of the results lack the quantitative basis that is necessary if leaf venation density is to be used as a climatic=environmental proxy. Moreover, the various researchers used different methods to determine venation density: either they measured the length of veins per area (e.g. Zalenski, 1902), or the distance between two veins (e.g. Wylie, 1951) or the number of veins per cm (Manze, 1968). Up till now it is not clear which of these methods gives the best results and whether the results obtained with these methods can be correlated. In addition, still very little is known about the natural variation of venation density within one leaf, one plant, one population=one locality or one species. For instance, Manze (1968) investigated the venation density variation in leaves from single trees, but his
data set is generally too small (less than 10 leaves per tree) to provide a realistic idea about the frequency distribution of leaf venation density within one tree. An even more serious problem relates to the size dependence of leaf venation density. As shown above, in many plants, leaf venation density decreases with increasing leaf size (which is obviously a consequence of the mode of leaf ontogeny; cf. Schuster, 1908). This leaf size effect has to be taken into account, if leaf venation density is correlated with parameters other than leaf size. However, almost all previous studies that try to link leaf venation density to environmental parameters largely neglect the leaf size effect (one of the few exceptions is given by Schuster (1908). Most of the previous studies have only considered dicotyledonous angiosperms. Correspondingly, almost nothing is known about the factors that influence leaf venation density in other plant groups such as ferns and cycads which dominated the terrestrial floras for a long time in the geological past. We also ignore whether the atmospheric carbon dioxide concentration influences the leaf venation density. One would possibly expect that leaf venation density decreases with increasing carbon dioxide because in many cases stomatal density and leaf venation density react in a similar way to changes in environmental factors (cf. Table 1). Based on new data some of these questions and problems will be addressed in the following sections.
4. Material and methods We have studied the leaf venation density of both fossil and recent plants. For our analyses of recent leaves we have concentrated on two species, i.e. Acer monspessulanum L. and Quercus petraea Liebl., both are common deciduous trees of temperate to warm-temperate Western Europe. To analyse whether the leaf venation density has changed over the last 150 years as a result of an increasing atmospheric carbon dioxide content we have used herbarium material of Acer monspessulanum L. and Quercus petraea Liebl., collected since 1840 from the Palatinate and the Nahe area (both southwestern Germany); the material was obtained from the Pfalzmuseum fu¨r Naturkunde (Bad Du¨rkheim, southwest-
D. Uhl, V. Mosbrugger / Palaeogeography, Palaeoclimatology, Palaeoecology 149 (1999) 15–26
ern Germany). We restricted the herbarium material to the Palatinate and Nahe area to minimize the climatic variation between the various collecting sites which could conceal the influence of changing carbon dioxide contents on the leaf venation densities. In addition, leaf material of Acer monspessulanum L. was collected in June 1996 in the Nahe area (southwestern Germany). Quercus petraea material, grown under 350 and 700 ppm CO2 in a gas chamber, was obtained from Dr. W.M. Ku¨rschner, University of Utrecht (for additional information, see Ku¨rschner, 1996). As fossil leaves we studied Cordaites, Taeniopteris and Compsopteris from a Permian flora of China, that covers the Lower Permian and the lower Upper Permian (Mosbrugger, 1987); the material was obtained from the Swedish Museum of Natural History in Stockholm (Sweden). We used a two-dimensional and one-dimensional approach to describe the leaf venation density. The two-dimensional approach is based on camera lucida drawings of the leaf venation.The drawings were scanned and the vein length per area was analysed with the commercial image analysis system OPTIMAS 4.101. In addition, various other venation parameters were measured (e.g. areole size, number of free veinlets=area) but in the following we will largely concentrate on the parameter vein length per area.The two-dimensional approach is very timeconsuming. It has only rarely been used in previous studies and in these few cases vein length per area is generally the only parameter that has been measured (e.g. Zalenski, 1902, 1904). The one-dimensional approach for the determination of the leaf venation density is based on the (mean) distance between veins. This parameter was obtained directly from the leaves with the aid of a binocular and an ocular micrometer: for each leaf ten measurements of the distances between 11 veins, that crossed a line parallel to the primary vein, where done. The distance between veins (one-dimensional approach) is much easier to determine than vein length per area (two-dimensional approach) and corresponds to the venation density parameter that was mostly used in previous studies (e.g. Wylie, 1951; Manze, 1968; cf. Section 2). To get information about the variation of leaf venation density within a leaf we measured venation parameters for four different sites, i.e. base, cen-
19
Fig. 1. Sketch of a leaf of Acer monspessulanum. The four different sites, where measurements were carried out, are indicated by squares. A D base; B D centre; C D top; D D margin of the leaf.
tre, top and margin, of Quercus petraea and Acer monspessulanum leaves (Fig. 1). In some parameters (e.g. vein length per area; Fig. 2A) there is no significant variation between the four sites, others (e.g. area of areoles; Fig. 2B) show a more pronounced within-
Fig. 2. Relative vein length per area (A), and relative area of areoles (B) of 11 leaves of Acer monspessulanum in %. For all the leaves measurements were done at four different sites; base (Ž), centre (), top (M) and margin (r). Mean was set at 100%. In (A) there are only slight differences between the different leaf sites, whereas in (B) differences are greater.
20
D. Uhl, V. Mosbrugger / Palaeogeography, Palaeoclimatology, Palaeoecology 149 (1999) 15–26
leaf variation. In none of the parameters, however, did we observe that some leaf sites had consistently higher values than others. In the following sections all measurements refer to the centre of a leaf (cf. Fig. 1).
5. Results 5.1. Measuring venation density: vein length per area versus distance between veins As the first problem we analysed whether the two-dimensional and the one-dimensional approach to estimate leaf venation density are correlated and provide equally good estimates of leaf venation density. As illustrated in Fig. 3A, B, there is indeed a pretty good correlation between vein length per area
(two-dimensional approach) and distance between veins (one-dimensional approach) in both species analysed (i.e. Acer monspessulanum and Quercus petraea). Hence, obviously both methods can be used for measuring leaf venation density. We would expect, however, that vein length per area provides the more reliable parameter because it considers the entire two-dimensional venation network. 5.2. Venation density and leaf size As described in Section 2, data from the literature about the influence of leaf size on venation density are controversial (Schuster, 1908; Gupta, 1961; Manze, 1968; Kull and Herbig, 1994, 1995). We measured the venation density in Acer monspessulanum and Quercus petraea as a function of leaf size, using the two-dimensional and the one-dimensional approach of measuring leaf venation density (Figs. 4 and 5). In Acer monspessulanum vein length per area decreases with increasing leaf size (Fig. 4A, r D 0:689) whereas the distance between veins shows no correlation with leaf size (Fig. 4B, r D 0:097). This emphasizes the fact mentioned already in the previous section that measuring the vein length per area is the more reliable (although much more laborious) method of estimating venation density. A different behaviour is observed in Quercus petraea. In this species there is no correlation between leaf size and vein length per area (Fig. 5A) or between leaf size and distance between veins (Fig. 5B). 5.3. Venation density and atmospheric CO2 To study the influence of atmospheric carbon dioxide on leaf venation we used two different approaches.
Fig. 3. Correlations between distance between veins (µm) and vein length per area (mm=mm2 ) in leaves of Acer monspessulanum (A) and Quercus petraea (B). Data are from the same leaves as in Figs. 4 and 5. Significance of regression lines: P < 0:1%.
5.3.1. Influence of atmospheric CO2 on venation density in controlled experiments We analysed the venation density of leaves of Quercus petraea which were grown under 350 and 700 ppm carbon dioxide, respectively. As shown in Fig. 5, plants grown at 350 and 700 ppm carbon dioxide do not differ significantly in their venation density measured as vein length per area (Fig. 5A) or distance between veins (Fig. 5B). For the same leaf
D. Uhl, V. Mosbrugger / Palaeogeography, Palaeoclimatology, Palaeoecology 149 (1999) 15–26
21
Fig. 4. Correlation between leaf size and venation density, measured as vein length per area (A) and distance between veins (B). Data are from 20 different leaves from a single Acer monspessulanum tree. Only in (A) is there a correlation between vein length per area and leaf size according to the regression equation (y D 0:001946x C 15:9282; significance: P < 0:1%).
Fig. 5. Correlation between leaf size and venation density, measured as vein length per area (A) and distance between veins (B). Data are from 47 different leaves of Quercus petraea and are subdivided into two subsets which were grown under 350 ppm CO2 (Ž) and 700 ppm CO2 (ž), respectively. The two subsets do not differ significantly in their regression line (P > 0:2).
material (cf. Section 4) Ku¨rschner (1996) has studied the stomatal density and found a significantly lower stomatal density and stomatal index in leaves grown at 700 ppm carbon dioxide.
pendent venation index VI according to
5.3.2. Changes of venation density during the last 150 years of growing atmospheric CO2 concentrations The herbarium material of Acer monspessulanum and Quercus petraea, collected since 1840 (cf. Section 4), was analysed with respect to the leaf venation density. To do this we used the distance between veins and the vein length per area as estimates of venation density. Because the vein length per area of Acer monspessulanum proved to change with leaf size (cf. Section 5.1) we also calculated a size-inde-
VI D .vein length per area
C/=aL;
with a D slope, C D constant of the regression equation of Fig. 4A, and L D leaf length. Fig. 6 illustrates for Acer monspessulanum how the parameters distance between veins and venation index changed in the last 150 years during which the atmospheric carbon dioxide concentration increased from 280 to 350 ppm. For both parameters there is no significant correlation with the atmospheric carbon dioxide content. The same can be seen in Quercus petraea (Fig. 7); here only the parameter distance between veins has been used which shows no correlation with leaf size (cf. Section 5.1).
22
D. Uhl, V. Mosbrugger / Palaeogeography, Palaeoclimatology, Palaeoecology 149 (1999) 15–26
Fig. 7. Venation density (measured as distance between veins) of leaves from herbarium material of Quercus petraea collected since 1840. Venation density does not change significantly with increasing atmospheric carbon dioxide concentration.
Fig. 6. Venation density of leaves from herbarium material of Acer monspessulanum collected since 1840. Venation density is measured as distance between veins (A) and as venation index (B) (for explanation see text). The atmospheric carbon dioxide concentration for the last 150 years are taken from Climate Change (1995). Venation density does not change significantly with increasing atmospheric carbon dioxide concentration. In (A), if more than one leaf was measured, the mean value together with the maximum and minimum value and the number of measured leaves are given.
5.4. Venation density in some plants from the Permian of China As stated in Section 2, only two studies (Zeuner, 1932; Manze, 1968) have used leaf venation density as a palaeoenvironmental proxy and both concentrate on the Tertiary. Apart from these exceptions, leaf venation density is mostly considered to be a taxonomic character, in particular in Palaeozoic and Mesozoic non-angiosperms. To get some information about the natural variation of leaf venation density in these plant groups we analysed
some plants from the Baode flora of northwestern Shanxi (China) that covers most of the Permian in five plant-bearing horizons (Kaiser, 1976; Mosbrugger, 1987): horizons F and A correspond to the lower Lower Permian (Shanxi Formation), horizons G, I to the upper Lower Permian (Lower Shihezi Formation) and horizon J to the lower Upper Permian (Upper Shihezi Formation). All plant-bearing horizons correspond to a fluvial flood-plain environment. Climatically, the stratigraphic interval is characterized by a trend to increasing aridity as documented by the decreasing abundance of coal and by the more common occurrence of red beds (cf. Liu, 1990). As plant taxa we have chosen the seed-plant genera Cordaites, Compsopteris and Taeniopteris, in which leaf venation density is commonly used as a taxonomic character. Cordaites is one of the precursors of conifers and has tongue-shaped leaves with parallel venation. Compsopteris and Taeniopteris are more closely related to the cycadophytes. The leaves of Compsopteris are pinnate, their leaflets have an open pinnate venation with a well pronounced midvein. In Taeniopteris the leaves are simple and resemble the leaflets of Compsopteris in overall shape and venation. Fig. 8 illustrates the variation of leaf venation density (measured as distance between veins) for the
D. Uhl, V. Mosbrugger / Palaeogeography, Palaeoclimatology, Palaeoecology 149 (1999) 15–26
23
Fig. 9. Distribution of venation density (measured as distance between veins) in Taeniopteris sp. from horizon G (cf. Fig. 8).
Fig. 8. Venation densities (measured as distance between veins, ‘error bars’ indicate maximum and minimum values) of Cordaites sp. (A), Compsopteris sp. (B) and Taeniopteris sp. (C) from the Permian of Baode (northwestern Shanxi, China). Further explanations in the text.
The question remains whether the above mentioned trend to increasing aridity during the Permian of China is also reflected in the leaf venation density. Compsopteris cannot contribute to this problem because of the lack of sufficient data. In Taeniopteris and Cordaites, however, a trend to slightly decreased distance between veins in higher stratigraphic levels is observed, if only those leaves are considered that belong to one species.
6. Discussion three genera and the five horizons. In most cases, the variation of leaf venation density (maximum value minus minimum value) within one horizon is between 100 and 150 µm and thus of the same order of magnitude as in angiosperms (cf. Figs. 3–7). Only in Taeniopteris of horizon G is there an extremely high variation, but an analysis of the frequency distribution of the leaf venation density in this horizon (Fig. 9) reveals a bimodal distribution with means around 250 and 500 µm, respectively. Presumably, this bimodal frequency distribution of leaf venation density corresponds to two different Taeniopteris species. The alternative interpretation that the bimodal distribution belongs to a single species but reflects the mixing of ecologically very different biotopes, is less probable because the overall variation of the venation density is larger than is normally observed in a single species. For the same reason, the Cordaites leaves of horizon A presumably represent a species different from that of the other horizons.
In Section 2 we have reviewed previous studies that have analysed the dependence of leaf venation density on various environmental parameters and in Section 3 we have summarized a few questions that remain open. With our analysis we have addressed some of these questions and can provide some answers. First of all we have demonstrated that the distance between veins as well as the vein length per area are reasonable estimates of venation density in leaves with a vein network. However, the parameter ‘distance between veins’ is less reliable in detecting trends because its background variation is higher than for the parameter ‘vein length per area’ (cf. Fig. 3). Unfortunately, the determination of vein length per area is much more time consuming than measuring the distance between veins. This means that in a first analysis or in order to get a large data set it is reasonable to use the more rapid technique, i.e. to measure the distance between veins. If, however, these data do not reveal a signal with respect to
24
D. Uhl, V. Mosbrugger / Palaeogeography, Palaeoclimatology, Palaeoecology 149 (1999) 15–26
leaf venation density, it does not necessarily imply that there is no signal; sometimes this signal is only revealed using the parameter ‘vein length per area’ (cf. Fig. 4). We also found that the leaf venation density may or may not depend on leaf size: In Acer monspessulanum there is a decrease of vein length per area with increasing leaf size, whereas the parameter distance between veins shows no correlation with leaf size; in Quercus petraea, however, there is no correlation with leaf size for both parameters. Hence, whenever leaf venation density is to be used either as a palaeoenvironmental proxy or as a systematic character, the first question to be answered is whether leaf venation density is size dependent. If leaf venation density varies with leaf size, then the regression equation has to be used to determine a size independent venation index, which may be based on vein length per area or on distance between veins (cf. Section 5.3). Unfortunately, the possible effect of leaf size on leaf venation density has been almost entirely neglected in previous studies. As a third result we observed no significant influence of atmospheric carbon dioxide concentration on leaf venation density in Quercus petraea and Acer monspessulanum. The result is surprising because at least in Quercus petraea the stomatal density and stomatal index decreases with increasing atmospheric carbon dioxide content, both in gas chamber experiments as well as in herbarium material of the last 150 years (Ku¨rschner, 1996). Our result could indicate that despite an increase of the atmospheric carbon dioxide content and a corresponding decrease of the stomatal density over the last 150 years, there was no significant change in transpiration that would have affected the leaf venation density. The overall transpiration may indeed have remained more or less unchanged because the increase of the atmospheric carbon dioxide content since the early 19th century induced not only a decrease of the stomatal density and thus of the stomatal transpiration, but also a slight warming (cf. Climate Change, 1995) that may have intensified the transpiration of leaves. Our analysis of some Permian gymnosperm leaves indicates that their leaf venation density has a natural variability that is similar to that observed in angiosperms. Moreover, there are hints that even in these plants leaf venation density can be used as
an environmental proxy: In the Permian of China we observe a slight increase of the leaf venation density that parallels a trend to increasing aridity. On the other hand, the increase in venation density is not very conspicuous, possibly because the plants studied represent floodplain plants which profit from a high groundwater table and thus are somewhat buffered against changes in precipitation. All in all there is evidence that in seed plants, at least since the Permo-Carboniferous, leaf venation density is a morphological character that can be adapted in a self-organization process to a changing environment. Many environmental factors may influence the leaf venation density, but it is most directly linked to transpiration. Thus leaf venation density can be used as an environmental proxy in palaeoecology and palaeoclimatology, but its frequent dependence on leaf size has to be taken into account. The potential of this proxy, however, is not yet fully explored.
Acknowledgements We thank the Pfalzmuseum fu¨r Naturkunde in Bad Du¨rkheim for the loan of herbarium specimens, and the Swedish Museum of Natural History in Stockholm for the loan of the material from the Permian of China. Furthermore we thank W. Ku¨rschner, Utrecht for the opportunity to investigate the greenhouse specimens of Quercus petraea. Particular thanks go to D. Ferguson (Vienna) and to one anonymous reviewer for critically reading the manuscript and for their valuable suggestions.
References Barykina, R.P., 1967. An anatomical analysis of the vegetative organs of Amygdalus nana L. Vestn. Moskov. Univ. 22 6. Ser. 3, 42–56, in Russian. Beerling, D.J., Chaloner, W.G., 1993a. Stomatal density responses of Egyptian Olea europaea L. leaves to CO2 change since 1327 BC. Ann. Bot. 71, 431–435. Beerling, D.J., Chaloner, W.G., 1993b. Evolutionary responses of stomatal density to global CO2 change. Biol. J. Linn. Soc. 48, 343–353. Beerling, D.J., Chaloner, W.G., 1994. Atmospheric CO2 changes since the last glacial maximum: evidence from the stomatal density record of fossil leaves. Rev. Palaeobot. Palynol. 81,
D. Uhl, V. Mosbrugger / Palaeogeography, Palaeoclimatology, Palaeoecology 149 (1999) 15–26 11–17. Beiguelman, B., 1962. Cerrado: Vegetacao oligotro´fica. Ci. Cult. 14, 99–107. Birks, H.J.B., Birks, H.H., 1980. Quaternary palaeoecology. Univ. Park Press, Baltimore, MD, 289 pp. Bolton, J.K., Brown, R.H., 1980. Photosynthesis of grass species differing in carbon dioxide fixation pathways. V. Response of Panicum maximum, Panicum milioides, and tall fescue (Festuca arundinacea) to nitrogen nutrition. Plant Physiol. 66, 97–100. Chaloner, W.G., Creber, G.T., 1990. Do fossil plants give a climatic signal? J. Geol. Soc. 147, 343–350. Chonan, N., 1967. Studies on the photosynthetic tissues in the leaves of cereal crops. III. The mesophyll structure of rice leaves inserted at different levels of the shoot. Proc. Crop Sci. Soc. Jpn. 36, 291–296, in Japanese with English abstract. Climate Change, 1995. The science of climate change. In: Houghton, J.T., Meira Filho, L.G., Callander, B.A., Harris, N., Kattenberg, A., Maskell, K. (Eds.), Contribution of Working Group 1 to the Second Assessment Report of the Intergovernmental Panel of Climate Change. Cambridge Univ. Press, 572 pp. Critchfield, W.B., 1960. Leaf dimorphism in Populus trichocarpa. Am. J. Bot. 47, 699–711. Eberhard, P.H., 1903. Influence de l’air sec et de l’air humide sur la forme et sur la structure des vegetaux. Ann. Sci. Nat. Bot. 8e Se´r. 18, 61–152. Ettingshausen, C. v., 1861. Die Blatt-Skelete der Dikotyledonen mit besonderer Ru¨cksicht auf die Untersuchung und Bestimmung der fossilen Pflanzenreste. Staatsdruckerei, Vienna, XLV C 308 pp. Frielingsdorf, J., 1992. Klimamarken in neogenen Ho¨lzern vom Niederrhein. Sondervero¨ff. Geol. Inst. Univ. Ko¨ln 86, 118 pp. Gorysina, T.K., Can-Czin, Ti, Chan’, De-cun, 1961. An attempt of comparative studying of water regime and leaf structure in different layers of the oak-tree forest. Vestn. Leningrad. Univ. 15 (3), 21–31, in Russian. Grace, J., Russell, G., 1977. The effect of wind on grasses. III. Influence of continuous drought or wind on anatomy and water relations in Festuca arundinacea SCHREB. J. Exp. Bot. 28, 268–278. Greenwood, D.R., 1992. Taphonomic constraints on foliar physiognomic interpretations of Late Cretaceous and Tertiary palaeoclimates. Rev. Palaeobot. Palynol. 71, 149–190. Gupta, B., 1961. Correlation of tissues in leaves. 1. Absolute vein-islet numbers and absolute veinlet termination numbers. Ann. Bot. 25 (97), 65–70. ¨ ber experimentelle Beeinflussung der Heilbronn, A., 1926. U Blattnervatur. Biol. Zentralbl. 46, 477–480. Herbig, A., Kull, U., 1991. Leaves and ramification. Mitt. SFB 230 Nat. Konstr. 7, 109–117. Ho¨germann, C., 1990. Untersuchungen zur Korrelation zwischen Leitbu¨ndeldichte und Waldschichtung an tropischen Vertretern von 79 Species aus dem Regenwald von VenezolanischGuayana. Flora 184, 1–14. Hughes, A.P., 1959. Effects on the environment on leaf develop-
25
ment in Impatiens parviflora DC. J. Linn. Soc. London Bot. 56, 161–165. Kaiser, H., 1976. Die permische Mikroflora der CathaysiaSchichten von Nordwest-Schansi, China. Palaeontographica B 159, 83–156. Ka¨mpf, A.N., Lange, P., 1983. Der Einfluß der Blattdu¨ngung auf morphologische und anatomische Merkmale bei Bromeliaceen. Angew. Bot. 57, 9–18. ¨ ber den anatomischen Bau du¨rre- und Keller, B.A., 1933. U hitzeresistenter Bla¨tter. Ber. Dtsch. Bot. Ges. 51, 514–522. Kull, U., Herbig, A., 1994. Leaf venation patterns and principles of evolution. Mitt. SFB 230 9, 167–175. Kull, U., Herbig, A., 1995. Das Blattadersystem der Angiospermen: Form und Evolution. Naturwissenschaften 82, 441–451. Ku¨rschner, W.M., 1996. Leaf stomata as Palaeoatmospheric CO2 levels. LPP Contrib. Ser. 5, Utrecht. Lalanne, G., 1890. Recherches sur les caracte`res anatomiques des feuilles persistantes des Dicotyle´dones. The`se Sci., Paris. Lebedincev, E., 1927. Physiologische und anatomische Besonderheiten der in trockener und feuchter Luft gezogenen Pflanzen. Ber. Dtsch. Bot. Ges. 45, 83–96. Liu, G., 1990. Permo-Carboniferous paleogeography and coal accumulation and their tectonic control in the North and South China continental plates. Int. J. Coal Geol. 16, 73–117. MacGinitie, H.D., 1974. An early Middle Eocene flora from the Yellostone–Absaroka Volcanic Province, Northwestern Wind River Basin, Wyoming. Univ. Calif. Publ. Geol. Sci. 108, 103 pp. Mai, D.H., 1995. Tertia¨re Vegetationsgeschichte Europas. Fischer, Jena, Stuttgart, New York. Manze, U., 1968. Die Nervaturdichte der Bla¨tter als Hilfsmittel der Pala¨oklimatologie. Sondervero¨ff. Geol. Inst. Univ. Ko¨ln 14, 103 pp. Marenkov, A.J., Titarenko, E.M., Bogdasova, V.P., 1971. Effect of mineral nutrition on modification of xeromorphic property in the leaves of winter wheat. Fiziol. Rast. 18, 135–141, in Russian, with English abstract. Maximov, N.A., 1929. The Plant in Relation to Water. George Allen and Unwin, London. McElwain, J., Mitchell, F.J.G., Jones, M.B., 1995. Relationship of stomatal density and index of Salix cinerea to atmospheric carbon dioxide concentration in the Holocene. Holocene 5, 216–219. Mory, B., 1992. Zur Blattnervatur antillanischer Arten der Gattung Crossopetalum R. Br. (Celastraceae). Flora 187, 17–36. Mosbrugger, V., 1987. The Permian taphoflora of Baode (NW Shanxi, China) — a preliminary report. C. R. XIe Congr. Int. Stratigr. Ge´ol. Carbonife`re, Beijing, 1987, 3, 135–142. Mosbrugger, V., Utescher, T., 1997. The coexistence approach — a method for quantitative reconstructions of Tertiary terrestrial palaeoclimate data using plant fossils. Palaeogeogr., Palaeoclimatol., Palaeoecol. 139, 61–86. Napp-Zinn, K., 1984. Handbuch der Pflanzenanatomie, Bd. VIII, 2B, Teil 1. Borntra¨ger, Berlin, Stuttgart. Napp-Zinn, K., 1988. Handbuch der Pflanzenanatomie, Bd. VIII, 2B, Teil 2. Borntra¨ger, Berlin, Stuttgart. Philpott, J., 1956. Blade tissue organization of foliage leaves
26
D. Uhl, V. Mosbrugger / Palaeogeography, Palaeoclimatology, Palaeoecology 149 (1999) 15–26
of some Carolina shrub-bog species as compared with their Appalachian mountain affinities. Bot. Gaz. 118 (2), 88–105. Pyykko¨, M., 1979. Morphology and anatomy of leaves from some woody plants in a humid tropical rainforest of Venezuelan Guayana. Acta Bot. Fenn. 112, 1–41. Rippel, A., 1919. Der Einfluss der Bodentrockenheit auf den anatomischen Bau der Pflanzen, insbesondere von Sinapis alba L. und die sich daraus ergebenden physiologischen und entwicklungsgeschichtlichen Fragen. Beih. Bot. Centralbl. Abt. I 36, 187–260. ¨ kologisch-taxonomische UntersuchunRoth, I., Yee, S., 1991. O gen u¨ber Leitbu¨ndeldichte und Leitbu¨ndelmuster bei Angiospermen-Laubbla¨ttern. Bot. Jahrb. Syst. 113 (1), 7–71. ¨ ber die anatomischen Jugendformen der Schramm, R., 1912. U Bla¨tter einheimischer Holzpflanzen. Flora 104, 225–295. Schuster, W., 1908. Die Blattaderung des Dicotylenblattes und ihre Abha¨ngigkeit von a¨ußeren Einflu¨ssen. Ber. Dtsch. Bot. Ges. 26, 194–237. Thiraporn, R., Geisler, G., 1978. Untersuchungen zur Entwicklung morphologischer und anatomischer Merkmale von Maisinzuchtlinien in Abha¨ngigkeit von der Temperatur. Z. Acker Pflanzenbau 147, 300–308. Whitehead, F.H., 1964–1965. Phenotypic adaptation in wind exposed plants. Sci. Hortic. 17, 31–39. Wiemann, M.C., Manchester, S.R., Dilcher, D.L., Hinojosa, L.F.,
Wheeler, E.A., 1998. Estimation of temperature and precipitation from morphological characters of dicotyledonous leaves. Am. J. Bot. 85, 1796–1802. Wilf, P., 1997. When are leaves good thermometers? A new case for leaf margin analysis. Paleobiology 23, 373–390. Wilf, P., Wing, S.L., Greenwood, D.R., Greenwood, C.L., 1998. Using fossil leaves as palaeoprecipitation indicators: an Eocene example. Geology 26, 203–206. Wolfe, J.A., 1977. Paleogene floras from the Gulf of Alaska Region. Geol. Surv. Prof. Pap. 997. Wolfe, J.A., 1993. A method of obtaining climatic parameters from leaf assemblages. U.S. Geol. Surv. Bull 2040. Wylie, R.B., 1951. Principles of foliar organization shown by sun-shade leaves from ten species of deciduous dicotyledonous trees. Am. J. Bot. 38, 355–361. Zalenski, W. v., 1902. Ueber die Ausbildung der Nervation bei verschiedenen Pflanzen. Ber. Dtsch. Bot. Ges. 20, 433–440. Zalenski, W. v., 1904. Materials for the study of the quantitative anatomy of different leaves of the same plant. Me´m. Polytech. Inst. Kiev. 4, 1–203, in Russian. ¨ ningen und Zeuner, F.E., 1932. Die Nervatur der Bla¨tter von O ihre methodische Auswertung fu¨r das Klimaproblem. Centralbl. Mineral. Geol. Pala¨ontol. B, 260-264. Zeuner, F.E., 1936. Palaeobiology and climate of the past. Probl. Palaeontol. 1, 199–216.