Competitive Interactions of Flowering Rush (Butomus umbellatus L.) Cytotypes in Submersed and Emergent Experimental Aquatic Plant Communities

The ability to invade communities in a variety of habitats (e.g., along a depth gradient) may facilitate establishment and spread of invasive plants, but how multiple lineages of a species perform under varying conditions is understudied. A series of greenhouse common garden experiments were conducted in which six diploid and four triploid populations of the aquatic invasive plant Butomus umbellatus L. (Butomaceae) were grown in submersed or emergent conditions, in monoculture or in a multispecies community, to compare establishment and productivity of cytotypes under competition. Diploid biomass overall was 12 times higher than triploids in the submersed experiment and three times higher in the emergent experiment. Diploid shoot:root ratio was double that of triploid plants in submersed conditions overall, and double in emergent conditions in monoculture. Relative interaction intensities (RII) indicated that triploid plants were sixteen times more negatively impacted by competition under submersed conditions but diploid plants were twice as impacted under emergent conditions. Recipient communities were similarly negatively impacted by B. umbellatus cytotypes. This study supports the idea that diploid and triploid B. umbellatus plants are equally capable of invading emergent communities, but that diploid plants may be better adapted for invading in submersed habitats. However, consistently lower shoot:root ratios in both monoculture and in communities suggests that triploid plants may be better-adapted competitors in the long term due to increased resource allocation to roots. This represents the first examination into the role of cytotype and habitat on competitive interactions of B. umbellatus.

The characteristics of introduced species and recipient communities that contribute to establishment and spread of invaders has been an active area of research for a few decades, with considerable effort put towards an understanding of the role of biotic resistance (i.e., negative Experiments were conducted in temperature-controlled greenhouses at the ERDC. Prior to experiments, neighbor species were collected from cultures, their rhizomes (emergent species) or apical meristems (submersed species) were harvested from culture pots and floated in water for several days, then planted and allowed to root for one week. Propagule size was standardized within species. For emergent species, rhizome pieces were approximately 5-10 cm long and for submersed species, 10-15 cm apical stem cuttings were used. One week after planting the experimental communities, B. umbellatus propagules were added to the containers. B. umbellatus plants were grown from rhizome fragments (4-6 cm; triploid plants) or bulbils (diploid plants) to a similar size, and weighed before planting. The difference in planting times (one week) between neighbor species and B. umbellatus was enough to promote establishment of experimental community species, particularly in the warm greenhouses. For the emergent experiment, plants were potted in 6 L nursery containers with 4 L commercial topsoil amended with a 10 g fertilizer pellet (20-10-5; Scotts Agriform™, Marysville, OH, USA). Pots were placed individually within 20 L plastic buckets, then placed in 1200 L fiberglass tanks (interior dimensions: 1.5 m × 0.94 m × 0.92 m) within greenhouses. For the submersed experiment, Diversity 2020, 12, 40 4 of 15 plants were potted in 1 L plastic cups, in approximately 500 mL commercial topsoil, amended with a 5 g fertilizer pellet (20-10-5; A.M. Leonard, Piqua, OH, USA). A 2 cm cap of masonry sand was added over the topsoil to prevent resuspension of soil and loss of nutrients into the water column. Pots were placed individually in 48 L aerated aquaria in greenhouse tanks. Prior to placing pots in aquaria, a one-time dose of nutrient solution [66] was added to each aquarium and allowed to equilibrate for several days. Aquaria were aerated prior to introducing plants and aeration was maintained for the duration of the experiment. Diatom filters were used periodically to maintain clear water and remove algae in aquaria. Air temperature was 28.3 ± 3.1 • C (Mean ± SE) in the emergent study (measured in the center of the greenhouse) and water temperature in the submersed study (averaged between three tanks, measured 1 m deep) was 21.0 ± 2.8 • C over the course of the experiments. Experimental design for both experiments was a partial incomplete block design (Cochran and Cox 1968), following Bose, et al. [67], with 21 treatment combinations, 6 replicates per treatment, and 18 blocks. Experimental units (individual pots) were arranged with seven per tank (block), 18 tanks overall.
Plant biomass was harvested after 12 (emergent) and 16 (submersed) weeks. Harvests for both studies were conducted identically. Aboveground (shoots), belowground (roots), and reproductive tissue (bulbils or rhizome buds) biomass was separated for each species and washed of debris, then placed individually in paper bags, dried at 70 • C for one week, then weighed to the nearest 0.01 g. Biomass for all species (i.e., B. umbellatus and neighbors) was measured separately.

Statistical Approach
The same statistical approach was used to analyze data from both (emergent, submersed) experiments. First, it was determined whether biomass variables (total biomass, reproductive tissue biomass, shoot:root ratio) varied between B. umbellatus cytotypes and between growth in monoculture or in a multispecies community. For this, we used mixed effects models with neighbor presence (two levels: monoculture, multi-species) and cytotype (two levels: diploid, triploid) as fixed effects. Additionally, population nested within cytotype was included as a random effect. Initially, tank was included as a block effect (i.e., location within the greenhouses) in all models but was highly insignificant (p > 0.6), so was ultimately removed from most analyses. Tank was retained as a random variable in RII analyses only (see below). Additionally, initial propagule weight was included in all models as a covariate. Submersed and emergent experiments were analyzed separately.
To quantify the effect of neighbor presence (a putative competitive interaction) on flowering rush and neighbor biomass, relative interaction intensities (RII) were calculated [68]. RII has a value between −1 and +1, with negative values indicating competition and positive values indicating facilitation. RII is calculated as: where B w is the mass of a plant grown in the presence of other species and B o is the mean mass of plants of a single species grown alone [69]. Differences in RII between cytotypes were determined with a mixed effects model, where cytotype was a fixed effect and population nested within cytotype and tank (block) were random effects. RII was calculated both for flowering rush plants grown in recipient communities and also for the recipient community (as summed biomass of neighbor species). All statistical tests were performed with SAS ver. 9.4.

Results
Establishment and growth of Butomus umbellatus in experimental communities largely differed by cytotype. Regardless of water depth, diploid plants outperformed triploid plants in biomass production (both overall and reproductive biomass). However, the relative effect of competition on plant growth varied by cytotype and by experiment.

Submersed Interaction Experiment
In the submersed interaction experiment, we detected no significant effect of competition (i.e., presence of neighbors) on biomass (total or reproductive) variables (Table 2, Figure 1). However, diploid plants produced sixty times more total biomass than triploid plants across treatments. Likewise, reproductive biomass in B. umbellatus was not negatively affected by neighbors, and the cytotype effect was marginally significant, with triploid plants not producing any rhizome buds during the experiment. Shoot:root ratio differed both by neighbor presence and cytotype identity; overall, shoot:root ratio for diploid plants was seven times greater than triploid plants (diploid: 2.65 ± 0.51, triploid: 0.43 ± 0.62).
In contrast to the mixed model results above, there was a significant effect of competition on triploid but not diploid plants, detected based on RII calculations ( Figure 2A). Additionally, RII for triploid (RII = −0.64 ± 0.11) plants was thirteen times stronger than for diploid (RII = −0.048 ± 0.09) plants in the submersed experiment. B. umbellatus did not have a significant effect on recipient community RII, regardless of cytotype ( Figure 2A).

Submersed Interaction Experiment
In the submersed interaction experiment, we detected no significant effect of competition (i.e., presence of neighbors) on biomass (total or reproductive) variables (Table 2, Figure 1). However, diploid plants produced sixty times more total biomass than triploid plants across treatments. Likewise, reproductive biomass in B. umbellatus was not negatively affected by neighbors, and the cytotype effect was marginally significant, with triploid plants not producing any rhizome buds during the experiment. Shoot:root ratio differed both by neighbor presence and cytotype identity; overall, shoot:root ratio for diploid plants was seven times greater than triploid plants (diploid: 2.65 ± 0.51, triploid: 0.43 ± 0.62).
In contrast to the mixed model results above, there was a significant effect of competition on triploid but not diploid plants, detected based on RII calculations ( Figure 2A). Additionally, RII for triploid (RII = −0.64 ± 0.11) plants was thirteen times stronger than for diploid (RII = −0.048 ± 0.09) plants in the submersed experiment. B. umbellatus did not have a significant effect on recipient community RII, regardless of cytotype ( Figure 2A).

Emergent Interaction Experiment
Under emergent conditions, diploid plants produced five times more total biomass than triploid plants overall but the impact of neighbor-presence on biomass differed between cytotypes (i.e., a significant neighbor presence × cytotype interaction; Figure 3; Table 2). In monoculture and with neighbors, diploid plants produced four and nearly seven times more total biomass, respectively. Diploid plants produced twenty times more reproductive biomass when grown alone and nearly three times as much under competition. Bulbil production in diploid plants was four times higher in monoculture than under competition (9.2 ± 0.88 g vs. 2.3 ± 0.89 g). A significant neighbor presence x cytotype interaction was detected for shoot:root ratio (Table 2), indicating differential effects of competition on shoot:root ratio for diploid and triploid B. umbellatus plants. Shoot:root ratio was higher in diploid plants in the monoculture treatment (1.16 ± 0.08 vs. 0.65 ± 0.1), whereas shoot:root ratio was similar between cytotypes when grown under competition (0.14 ± 0.1 vs. 0.09 ± 0.08).
The competitive effect in the emergent experiment, calculated as RII, was significant for both B. umbellatus and the recipient community ( Figure 2B). Although mean relative interaction intensity (RII) in diploid plants was nearly twice that calculated for triploid plants (−0.67 vs. −0.36), the cytotype difference was marginally insignificant in the emergent experiment (p = 0.07). Although not as strong as the effect on B. umbellatus from the recipient community, B. umbellatus did exert a significant negative competitive effect (i.e., confidence intervals do not overlap zero) on the recipient community, regardless of cytotype ( Figure 2B). between cytotype means (i.e., their 95% CI do not overlap) are indicated by an asterisk. If confidence intervals of a mean do not overlap zero (e.g., neighbor means in panel B), the interpretation is that of a significant competitive effect for that treatment. For example, in panel (A), triploid and diploid RII means are significantly different to each other, but diploid RII is not significantly different to zero (i.e., no effect).

Emergent Interaction Experiment
Under emergent conditions, diploid plants produced five times more total biomass than triploid plants overall but the impact of neighbor-presence on biomass differed between cytotypes (i.e., a significant neighbor presence × cytotype interaction; Figure 3; Table 2). In monoculture and with neighbors, diploid plants produced four and nearly seven times more total biomass, respectively. Diploid plants produced twenty times more reproductive biomass when grown alone and nearly three times as much under competition. Bulbil production in diploid plants was four times higher in monoculture than under competition (9.2 ± 0.88 g vs. 2.3 ± 0.89 g). A significant neighbor presence x cytotype interaction was detected for shoot:root ratio (Table 2), indicating differential effects of competition on shoot:root ratio for diploid and triploid B. umbellatus plants. Shoot:root ratio was higher in diploid plants in the monoculture treatment (1.16 ± 0.08 vs. 0.65 ± 0.1), whereas shoot:root ratio was similar between cytotypes when grown under competition (0.14 ± 0.1 vs. 0.09 ± 0.08).
The competitive effect in the emergent experiment, calculated as RII, was significant for both B. umbellatus and the recipient community ( Figure 2B). Although mean relative interaction intensity (RII) in diploid plants was nearly twice that calculated for triploid plants (−0.67 vs. −0.36), the cytotype difference was marginally insignificant in the emergent experiment (p = 0.07). Although not as strong as the effect on B. umbellatus from the recipient community, B. umbellatus did exert a significant negative competitive effect (i.e., confidence intervals do not overlap zero) on the recipient community, regardless of cytotype ( Figure 2B).

Discussion
Butomus umbellatus establishment and spread in North American aquatic habitats has generated negative ecological and economic impacts, leading to increased interest in predicting invasiveness, potential distribution (Banerjee et al., in prep), and development of effective management tools for the species [59,70,71]. However, research on B. umbellatus management has largely failed to take into account genetic variation and the implications of differences in ploidy in the invaded range (but see  [72][73][74][75][76][77][78][79][80]. However, it is rare that genetic identity of an invader is paired with plasticity to environmental variables and interspecific competition in order to assess invasive potential. The results from the current research demonstrate that genetic variation and habitat heterogeneity can interact to influence plant invasion (e.g., biomass production) in experimental communities.
Similar to previous work (Harms et al. in review), diploid B. umbellatus plants outperformed triploid plants under most experimental conditions. Diploid plants, when grown alone, produced nearly 10 and 4 times more biomass in submersed and emergent experiments, respectively. However, the effect of competition largely negated biomass differences between cytotypes, at least in the emergent experiment. Diploid plants produced 17 times more biomass than triploid plants under competition in submersed conditions but only 1.3 times more biomass in the emergent competition treatment. The disparity between performance in monoculture and under competition for diploid plants signals that superior competitive ability, at least in emergent habitats, is likely not driving invasion success of diploid B. umbellatus in North America. This is also corroborated by RII calculations; diploid plants, despite higher biomass production overall, are twice as impacted by competition than triploid plants in the emergent experiment (−0.67 vs. −0.36.), whereas RIIs in the submersed experiment indicate far more impact of competition to triploid than diploid plants (−0.64 vs. −0.048). In fact, the effect of competition on diploid plants in the submersed experiment were essentially zero (Figure 2). Implications of these findings (diploid superiority in monoculture and in submersed environments) are that (1) diploid plants may be well suited for highly disturbed habitats with few other species, (2) diploid plants may be better at invading submersed communities, and that (3) both cytotypes are equally capable of establishing in emergent communities.
One major phenotypic difference between cytotypes in the current study was in shoot:root ratio and its variation in response to treatment conditions. Disproportionate biomass allocation to

Discussion
Butomus umbellatus establishment and spread in North American aquatic habitats has generated negative ecological and economic impacts, leading to increased interest in predicting invasiveness, potential distribution (Banerjee et al., in prep), and development of effective management tools for the species [59,70,71]. However, research on B. umbellatus management has largely failed to take into account genetic variation and the implications of differences in ploidy in the invaded range  [72][73][74][75][76][77][78][79][80]. However, it is rare that genetic identity of an invader is paired with plasticity to environmental variables and interspecific competition in order to assess invasive potential. The results from the current research demonstrate that genetic variation and habitat heterogeneity can interact to influence plant invasion (e.g., biomass production) in experimental communities.
Similar to previous work (Harms et al. in review), diploid B. umbellatus plants outperformed triploid plants under most experimental conditions. Diploid plants, when grown alone, produced nearly 10 and 4 times more biomass in submersed and emergent experiments, respectively. However, the effect of competition largely negated biomass differences between cytotypes, at least in the emergent experiment. Diploid plants produced 17 times more biomass than triploid plants under competition in submersed conditions but only 1.3 times more biomass in the emergent competition treatment. The disparity between performance in monoculture and under competition for diploid plants signals that superior competitive ability, at least in emergent habitats, is likely not driving invasion success of diploid B. umbellatus in North America. This is also corroborated by RII calculations; diploid plants, despite higher biomass production overall, are twice as impacted by competition than triploid plants in the emergent experiment (−0.67 vs. −0.36.), whereas RIIs in the submersed experiment indicate far more impact of competition to triploid than diploid plants (−0.64 vs. −0.048). In fact, the effect of competition on diploid plants in the submersed experiment were essentially zero (Figure 2). Implications of these findings (diploid superiority in monoculture and in submersed environments) are that (1) diploid plants may be well suited for highly disturbed habitats with few other species, (2) diploid plants may be better at invading submersed communities, and that (3) both cytotypes are equally capable of establishing in emergent communities. One major phenotypic difference between cytotypes in the current study was in shoot:root ratio and its variation in response to treatment conditions. Disproportionate biomass allocation to underground tissues has been implicated as a trait common to successful competitors [81][82][83][84][85], and B. umbellatus, in particular [86]. In a recent study [86], it was demonstrated that shoot:root ratio varied between introduced B. umbellatus cytotypes and was further influenced by nutrients [72], but the pattern was somewhat counterintuitive: plants increased root biomass allocation with increased availability of nitrogen, and triploid plants had consistently lower shoot:root ratios regardless of nutrient treatment. In the current study, B. umbellatus plants, regardless of cytotype, generally had three times larger shoot:root ratios in the submersed versus the emergent experiment. Furthermore, shoot:root ratios were higher in monoculture in both submersed and emergent experiments regardless of cytotype, indicating increased allocation towards vegetative structures. In contrast, shoot:root ratios were lower across the board in competition treatments, suggesting plasticity to competition in the form of biomass allocation to roots. Differences between cytotypes in biomass allocation from aboveground to belowground tissues may be critical to long-term persistence in a variety of habitats. Ren et al. (2019) found that increased allocation to roots in response to increased nitrogen was a successful strategy for Solidago canadensis L. (Asteraceae), whereas a native focal species did not respond similarly. The implications of their work are that S. candensis is likely to perform better in low resource environments, especially under elevated nutrient deposition. The different allocation patterns between B. umbellatus cytotypes in this and other studies raises interesting questions about the types of habitats that may be invadible by each and whether they are likely to invade similar habitats. Diploid plants tended to have larger shoot:root ratios which may benefit them in environments in which competition for light is high. In fact, the much larger shoot:root ratios of diploid plants in the submersed experiment, and reduced impact (relative to the emergent experiment) on shoot:root ratios due to competition with the canopy forming species M. spicatum and H. dubia, may further reflect adaptation to that environment.
Although the main objective of this work was to compare cytotype (i.e., between-cytotype) mean biomass variables grown alone or with neighbors, within cytotype (i.e., between-population) variation displayed an interesting pattern. Diploid populations were overall more variable when grown in monoculture, but not in competition, than triploid populations. However, this pattern was evident only in the submersed experiment (submersed experiment, coefficient of variation for total biomass in monoculture: diploid = 69.5, triploid = 12.0; emergent experiment, coefficient of variation for total biomass in monoculture: diploid = 30.4, triploid = 39.7). This may have been at least in part influenced by an increased genetic diversity among diploid populations (i.e., three unique genotypes) relative to triploid populations (one genotype). For instance, the Springbrook pond, IL (SB) population used here is diploid but it is a unique multilocus genotype (G3; Gaskin, unpublished data), apart from the common diploid genotype G4 (Table 1). In a previous study to examine differences in nutrient response between cytotypes, this diploid G3 population responded more similarly to triploid than diploid plants in most measured phenotypic responses, including biomass allocation and tissue chemistry [86]. In the current study, SB plant biomass consistently mirrored triploid rather than diploid plants, producing the least biomass of all diploid populations and also lowest shoot:root ratios when grown in monoculture (e.g., Figure 1). Although the goal of this study was not to compare populations, but rather cytotypes, this result demonstrates the value of presenting multiple genetic levels (i.e., population-level and cytotype-level) of results.
Several limitations to this work are worth discussing. First, competition experiments took place under relatively constant greenhouse conditions, whereas real-world interactions would certainly occur in a fluctuating environment with myriad contrasting and interacting biotic and abiotic variables. Furthermore, plant responses to changing conditions may vary in their magnitude or duration [87]. Future experiments might be conducted under realistic field conditions or at least in outdoor mesocosms. In addition, the location of the experiments was considerably south of the most southern natural population of B. umbellatus in North America [88]. However, most of the southeastern states have suitable climates for B. umbellatus establishment (Banerjee et al., in review) and experiments were conducted early in the year (February through May) to more closely reflect northern locations during summer months. Whether the location of this experiment (Mississippi) had any bearing on the results of competition in different water depths is unknown, but future experiments could be conducted across climate or other geographical gradients to further explore how invasion may succeed in different areas.
As demonstrated here and elsewhere, genetic variability in invasive species may be sufficient to generate patterns where some areas are more prone to invasion than others. In North America, B. umbellatus is represented by several genotypes within two cytotypes which vary in a number of important traits, including pathogen susceptibility [59], nutrient response [86], and competitive ability (this study). Future research into ecosystem impacts or to develop effective management strategies for B. umbellatus should take this variability into account. At the very least, future research reporting results should clearly state which genotype was used/observed in the study. If genotype determination is not possible, experimental plant vouchers should be retained for future analyses. Because the genetic variation present in populations of B. umbellatus can manifest to reflect important ecological variation (i.e., competitive ability, nutrient response), it will be critical to consider in future management activities.  (Table A1). Sites were selected through discussion with state or local personnel, online database searches (e.g., EDDMapS.org) or random encounters during transit. At each site, a number of site characteristics were recorded. Of primary interest for the current research is the variable water depth. Water depth was recorded in the middle of the infestation, the extent of which was determined through visual searching at the surface and upper portion (0.5m) of the water column. Ploidy was subsequently determined for populations (John Gaskin, USDA ARS; unpublished data). Least square means of water depth for cytotypes are shown below ( Figure A1). Mean water depths were determined by using a linear mixed model with population nested in cytotype and year as random variables and cytotype as a fixed variable.