Introduction

The phenomenon of carnivory by plants has been recognised and studied since Charles Darwin [1], and is known to have evolved at least 6 times [2]. Its advantages to the plant are thought to involve the gain of significant amounts of nitrogen and phosphorus, explaining in part why the true carnivory is typically found in perennial plants of acid, nutrient-poor boggy soils [2], [3], [4]. However, intermediate states are known between normality and full carnivory, and most wild plants will increase their growth (hence, potentially, reproductive output) given the additional nitrogen and phosphorus from decaying animal remains. Here we report on evidence for reproductive benefits from carnivory in a plant showing none of the ecological or life history traits of standard carnivorous species.

The teasel, Dipsacus fullonum (Dipsacaceae) is a biennial herb, forming a rosette in its first year before growing quickly in the spring of its second year to a flowering height of between 0.5–2.5 m. [5]. Dipsacus has long been observed to catch insects and other invertebrates in their leaf basins which fill with rainwater, leading to speculation about it benefitting from carnivory [6], [7], [9], although textbooks on carnivorous plants have not considered Dipsacus [3], [8]. Dipsacus is unlike typical carnivorous plants in being associated with dry disturbed ground, often calcareous and nitrogen enriched - conditions lethal to true carnivorous plants [10]. It scores 7 on the Ellenberg scale for nitrogen [11], is biennial, and its size greatly exceeds most carnivorous plants. Here we report experimental tests of the evolutionary benefit of carnivory to Dipsacus, using total seedset as the best simple estimate of reproductive output of a biennial plant [12].

Methods

Results

Raw data are supplied in the supplementary excel file Table S1. Throughout the growing season, Dipsacus were observed to collect dead insects in leaf bases (mainly coleoptera, hemiptera, lepidoptera and diptera). Experimentally applied maggots took 2–4 weeks to disappear, with decayed remains of the previous maggot usually visible when a new one was added. Data on above-ground biomass and seed production are summarised in Table 1, with anova results in Table 2 (suppressing the higher order interaction terms). Post-harvest, the biggest source of variation in both the seed mass and biomass was between sites, with plants from site 2 having over five times the seed and biomass as plants from site 1 (p<0.001; Table 1). This had been evident from their smaller over-wintering rosettes with fewer leaves (15.9 sd 3.2 against 27.9 sd 9.7; p<0.01). Number of overwintering leaves correlated well with eventual biomass (r = 0.84, p<0.01) and seed production (r = 0.66 p<0.01). The proportion of final biomass that was seeds did not alter with plant size or site. There were no significant treatment effects on biomass, nor any of cutting drainage holes in each leaf basin. By contrast the effect of the supplementary insect feeding treatment on seed production, and the seed mass:biomass ratio, were both significant (both p<0.05), with highest values in maggot-supplemented plants (Table 1, Figure 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. A graph of seed mass against vegetative biomass for Dipsacus plants, showing feeding treatment.

https://doi.org/10.1371/journal.pone.0017935.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Effects of feeding treatments on biomass and seed production in teasels Dipsacus fullonum (mean+-sd ); data have been pooled over two populations and two water removal treatments, and (for biomass and seed mass) were log-transformed, averaged then back-transformed.

https://doi.org/10.1371/journal.pone.0017935.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. F values from Anova tests on biomass and seed production in teasel Dipsacus fullonum: significant results are in bold.

https://doi.org/10.1371/journal.pone.0017935.t002

Discussion

For a plant to be considered carnivorous, a key criterion is that experimental manipulations of its insect food supply can be shown to produce growth or developmental responses. Increases in size and biomass have been shown for Pinguicula, Drosera and Utricularia supplemented with appropriate animal bodies [10], [13]–[15], a result that was not duplicated here.

The dominant pattern in Dipsacus biomass was attributable to the natural variation between sites, initially manifest as differences in the sizes of over-wintering rosettes and probably explicable by unquantified differences in soil chemistry. The correlation between number of rosette leaves and final biomass was highly significant (r = 0.84, p<0.01), agreeing with the model that the plant's final size is largely determined by energy capture in the previous growing season [16]. Surprisingly, cutting each leaf base to prevent water buildup had no effect on any growth parameter. Insect nutrition had no detectable effect on biomass, either in absolute terms or as deviation from size predicted from its over-wintering rosette. However, the seed production and the seed mass:biomass ratio differed between insect-feeding treatments increasing as expected if insects were supplying mineral nutrition (Table 1). Similarly, Thum [15] listed increased seedset along with other indices of overall size when Drosera were fed supplemental flies (although based on a pseudo-replicated design). Wakefield [17] found that supplemental feeding of Sarracenia with flies did not increase size or biomass but did increase their nutrient content: in this case seed set was not quantified. Meyers [18] reported Utricularia's growth is reduced by half in the absence of prey, but again seedset was not quantified.

These data allow an estimate of nitrogen fluxes. The total animal biomass added over the season was 0.33 g dry mass per plant; assuming diptera have a mean nitrogen content at 10% [19] this equates to approximately 33 mg nitrogen as animal tissue added to a plant over the season. It has not been possible to find a published value for the nitrogen content of Dipsacus seeds, but Mattson [20] gives a range of 1–6% for seeds in general with lower values for non-legumes. Even assuming a low value of 15 mg/g nitrogen in seeds, the supplemental feeding supplied enough nitrogen for less than 2.5 g of seeds, while the regression lines predicted a difference between fed and control plants' seedset of approximately 7 g for a 300 g plant (Figure 1), implying that the apparent seed gain involved more nitrogen than was added in food. This may be from some other nutrient being limiting, or a statistical artefact; either way the result needs duplicating.

These results provide the first empirical evidence for Dipsacus displaying one of the principal criteria for carnivory given by Juniper et al (3); the use of products absorbed from prey to increase fitness. The result needs to be duplicated, and there remain other criteria of carnivory still to be demonstrated in Dipsacus; does it actively attract insects to its basins, how are insects digested / broken down, and are there any specialist structures such as waxy scales which cause insects to slip?

Supporting Information

Acknowledgments

Thanks are due to the Wimbledon Common rangers for permission to work on the common, Amanda Morgan for technical help, and anonymous referees for constructive feedback.