Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Biodiversity action plan wikipedia , lookup
Occupancy–abundance relationship wikipedia , lookup
Latitudinal gradients in species diversity wikipedia , lookup
Introduced species wikipedia , lookup
Island restoration wikipedia , lookup
Molecular ecology wikipedia , lookup
Coevolution wikipedia , lookup
Storage effect wikipedia , lookup
Research Paper 224 Growth Rings in the Roots of Temperate Forbs are Robust Annual Markers G. von Arx and H. Dietz Institute of Integrative Biology ETH, Swiss Federal Institute of Technology, ETH Zentrum CHN, G 35.1, 8092 Zürich, Switzerland Received: June 10, 2005; Accepted: November 12, 2005 Abstract: There is increasing interest in the analysis of annual growth rings in the secondary root xylem of perennial forbs (herb-chronology). Therefore, we need to verify whether these growth rings are always formed annually. To investigate the formation of root rings we performed common garden experiments at two distinct sites in Switzerland. We grew nine unrelated forb species from seed and subjected them to competition and clipping treatments. Anatomical developments in the roots of the individuals were tracked during five growing seasons. Across all species and treatments at least 94% of the expected growth rings associated with full growing seasons were identifiable and the development of the anatomical patterns was consistently seasonal. While the distinctness of annual rings varied somewhat between species and sites, the treatments had no effect on the presence of annual rings. In no case were false rings developed. The results of this study demonstrate that the growth rings in the roots of northern temperate forbs represent robust annual growth increments and, hence, can reliably be used in herb-chronological studies of age- and growth-related questions in plant ecology. Key words: Age determination, anatomical pattern, annual ring, growth increment, herb-chronology, perennial forbs. Introduction A thorough understanding of the processes underlying plant population and community development requires not only knowledge on the present state of vital parameters like plant size, fitness, and population density but also on their change over time. Limited access to understanding such changes is possible by interpreting static, age-classified life tables (e.g., Dietz and Ullmann, 1998; Brys et al., 2004; Freville and Silvertown, 2005). However, age-related parameters such as birth and death rate, the age structure of the population or age-dependent size and fitness are relatively difficult to obtain, especially for perennial forbs (cf. Harper, 1977; Vorontzova and Zaugolnova, 1985; Menges, 2000). Plant Biol. 8 (2006): 224 – 233 © Georg Thieme Verlag KG Stuttgart · New York DOI 10.1055/s-2005-873051 ISSN 1435-8603 A promising methodological approach to investigate age-related patterns in forbs is herb-chronology (i.e., the ecologicallyoriented analysis of annual rings in the secondary root xylem of perennial forbs), which has been developed during the past decade (e.g., Dietz and Ullmann, 1997, 1998; Dietz and Fattorini, 2002; Dietz and von Arx, 2005; von Arx and Dietz, 2005). The anatomical basis of annual growth increments (hereafter referred to as “annual rings”) in the roots of forbs is the formation of earlywood vessels with large lumina in spring and latewood vessels with smaller lumina in late summer and autumn (Dietz and Ullmann, 1997; von Arx and Dietz, 2005). Similar to the better known tree rings (e.g., Larson, 1994; Schweingruber, 2001), discernible annual rings in the secondary root xylem have proven to be a common phenomenon in many perennial forb species with persistent main roots (Zoller, 1949; Bakshi and Coupland, 1960; Werner, 1978; Humulum, 1981; Boggs and Story, 1987; Dietz and Ullmann, 1998; Schweingruber and Dietz, 2001; Dietz and Fattorini, 2002; Dietz and Schweingruber, 2002; Dietz and von Arx, 2005). Counting the number of annual rings in the roots of forbs has been used for age determination, demographic analyses, and investigations into population development (e.g., Boggs and Story, 1987; Dietz and Ullmann, 1998; Dietz et al. 1999; Dietz, 2002). Furthermore, determination of plant age along environmental gradients allows us to make inferences on how plant life histories respond to different growth conditions (Erschbamer and Retter, 2004; Rixen et al., 2004; von Arx et al., 2006). Patterns of ring width have been used to study how plant growth responds to variation in environmental conditions (Dietz and Ullmann, 1998; Dietz et al., 2004; von Arx et al., 2006) or unpredictable events such as climatic fluctuations (Dietz and von Arx, 2005). Although the annual nature of growth rings could be verified for one or several individuals of known age in about twenty species (Dietz and Schweingruber, 2002), so far, it has not been tested whether distinct anatomical implementations of growth rings in unrelated species all represent true annual rings and whether annual rings are reliably formed under a wide range of growth conditions (Schweingruber and Dietz, 2001). For example, an individual may fail to form one or several annual rings when its aboveground growth is suppressed by strong competition (Dietz and Ullmann, 1998) or permanent snow cover (Wijk, 1980). Furthermore, under unfavourable growth conditions growth rings may become indistinct Annual Root Rings in Forbs Plant Biology 8 (2006) Table 1 Characterization of the study species. All species occur in Europe (native area) and in North America (introduced area; Tutin et al., 1964 – 1983; Natural Resources Conservation Service, 2005). Characterization of plant size and habitat preference follows Oberdorfer, 1994. Table modified after Dietz and Schweingruber, 2001, courtesy of the Bulletin of the Geobotanical Institute ETH Species Family Life-form Habitat Root architecture Type of species* Anchusa officinalis L. Boraginaceae semi-rosette forb, deep-rooting ruderal and other disturbed sites branched taproot 1A Cichorium intybus L. Asteraceae semi-rosette forb, deep-rooting ruderal and other disturbed sites branched taproot 1A Trifolium pratense L. Fabaceae low-growing forb ruderal sites, meadows, pastures branched taproot 1A Digitalis grandiflora Mill. Scrophulariaceae tall-growing forb tall herb communities, forest edges fibrous roots 2A Sanguisorba minor L. Rosaceae semi-rosette forb open vegetation, often on raw, calcareous soils branched taproot 2A Silene vulgaris Garcke Caryophyllaceae semi-rosette forb open vegetation, often on raw soils branched taproot 2A Euphorbia esula L. Euphorbiaceae clonal forb with lateral roots, deep-rooting ruderal sites with relatively low soil moisture branched roots 3Z Galium verum L. Rubiaceae clonal forb with lateral roots, deep-rooting meadows and pastures on calcareous soils fibrous roots 3Z Hypericum perforatum L. Hypericaceae medium-sized forb meadows, pastures, road verges etc. on nutrient-poor soils branched roots 3Z * 1, short-lived species with distinct growth rings; 2, long-lived species with distinct growth rings; 3, long-lived species with diffuse ring pattern; A, used in all sites and treatments; Z, only used in control plots at Zurich. because of missing or scarce latewood between successive earlywood zones (Dietz and Fattorini, 2002). Conversely, if a plant is mown or heavily grazed by herbivores it may respond with regeneration growth and formation of large, earlywood-like vessels in the roots that may look like the beginning of a new (annual) ring. Besides resulting in inaccurate determination of plant age, such anatomical responses would blur the presence of ring width signals preserved in specific calendar years that were characterized by particularly good or bad growth conditions (Dietz and von Arx, 2005) by shifting the reading frame of the chronologies. Since short-lived forb species are generally more responsive to short-term changes in growth conditions (Grime, 2001), they may be particularly prone to forming false rings, for instance when root growth resumes after disturbance (e.g., Martinkova et al., 2004) or a temporary interruption due to unfavourable growth conditions. Conversely, in long-lived species rings may be hard to identify (pseudo-missing rings) when growth rings become narrow and tightly packed (Dietz and Fattorini, 2002). In species that show less distinct patterns of annual rings, environmental fluctuations may result in anatomical responses in the roots that prevent accurate identification of annual growth rings. In addition, patterns may vary depending on the seasonality of the climate (Schweingruber and Dietz, 2001). Here we report on the first experimental verification of the annual formation of growth rings in the secondary root xylem of perennial forbs under contrasting growing conditions. We addressed the following main questions: (i) How does the anato- my of annual rings develop over time and are there unambiguous criteria for detecting transitions between two consecutive annual rings? (ii) Are growth rings formed on a truly annual basis, irrespective of contrasting microsite conditions and anatomical and life history characteristics of the species? (iii) Is the distinctness of the annual ring anatomy related to variation in the seasonality of climate and associated local site factors? To investigate these questions we selected nine perennial forb species that represent a wide range of phylogeny, longevity, and distinctness of ring anatomy. Two common garden experiments were established, one located in the temperate lowland north of the Swiss Alps and the other one in the insubric southern foothills of the Swiss Alps. Plants were subjected to control, competition, and clipping treatments and tracked over five growing seasons. Materials and Methods Study species Nine species of perennial forbs, representing a wide phylogenetic and ecological range, were selected for the study (Table 1). Most of the species are widespread and abundant in Switzerland, and all develop persistent main roots that allow analysis of the development of annual rings over several years (Fig. 1). The species were selected so as to include three species from each of three types: i) short-lived species with clearly demarcated growth rings (hereafter referred to as “shortlived species”), ii) long-lived species with clearly demarcated 225 226 Plant Biology 8 (2006) Fig. 1 Examples of anatomical patterns in the secondary root xylem of the study species. (a) Anchusa officinalis, (b) Cichorium intybus, (c) Trifolium pratense, (d) Digitalis grandiflora, (e) Sanguisorba minor, (f) Silene vulgaris, (g) Euphorbia esula, (h) Galium verum, and (i) Hypericum perforatum as viewed through a dissecting microscope. Lignified structures are stained with phloroglucinol/HCl and appear dark in the images. All shown plants were grown from seed in summer 2000 and were har- G. von Arx and H. Dietz vested in early May 2004, except E. esula and H. perforatum which were harvested in late April 2003. Black markers denote transitions from latewood of the previous to earlywood of the following growing season, white markers denote presumable positions of uncertain annual ring borders that were partly classified as missing in the analysis. Scale bars = 500 μm. Annual Root Rings in Forbs Plant Biology 8 (2006) Table 2 Description of the two study sites. Table modified after Dietz and Schweingruber, 2001, courtesy of the Bulletin of the Geobotanical Institute ETH Zurich (north of the Alps) Oggio (southern foothills of the Alps) Location 47826′N, 8830′E 46805′N, 8859′E Altitude 520 m asl 590 m asl Climatic conditions temperate insubric (warm summers with high precipitation, relatively dry and mild winters) Mean annual temperature 9 8C 11 8C Mean annual precipitation 1 100 mm 1 800 mm Site characteristics flat, sun-exposed area sun-exposed terrace on south-facing slope Soil characteristics loamy rich organic forest soil with very good drainage Treatments control, competition, clipping control, clipping growth rings (hereafter referred to as “long-lived species”), and iii) long-lived species that show less clear, deviating anatomical patterns (hereafter referred to as “diffuse species”). Experimental sites and setup The experiment was established at a temperate site north of the Swiss Alps (Zurich) and at an insubric site south of the Swiss Alps (Oggio; Table 2). Zurich has a colder climate than Oggio and the moisture of the loamy soil at Zurich is usually higher than that of the well-drained, south-facing forest soil at Oggio, although mean annual precipitation is higher south of the Alps (Table 2). Furthermore, the Oggio site hardly experiences any frost during the vegetation period from March to October or any day with a mean temperature below freezing, even in the winter months (MeteoSchweiz, 2005). At the Zurich site, blocks with control plots (n = 12), clipping plots (n = 10), and competition plots (n = 9) were established, whereas at Oggio, 12 plots each of control and clipping treatments were established and randomized within a block. The three diffuse species were only grown at the Zurich site. Plants were raised from seeds obtained from different botanical gardens in central Europe and grown in the greenhouse for one month before planting into the plots in late September 2000. Within plots, plants of the same species were grouped together to minimize interspecific competition. More competitive species were grouped together within a plot to prevent them from overgrowing smaller species. In the control and clipping plots three individuals per species were planted at a distance of 25 cm from each other to maintain competition between individuals at low levels. In the clipping treatment, plants were cut 2 cm above the ground in September 2001 and May 2002. In the competition treatment, eight individuals per species were planted at a distance of 10 cm from each other to result in considerable, predominantly intraspecific com- petition. Plots were not weeded, except for elimination of strong spontaneous competitors in control and clipping plots about three to five times per growing season. The plants were not watered except occasionally during the summer heatwave of 2003 (Schaer et al., 2004) to avoid excessive drought/mortality. The experiment was run until May 2004 and covered three complete (2001 – 2003) and two incomplete (2000, 2004) growing seasons. Data collection Beginning in autumn 2001, one plot each per treatment and site was randomly selected and completely harvested three times per year (April, June, and October) to track the developing root xylem anatomy during the growth periods. Between autumn 2001 and spring 2004, a total of eight plots were harvested per treatment and site. There was a total mortality of 40 % across all plots and years. At Oggio mortality in both treatments was higher (60 %) than at Zurich (20 % in control and clipping plots and 40 % in competition plots). We therefore combined all remaining plots in the last harvest at Oggio to obtain a sufficient number of plants for analysis. Differences in mortality between species were pronounced, ranging from 13 % (Hypericum perforatum) to 57 % (Cichorium intybus). Thin cross-sections of the proximal ends of the main roots were obtained with a sledge microtome and stained with phloroglucinol-HCl to render lignified structures of the secondary xylem (vessels and, where present, lignified parenchyma cells) red. Cross-sections were photographed through the phototube of a dissecting microscope (Leica MZ8) using a digital camera (Nikon CoolPix 990) and the obtained images were used for further analysis (Dietz and Fattorini, 2002). Data analysis For each individual, the anatomical pattern, in particular vessel size at the growing periphery of the secondary xylem, was inspected for each harvest date and classified as consistent or inconsistent with the expected phenological development (Dietz and Ullmann, 1997). Consistency with the postulated phenological development was noted if the following criteria for the outer margin of the secondary xylem were met at the different harvest times: (i, April) earlywood zone present with large vessels fully developed (cf. Figs. 1 a – g, i, Fig. 2, spring); (ii, June) earlywood zone complete, latewood zone with small vessels developing (cf. Fig. 2, summer); or (iii, October) latewood zone fully developed (cf. Fig. 2, autumn). Furthermore, the number of detectable rings and, if possible, the calendar year of false and missing rings were determined for each individual to assess the reliability of annual ring formation. The incomplete rings of 2000 and 2004 were included in ring counting. The age of the plant samples was known to the observer, which might raise concern whether ring counting might have been biased. However, the objectivity of the applied criterion (see Introduction) was proven by an automated image analysis system (ROXAS, see von Arx and Dietz, 2005). To examine whether clipping in autumn 2001 or spring 2002 resulted in any irregularity in the developing anatomical patterns, the individuals from all treatments were classified as showing either i) no irregularities, ii) conspicuous irregularities in parenchyma lignification, or iii) intra-annual fluctuations in vessel size. 227 228 Plant Biology 8 (2006) Fig. 2 G. von Arx and H. Dietz Annual Root Rings in Forbs 3 Plant Biology 8 (2006) Fig. 2 Chronology of anatomical root development of the short-lived Cichorium intybus (a) and the long-lived Sanguisorba minor (b) harvested between autumn 2001 (second growth period) and spring 2004 (fifth growth period). Note varying degrees in the lignification of parenchyma cells among individuals, especially in C. intybus, while the pattern and arrangement of vessels are quite consistent within species. Markers as explained in Fig. 1. Scale bars = 500 μm. To test for dependence of the response variables on the factors site, growth treatment, and type of species (cf. Table 1), nominal logistic regression was performed on proportion data and log-linear analysis on count data (see Table 3 for details; Crawley, 2005). Significant relationships were assumed at p ≤ 0.05. Data of the different harvest dates were pooled prior to statistical analysis. For factors with more than two levels, Bonferroni-adjusted α levels were used for pairwise comparisons. Results for the three species with diffuse anatomical patterns were only used in comparisons between types of species. Differences between growth treatments were not statistically tested for the Zurich data to avoid problems of pseudo-replication (Hurlbert, 1984). Nominal logistic regressions were performed in SAS JMP 5.1 and log-linear analysis in StatSoft Inc. Statistica 5.5. Results Between autumn 2001 and spring 2004 a total of 686 individual plants (514 from Zurich, 172 from Oggio) were harvested and analyzed; these were expected to contain overall 2257 annual rings in the main roots (depending on harvest date, individuals were expected to show between two and five annual rings per plant). Phenology of annual ring formation Phenological development of annual rings corresponded to the expected, strongly seasonal pattern in 99 % of all plants, i.e., wide earlywood vessels were only formed in spring and vessel size decreased towards autumn (Figs. 1, 2). Only in H. perfora- tum (Zurich, control plots) was there a considerable proportion of individuals (38 %) that showed inconsistencies in the anatomical development. Although there were occasional intra-annual fluctuations in vessel size, these were generally easily discriminated against true annual ring borders, which were characterized by a sharp increase of vessel size between latewood of the previous and first earlywood of the following growing period. Deviations from this pattern were most pronounced in the diffuse species (cf. Table 1). H. perforatum developed comparably wide vessels within a band of weakly lignified parenchyma in the distal portion of the latewood (Fig. 1 i). In Euphorbia esula, annual ring borders were only weakly demarcated but corresponded to a distinct band of unlignified parenchyma (Fig. 1 g), whereas reliability and distinctness of annual rings varied considerably within individual plants in Galium verum (Fig. 1 h). The onset of earlywood development varied among species: while Trifolium pratense and Sanguisorba minor showed well-developed earlywood by April, the other species had only started to develop the earlywood by this time (Fig. 1). Presence of annual rings In 74 % of individuals, the number of observed rings was identical to the number of expected rings, corresponding to 91 % of all expected annual rings being observed across all species and treatments. The first annual ring, representing the short initial autumn growth period in 2000, was missing or indistinct (hereafter referred to as “missing”) in 20% of all plants, whereas annual rings other than the first one were missing in 6 % of the plants. Of these, about half had more than one missing ring. False rings were generally not observed. At Zurich, 82 % of the plants showed the expected number of annual rings, and in a further 13 % only the first (incomplete) ring was missing. At the insubric site near Oggio the proportion of individuals that lacked the first ring was significantly higher (23 %; df = 2, χ2 = 33.8, p < 0.001; Fig. 3 a). This site effect was mainly due to a strong response in only two species: at Table 3 Overview of data analysis used to test for dependence of the anatomical response variables on the factors site, growth treatment, and type of species. Proportion data were analyzed using nominal logistic regression and count data using log-linear analysis. The diffuse species were excluded for site comparison. Treatment effects were only tested for Oggio data, types of species were only compared for data of control plots at Zurich (see text and Table 1 for further explanations) Response variable Response levels Data type Model tested Phenology of root anatomy consistent with seasonal rhythm inconsistent with seasonal rhythm proportion –* Presence of annual rings all rings present first ring missing other ring(s) missing proportion site × type of species × species treatment × type of species × species type of species × species Calendar year of missing rings only 2000 (first ring incomplete) other calendar years missing count variable site × species treatment × species type of species Anatomical effects of clipping no effect irregularity in parenchyma lignification intra-annual fluctuation in vessel size proportion treatment × type of species × species * Data structure was inappropriate for statistical testing, see “Results”. 229 230 Plant Biology 8 (2006) G. von Arx and H. Dietz Fig. 3 Variation in the proportion of plant individuals with missing rings. (a) Differences between sites, treatments, and types of species (cf. Table 1; for the latter classification only individuals from control plots in Zurich were used) and (b) differences between species. Species with diffuse anatomical pattern were omitted from site and treatment comparisons, since they were only grown in control plots at Zurich. Different letters indicate significant differences (p = 0.05). Z, Zurich; O, Oggio; 0, control; CP, clipping; CM, competition; S, short-lived; L, long-lived; D, diffuse species; Ao, Anchusa officinalis; Ci, Cichorium intybus; Tp, Trifolium pratense; Dg, Digitalis grandiflora; Sm, Sanguisorba minor; Sv, Silene vulgaris; Ee, Euphorbia esula; Gv, Galium verum; Hp, Hypericum perforatum. Zurich 91 % of the Anchusa officinalis individuals showed the expected number of annual rings compared to only 47 % at Oggio (df = 2, χ2 = 49.8, p < 0.001). In Digitalis grandiflora the respective proportions were 43 % (Zurich) and 27 % (Oggio; df = 2, χ2 = 6.8, p = 0.04). In the other species, there was no or a similar trend, but the differences in the respective proportion were small (0 – 9 %). In contrast, plants from the two sites did not differ significantly in the low proportion of missing rings at positions other than the first position (28 % at Zurich vs. 21 % at Oggio; df = 1, χ2 = 0.8, p > 0.25). Among the clipped plants at Oggio, there was a higher proportion of individuals with the expected number of annual rings and a lower proportion with only the first ring missing than among the control plants (df = 2, χ2 = 20.4, p < 0.001; Fig. 3 a). In clipped plants at Oggio, 81 % of missing rings were the first ring, which was not significantly different from control plants (77 %; df = 1, χ2 = 0.0, p > 0.90). At Zurich, irrespective of treatments, 82 % of the plants showed the expected number of annual rings (Fig. 3 a), and 71 % of the remainder lacked only the first annual ring. There was significant variation in the proportion of individuals showing the expected number of rings among species and among different types of species (df = 4, χ2 = 19.4, p < 0.001; Fig. 3 a). The short-lived species showed the highest (84 %) and the diffuse species the lowest number of exact matches (58 %; Fig. 3 a). In all three types of species there were species which showed a very close match between the number of observed and the number of expected rings (Fig. 3 b). However, Annual Root Rings in Forbs whereas the three short-lived species had a homogeneously high proportion of individuals showing the expected number of rings (≥ 77 %), this proportion varied considerably in the long-lived and the diffuse species (36 % to 95 % and 30 % to 92 %, respectively). While in the short-lived species only 10% of all individuals lacked the first ring, both the long-lived and diffuse species had a much higher proportion of individuals in which the first ring was missing (≥ 25 %; Fig. 3 a). In contrast, the short-lived species showed a higher proportion of missing rings at other positions than the first position (49 %) than the long-lived (12 %) and diffuse species (28 %; df = 2, χ2 = 26.8, p < 0.001). Effects of clipping on root anatomy At Oggio, anatomical irregularities occurred significantly more frequently in clipped than in control plants (df = 2, χ2 = 27.8, p < 0.001). Whereas conspicuous changes in parenchyma lignification were more frequent in clipped than in control plants (26 % vs. 14 %), there was no considerable difference in intraannual fluctuations of vessel size (6 % vs. 8 %). At Zurich, between 4 % (competition) and 13 % (clipping) of the plants showed conspicuous irregularities in parenchyma lignification and between 4 % (competition) and 9 % (control and clipping) intra-annual fluctuations of vessel size. Discussion Overall, our results confirm previous claims that the growth rings in the roots of perennial forbs growing in the temperate seasonal zone are robust annual markers. Across all species and treatments at least 94 % of the expected growth rings associated with full growing seasons were identifiable. While the formation of annual rings corresponded to the postulated seasonal pattern in all but very few, exceptional cases, a small proportion of rings were hard to identify or not identifiable. However, most of these missing rings are attributable to problematic species or particular conditions as discussed below. Robustness of annual ring formation in perennial forbs In 76 % of all the plants with missing annual rings, these rings were barely or not identifiable due to an indistinct transition between xylem produced in the incomplete first and the second growing seasons. Plants which germinate late in the growing season may not attain a sufficient size to produce a detectable first growth ring. This is supported by the results of our study and by the results of a separate experiment at Zurich in which individuals of the short-lived Trifolium pratense and the long-lived Sanguisorba minor were germinated in spring and early summer and grown under conditions comparable to those of the control plots (von Arx, unpublished data). In this sample, all 70 plants used showed a clearly identifiable annual ring marking the transition from the first growing period to the second one. This result demonstrates that the first annual ring may be consistently formed if plants do not germinate late in the growing period. More critical are missing rings associated with later years, which were observed in 6 % of the individuals. However, an unknown percentage of these cases may be ascribed to a few plants, harvested in spring, that had not yet initiated formation of a new ring, or may be artefacts caused by an erroneous choice of a younger root instead of the main root, or by a spontaneous individual of the same species that Plant Biology 8 (2006) was harvested at the position of a dead experimental plant. The existence of erroneous samples is also suggested by the clarity of the other annual rings in several individuals that did not show the expected number of rings. Regional variation Differences in climate and soil conditions (cf. Table 2) were probably responsible for the higher distinctness of annual rings at the temperate Zurich than at the insubric Oggio site. The sun-exposed site at Oggio, which is located in a south-facing forest clearing, is more protected against strong climatic influences such as frost than the Zurich site. The observed difference may indicate a general trend of decreasing distinctness of annual rings along a gradient from the temperate to the subtropical zones (cf. Schweingruber and Dietz, 2001). For example, whereas individuals of Sanguisorba minor collected at 900 m asl on east-Mediterranean Cyprus (358N) showed distinct annual rings, in established individuals of the same species collected at 100 m asl on sub-tropical Tenerife (Canary Islands, 288N) there were hardly any growth rings in the roots (von Arx and Dietz, unpublished data). The distinctness of annual rings in response to different site conditions may also vary considerably between species. The short-lived species were more strongly affected by the differences between Oggio and Zurich which may be due to their greater growth plasticity (Grime, 2001). Variation due to local growth constraints Small-scale biotic and abiotic factors (vegetation density, herbivore attack, nutrient and light availability, etc.) often vary considerably within plant populations and may strongly influence plant growth (e.g., Crawley, 1997; Silvertown and Charlesworth, 2001). Although clipped plants or plants grown in high-density plots were 20 – 80 % smaller than those of control plots (data not shown), and irrespective of corresponding intra-specific differences in ring width (cf. Fig. 2), they showed a lower proportion of missing rings than control plants. Our results indicate that regeneration growth after aboveground biomass removal or reduction of growth by strong competition is not likely to result in missing or false rings. However, occasional intra-annual fluctuations in vessel size that may occur in regenerating plants require cautious anatomical interpretation and comparison with the anatomical characteristics of true ring borders in the same individual to avoid recording false rings. Consequently, the general earlywood-latewood pattern of annual rings proved to be a robust characteristic in forbs under a wide variety of growing conditions. Phylogenetic variation As assumed by our a priori classification of species, problems in observing the expected number of annual rings were most pronounced in the (diffuse) species with an abnormal anatomical pattern. Accordingly, tracking the formation of annual rings during several growing seasons yielded less consistent patterns than in the other types of species. This result supports previous evidence that there is no generally reliable criterion for identification of annual rings in the roots of forbs other than patterns of vessel size (Werner 1978; Dietz and Ullmann 1997; von Arx and Dietz 2005). However, our results also show that even in species with a problematic anatomical pattern 231 232 Plant Biology 8 (2006) (e.g., Euphorbia esula), annual rings may be identifiable (with the assistance of anatomical markers other than vessel size) if the phenology of ring formation is known. The reliability of such markers must be verified for each species. In some or most of the species (e.g., Anchusa officinalis, Cichorium intybus, T. pratense, Silene vulgaris, Galium verum, and Hypericum perforatum) the number of occasionally occurring bands of (stronger) lignified parenchyma does not always correspond to the number of annual rings and may therefore mislead the observer (Werner, 1978). The comparably uniformly low number of missing first rings in the (tap-rooted) short-lived species may result from their generally faster growth that may have allowed them to attain a sufficient size in the incomplete first growing season for producing an identifiable first annual ring, quite in contrast to the long-lived species with fibrous roots (cf. Table 1). Implications for herb-chronology The two major concerns with herb-chronological analyses have been that the method is destructive and that one needs to be sure that the growth rings in the roots of forbs are developed annually. Many previous studies (see Introduction) have demonstrated that the destructive approach of herb-chronological analysis does not seriously limit its value in plant ecology (but see Colling et al., 2002). Individuals in a small number of species may regenerate from the roots remaining in the soil after herb-chronological harvest and sample sizes in the order of ten to 100 individuals per population already produce reasonable results for most questions in population ecology (cf. Boggs and Story, 1987; Dietz and Ullmann, 1998; Dietz et al., 1999; Dietz, 2002; Dietz et al., 2004; Erschbamer and Retter, 2004; Münzbergovà et al., 2005; von Arx and Dietz, 2005). In most cases this sample size represents only a tiny fraction of the population size and, hence, should neither represent an unacceptable intervention into the viability of the population nor preclude future resampling of the same population to monitor changes in age or growth pattern over time (cf. Dietz, 2004). In addition, the results of our present study have experimentally confirmed existing evidence that clear growth rings in the roots of forbs are only induced by the seasonality of the climate in the temperate seasonal zone but not by other factors such as competition or regrowth following disturbance. A baseline proportion of about 1 – 2 % of full growing seasons may lack a detectable annual growth ring in the main roots of seasonal temperate forbs. For most studies, this small percentage will not pose serious problems. Neither the shape of population age structures nor ring width patterns will be considerably affected if an annual ring is occasionally missing. Even studies sensitive to shifts in the reading frame of ring width chronologies will produce acceptable results if low percentages of rings are absent, especially if one is interested in a ring width signal that is among the most recent years of the studied plants (Dietz et al., 2004; Dietz and von Arx, 2005). This is because the lower the number of rings one has to count (starting from the outer border) to locate the ring of interest, the lower the likelihood that a missing ring will result in the wrong identification of the target ring. G. von Arx and H. Dietz Acknowledgements We thank Fritz H. Schweingruber for help in planning and setting up the experiment, for technical advice and for comments on the manuscript. Sabine Güsewell is acknowledged for assistance with data analysis and Tino Fotsch for support in maintaining the common garden experiment. Jake M. Alexander and two anonymous reviewers improved a former version of the manuscript. This work was supported by a grant of the Swiss National Science Foundation (SNF) to HD (NF 3165426.01). References Bakshi T. S. and Coupland R. T. (1960) Vegetative propagation in Linaria vulgaris. Canadian Journal of Botany 38, 243 – 249. Boggs K. W. and Story J. M. (1987) The population age structure of spotted knapweed (Centaurea maculosa) in Montana. Weed Science 35, 194 – 198. Brys R., Jacquemyn H., Endels P., De Blust G., and Hermy M. (2004) The effects of grassland management on plant performance and demography in the perennial herb Primula veris. Journal of Applied Ecology 41, 1080 – 1091. Colling G., Matthies D., and Reckinger C. (2002) Population structure and establishment of the threatened long-lived perennial Scorzonera humilis in relation to environment. Journal of Applied Ecology 39, 310 – 320. Crawley M. J. (1997) Plant Ecology. Oxford: Blackwell Science. Crawley M. J. (2005) Statistical Computing: An Introduction to Data Analysis Using S-Plus. Chichester: John Wiley. Dietz H. (2002) Plant invasion patches – reconstructing pattern and process by means of herb-chronology. Biological Invasions 4, 211 – 222. Dietz H. (2004) Plant invasion patches revisited – changes in development. Weed Technology 18, 1381 – 1385. Dietz H. and Fattorini M. (2002) Comparative analysis of growth rings in perennial forbs grown in an alpine restoration experiment. Annals of Botany 90, 663 – 668. Dietz H., Fischer M., and Schmid B. (1999) Demographic and genetic invasion history of a 9-year-old roadside population of Bunias orientalis L. (Brassicaceae). Oecologia 120, 225 – 234. Dietz H. and Schweingruber F. H. (2001) Development of growth rings in roots of dicotyledonous perennial herbs: experimental analysis of ecological factors. Bulletin of the Geobotanical Institute ETH 67, 97 – 105. Dietz H. and Schweingruber F. H. (2002) Annual rings in native and introduced forbs of lower Michigan, USA. Canadian Journal of Botany 80, 642 – 649. Dietz H. and Ullmann I. (1997) Age-determination of dicotyledonous herbaceous perennials by means of annual rings: exception or rule? Annals of Botany 80, 377 – 379. Dietz H. and Ullmann I. (1998) Ecological application of “Herbchronology”: comparative stand age structure analyses of the invasive plant Bunias orientalis L. Annals of Botany 82, 471 – 480. Dietz H. and von Arx G. (2005) Climatic fluctuation causes largescale synchronous variation in radial increments of the main roots of northern hemisphere forbs. Ecology 86, 327 – 333. Dietz H., von Arx G., and Dietz S. (2004) Growth patterns in two Alpine forbs as preserved in annual rings of the roots: the influence of a snowbank gradient. Arctic, Antarctic, and Alpine Research 36, 591 – 597. Erschbamer B. and Retter V. (2004) How long can glacier foreland species live? Flora 199, 500 – 504. Freville H. and Silvertown J. (2005) Analysis of interspecific competition in perennial plants using life table response experiments. Plant Ecology 176, 69 – 78. Annual Root Rings in Forbs Grime J. P. (2001) Plant Strategies, Vegetation Processes, and Ecosystem Properties. Chichester: John Wiley. Harper J. L. (1977) Population Biology of Plants. London: Academic Press. Humulum C. (1981) Age distribution and fertility of populations of the arctic-alpine species Oxyria digyna. Holarctic Ecology 4, 238 – 244. Hurlbert S. H. (1984) Pseudoreplication and the design of ecological field experiments. Ecological Monographs 54, 187 – 211. Larson P. R. (1994) The Vascular Cambium. Development and Structure. Berlin: Springer. Martinkova J., Kocvarova M., and Klimesova J. (2004) Resprouting after disturbance in the short-lived herb Rorippa palustris (Brassicaceae): an experiment with juveniles. Acta Oecologica 25, 143 – 150. Menges E. S. (2000) Population viability analyses in plants: challenges and opportunities. Trends in Ecology and Evolution 15, 51 – 56. MeteoSchweiz (2005) Norm- und Mittelwerte 1961 – 1990 (http:// www.meteoschweiz.ch /de/Daten/Messwerte/IndexMesswerte. shtml). Münzbergova Z., Krivanek M., Bucharova A., Juklickova V., and Herben, T. (2005) Ramet performance in two tussock plants – do the tussock-level parameters matter? Flora 200, 275 – 284 Natural Resources Conservation Service, USDA (2005) Plants national database (http://plants.usda.gov/index.html). Oberdorfer E. (1994) Pflanzensoziologische Exkursionsflora. Stuttgart: Ulmer. Rixen C., Casteller A., Schweingruber F. H., and Stoeckli V. (2004) Age analysis helps to estimate plant performance on ski pistes. Botanica Helvetica 114, 127 – 138. Schaer C., Vidale P. L., Lüthi D., Frei C., Häberli C., Lingier M. A., and Appenzeller C. (2004) The role of increasing temperature variability in European summer heatwaves. Nature 427, 332 – 336. Schweingruber F. H. (2001) Dendroökologische Holzanatomie. Anatomische Grundlagen der Dendrochronologie. Bern: Paul Haupt. Schweingruber F. H. and Dietz H. (2001) Annual rings in the xylem of dwarf shrubs and perennial dicotyledonous herbs. Dendrochronologia 19, 115 – 126. Silvertown J. and Charlesworth D. (2001) Introduction to Plant Population Biology. Oxford: Blackwell Science. Tutin T. G., Heywood V. H., Burges N. A., Moore D. M., Valentine D. H., Walters S. M., and Webb, D. A. (1964 – 1983) Flora Europaea. Cambridge: Cambridge University Press. von Arx G. and Dietz H. (2005) Automated image analysis of annual rings in the roots of perennial forbs. International Journal of Plant Sciences 166, 723 – 732. von Arx G., Edwards P. J., and Dietz H. (2006) Evidence for life history changes in high-altitude populations of three perennial forbs. Ecology, in press. Vorontzova L. I. and Zaugolnova L. B. (1985) Population biology of steppe plants. In The Population Structure of Vegetation (White, J., ed.), Dordrecht: Dr W. Junk Publishers, pp.143 – 178. Werner P. A. (1978) On the determination of age in Liatris aspera using cross-sections of corms: implications for past demographic studies. The American Naturalist 112, 1113 – 1120. Wijk S. (1980) Plants and the snow cover. Fauna and Flora 75, 32 – 36. Zoller H. (1949) Beitrag zur Altersbestimmung von Pflanzen aus der Walliser Felsensteppe. Berichte über das Geobotanische Forschungsinstitut Rübel in Zürich 61 – 68. Plant Biology 8 (2006) G. von Arx Institute of Integrative Biology ETH Swiss Federal Institute of Technology ETH Zentrum CHN G 35.1 8092 Zürich Switzerland E-mail: [email protected] Editor: R. Monson 233