Striking between-population floral divergences in a habitat specialized plant

Abstract

When the habitat occupied by a specialist species is patchily distributed, limited gene flow between the fragmented populations may allow population differentiation and eventual speciation. 'Sky islands'-montane habitats that form terrestrial islands-have been shown to promote diversification in many taxa through this mechanism. We investigate floral variation in Impatiens lawii, a plant specialized on laterite rich rocky plateaus that form sky islands in the northern Western Ghats mountains of India. We focus on three plateaus separated from each other by ca. 7 to 17 km, and show that floral traits have diverged strongly between these populations. In contrast, floral traits have not diverged in the congeneric I. oppositifolia, which co-occurs with I. lawii in the plateaus, but is a habitat generalist that is also found in the intervening valleys. We conducted common garden experiments to test whether the differences in I. lawii are due to genetic differentiation or phenotypic plasticity. There were strong differences in floral morphology between experimental plants sourced from the three populations, and the relative divergences between population pairs mirrored that seen in the wild, indicating that the populations are genetically differentiated. Common garden experiments confirmed that there was no differentiation in I. oppositifolia. Field floral visitation surveys indicated that the observed differences in floral traits have consequences for I. lawii populations, by reducing the number of visitors and changing the relative abundance of different floral visitor groups. Our results highlight the role of habitat specialization in diversification, and corroborates the importance of sky islands as centres of diversification.

Fig 1. Morphological characters studied in Impatiens lawii flowers.

(A) Side view of the whole flower and (B) Front view of the flower, depicting the three-dimensional characters measured. (C) Quantification of dissected floral parts. Refer to Table 1 for the details of character abbreviations.

Fig 2. Flower size represented by flower width (FW) across the three wild populations from the three plateaus.

(A) I. lawii; and, (B) I. oppositifolia. Boxplots with the median and interquartile range for flower width are presented for 30 individuals from each species and plateau. An asterisk indicates a significant difference (p<0.05) in flower width between plateaus.

Fig 3. Principal component analysis of floral morphology.

Biplots are shown for the first two principal component (PC) axes based on the 15 floral traits examined in Impatiens lawii (A, C, E) and Impatiens oppositifolia (B, D, F), for 30 individuals each from the three plateaus (depicted by the different colours and shapes as shown in the legend). Ellipses indicate 95% confidence intervals. The details of the abbreviations used for floral traits are presented in Table 1.

Fig 4. Floral morphological variation represented by the composite estimates of the first three principal component axes (PC1 + PC2 + PC3) for the floral traits examined in Impatiens lawii (A, C, E) and Impatiens oppositifolia (B, D, F), for individuals from the three plateaus.

Floral traits of plants from the wild population (A, B) are contrasted with results from plants grown in common garden experiments where plants were propagated from seeds (C, D) or from early stage transplants (E, F). Boxplots with the median and interquartile range for flower width are presented for 30 individuals from each species and plateau. An asterisk indicates a significant difference (p<0.05) between plateaus based on Tukey’s post hoc tests.

Fig 5. Floral visitors to Impatiens lawii and I. oppositifolia in Kaas and Chalkewadi plateaus.

(A) The number of floral visitors belonging to the different putative pollinator groups. The number of floral visitors was normalized by the number of flowers observed. Lower case letters above the bars indicate significant differences (p < 0.05) in total floral visitation rates between plateaus and species based on Tukey’s post hoc tests. (B) Floral visitors depicted as the fraction of the total number of visitors.