Plastic response in amphibious plant functional traits along a latitudinal (climate) gradient

This dataset supports an investigation into the variation of vegetative and reproductive traits of six plant species characteristic of Mediterranean temporary ponds, Baldellia ranunculoides, Damasonium bourgaei, Damasonium polyspermum, Elatine brochonii, and Isoetes velata, across a 750-km north-south latitudinal (climate) gradient (31°-35° N) in Morocco (mean annual precipitation range 264 to 856 mm). Selected vegetative (height, number of leaves, leaf area) and reproductive (number of fruits, number of seeds/spores per fruit, number of seeds/spores per plant, seed/spore size and seed mass) traits of these six species were assessed on both field-collected specimens and on plants grown from seeds/spores in mesocosms under uniform experimental conditions. We hypothesized that the observed differences in functional traits (vegetative and reproductive) among populations mainly reflect phenotypic plasticity to variable environmental conditions. This work was possible thanks to the scientific permit (N° 241 HCEFLCD/DLCDPN/DPRN/CFF) granted by the Agence Nationale des Eaux et Forêts (Morocco).

Methods
Traits of plants from natural populations

For each species, 10 individuals per population were randomly collected at the end of their cycle (June–July 2014). Three vegetative traits (height, leaf number, and area) and five reproductive traits (fruit count, seeds per fruit, seeds per plant, seed size, and mass of 40 seeds) were measured for each individual. Here, "fruit" includes Isoetaceae sporangia, and "seeds" include spores.

Traits of plants grown in common garden 

A common garden experiment examined trait variation among populations of six species under uniform light, temperature, hydrology, and substrate. Plants grew from seeds (Spermatophytes) or sporangia (Lycopodiophytes) collected from 10 individuals per species per pond at the end of their cycle. Mean seed mass (n = 40) was measured before the experiment.

In transparent plastic pots (8 cm diameter, 12 cm height), a layer of synthetic cotton wool was placed at the bottom and then covered with 4 cm of temporary pond sediment that was first sterilized (120 °C during 24 hours). To increase the chance of having at least one germination, three seeds from the same population were planted in each pot which was flooded with 2 cm of tap water. For each species, 10 replicates per population were set up (270 pots total). Pots were randomly placed in a sunlit lab, watered daily (2 cm depth), and fully randomized weekly. Temperature was recorded weekly. Every three weeks, vegetative traits were measured on three individuals per pot, then averaged. After nine months, fruits per plant were counted. To assess other reproductive traits, four dried fruits per plant were analyzed for seed count, size, and mass (40 seeds).

Data analysis

Trait associations in natural populations were analyzed using a correlation matrix. Spearman correlations examined relationships between traits and pond latitude for both field-collected and laboratory-grown plants across all six species. A similar approach was used to assess the relationship between traits and rainfall, separately for individuals collected in the field and for plants cultivated in the laboratory.

To visualize trait variation in each of the six species across regions, principal component analyses (PCA) were conducted separately for field and cultivated plants. Spearman correlations assessed the relationship between population coordinates on PCA axis 1 and latitude. Functional trait differences between regions were tested using one-way ANOVA for each species.

To quantify the phenotypic plasticity of the plants investigated in this study, we used the Relative distance plasticity index (RDPI) which allows statistical comparisons of plasticity between species for each trait (Valladares et al., 2006). The function ‘rdpi’ of the R package ‘Plasticity’ was adapted for phenotypic plasticity computing on the basis of phenotypic distances among individuals of the same species under different environments.

RDPI ranges from 0 (no plasticity) to 1 (maximal plasticity) and is calculated as follow:

RDPI = ∑(d ij →i′j′/(x i′j′ + x ij ) )/n

where:

xij: is matrix of a specific trait x (i row and j column)

 i: is a given level of the environmental treatment (latitude)

 j: is to the individual number identification

 d ij→i′j′: absolute value of distances between all pairs of individuals from the same species under different environments

x i′j′ + x ij:  sum of trait values of all pairs of individuals from the same species under different environments

n:  is the total number of distances

Differences of RDPI values for each trait between species were evaluated with one-way ANOVA and post hoc Tukey mean comparison test using R software.

Species' investment preference (reproductive or vegetative) was analyzed using Spearman correlation between latitude and the seed mass/plant height ratio. These traits were chosen as the strongest indicators of reproductive and vegetative investment (Table 2).

Multivariate and univariate analyses were performed using STATISTICA 10 (StatSoft Inc., Tulsa OK, USA).