Functional traits of broad-leaved monocot herbs in the understory and forest edges of a Costa Rican rainforest

Relative species abundance

The understory flora of broad-leaved monocot herbs includes a mix of abundant, infrequent, and relatively rare species. Our qualitative classification ranked nine of our 21 study species as abundant, including at least one species in each of the six focal families (Table 1). These were Dieffenbachia longispatha, Philodendron grandipes, and Spathiphyllum fulvovirens (Araceae), Costus scaber (Costaceae), Asplundia uncinata (Cyclanthaceae), Heliconia irrasa (Heliconiaceae), Goeppertia cleistantha and G. micans (Marantaceae), and Renealmia cernua (Zingiberaceae).

We ranked eight species of forest edge and large gap habitats as abundant in our qualitative classification. These were Costus malortieanus, Carludovica sulcata, Cyclanthus bipartitus, Heliconia imbricata, H. latispatha, H. mathiasiae, Goeppertia maranthifolia, and Pleiostachya leucostachya (Table 1). Terrestrial species of Araceae were absent from these habitats with the exception of Spathiphyllum friedrichsthallii, an open bog specialist.

Morphological, architectural, and ecophysiological, life history traits

Leaf arrangement between cauline and basal leaves had a strong phylogenetic link to family in most cases. All of the Araceae, Cyclanthaceae included in our study had basal leaf arrangements, while all of the Costaceae and Zingiberaceae had cauline leaves. One exception to all basal leaves was present in Heliconia with H. mathiasiae having cauline leaves. Most of the Marantaceae had basal or subbasal leaves but exceptions were Goeppertia marantifolia and G. warscewiczii with cauline and Ischnosiphon inflatus with a branched growth form.

In stature understory herbs were relatively short with a mean height of 1.2 m, driven in part by phylogenetic relationships. Three species of Goeppertia, the smallest understory species, were at or below 0.5 m in mean height, while six of the eight species of Araceae were below 0.8 m in mean height (Table 1). Relatively robust and taller species were also present in the understory with Dieffenbachia longispatha, Asplundia sleeperae, Heliconia umbrophila, and Ischnosiphon inflatus all averaging at least 1.8 m in height. Forest edge and gap species were on average 2.9 m in height, more than twice as tall as understory species with a number of species of Heliconia reaching to or above 4 m in height. Only Costus malorieanus and the rheophyte specialist Dicranopygium wedelii averaged less than 1 m in height.

Leaf number on mature plants was highly variable for both understory and forest edge species, due in part to canopy architecture. Despite the large difference in stature, the mean number of leaves per plant did not differ significantly between understory and forest edge and gap species, with 20.5 and 21.5 leaves, respectively (Fig. 4). Leaf number varied from a low of 3.6 in Anthurium ochranthum in the understory and 5.4 for Costus malortieanus at the forest edge to a high of 79.7 cauline leaves in the understory Costus scaber and 148 for the forest edge Costus laevis. Mean leaf number, although low in variability, was significantly related to plant height (Fig. 5).

Mean area of individual mature leaves was 787 cm² for the understory species and three times greater at 2,517 cm² for the forest edge species. The large leaf size of Heliconia and Calathea species heavily influenced to a high mean leaf area for this latter group. Again, however, the range of values for each group were highly variable. For forest understory species the sizes ranged from only 11 cm² in Goeppertia micans to 1,968 cm² in Dieffenbachia longispatha. The range for forest edge species was 264 cm² for Costus malortieanus to 8,054 cm² in Heliconia latispatha.

Total leaf area of individual plants paralleled the data for individual leaf size and was 3.5 times higher in the forest edge species but again with highly variable ranges of values. Mean total plant leaf area 1.23 m² for understory herbs, ranging from only 0.01 m² in the small Goeppertia micans to 4.72 m² in Asplundia sleeperae. For forest edge species the mean plant leaf area was 4.25 m², ranging from 0.14 m² in Costus malortieanus to 12.31 m² in Calathea lutea. Leaf canopy area was significantly related to plant height (Fig. 5).

Specific leaf area did not differ significantly between the two groups (Fig. 4). Understory species had a mean of 231 cm²g⁻¹, ranging from 157 cm²g⁻¹ in the thick leaves of Dieffenbachia longispatha to 413 cm²g⁻¹ in Goeppertia micans. The mean SLA for forest edge taxa was 197 cm²g⁻¹, with a low of 93 cm²g⁻¹ in Costus laevis and Heliconia wagneriana and a high of 285 cm²g⁻¹ in Calathea similis.

Despite the significant differences in leaf size (Fig. 4), leaf water content was not significantly different between the two groups, with individual species expressing a wide range of values. Understory species had a mean water content of 5.76 g g⁻¹ with a range from 3.73 g g⁻¹ in Ischnosiphon inflatus to 10.43 g g⁻¹ in Goeppertia leaucostachys. Mean water content was 4.70 g g⁻¹ for forest edge species with values from 1.84 g g⁻¹ in Heliconia wagneriana to 12.83 g g⁻¹ in Costus malortieanus.

Leaf nitrogen content also did not differ significantly between the two groups of species (Fig. 4) but showed some strong phylogenetic patterns. Understory species had a mean nitrogen content of 2.57%, with a low species mean of 2.31% in Asplundia uncinata and a high of 4.65% in Spathiphyllum fulvovirens. Mean values for forest edge species were similar in breadth from a low of 2,24% in Calathea similis to 4.44% in Spathiphyllum friedrichsthallii. Taken overall, Araceae tended to have higher than average nitrogen content while Marantaceae in both habitats tended to be lower than overall means.

Despite the relatively high leaf nitrogen contents of understory herb species, light-saturated rates of net photosynthetic assimilation were quite low. These ranged from 1.60 µmol m⁻² s⁻¹ in Costus scaber to 3.18 in Heliconia irrasa, with a mean of 2.39 µmol m⁻² s⁻¹ for the nine species measured. In contrast, rates for forest edge species ranged from maximum 4.35 µmol m⁻² s⁻¹ in Calathea lasiostachya, a gap species, to 14.09 µmol m⁻² s⁻¹ in Heliconia latispatha with a mean of 8.1 µmol m⁻² s⁻¹.

Mean values of stable carbon isotope ratio (δ¹³C) differed significantly between the understory and forest edge species reflecting their exposure to sun and relative humidity levels. The mean for understory species was -36.0 o/oo, with a range from -37.3 o/oo in Spathiphyllum fulvovirens and Goeppertia micans to -32.7 o/oo in Dieffenbachia tanduzii. The mean δ¹³C for forest edge species was -33.3 o/oo, ranging from -28.5 o/oo in Heliconia wagneriana to -36.4 o/oo in Carludovica sulcata.

Ecology versus phylogeny

We find that the morphological and life history traits of the 40 species in this study reflect a combination of phylogenetic canalization and ecological selection for specific habitats. For example, most forest understory species are low-growing and basal in their arrangement of leaves, suggesting that this might be a habitat-specific trait. Exceptions to this pattern were present with phylogenetically clustered taxa, occurring with species of Costus and Renealmia. Both genera uniformly exhibited family traits of multiple smaller cauline leaves. These species also lie at the upper end of height for understory herbs (Table 1).

Under higher light conditions at forest edges and in large gaps, environmental selection is strong. Herb species characteristic of these habitats were tall and generally attained their heights by adopting a cauline or at least sub-basal form of canopy architecture. This is well illustrated in the cauline growth forms present in four out of seven species of forest edge and gap species of Calathea s.l., in contrast to the basal leaves of most understory species (Table 1). The mean height of the two understory species of Heliconia was 1.1 and 1.8 m, while all but one of the five high light species have mean heights above 3.3 m. The only relatively diminutive species in the latter group of species was H. sarapiquensis with a mean height of 2.2 m. This is a relatively rare species at La Selva and seems to grow in small gaps. Regardless of light environment, cauline leaves are generally much smaller than basal leaves in all of the broad-leaved monocot studied. These leaves had mean leaf areas of 80–537 cm² in the nine species measured (Table 1).

It is notable that leaf water content and specific leaf area were highly variable in the study species, exhibiting neither environmental nor phylogenetic patterns (Fig. 6). There were, however, elements of phylogenetic patterns in leaf nitrogen contents. All of the understory species of Araceae had leaf nitrogen contents above the average for all understory species (Table 2). There are a variety of ways in which leaf nitrogen concentration may indicate ecophysiological strategies related to relative nitrogen partitioning, the electron transport capacity per unit of chlorophyll, and/or the specific activity of RuBP carboxylase (Evans, 1989).

The photosynthetic capacity of understory herbs was low, despite high leaf nitrogen contents, with limited exposure to high levels of solar irradiance. Neither leaf nitrogen content nor specific leaf area showed any significant correlation with photosynthetic assimilation nor is a strong phylogenetic signal present (Fig. 6). Survival in shaded environments does not require maximizing photosynthetic capacity for a given leaf nitrogen content. Species that are well adapted to understory environments are known to partition relatively larger amounts of nitrogen into the thylakoids compared to sun species, a trait which produces lower photosynthetic capacity per unit of nitrogen (Evans, 1989). However, leaf nitrogen content and maximal rates of net photosynthetic assimilation among species do not scale well with photosynthetic capacity or leaf nitrogen content per unit leaf area. Photosynthetic rates, through their effects on growth, can strongly influence the population dynamics of plants in low and variable light environments, but the magnitude of this effect varies between species. In species in which fitness is independent of maximum rates of net photosynthesis there would be little opportunity for selection on photosynthetic rates (Westerband & Horvitz, 2017).

Although all of the species in this study exhibited at least some degree of vegetative reproduction, clonality is strongly developed in a number of species. This evolution of active vegetative reproduction does not show strong phylogenetic patterns, however, but rather a general but inconsistent relationship to habitat. For example, all of the forest edge species of Heliconia are strongly clonal, as is the understory species H. irrasa. However, the other understory Heliconia, H. umbrophila, is only weakly clonal. Costus scaber in the forest understory does not have active vegetative reproduction, while C. malortieanus and C. laevis growing along forest edges readily reproduce vegetatively from underground rhizomes.

Adaptive tradeoffs

Why do understory herbs not grow taller, hypothetically exposing their leaves to a higher light environment? Clearly, there are “tradeoff” strategies to optimize fitness of these species. A competitive advantage accrues to herbs in open communities that grow taller and shade their close neighbors and capturing more light (Givnish, 1982). There is a selective tradeoff, however, as increased height requires reinforced structural tissues and thus a tradeoff in carbon allocation between height and support. In the low light environment of forest understories, there appears to be little selection to grow taller than heights of 0.5–2 m, perhaps because the added carbon gain associated with height beyond this level does not justify the added cost of structural tissues. Except in local areas where understory palms occur densely, understory herbs do not seem to strongly compete with light with their herb neighbors. In high light environments, however, there is a competitive advantage in greater height growth to overtop competitors (Valladares & Niinemets, 2008).

Another tradeoff exists in carbon allocation between structural tissues and photosynthetic tissues. This allocation is particularly critical in low-light understory environments where the benefit of a greater photosynthetic surface area requires carbon investment in structures to support that amount of biomass (Cooley, Reich & Rundel, 2004). For tropical understory palms at La Selva, support tissues have been shown to increase plant resilience to damage from falling debris, indicating tradeoffs between photosynthetic surface area and reinforcement of support tissues have been documented in Chazdon (1986a). Falling debris is a major cause of mortality in tropical understory plants including tree seedlings (Clark & Clark, 1989) and understory herbs (Gartner, 1989; Sharpe, 1993). Despite the ecological significance of falling debris, it has been shown that understory herbs do not exhibit the same degree of biomechanical strength in their petioles and stems as understory palms (Cooley, Reich & Rundel, 2004).

The biomechanical properties of Asplundia uncinata are intriguing in contrast to those of understory palms as this species is morphologically similar enough to the palms Asterogyne martiana and Geonoma cuneata to confuse a casual observer. Like these two understory palms, Asplundia uncinata occurs in large clonal clusters in deep shade, often excluding all other ground plants. Understory palm taxa form multi-ramet clones connected by thick aboveground rhizomes. Individual ramets exhibit slow and steady growth, attaining similar heights and displaying similar numbers of large bifid leaves. Despite obvious morphological similarities, Asplundia uncinata differs from the palms in the flexibility and strength of its petioles, with biomechanical traits that are like those of other understory herbs (Cooley, Reich & Rundel, 2004). These trait differences may reflect important growth strategies that enable the continued coexistence of morphologically similar species.

While the relative allocation of carbon to structural and photosynthetic tissues has obvious significance, there are clearly multiple successful strategies of arraying the architecture of photosynthetic surface area. Valladares, Skillman & Pearcy (2002) demonstrated this very clearly in a study of the efficiencies of light capture among 24 species of understory tree seedlings, low shrubs, and herbs representing a diversity of families, growth forms, and leaf attributes. Broad-leaved monocot herbs and other understory species showed convergence on a narrow range of light absorption efficiencies with ratios of absorbed light to available light of 0.60–0.75. Other studies have shown than that comparative growth by understory herbs and shrubs can be accomplished through alternative character syndromes, and leaf longevity may not be correlated with photosynthetic capacity in shade adapted plants related to leaf nitrogen content, SLA, and leaf longevity (Mulkey, Smith & Wright, 1991; Mulkey, Wright & Smith, 1992).

Do functional traits predict species abundance?

Some species of broad-leaved terrestrial monocot herbs occur in great abundance while others are relatively rare. Can functional traits help explain their relative differences in abundance? Of the functional traits included in this study, none were predictive of species abundance. One possibility, of course, is that there are other critical functional attributes of ecophysiological limiting factors or reproductive traits that we were not measuring (Horwitz, 1993; Westerband & Horvitz, 2015; Tokarz et al., 2019). The suggestion has been made that there may exist a pool of functionally equivalent associated species present in a community confer resilience in ecosystem function (Walker, Kinzig & Langridge, 1999).