Comparative Analysis of Growth Rings in Perennial Forbs Grown in an Alpine Restoration Experiment

Abstract

Recent studies have demonstrated that growth rings are widespread in the roots of forbs, and there is evidence that the rings are formed annually. However, the annual nature and development of the growth rings has not yet been examined in comparative experimental studies. In this study growth rings were analysed in the main roots of four alpine forbs (Lotus alpinus, Trifolium thalii, Silene willdenowii and Potentilla aurea) that were grown in an alpine restoration experiment for 6 years. All individuals of L. alpinus and T. thalii, and some individuals of S. willdenowii showed six clearly demarcated growth rings, demonstrating that the rings were formed annually. P. aurea did not show distinguishable growth rings. In L. alpinus and T. thalii there were fluctuations in growth ring width that were consistent between individuals and also between species, and matched variations in climatic growth conditions. Results of the present study indicate that conclusions drawn from previous studies suggesting that growth rings in the roots of forb species are most likely formed annually are also valid for alpine plants. In terms of annual ring width patterns, this study also provides the first strong evidence for consistent responses of different forb species and individuals to commonly experienced variations in habitat conditions.

INTRODUCTION

Growth rings in the secondary xylem of roots are widespread among dicotyledonous perennial herbs of the strongly seasonal zone in the Northern hemisphere (Dietz and Ullmann, 1997; Schweingruber and Dietz, 2001; Dietz and Schweingruber, 2002). Phenological observations indicate that growth rings in forbs are formed annually (Dietz and Ullmann, 1997). In addition, for a limited set of species, individuals of known age could be verified to possess annual rings in the roots (Dietz and Ullmann, 1997; Dietz and Schweingruber, 2001). However, comparative experimental investigations that provide more rigorous and extensive tests of the suggested annual nature of growth rings in forbs are still lacking.

If growth rings are present in the roots of forbs there may be considerable fluctuations in ring width (Dietz and Ullmann, 1998), but there is almost no information available on the ecological processes underlying these patterns. To our knowledge, there is just one study on a population of the invasive forb Bunias orientalis L., growing in a gradient of light supply, that suggests annual growth increments in the roots respond sensitively to growth conditions (Dietz and Ullmann, 1998). If fluctuations in annual growth ring width are indeed generally related to variations in growth conditions, such relationships could be widely used in plant (population) ecology in herbaceous vegetation. For example, a posteriori analyses of growth ring patterns may reveal whether site conditions for plant growth have improved or deteriorated in previous years, or whether and how variations in microsite conditions affect plant growth as compared with differences between sites or habitats.

To provide the basis for unequivocal interpretation of growth ring patterns in forbs, experimental investigations are needed that include tests of the annual nature of the growth rings and analyses of patterns of growth ring width under a range of different treatments and for different species. Corresponding experiments were set up recently for different lowland forb species (Dietz and Schweingruber, 2001). While such experiments run over several years and, therefore, do not yield immediate results, there may also be the chance to sample individuals in existing, long‐term experiments.

In this paper we report on patterns of growth rings formed in four species of alpine forbs that were grown from seed and were planted in a restoration experiment on an alpine ski run in the Swiss Alps (Fattorini, 2001). This experiment provided an opportunity for comparative analysis under the conditions of the seasonal alpine environment that is characterized by short growth periods and mostly low temperatures (Körner, 1999). Individuals were sampled in the seventh year of the experiment. Thus a reasonably long period of time had elapsed in which to detect possible effects of variation in growth conditions or of ontogenetic variation on the number and width of the growth rings produced.

The objectives of the study were (1) to test whether growth rings occur in the secondary root xylem of the species and whether they represent true annual rings; and (2) to determine whether the species show fluctuations in growth ring width and, if so, whether these fluctuations are consistent among individuals or species, i.e. a result of common experience of varying growth conditions.

MATERIALS AND METHODS

Preparation of the species

Four perennial alpine forb species with permanent main roots, Lotus alpinus (DC.) Ramond, Trifolium thalii Vill. (both Fabaceae), Silene willdenowii Sweet (Caryo phyllaceae) and Potentilla aurea L. (Rosaceae) were used in the present study [nomenclature follows Hess et al. (1976–1980)]. Seeds of the species were collected from alpine grasslands in the surroundings of Davos (northeastern Swiss Alps) prior to the restoration in 1995. Seeds were germinated in Petri dishes in January 1995. Seedlings and juveniles were grown in trays for approx. 20 weeks in climate chambers and in glasshouses. Each tray contained 225 compartments (Rootrainers; Spencer‐Lemaire, Edmonton, Alberta, Canada) filled with approx. 32 cm³ of ‘garden’ soil (70 % peat, 14 % brown earth, 6 % sand, 10 % perlite granules). Plants were irrigated regularly with tap water but were not fertilized. Before transplanting to the field site, plants were transferred to approx. 1900 m a.s.l. to acclimatize for about 3 weeks.

Study site

The research area is located within the alpine belt of the Jakobshorn Mountain (2596 m a.s.l.) near Davos (46°47′N, 9°49′E), where snowfall and frost may occur at any time throughout the summer, and the vegetation period is confined to 3 months. A machine‐graded downhill ski run was constructed in 1970 across alpine sedge grassland on a north‐facing slope (35°) on silicate. No soil was left, and the substrate was rocky with very little organic matter and nitrogen content.

At the beginning of July 1995, restoration plots were set up on the ski run at approx. 2450 m a.s.l. Plots were either supplied with biodegradable wood fibre mats or small amounts of garden soil (1 kg m^–2) to test whether these treatments affect restoration success. Between 10–13 Jul. 1995, 480 individuals each of L. alpinus and T. thalii, and 240 individuals each of S. willdenowii and P. aurea were transplanted into the plots (distance between neighbouring plants approx. 15 cm). Further details of the experimental design are given in Fattorini (2001).

Data collection and analysis

On 13 Aug. 2001, i.e. in the seventh growth period, six well‐developed individuals per species were carefully excavated (sample size was kept low to preserve the restoration experiment for future investigations). The individuals were collected from differently treated plots (three per plot) to introduce environmental variation to the data set. Because of the small sample size, statistical results are pooled over the different treatments.

In the laboratory, the main roots were cut using a sledge microtome to produce thin transverse sections (approx. 30 µm). The roots were cut near to the proximal end to avoid missing growth rings (Schweingruber and Dietz, 2001). Cross‐sections of the roots were stained with Phloroglucinol/HCl (causing reddish colouration of the cell walls of vessels and of lignified parenchyma cells), and examined under a dissecting microscope (Leica MZ 8). Cross‐sections were photographed through the photo tube of the microscope using a digital camera (Nikon CoolPix 990, resolution 2048 × 1536 pixels) and images were transferred to the computer for image processing and analysis.

Secondary xylem was analysed for the presence and pattern of growth rings that form by the alternation of earlywood (large vessels) with latewood (small vessels) in concentric rings. If growth rings were present their number was counted and ring widths were measured manually using linear measurement tools in a technical drawing and image processing program (Deneba Canvas 7 for Windows, cf. Fig. 1). Three separate radii were analysed for each individual to account for tangential variation in growth ring width. Means of the three measurements were used as the width of the respective growth rings.

Fig. 1. Patterns of growth rings in the secondary root xylem of Lotus alpinus (A and B), Trifolium thalii (C), Potentilla aurea (D) and Silene willdenowii (E and F). Markers in A–C and E denote transition from latewood of the previous growing period to earlywood of the following one. The distance between consecutive markers represents the width of the enclosed growth ring at that radius. Bars = 1 mm.

Statistical analysis of the ring width patterns was possible for L. alpinus and T. thalii because all individuals of both species showed growth rings. The growth ring widths of each individual were standardized, i.e. the width values were subtracted from the sample mean and the differences divided by the standard deviation to bring all values to compatible units from a distribution with a mean of zero and a standard deviation of one. In all species the first (1995) growth increment, which was strongly influenced by culture conditions in the glasshouse, was particularly wide and was omitted from analysis.

A permutation test was used to determine whether particularly wide or narrow growth rings were significantly over‐represented in specific growth periods within the period 1996–2000. In the null model particularly wide or narrow growth rings are expected to occur with equal probability in every year, so that random patterns should emerge if several individuals are compared. To test for deviation from this null model, standardized values for each year were summed over all six individuals per species and divided by six to obtain a mean score for a given year. Each of these scores (one for each of the five consecutive years) was compared with a distribution of 999 scores obtained from the same data set but with the ring width values randomly reshuffled within individuals (null distribution). Significant deviations were detected by comparison of the observed scores with the scores of the null distribution (Manly, 1999). As five different years (1996–2000) were tested simultaneously Bonferroni adjustment was used (α = 0·01).

RESULTS

Presence and number of growth rings in the secondary xylem

In two of the species tested, Lotus alpinus and Trifolium thalii, clearly demarcated growth rings could be observed in all sampled individuals (Fig. 1A–C). In Silene willdenowii relatively clearly or clearly demarcated growth rings were found only in three individuals (collected from one of the two treatment plots, Fig. 1E). The other three individuals showed a relatively diffuse anatomical pattern (Fig. 1F). Likewise, all six individuals of Potentilla aurea showed a rather diffuse xylem anatomy lacking clear growth rings (Fig. 1D).

If growth rings were present, the rings were formed by moderate to strong differences in vessel diameter between earlywood and latewood and a higher density of vessels in the earlywood (Fig. 1). In these cases, i.e. in 15 individuals from three species, the number of growth rings matched exactly the age of the plants: each individual showed six growth rings corresponding to the number of growth periods from 1995 to 2000 (an example is marked in Fig. 1A). In addition, each individual showed the beginning of the seventh growth ring at the outer margin of the xylem (earlywood of 2001). These results provide strong evidence that growth rings in L. alpinus, T. thalii and S. willdenowii were formed annually.

Patterns of growth ring width

In all three species showing growth rings, the first ring (1995) was particularly wide, accounting for 19–37 % of the total xylem radius of L. alpinus and T. thalii (excluding the 2001 growth increment), and for 57–72 % in S. willdenowii (Fig. 2). In contrast, all later rings had a width varying between 8 and 21 % of the total xylem radius in L. alpinus and T. thalii, and between 4 and 9 % (16 %) in S. willdenowii. For all species, the pattern of fluctuations in ring width was fairly consistent (Fig. 2).

Fig. 2. Fluctuations in relative mean annual ring width in the secondary root xylem in Lotus alpinus (A), Trifolium thalii (B) and Silene willdenowii (C). Each line represents one individual. Different line styles indicate individuals from different treatments. See text for further information.

Despite inter‐individual fluctuations, all plants of L. alpinus and T. thalii showed a largely consistent pattern of ring widths, irrespective of the treatments under which the individuals were grown. In all cases the 1998 growth increment was larger than that of the previous or the following year and, except for one individual of T. thalii, was the largest growth ring in the whole period from 1996 to 2000 (Fig. 2). For both species the positive deviation of the 1998 growth ring width from average ring width was highly significant (Table 1). Conversely, the 1999 growth increment was smaller than that of the previous and of the following year (Fig. 2) and, in most cases, was the smallest in the period from 1996 to 2000. The negative deviation of the 1999 ring width was also highly significant (Table 1). For 1996, 1997 and 2000 there were no significant deviations but there was a trend for narrow rings in T. thalii in 2000. In S. willdenowii no substantial and consistent variation in the width of the narrow rings could be observed (Fig. 2).

Table 1.

Year‐specific deviation (s.d. units) of mean standardized ring widths from mean scores of the null distribution and their corresponding significance (P‐values)

	Lotus alpinus		Trifolium thalii
Year	Deviation	P	Deviation	P
1996	–0·25	0·58	0·18	0·64
1997	0·06	0·85	0·25	0·50
1998	1·08	0·00	1·27	0·00
1999	–1·05	0·00	–1·01	0·00
2000	0·16	0·60	–0·68	0·05

DISCUSSION

Presence of annual rings

This is the first comparative experimental study to demonstrate the existence of annual rings in roots of perennial herbaceous plant species and to verify the annual nature of growth rings for several alpine forb species. Three of the four species surveyed showed clearly demarcated annual growth rings in all or in some of the individuals. In all cases the individuals that exhibited growth rings had six complete rings that correspond to the number of growth periods from 1995 to 2000. Thus, there was no evidence for missing or false rings.

However, rings may be missing under adverse growth conditions, i.e. growth increments in the root may be too weak to produce clearly distinguishable growth rings even if annual growth increments are not actually missing in the secondary xylem (pseudo‐missing rings). This is indicated by Silene willdenowii where some of the individuals showed annual rings and others did not. Apparently, even moderate differences in growth conditions can be sufficient to determine whether distinguishable growth rings are produced in S. willdenowii or not. Closer inspection of the anatomical pattern of S. willdenowii individuals lacking growth rings (Fig. 1F) suggests that the rings of earlywood, comprising wider vessels, are tightly packed, obstructing the discrimination of annual rings in the distal portion of the secondary xylem.

A similar pattern was found in all individuals of Potentilla aurea (Fig. 1D). This species was stunted in the experimental plots compared with individuals growing in more productive alpine pastures nearby, where individuals of the species did show annual rings (C. Rixen, pers. comm.). It appears that a more or less diffuse anatomical pattern of predominantly wide vessels is rather widespread among forbs that are strongly limited in growth, e.g. in harsh alpine environments, in dry habitats or under intense competition (cf. Dietz and Ullmann, 1998; Dietz and Schweingruber, 2002).

True missing rings, i.e. missing growth increments, may occur in some years when plant growth is extremely reduced (e.g. due to permanent snow cover; Wijk, 1980). In dendrochonological studies of woody plants often only some xylem radii (incremenent cores) are analysed (Schweingruber, 1988; Wimmer and Vetter, 1999). In contrast, the whole cross‐section of the root is usually inspected in herb‐chronology. This prevents growth rings that are only partially developed being ignored.

The two Fabaceae species Lotus alpinus and Trifolium thalii had noduled roots and apparently profited from symbiosis with nitrogen‐fixing bacteria (cf. Jeffries et al., 1981; Walker, 1993). Increased nitrogen supply may at least partly explain their greater and clearer annual growth increments on the gravel slopes of the alpine ski run, where availability of nutrients was low (Fattorini, 2001).

Relationship between ring width patterns and growth conditions

For trees, it is well known that climatic variations can result in corresponding variations in annual growth increments in the trunks (e.g. Schweingruber, 1996; Meldahl et al., 1999; Parish et al., 1999). Similar to these observations in trees, in the present study there was high consistency in annual ring width within and among forb species during favourable and unfavourable years. These are the first results indicating the use of herb‐chronology for the detection of synchronous variation in growth among individuals and species of forbs in response to changes in growth conditions at the scale of a site. In both L. alpinus and T. thalii the 1998 growth rings were particularly wide whereas those of the following year were particularly narrow. In the northern and central Alps, 1998 was a warm and mostly sunny year with low precipitation. The higher altitudes in the Alps received record high temperatures in February 1998 (SMA MeteoSchweiz, 1998). High temperatures or low precipitation in the early months of the year prevented the accumulation of much snow in the northern Alps. In contrast, 1999 was a rather cloudy and wet year in the Alps. At the beginning of the year, during three periods of heavy precipitation in January and February, there was a record accumulation of snow at some places in the Alps resulting in avalanche catastrophes (SMA MeteoSchweiz, 1999). In that year the onset of the growth period for plants growing in the restoration experiment was probably delayed due to prolonged snow pack. Thus, for L. alpinus and T. thalii, climatic fluctuations in alpine growth conditions appear to be reflected in corresponding fluctuations in the annual growth increments in the roots. Unlike the two Fabaceae species, S. willdenowii did not show considerable fluctuations in annual ring width which may be attributed to the narrowness of the growth increments in this species.

We conclude that growth rings in the xylem of permanent main roots in alpine forbs are probably annual rings. This corresponds to previous suggestions for lowland plants (e.g. Dietz and Ullmann, 1997). Moreover, this study indicates that fluctuations in the width of annual rings reflect changes in growth conditions rather than, or in addition to, ontogenetic growth patterns. Further studies similar to that of Dietz and Schweingruber (2001) are needed to investigate the effects of a wide variety of factors such as climatic fluctuations, microsite conditions, herbivory and phylogenetic relationships on patterns of annual rings in forbs. If present, these kinds of relationships could prove very helpful in reconstructive studies on the ecology of perennial forbs.

Received: 6 June 2002; Returned for revision: 1 July 2002; Accepted: 15 August 2002 Published electronically: 2 October 2002