Niche comparison among two invasive leafminer species and their parasitoid Opius biroi: implications for competitive displacement

Fundamental to competitive displacement in biological invasion is that exotic species occupy the ecological niches of native species in novel environments. Contrasting outcomes of competitive displacement have occurred between Liriomyza trifolii and L. sativae in different geographical regions following their introduction. Various factors have been advanced in an attempt to explain these different competitive outcomes, although none of these explanations have addressed the effects of niche differences. We conducted field cage experiments to compare the feeding and habitat niches of the two leafminer species and their primary parasitoid, Opius biroi, when occurring together on kidney bean. A wider spatiotemporal niche breadth was found in L. trifolii (0.3670) than in L. sativae (0.3496). With respect to the parasitoid, the proportional niche similarity between L. sativae and the parasitoid was 0.3936 but only 0.0835 for L. trifolii, while similar results were found for niche overlap, indicating that stronger trailing behaviour and parasitic effects of O. biroi occurred in L. sativae. In conclusion, L. trifolii has outperformed L. sativae in occupying the ecological niche and is superior to L. sativae in avoiding parasitization by the pupal parasitoid, O. biroi.

factors responsible for the rapid displacement 17,18 , while in Japan, the displacement has been attributed to the higher fecundity of L. sativae coupled with different effects of parasitoids 19 . However, niche effects, which are fundamental to interspecific competition 1 , were not considered in any of the reports explaining the displacement of the two leafminers.
Natural enemies, which are often thought to be a major factor in maintaining pests at low levels, can potentially reduce interspecific competition 20 . However, they may also be a significant factor in mediating interspecific interactions 21 . By experimentally removing the predators of two aphid species, increases in the density of the two species and the competitive interactions between them occurred 22 . A similar phenomenon occurred when parasitoids of hemlock scale insects were excluded 23 . According to Denno's assessment, approximately 17% of the interspecific interactions of herbivores were mediated by natural enemies 7 . In addition to influencing the intensity of interspecific competition, natural enemies can also alter the outcome of competition between two species and, depending on whether they promote coexistence or accelerate exclusion, will determine which of the herbivores suffers the most predation 7 . The effects of parasitoids on the competitive displacement that has occurred between the two agromyzid leafminers are still not thoroughly understood, although one braconid parasitoid, Dacnusa sibirica, has been reported to be more likely to control L. trifolii 24 .
The definition of a niche is complex; it includes the spatial niche, functional niche and n-dimensional hypervolume niche, etc. 25 . In this study, we conducted field cage experiments in an attempt to explore the niche (spatial niche throughout a year) differences between L. sativae and L. trifolii and to evaluate the asymmetrical effects of a pupal parasitoid (Opius biroi Fischer, which is an indigenous and prevalent parasitoid of leafminers in China) 26 on the two competitive, closely related and highly invasive leafminer species when they occurred together on kidney bean.

Results
Population dynamics and spatial distributions of L. trifolii, L. sativae and O. biroi. The populations dynamics of L. trifolii and L. sativae revealed similar tendencies from May to June. However, densities of L. trifolii were higher than those of L. sativae from 24 May to 5 June, 2013. Populations of O. biroi emerged on 5 June, and their densities first rose and then fell (Fig. 1).
Niche analysis of L. trifolii, L. sativae and O. biroi. We performed niche analysis based on the population surveying results for the two leafminer species and the braconid pupal parasitoid, O. biroi, on caged kidney beans.
Niche comparison between the two leafminer species and their parasitoid. The niche overlaps between L. sativae and O. biroi were obviously higher than that between L. trifolii and O. biroi (the temporal overlaps were 0.6051 and 0.1126, respectively, the spatial overlaps were 0.9886 and 0.9385, respectively, and the spatiotemporal overlaps were 0.5982 and 0.1057, respectively) (Tables 1-3).
The spatiotemporal similarities between L. trifolii and O. biroi were obviously higher than that between L. sativae and O. biroi (0.3936 and 0.0835, respectively). The temporal similarities of them were 0.4238 and 0.1112, respectively, while the spatial similarities of them were 0.9287 and 0.7512, respectively (Tables 1-3). Competition analysis among L. sativae, L. trifolii and O. biroi. The spatiotemporal competition coefficient between L. sativae and L. trifolii was 0.5871, which was obviously higher than the coefficients for intraspecific competition in L. sativae (0.3496) and in L. trifolii (0.3670) ( Table 4).
The spatiotemporal competition coefficient between L. sativae and O. biroi (0.4808) was higher than that between L. trifolii and O. biroi (0.0901). This was also true for the spatial competition coefficient (0.9872 and 0.8418, respectively) and the temporal competition coefficient (0.4870 and 0.1071, respectively) ( Table 4).

Discussion
The importance of the ecological niche and competition in modern ecology and population biology is reflected in the fact that niche theory and competition theory have often been closely associated 27,28 . L. trifolii and L. sativae are two closely related pest leafminers, when found occurring sympatrically, often exhibit serious competitive   Table 3. Temporal niche of the two leafminer species, Liriomyza sativae (LS) and L. trifolii (LT) and the pupal parasitoid, Opius biroi (OB).
interactions 14 . In the present study, a wider spatiotemporal niche breadth and lower parasitism rate were found in L. trifolii than in L. sativae.
Interspecific competition is an almost unavoidable consequence of niche overlap 27 . The niche overlap between L. sativae and L. trifolii on our experimental bean plants was ca. 0.6, which showed that relatively serious resource competition was occurring between the two leafminer species. Niche breadth is the most important metric in reflecting inter-and intraspecific interactions 29 . Based on our analysis, a wider spatiotemporal niche was found in L. trifolii than in L. sativae when reared on kidney beans, indicating that the former leafminer species had an advantage over the latter in exploiting the limited available resources. The wider spatial niche of L. trifolii indicated that this species had a greater ability to occupy spatial cells, while the larger temporal niche of L. sativae showed that it occurred more frequently during plant growth. In addition, a wider niche breadth of L. trifolii (0.5002) than L. sativae (0.3879) on field bean plants was also found in Sanya, Hainan, during 2012-2013, which provides support for this study in terms of evidences from the field (unpublished).
Natural enemies play an important role in interspecific interactions by influencing competitive interactions through changes in the population densities of competing species 7 . We found that O. biroi had higher niche overlap with the leafminer L. sativae than its closely related species, L. trifolii, when they were both reared on kidney bean, showing that O. biroi preferred to seek out and parasitize L. sativae larvae more than those of L. trifolii. If asymmetric effects occur in both of the competitors, this could indirectly affect the direction of competition. Therefore, the asymmetric parasitoid effect found in the present study should favour L. trifolii when the three species occurring sympatrically.
The coefficients of interspecific competition between L. sativae and L. trifolii were higher than those for intraspecific competition, which could potentially drive competitive interaction between the two species. The relative strengths of interspecific and intraspecific competition can determine the direction of species coexistence or exclusion 30 25 . Beyond these environmental factors, the inherent mechanisms of this displacement are still unclear. In this study, we explored the competitive interactions between two closely related species based on niche data; the results will help to enrich the theory of species displacement and can ultimately provide timely information that is useful for the management of future biological invasions.

Materials and Methods
Ethics statement. Our     Surveying Liriomyza spp. and their parasitoid. We began to survey the populations of Liriomyza spp.
and their parasitoids on 24 May. After ten healthy and uniform plants were selected for the survey (there were 5 replicates, and 2 plants within each replicate), the vertical space of each plant was divided into five grid cells: 0~30 cm, 30~60 cm, 60~90 cm, 90~120 cm and higher than 120 cm above the ground. We sampled 3 leaves in each cell for each survey (a total of 150 leaves were collected). The excised leaves were taken into the incubator, where identifications were conducted of the Liriomyza leafminer species and their parasitoids after their eclosion. The plants were surveyed every 4 days until the beans were harvested. No pesticides were used at any time during the survey period.
Calculation of niche parameters and data analysis. Niche breadth was calculated according to Levins (1968) 34 : where S represents the resource states (i.e., spatial or temporal cells are available in this study), and P i is the proportion of individuals of a species that is associated with S.
Niche overlap was also calculated from Levins (1968) 34 : where P ih and P jh are the proportions of the individuals of species i and j, respectively, that are associated with resource state h (i.e., one of the spatial or temporal cells in this study), and B i is the niche breadth of species i.
Niche proportional similarity was calculated according to Colwell & Futuyma (1971) 35 : where P ih and P jh are the same as above.
The competition coefficient was calculated according to May (1975) 36 : where P i is the same as in eq. (1), and P j is the proportion of individuals of species j that are associated with the resources state. The two-dimensional niche was calculated according to Cody (1974) 37 and May (1975) 36 :

TS T S
where B T and B S are the temporal and spatial niches of the species, respectively. All the niche parameters were analysed based on the population dynamics of the whole surveyed field [38][39][40] . Data analyses were conducted using IBM SPSS Statistics 19.0 (SPSS Inc., Chicago, IL, USA). We used one-way analysis of variances (ANOVAs) with Tukey's HSD test to analyse the data. All percentage data were arcsine square root transformed before being analysed.