Introduction

Disturbance is a common feature of natural habitats and occurs at various spatial and temporal scales (Klimešová and Klimeš 2003, 2008; Martínková et al. 2004; Bornette et al. 2008; Puijalon et al. 2008; Latzel and Klimešová 2009). Severe disturbance can cause destruction of above-ground plant parts (Klimešová and Klimeš 2003, 2008) and fragmentation of clones formerly interconnected by e.g. stolons, rhizomes or roots (Sainty et al. 1998; Stuefer and Huber 1999; Jia et al. 2008). In frequently disturbed habitats, therefore, plants must have developed some strategies to cope with disturbance (Klimešová and Klimeš 2003, 2008; Martínková et al. 2004; Bornette et al. 2008; Puijalon et al. 2008; Latzel and Klimešová 2009).

Stoloniferous herbs represent a large group in clonal plants and are distributed in a wide range of habitats (Klimeš et al. 1997). When remaining intact, stolon internodes play important roles in clones’ foraging responses (Birch and Hutchings 1994; Wijesinghe and Hutchings 1999) and resource sharing (Alpert 1999; Alpert et al. 2003), thereby contributing greatly to their fitness (Alpert 1999; van Kleunen and Stuefer 1999; Alpert et al. 2003; Roiloa et al. 2007; Wang et al. 2008; Yu et al. 2009).

Reserves in plant organs may help plants survive and regrowth after damage caused by various disturbances such as grazing, mowing, trampling, water currents and human activities (Lawson et al. 2000; Goulas et al. 2001; Klimešová and Klimeš 2003, 2008; Martínková et al. 2004; Puijalon et al. 2008). In natural habitats stoloniferous herbs face frequently with severe disturbance that often results in fragmentation of the clones (Sainty et al. 1998; Stuefer and Huber 1999; Jia et al. 2008). After fragmentation stolon internodes may play an important role in ramet survival and growth because they may store substantial amount of non-structural carbohydrates and soluble proteins that can be mobilized and used by the ramets (Fry 1993; Bouchart et al. 1998; Stuefer and Huber 1999; Suzuki and Stuefer 1999; Lawson et al. 2000; Goulas et al. 2001). Such a role of stolon internodes may be especially important for juvenile ramets that have not rooted and thus lack the capacity to acquire water and nutrients from soils (Stuefer and Huber 1999).

Because disturbance occurs at different scales with different degrees of severity (Klimešová and Klimeš 2003, 2008; Bornette et al. 2008), stolons of clonal plants may be subjected to different degrees of damage. After fragmentation, therefore, a ramet can be attached (a) with various lengths of stolon internodes, (b) with a proximal internode (before and thus developmentally older than the ramet) only, a distal internode (after and thus developmentally younger than the ramet) only or both, and (c) with or without leaves. If the reserve in internodes contributes to the fitness of individual ramets (Stuefer and Huber 1999), then increasing length of the stolon internodes attached to the ramets will increase ramet survival and growth.

Besides the main function of photosynthesis, leaves can also store a certain amount of carbohydrates and proteins (Granstedt and Huffaker 1982; Danckwerts and Gordon 1989; Corre et al. 1996). After fragmentation, storage in leaves may also contribute to the survival and growth of unrooted ramets, especially when the storage in other plant parts (e.g. stolon internode) is insufficient. We, therefore, hypothesize that presence of leaves will contribute to ramet survival and growth and such effects will be larger when the stolon internode attached to the ramet is shorter.

Studies have shown that structural and physiological constraints can greatly affect resource mobilization within a clone (Watson and Casper 1984; Price and Hutchings 1992; Price et al. 1996; Marshall and Price 1997). For instance, in Glechoma hederacea vascular architecture causes sectorial patterns of resource distribution, and in some species discrete integrated physiological units (i.e. sectorial compartments where the metabolite movement is confined; Watson and Casper 1984; Marshall and Price 1997) can be formed. Tracer studies have also shown that the efficiency of resource mobilization differs between basipetal and acropetal directions (Marshall 1990; Jonsdottir and Watson 1997). Moreover old (proximal) and young (distal) stolon internodes may differ in the function of carbohydrate storage (Turner and Pollock 1998). Thus, it is hypothesized that the positions (proximal, distal or both) of the stolon internode attached to the ramet will affect ramet survival and growth.

We tested these hypotheses in two greenhouse experiments with the stoloniferous, invasive herb Alternanthera philoxeroides (alligator weed). In one experiment, we measured the survival rate and growth of the ramets either without stolon internode (0 cm in length) or attached with internodes of 2, 4, 6 and 8 cm and either without or with two opposite leaves. We predict that increasing length of the internode attached to the ramet and presence of leaves will increase the survival and growth of A. philoxeroides. In the other experiment, we measured the survival rate and growth of the ramet attached with a proximal internode, a distal internode or both, keeping the total length of the internode segment the same. We predict that stolon internode position will significantly affect the survival and growth of A. philoxeroides.

Materials and methods

The species

Alternanthera philoxeroides (Mart.) Griseb. is a perennial herb of the Amaranthaceae family and originates from South America (Sainty et al. 1998; Pan et al. 2007). It is listed as one of the worst invasive weeds in many countries, including China (Holm et al. 1997; Wang et al. 2008). A.philoxeroides produces monopodially growing stolons, and relies mainly on clonal growth to sustain and renew its populations (Pan et al. 2007; Yu et al. 2009). A ramet of A.philoxeroides consists of a node that can root at moist conditions, two opposite leaves, and two axillary buds that can develop into new stolons and ramets or, at some conditions, inflorescences. A.philoxeroides forms white, capitate inflorescences of 1–2 cm in diameter from leaf axils, and produce viable seeds in its native range (Pan et al. 2007). However, in China it does not produce viable seeds under field conditions and genetic diversity of the populations is extremely low, likely originated from a single genotype (Ye et al. 2003; Wang et al. 2005; Pan et al. 2007). New stolons are formed throughout the growth season, whereas inflorescences are formed during summer and autumn. A.philoxeroides is amphibious and grows vigorously in both aquatic (e.g. still water and flowing stream) and terrestrial (e.g. roadside, cropland and lawn) environments (Sainty et al. 1998; Ma and Wang, 2005; Pan et al. 2006; Jia et al. 2008). Morphology, anatomy and growth of A.philoxeroides in aquatic conditions differ significantly from those in dry, terrestrial habitats, showing respective adaptations to local environments (Tao et al. 2004; Tao and Jiang 2004). But in China these differences are most likely plastic responses rather than genetically based changes (Ye et al. 2003; Wang et al. 2005; Pan et al. 2007). In aquatic situations, A.philoxeroides can occupy the entire water surface, restrict the growth of fish and inhibit boat movement, where it also provides a habitat for mosquitoes to breed. In terrestrial situations, A.philoxeroides can displace native plants, and block irrigation and drainage systems of farmland (Sainty et al. 1998; Ma and Wang 2005; Pan et al. 2006; Wang et al. 2008).

Stolon internodes of A. philoxeroides can be easily broken and its leaves lost due to disturbance such as herbivory, mowing, trampling, herbicides, strong water current, fishing and transportation boats, and the whole clones can be readily split into segments of different size (Sainty et al. 1998; Schooler et al. 2006; Julien et al. 2007). Because each node (ramet) has the potential to survive independently and to produce new stolons and ramets, A. philoxeroides can spread vigorously (Holm et al. 1997; Sainty et al. 1998).

Experimental design

Ramets of A. philoxeroides were collected from Jiangxi province, China in mid April, and vegetatively propagated in an artificial pond (7.0 m × 1.2 m × 1.2 m, length × width × depth) in the Botanical Garden of the Institute of Botany, the Chinese Academy of Sciences in Beijing. After about 5 months of cultivation, the surface of the pond was fully occupied by the stolons and ramets of A. philoxeroides. On 14 October, about 3,000 stolon segments of similar stolon diameter were cut off from this population. Each stolon segment contained an unrooted ramet (i.e. a node with two leaves and two axillary buds), the proximal internode (connected to its adjacent, older ramet) and the distal internode (connected to its adjacent, younger ramet). Both proximal and distal internodes were at least 5 cm long. Of the 3,000 stolon segments, 1,260 were selected and used for the experiments and 45 for initial measurements.

The study consisted of two experiments. The first experiment (Exp. 1, Table 1) was to test the effects of stolon internode length and presence of leaves on the survival and growth of the A. philoxeroides ramets. There were five treatments of stolon internode length (0, 2, 4, 6, and 8 cm), combined with two treatments of leaf presence status (with or without the two leaves). In this experiment, the stolon internode (segment) consisted of both proximal and distal internodes, each having half of the total length. For instance, for the 4 cm length treatment, both the distal and the proximal internode of the segment were cut into 2 cm long. The second experiment (Exp. 2, Table 1) was mainly to test the effects of stolon internode position and its potential interactive effects with stolon internode length and with presence of leaves. In this experiment, there were three levels of stolon internode position (the internode was located only in the proximal part—before the ramet, only in the distal part—after the ramet, or in both), combined with two levels of internode length (2 and 4 cm) and two levels of leaf presence status (with or without the two leaves). Because the four treatments in which the ramets were attached with both a proximal and a distal internode (Table 1, shaded portion) were already included in Exp. 1, we omitted them in Exp. 2. Thus, these four treatments were shared by the two experiments (Table 1, shaded portion).

Table 1 Design of the two experiments
Full size table
Table 2 ANOVA results of effects of stolon internode length (0, 2, 4, 6 and 8 cm), presence of leaves and the interaction on the survival and growth of the Alternanthera philoxeroides plants at harvest (on week 9)
Full size table

In the two experiments, 63 plastic containers (39 cm × 26 cm × 10.5 cm; length × width × height) were used, and each container was divided into two equal parts, corresponding to two boxes (39 cm × 13 cm × 10.5 cm). The substrate used in the container was a mixture of sand and peat at a volume ratio of 1:1. Each container was planted with a total of 20 stolon segments: ten stolon segments with leaves were planted in one box and ten without leaves (leaves removed) were in the other. The ten segments in each box were placed horizontally in ten evenly spaced positions and then covered by 1 cm deep vermiculite, which was sufficient to completely cover all the segments and the leaves. Thus the treatments mimicked the situations faced by A. philoxeroides in frequently disturbed terrestrial habitats. Covering with vermiculite had three advantages: (1) it could prevent light to reach the segments and the leaves so that the confounding effects of photosynthesis by them were avoided, (2) it would not produce a significant physical barrier as was done by the real soil, and (3) it could make moisture conditions more homogeneous and comparable across the treatments. In frequently disturbed habitats, it is also possible that the ramets are covered by soils after clone fragmentation. During the experiments, the containers were systematically repositioned to avoid effects of possible environment patchiness within the greenhouse. To keep the soil moist, tap water was supplied once a day. There were seven replicates of boxes for each of the 18 treatments in the two experiments (Table 1).

The experiments started on 15 September 2009, ended 9 weeks later on 20 November, and were conducted in a heated greenhouse at Forest Science Co, Ltd of Beijing Forestry University. During the experiments mean temperature in the greenhouse were 17.4°C, as measured hourly by two Hygrochron temperature loggers (iButton DS1923; Maxim Integrated Products, USA).

At the start of the experiments, the average diameter of the stolon segments was 1.92 ± 0.24 mm (mean ± SD, N = 40). The dry weight of the 0-, 2-,4-,6- and 8-cm-long stolon segments were 6.9 ± 2.0, 8.4 ± 2.9, 12.1 ± 4.1, 15.8 ± 2.5, 17.0 ± 2.0 mg (mean ± SD, N = 5), respectively.

Measurements

On 4–5 October (week 3), 31 October–1 November (week 6) and 18–19 November (week 9), number of ramets and number of leaves were counted and total stolon length was measured for each A. philoxeroides plant that had emerged from the soil layer. On 20–22 November, each living plant was harvested and separated into leaves, stolons and roots. Total leaf area of each plant was measured by WinFOLIA Pro 2004a (Regent Instruments Inc., Canada). Then all plant parts were oven-dried at 70°C for 48 h and weighed. Whether the initial internode still existed or was depleted was also noted during harvest.

Data analysis

Because each box was initially planted with ten stolon segments, we could calculate the survival rate of the ramets in each box at harvest. We also calculated the average number of ramets, number of leaves, stolon length, leaf area and biomass of the emerged, living A. philoxeroides plants in each box. These data (survival rate and mean values of the variables measured in each box) were used in the analyses described below.

For Exp. 1, two-way ANOVA was used to investigate the effects of stolon internode length (0, 2, 4, 6 and 8 cm) and presence of leaves (with or without) on the survival rate, biomass and total leaf area of the A. philoxeroides plants. For Exp. 2, three-way ANOVA was conducted to test the effects of stolon internode position (proximal, distal and both), internode length (2 and 4 cm) and presence of leaves on the survival rate, biomass and leaf area of the A. philoxeroides plants. Data on number of ramets, stolon length and number of leaves measured on week 3, 6 and 9 were analyzed using repeated measures ANOVA, with measuring week as the repeated variable. SPSS 17.0 software (SPSS, Chicago, IL, USA) was used for all analyses.

Results

Effects of internode length and presence of leaves: Exp. 1

Stolon internode length and presence of leaves significantly affected the survival rate, biomass, total leaf area, number of ramets, total stolon length and number of ramets of the A. philoxeroides plants (Tables 2 and 3, significant effects of L and Le). All these variables increased markedly with increasing stolon internode length, and were significantly larger when the leaves were present (Figs. 1 and 2). There was no interaction effect of internode length by presence of leaves (Tables 2 and 3, no significant effect of L × Le or W × L × Le).

Table 3 ANOVA results of effects of stolon internode length (0, 2, 4, 6 and 8 cm), presence of leaves, measuring week (3, 6 and 9) and their interactions on number of ramets, stolon length and number of leaves of the Alternanthera philoxeroides plants
Full size table
Fig. 1
figure 1

Effects of stolon internode length and presence of leaves on a survival rate, b biomass and c total leaf area of the Alternanthera philoxeroides plants at harvest (on week 9). Means + 1 SE are given

Full size image
Fig. 2
figure 2

Effects of stolon internode length and presence of leaves on a–c number of ramets, d–f total stolon length and g–i number of leaves of the Alternanthera philoxeroides plants measured on week 3, 6 and 9, respectively. Means + 1 SE are given

Full size image

Effects of internode position: Exp. 2

Stolon internode position significantly affected the survival rate, biomass, total leaf area, number of ramets, total stolon length and number of leaves of the A. philoxeroides plants (Tables 4 and 5, significant position effect). The survival rate was significantly smaller when the ramets were attached with a distal internode than when attached with a proximal internode or both (Fig. 3a). However, biomass and total leaf area were the largest when the ramets were attached with a distal internode, smallest when attached with a proximal internode and intermediate when attached with both (Fig. 3b–c). On week 3, number of ramets, total stolon length and number of leaves were the largest when the ramets were attached with both internodes (Fig. 4a, d, g), but on week 6 and 9, they were the largest when the ramets were attached with a distal internode (Fig. 4b–c, e–f, h–i).

Table 4 ANOVA results of effects of stolon internode position, length (2 and 4 cm), presence of leaves and their interactions on the survival and growth of the Alternanthera philoxeroides plants at harvest (on week 9)
Full size table
Table 5 ANOVA results of effects of stolon internode position, length (2 and 4 cm), presence of leaves, measuring week (3, 6 and 9) and their interactions on number of ramets, stolon length and number of leaves of the Alternanthera philoxeroides plants
Full size table
Fig. 3
figure 3

Effects of stolon internode position on a survival rate, b biomass and c total leaf area of the Alternanthera philoxeroides plants at harvest (on week 9). Bars sharing the same letters are not different at P = 0.05 (by Duncan’s test). Means + 1 SE are given

Full size image
Fig. 4
figure 4

Effects of stolon internode position on a–c number of ramets, d–f total stolon length and g–i number of leaves of the Alternanthera philoxeroides plants measured on week 3, 6 and 9, respectively. Bars sharing the same letters are not different at P = 0.05 (by Duncan’s test). Means + 1 SE are given

Full size image

Interactive effects of internode position by length and by presence of leaves: Exp. 2

Effects of stolon internode position on the survival rate of the A. philoxeroides plants did not depend on presence of leaves or internode length (Fig. 5a, b, Table 4, no significant P × Le effect). On the other hand, the effects of internode position on biomass, total leaf area, total stolon length and number of leaves of the A. philoxeroides plants depended significantly on both internode length (Tables 4 and 5, significant P × L effects) and presence of leaves (Tables 4 and 5, significant P × Le effects). The effects were smaller when the stolon internode was shorter or when the leaves were removed (Fig. 5c–l).

Fig. 5
figure 5

Interactive effects of stolon internode position by length (a, c, e, g, i and k) and position by presence of leaves (b, d, f, h, j and l) on the survival (a–b) and growth (c–l) of the Alternanthera philoxeroides plants at harvest (on week 9). Means + 1 SE are given

Full size image

Discussion

Few studies directly tested the roles of stolon internodes and internode positions in ramet survival and growth after severe disturbance (Stuefer and Huber 1999; Suzuki and Stuefer 1999). We found that attachment with internodes and leaves contributed greatly to ramet survival and growth of A. philoxeroides, and more intriguingly, the contribution by proximal and distal internodes differed significantly and also depended on internode length and presence of leaves.

The survival rate was on average 46% for the A. philoxeroides ramet without stolon internode and increased to 96% when it was attached with an 8-cm-long internode (Fig. 1a). Also, all growth measures increased markedly with increasing internode length (Figs. 1b, c and 2). These results agree with previous findings on Potentilla anserina (Stuefer and Huber 1999) and suggest that storages in stolon internodes increase the fitness of juvenile ramets. By monitoring changes in carbohydrate and protein contents in stolon internodes, it has been shown that substantial amount of non-structural carbohydrates and soluble proteins stored in stolon internodes could be readily mobilized and used for the regrowth of damaged tissues (Danckwerts and Gordon 1989; Corre et al. 1996; Baur-Höch et al. 1999; Lawson et al. 2000; Goulas et al. 2001), and that at low temperature the survival rate of the Eremochloa ophiruoides stolons was positively correlated with the carbohydrate content in stolon internodes (Fry 1993). Although we did not measure the mobilization of carbohydrates and nitrogen in A. philoxeroides, we found that 87.2% of the internodes attached to the ramets were depleted completely at the time of harvest (Dong and Yu unpubl. data). Thus, it is very likely that the carbohydrates and nitrogen stored in the stolon internodes were mobilized and used by the juvenile, unrooted ramets of A. philoxeroides, which greatly increased their survival and growth. In frequently disturbed habitats, therefore, efficient use of the reserves in stolon internodes may be an adaptive strategy for stoloniferous plants to cope with severe disturbance that often causes fragmentation of clones.

Presence of leaves increased the survival and growth of A. philoxeroides, but it did not interact with stolon internode length (Table 2, Figs. 1 and 2), suggesting that storages in leaves contribute to the fitness of juvenile ramets and such effects do not depend on the amount of the reserves in the internodes of the ramets. There is evidence that leaves can store some non-structural carbohydrates and soluble proteins (Granstedt and Huffaker 1982; Danckwerts and Gordon 1989; Corre et al. 1996), and these substances stored in the two opposite leaves may have been used by the ramets of A. philoxeroides to sustain its survival and growth. Therefore, in addition to the main function as a primary carbohydrate producer through photosynthesis, leaves may also serve as a short-term reserve that can be used by the ramets bearing them to deal with severe disturbance.

At harvest all growth measures of A. philoxeroides were larger when the ramets were attached with a distal internode than when they were attached with a proximal internode (Figs. 3b, c and 4). Also, such effects were larger when the stolon internodes were longer or when the leaves were present (Fig. 5c–l), suggesting that the effects increased with increasing the total amount of the reserves. Of the 840 ramets attached with either a proximal internode or both a proximal and a distal internode, 25% still had a visible, healthy proximal internode and 75% of the proximal internodes were depleted completely at harvest; in contrast, of the 840 ramets attached with either a distal internode or both, only five (0.6%) still had a visible, healthy distal internode at harvest (Dong and Yu unpubl. data.). One explanation for the observed position effects on plant growth is, therefore, that the carbohydrate and nitrogen reserves in the proximal internodes could not be used as efficiently by the juvenile, unrooted ramets of A. philoxeroides as those in the distal internodes, and some reserves in the proximal internodes could not be readily mobilized to the ramets. Previous studies have shown that structural constraints such as sectoriality and physiological constraints such as formation of discrete integrated physiological units (IPUs) can confine the mobilization of carbohydrates and nutrients within a clone (Watson and Casper 1984; Mashall and Price 1997). Therefore, there may be larger structural or physiological constraints between the ramet (node) and its attached proximal internode, which restricted the mobilization of the carbohydrates and soluble proteins stored in the proximal internodes so that they could not be efficiently used for the growth of the ramets. Further studies that compare in detail the anatomical structure of the stolon internodes at different parts (i.e. proximal and distal) and trace carbon and nitrogen movements along the stolons of A. philoxeroides may help to elucidate the exact mechanisms behind.

Another possible explanation for the observed position effects on plant growth is that carbohydrate support from stolon internodes was important only at the beginning of regrowth of the plants (fragments; Richards and Caldwell 1985). When new ramets (new leaves and roots) were produced, the plants did not continue to depend on storage but could attain sufficient resources through assimilation by new regenerated tissues. At that time older tissues started to bring more costs than benefits because their maintenance needed carbohydrates (for respiration). Given that all the distal internodes were already dead (i.e. completely depleted) at the early stage of the experiment (i.e. within 3 weeks), presence of proximal internodes would greatly reduce the growth of the plants at the time of harvest (at week 9). Contrary to growth, survival of ramets attached with a distal internode had a lower survival rate (Fig. 3a).

We conclude that in frequently disturbed habitats such as rivers, ditches, lawns, croplands and roadsides, where A. philoxeroides is distributed, juvenile, unrooted ramets of A. philoxeroides can use the reserves in their attached internodes and leaves to regenerate after clone fragmentation. This strategy may contribute greatly to the invasiveness of A. philoxeroides, and facilitate its invasion into frequently disturbed, local communities.