Introduction

Invasion by exotic plants is a major threat to biodiversity and ecosystem integrity and function (Dukes and Mooney 2004; Gratton and Denno 2006). Arthropod biodiversity and abundance are often lower in invasive than in native plant communities (Wu et al. 2009; also see the review by Bezemer et al. 2014). Some insect herbivores, however, can rapidly adapt or even evolve and thereby exploit exotic plants as new hosts because their native habitats are seriously disrupted (Agrawal and Kotanen 2003), which is often overlooked or ignored (Bezemer et al. 2014).

Previous studies have shown that differences in morphology, nutrition, and allelochemicals of exotic and native plants, as well in the predators and parasitoids that they support, may affect the preference of herbivorous insects for host plants in invaded areas (Bezemer et al. 2014). Although plant phenology is also known to greatly affect insect performance and population dynamics (Hunter 1992), there is little evidence concerning the effects of phenological differences between exotic and native plants on insect herbivores. The phenological effects of invasive plants, therefore, have been receiving increasing attention in food web ecology and invasion biology (Wolkovich and Cleland 2011).

We here report on how the abundance and diet of a tussock moth, Laelia coenosa, are affected by the invasive plant Spartina alterniflora (Spartina hereafter) in a salt marsh previously dominated by the moth’s native host plant, Phragmites australis (Phragmites hereafter). The salt marsh is located in the Yangtze River estuary of China and has been heavily invaded by Spartina, leading to a replacement of native Phragmites and other natives. This invasion has caused multiple consequences to the native ecosystems (for a review, see Li et al. 2009), one of which is a shift in the habitat of certain native insect herbivores and/or their diet from Phragmites to Spartina in the salt marsh (Wu et al. 2009). The tussock moth L. coenosa, for example, has become increasingly prevalent in Spartina over the past 5 years (Fig. 1a, b), although Phragmites has historically been the main host for this moth in China (Zhao 2003).

Fig. 1
figure 1

Laelia coenosa and seasonal development of Spartina and Phragmites at the Chongming Dongtan National Nature Reserve, China. a Larva of L. coenosa on Spartina; b egg, adult, and cocoon with pupa of L. coenosa on Spartina; c, e, g Spartina in late June, late July, and late September, respectively; d, f, h Phragmites in late June, late July, and late September, respectively

Full size image

Considering that Spartina and Phragmites differ in phenology (Liao et al. 2007), we hypothesized that this difference might have resulted in differences in L. coenosa performance between the exotic and native plants. To test this hypothesis, we studied the abundance and diet of L. coenosa larvae on Phragmites and Spartina from summer to autumn in the salt marsh. Our aims were to determine whether the longer growing season of Spartina versus Phragmites affects the native herbivore’s abundance and diet in the invaded salt marsh and what are the possible consequences to the native ecosystems if any.

Methods

Study site

This study was conducted at the Chongming Dongtan National Nature Reserve in the Yangtze River estuary, Shanghai, China (31°25′–31°38′N, 121°50′–122°05′E). The whole wetland occupies an area of 32,600 ha and is dominated by native Scirpus spp. and Phragmites and by exotic Spartina. Over the study area, where there is no Scirpus spp., Phragmites and Spartina exist as either their respective monocultures or mixtures (Li et al. 2014).

Monitoring of population dynamics of L. coenosa

We monitored the population dynamics of L. coenosa at two wetland sites (A: 31°30.930′N, 121°57.411′E; B: 31°30.961′N, 121°57.438′E) 72 m apart. These two collecting sites were selected because numbers of L. coenosa were found to be much higher than in other regions in the last 5 years, and also because there were three habitat patterns in which four types of hosts could be investigated: (1) Phragmites in its monoculture (PMC); (2) Spartina in its monoculture (SMC); (3) Phragmites in a Phragmites–Spartina mixture (PMT); and (4) Spartina in a Phragmites–Spartina mixture (SMT). At each site, we investigated all of the four host types, i.e. PMC, SMC, PMT and SMT. The mixed habitat (PMT, SMT) was transitional in that Phragmites was being gradually replaced by the invasive Spartina. Five replicate plots (1 × 1 m per plot) were evaluated for each host type. In each replicate, we randomly selected 20 individual shoots of Spartina or Phragmites in their respective monocultures, 20 Phragmites individuals in the PMT, and 20 Spartina individuals in the SMT. The number of L. coenosa larvae on these plants were counted. In 2011, we surveyed all plots three times (23 June, 26 July, and 20 September) based on the moth’s generation numbers in the year and the greatest abundance for each generation’s larva. Both Phragmites and Spartina were growing healthily and had green leaves in June and July. From late September, however, Phragmites started to lose its leaves but Spartina still had more healthy green leaves. The first and second generations of L. coenosa appeared in June and July, respectively, while the third generation appeared and began to prepare for overwintering from late September till late October.

Isotope analysis of the diet of L. coenosa

To determine the diet of L. coenosa, we randomly collected five larvae from each mixed habitat at site A and B on 23 June 2011. The specimens were divided into two groups: one group collected from Phragmites, and the other from Spartina. The samples were dried at 75 °C for 72 h and then ground into a fine powder with a mortar and pestle; the powder was passed through a 0.15-mm screen. The powders (0.5–1.0 mg dry mass per sample) were subjected to a stable isotope analysis (Wu et al. 2009) with an isotope ratio mass spectrometer (DELTA plus Advantage, Thermo Scientific, USA) at Nanjing Institute of Geography and Limnology, Chinese Academy of Sciences, Jiangsu Province. Stable isotope patterns for Phragmites and Spartina plants were analyzed by the same method. Leaf tissues of Phragmites or Spartina for isotope analysis were collected from 10 randomly selected plants from each of sampled plot, immersed and rinsed with deionized water before use (Wu et al. 2009). Each treatment was replicated four times.

Data analysis

All data for different habitat treatments at the same collecting time and site and for isotope analysis were subjected to one-way ANOVAs. Before analyses, all data were checked for normality and equal variance. The results are represented as means and SE, and means were compared with Tukey’s test. The analyses were performed with the statistical package SPSS NLN, 15.0 (SPSS Inc., Chicago, USA). All data of isotope analysis were multiplied by 10 and then log-transformed before the statistical analysis.

The stable isotope ratios were calculated by the following equation:

$$\delta \left( X \right) = \left[ {\left( {\frac{{R_{\text{sample}} }}{{R_{\text{standard}} }}} \right) - 1} \right] \times 1000$$
(1)

where X is 13C, R is the ratio of 13C/12C, and δ is a notation for the ratios and expressed as ‰. The analytical precision of the measurement is 0.2 ‰ for δ 13C.

Results

At each site, the larval density of each plant type decreased with plant growing season prolonging, but the decrease was greater on Phragmites than on Spartina (Fig. 2). In summer (June–July), L. coenosa larval densities were similar on Spartina and Phragmites in their either monocultures or mixtures, although monocultures support higher densities than mixtures (Fig. 2a–d). In autumn (late September), however, the densities were much higher on Spartina than on Phragmites both in monocultures and mixtures (Fig. 2e, f).

Fig. 2
figure 2

Larval densities of L. coenosa on Spartina and Phragmites as affected by host type, collecting time, and site. Values are means + SE. Within each panel, means with the same letters are not significantly different (Tukey’s test: P < 0.05). The host types SMC, PMC, SMT, and PMT represent, respectively, Spartina in its monoculture, Phragmites in its monoculture, Spartina in a mixed habitat of Spartina and Phragmites, and Phragmites in a mixed habitat of Spartina and Phragmites. (a, b), (c, d), and (e, f) indicate densities on 23 June, 26 July, and 20 September, respectively. (a, c, e) and (b, d, f) indicate densities at site A and site B, respectively

Full size image

Stable isotope patterns of Phragmites, Spartina, and L. coenosa collected from Phragmites and Spartina were significantly different (F 3, 15 = 82.3, P < 0.001) (Fig. 3). The δ 13C value of L. coenosa collected from Phragmites plants was almost the same as that of Phragmites, while that of L. coenosa collected from Spartina plants was intermediate with respect to the values of Phragmites and Spartina (Fig. 3), indicating that in the mixed habitat, L. coenosa collected from Phragmites still had selected Phragmites as their basal dietary resource, but those collected from Spartina had fed not only on Spartina but also on Phragmites.

Fig. 3
figure 3

Distribution of δ 13C isotope in Phragmites plants (PH), Spartina plants (SP), L. coenosa larvae that were collected from either Spartina (LS) or Phragmites (LP) in the mixtures of PH and SP. Values are means + SE. Means with the same letters are not significantly different (Tukey’s test: P < 0.05)

Full size image

Discussion

Invasive plants disrupt many trophic interactions in native communities (van Hengstum et al. 2014). Although effects of exotic plants on native arthropods have been frequently reported (e.g., Wolkovich 2010; Tang et al. 2012), the impact of invasive plants on the population dynamics and diets of native insects has rarely been examined (Bezemer et al. 2014). In the present study, we monitored the population dynamics of native L. coenosa on the native Phragmites and the invasive Spartina and examined its use of the native and the invasive plants as food sources. We found that although some L. coenosa still depended on Phragmites, many L. coenosa larvae lived in the exotic plants in both Spartina monocultures and mixed habitats of Spartina and Phragmites. Isotope analysis indicated that the larvae consumed Spartina tissue. These results suggest that Spartina may have been a new host species for L. coenosa.

Temporal variation in the quantity of the plant resources may support different population sizes of herbivores (Hunter and Price 1992). As such, difference in food supply between Spartina and Phragmites may be an important factor determining which plant species L. coenosa lives on and consumes, although both plants can be hosts for the insect. In eastern China, Spartina leaves usually start to develop in early March and senesce in late November, while Phragmites leaves emerge in late March and wither in early October. Because the length of the growing season is 270 days for Spartina but only 220 days for Phragmites in the Yangtze River estuary (Liao et al. 2007), the temporal differences in the population densities of L. coenosa in the monocultures and mixtures may have been caused by phenological differences between the two plant species. In the summer growing season (June and July), both Spartina and Phragmites produce abundant green, soft leaves (Fig. 1c–f) that supply adequate nutrients to insect herbivores and thus, generate the same higher abundance of L. coenosa on the exotic and native plants. In autumn (late September), however, Spartina still provides sufficient green leaves for L. coenosa but Phragmites does not because the latter senesces (Fig. 1g, h), and therefore, Spartina can support more larvae than Phragmites, as temperature is still suitable for the larval development (Zhao 2003).

On the other hand, as suggested by previous researchers (e.g., Feeny 1970; Coley 1980), phenological changes in plant foliar quality, such as seasonal increases in leaf toughness and concentrations of secondary chemicals, may also play an important role in explaining variations in insect abundance among different plants or seasons. In this study, larval density of L. coenosa decreased with sampling date, but the decrease was greater on Phragmites than on Spartina. Other than effects of temperature, seasonal changes in L. coenosa abundance between Spartina and Phragmites may be further explained by effects of different physical and chemical characteristics between the two plant species in different seasons. In addition, larval densities of L. coenosa in summer were much higher in monocultures than in mixtures, which may be caused by differences in disturbances (e.g., allelochemicals, predators and parasitoids) between the two community patterns (Bezemer et al. 2014).

Stable isotope analysis provides a relatively quick and logical result for exploring what resources are being used by consumers (Hobson 1999). The stable isotopes of 13C and 12C (expressed as δ 13C) are relatively conservative as ‘carbon’ is transferred from plants to herbivores via a food web (Gratton and Denno 2006) and thus, the C isotope signatures can be used to determine the food sources for herbivores (Vander Zanden et al. 1999). In our study, Phragmites (C3 plant) and Spartina (C4 plant) have distinguishing δ 13C isotope signatures (Gratton and Denno 2006) that allow us to identify the food sources of L. coenosa through comparing their δ 13C signatures among different treatments (Wu et al. 2009). Our results indicate that the insects collected from Phragmites still use Phragmites as their food resource, but those living in Spartina community have fed on both Spartina and Phragmites, suggesting that L. coenosa collected from Spartina in the mixture may have come from those feeding on Phragmites. Although we did not test δ 13C signatures in the monocultures, we predict that the insect may feed on both Spartina and Phragmites in their respective monocultures. Consistent with our findings, previous studies have shown that a native butterfly, Euphydryas editha, can shift its diet from its native host plant (Collinasia parviflora) to an exotic plant (Plantago lanceolata); E. editha evolved such that the proportion that preferred the exotic host to the native host rapidly increased due to ten-year feeding on P. lanceolata (Thomas et al. 1987; Singer et al. 1993). Our earlier study (Wu et al. 2009) also predicted that if native Phragmites was widely replaced by exotic Spartina in China, the native insects that feed on Phragmites might expand their diets to include Spartina and might even also evolve to prefer Spartina.

In the current situation, two issues need to be considered. First, as the phenology of Spartina extends, the overwintering of L. coenosa may be delayed and its generation number may increase due to the extended period of food resources the insect can use (referred to as “resource effect” of Spartina). Second, Spartina has standing litter throughout the year while Phragmites is often harvested for industrial use. The latter leaves only the litter of culm in winter even if it is not harvested. Therefore, L. coenosa can be sheltered effectively by Spartina in the overwintering period when the tide comes (referred to as “pool effect” of Spartina). If the “resource” and “pool” effects are additive, we can predict that L. coenosa abundance may increase considerably, which will ultimately alter interactions between native consumers and native plants.

Nonetheless, the efficiency of the insect’s control over Spartina may be devalued because the third generation of L. coenosa occurs in the late growing season of Spartina (i.e. at the mature stage of its seeds), which leads to little regulating influence on Spartina at population level. Till now, the ability of L. coenosa feeding on Phragmites is still stronger than on Spartina in the salt marsh (personal observation by the authors), indicating that although the exotic plant supports its higher abundance, the insect cannot effectively control Spartina at population level. However, because Phragmites is the traditional natural host for the insect (Zhao 2003), the plant has been found to be seriously damaged by L. coenosa in eastern China including Chongming Dongtan. The differences in damaging ability between the invasive and native plants by the moths may lead to “apparent competition” (Holt and Kotler 1987; Noonburg and Byers 2005) that will further facilitate Spartina to displace Phragmites via increasing the pressure of local consumers to native plants (i.e. indirect invasion of Spartina). Effect by apparent competition may create an advancing invasion front unlikely to retreat, and could interact with other mechanisms of biological invasion (Orrock et al. 2008), hindering the conservation and restoration of the invaded ecosystem.

In conclusion, our results suggest that Spartina invasion has altered the population abundance and has extended the diet of the indigenous herbivorous insect, L. coenosa that previously feeds on native Phragmites, which may be caused by phenological differences between the invasive and native plants. Such differences can be predicted to change the trophic interactions of the invaded ecosystem and may facilitate further invasion of exotic Spartina curtailing the reestablishment of native Phragmites.