Journal of Plant Research

, Volume 130, Issue 2, pp 311–325 | Cite as

Leaf structure affects a plant’s appearance: combined multiple-mechanisms intensify remarkable foliar variegation

  • Yun-Shiuan Chen
  • Peter Chesson
  • Ho-Wei Wu
  • Shang-Hung Pao
  • Jian-Wei Liu
  • Lee-Feng Chien
  • Jean W. H. Yong
  • Chiou-Rong Sheue
Regular Paper


The presence of foliar variegation challenges perceptions of leaf form and functioning. But variegation is often incorrectly identified and misinterpreted. The striking variegation found in juvenile Blastus cochinchinensis (Melastomataceae) provides an instructive case study of mechanisms and their ecophysiological implications. Variegated (white and green areas, vw and vg) and non-variegated leaves (normal green leaves, ng) of seedlings of Blastus were compared structurally with microtechniques, and characterized for chlorophyll content and fluorescence. More limited study of Sonerila heterostemon (Melastomataceae) and Kaempferia pulchra (Zingiberaceae) tested the generality of the findings. Variegation in Blastus combines five mechanisms: epidermal, air space, upper mesophyll, chloroplast and crystal, the latter two being new mechanisms. All mesophyll cells (vw, vg, ng) have functional chloroplasts with dense thylakoids. The vw areas are distinguished by flatter adaxial epidermal cells and central trichomes containing crystals, the presence of air spaces between the adaxial epidermis and a colorless spongy-like upper mesophyll containing smaller and fewer chloroplasts. The vw area is further distinguished by having the largest spongy-tissue chloroplasts and fewer stomata. Both leaf types have similar total chlorophyll content and similar  F v/F m (maximum quantum yield of PSII), but vg has significantly higher F v/F m than ng. Variegation in Sonerila and Kaempferia is also caused by combined mechanisms, including the crystal type in Kaempferia. This finding of combined mechanisms in three different species suggests that combined mechanisms may occur more commonly in nature than current understanding. The combined mechanisms in Blastus variegated leaves represent intricate structural modifications that may compensate for and minimize photosynthetic loss, and reflect changing plant needs.


Blastus cochinchinensis Chloroplast Crystal Kaempferia pulchra Physical color Sonerila heterostemon 


In nature, variegated leaves are relatively commonly found in forest understories in tropical and subtropical environments (Givnish 1990). Interestingly, some plants continue to produce and maintain variegated leaves throughout their lives. Others have variegated leaves only in the juvenile phase. Plants with persistent variegated foliage, such as Begonia and Aglaonema, are often popular as ornamentals and have contributed the bulk of scientific knowledge of foliar variegation (Cui and Guan 2013; Kiew 2005). In contrast, variegation restricted to juvenile plants is much less studied because of the small size of the plants, and their tendency to be hidden in the deep shade environments of forests. Indeed, understanding of this form of variegation is poor.

The presence of variegation in plants challenges perceptions of leaf form and functioning. Variegation can arise from several distinct mechanisms (Hara 1957), and comparisons of these mechanism can potentially lead to much insight. For example, variegated leaves caused by reduced chlorophyll can be artificially induced in plants, but may come with serious cost to the plant (Sheue et al. 2012). However, variegation produced by structural color may be more common than has been thought previously, as revealed by several recent studies in Begonia (Sheue et al. 2012; Zhang et al. 2009), two cultivars of Aglaonema nitidum Kunth (Fooshee and Henny 1990), and Schismatoglottis calyptrata (Roxb.) Zoll. & Moritzi (Tsukaya et al. 2004). Moreover, structural variegation can potentially occur with little cost to the plant. A case study of Begonia showed that different taxa of Begonia, although presenting a wide range of variegation patterns, use the same physical mechanism of variegation (Sheue et al. 2012), namely extra-intercelluar air space between the chlorenchyma and adaxial epidermis or hypodermis (water-storage tissue), if present. Since the variegated forms often can persist in local populations, potentially competing with non-variegated forms of the same species, presumably some adaptive advantage of foliar variegation allows them to persist and prevail locally (Pao et al. 2014; Sheue et al. 2012). For example, in Hydrophyllum virginianum L., leaf variegation has been associated with reduced herbivore damage (Campitelli et al. 2008). However, the adaptive significance of foliar variegation is not fully understood, although several hypotheses and studies have been proposed and conducted (Campitelli et al. 2008; Smith 1986; Soltau et al. 2009; Tsukaya et al. 2004).

Blastus cochinchinensis Lour. (Melastomataceae) is a common native shrub occurring in subtropical montane forests in East Asia (Huang and Huang 1993). Interestingly, both variegated and non-variegated seedling forms often appear sympatrically in deep shade environments. Its remarkable variegation attracts the attention of local people who collect the variegated seedlings as ornamentals, and speculate about its causes on social media. Collectors are disappointed because this variegation trait is heteroblastic, only present at the juvenile stage. Despite speculation, scientific understanding of the cause of the variegation in B. cochinchinensis has not been available.

Mechanisms of variegation have been classified into two major mechanism categories: pigment-related variegation (chlorophyll and pigment, chemistry) and structural variegation (air space and epidermis, physical) (Hara 1957; Sheue et al. 2012; Zhang et al. 2009). Here we show that the remarkable juvenile variegation of Blastus is caused by a combination of mechanisms from both chemical and physical categories including two new mechanisms first reported here. As a test of generality, we study also Sonerila heterostemon Naudin (Melastomataceae, producing variegated leaves until just before flowering) and Kaempferia pulchra Ridl. (Zingiberaceae, peacock ginger). Both also have multiple mechanisms of variegation.

Materials and methods

Plant materials

Two types of seedlings, variegated and green seedlings (non-variegated) of B. cochinchinensis (Melastomataceae) (Fig. 1a–d) were collected from the Huisun Forest of National Chung Hsing University in central Taiwan during 2013–2014. An herbaceous species native to Singapore, S. heterostemon was obtained from Singapore nurseries in 2014 (Fig. 1f). Both taxa have distinctive juvenile variegated leaves. A further example of combined mechanisms came from the ornamental plant with variegated leaves, Kaempferia pulchra, cultivated in the Dr. Cecelia Koo Botanic Conservation Center in Pingtung County, Taiwan (Fig. 1g). For brevity, we refer to these species below as Blastus, Sonerila and Kaempferia.

Fig. 1

Habitats and seedlings of Blastus cochinchinensis (ae), Sonerila heterostemon (f) and Kaempferia pulchra (g). a Two types of Blastus seedlings, variegated and green, at the early juvenile stage (c. 2 cm tall) growing together in the deep shade understorey of a forest. b A variegated seedling with remarkable white patches on the adaxial surface of the early juvenile leaves. c A variegated seedling having made the transition to green leaf production showing two pairs of new green leaves (top 1st, 2nd nodes) above variegated juvenile leaves (3v–5v). This plant will now only produce green leaves. d Two green seedlings. Note these individuals produce green leaves only. e A mature Blastus shrub with flowers, up to 2–3 m tall. f Seedlings of Sonerila with variegated juvenile leaves growing in a deep shade environment (arrows). The inset showing a top view of a mature Sonerila with flowers. This plant transitioned from variegated leaves to green leaves before flowering. Note that the three pairs of remaining variegated leaves are faded (3v–5v). g A K. pulchra individual with variegated leaves and pink flowers. Note that some leaves have less distinctive variegated patterns. The inset shows a close view of a variegated leaf

The variegated seedlings of Blastus produce several pairs of variegated leaves after germination and later transform to produce normal green leaves at their early juvenile stage (usually less than 15 cm tall) (Fig. 1a–c). Once transformed, the variegated seedling produces only green leaves and eventually reaches a height of 2–3 m as a shrub (Fig. 1e). Sonerila is a small herb up to 50 cm tall, occurring naturally in moist deep shade conditions in old growth lowland dipterocarp forests of Borneo, Sumatra, Peninsular Malaysia and Singapore (e.g. Bukit Timah) (Davison et al. 2008). The earlier leaves are dark green with white spots, while the top leaf pairs below the flowers are fully green, without white spots (Fig. 1). There are normally two such pairs, which we term ‘mature green leaves’. Kaempferia pulchra is a deciduous herb up to 30 cm tall. It is native to South East Asia, growing in the shaded understory of forests. Voucher specimens of Blastus and Sonerila were deposited in the TCB and TNM herbaria in Taichung, Taiwan.

Variegation properties and micromorphological features

The adaxial surface of one fresh variegated leaf from each of three variegated seedlings of Blastus, and the flowering individuals of Sonerila and Kaempferia were observed using both transmitted and reflected light with a stereoscope (Leica S8AP0, Wetzlar, Germany, with a LED ring light) equipped with a digital camera (EOS 700D, Canon, Tokyo, Japan). Leaf optical features were compared between the white area (vw) and the green area (vg) of a variegated leaf. These leaves were also observed with a light microscope (Olympus, BH-2, Tokyo, Japan) provided with an external halogen lamp to view possible crystals in the adaxial dermal cells.

Both vw and vg areas of these variegated leaves of Blastus and Sonerila were cut to approximately 1 × 1 cm2, placed on a cold stage (prefrozen by liquid nitrogen) for 1–2 min, and observed with a tabletop scanning electron microscope (TM3000, Hitachi, Tokyo, Japan).

Ten variegated and ten green leaves were collected one from each of ten variegated and ten green seedlings of Blastus, for the observations of micromorphological traits, including the shape and size of epidermal cells of both sides, the size of guard cells and subsidiary cells, the type and density of trichomes of both sides. The trichome terminology of this study follows Hsiao (2008). ImageJ applied to the images taken at 300× were used for the comparisons of cell size, stomatal and trichome densities. Quantitative data were analyzed with ANOVA, and tested with Tukey’s honest significant difference (HSD) to compare vw and vg of the variegated leaves and the green leaves of green seedlings [normal green leaves (ng)].

Leaf structure and chloroplast ultrastructure

For investigating leaf thickness, internal leaf color, chloroplast distribution and size in fresh leaves of Blastus, one variegated or one green leaf from each of five individual seedlings of both types were sampled. The leaf pieces (c. 1 × 1 cm2) of vw, vg and ng areas were temporarily held within a block of radish parenchyma tissue, and cut with a rotary microtome (20 μm thick, Microslicer DTK-1000, Ted Pella, USA). These sections were observed and photographed with a light microscope (Olympus, BH-2, Tokyo, Japan) equipped with a digital camera (EOS 700D, Canon, Tokyo, Japan). In each section image, six chloroplasts from both upper and lower mesophyll cells were randomly selected to measure chloroplast lengths with ImageJ (n = 30).

Small pieces (0.5−2.0 × 0.5−2.0 mm2) of the leaf areas of Blastus and Sonerila between primary veins of the middle leaf from vw, vg and ng were sampled. A general electron microscopy protocol was followed to obtain semithin (1 μm thick) and ultrathin (75 nm thick) sections (Sheue et al. 2012). Prepared semithin and ultrathin sections were observed with a light microscope (Olympus, BH-2, Tokyo, Japan) and a TEM (H 600, Hitachi, Tokyo, Japan, or JEM-2000 EXII, JEOL, Tokyo, Japan), respectively. One variegated leaf from each of two individuals of the limited available material of Kaempferia was free hand sectioned and observed (transverse view).

The clearing technique was applied for understanding the types and distribution of crystals of Blastus leaves. The middle parts of five leaves of both leaf types were sampled and treated following the same protocol as Sheue et al. (2014) to observe crystal characters. For a further comparison of the crystal distribution in the adaxial epidermal cells among leaf types, we observed three samples of the vw and vg areas from each variegated leaf and three samples from each green leaf. Samples were randomly selected for observing crystal presence, and each consisted of ten sequential cells in a straight line in a single field at 400× (15 samples, observed cell no. =150).

Chlorophyll content and chlorophyll fluorescence

A fully expanded leaf (variegated or green) of the top 2nd or 3rd node from each of ten seedlings of each type of Blastus was collected. Chlorophyll a and b concentration were analyzed with a spectrophotometer (V-530 Spectrophotometer, JASCO, Tokyo, Japan) after extraction with 80% acetone from 1 g fresh leaf samples (Arnon 1949). Pigment concentrations were calculated based on Mackinney–Arnon equations (Hall and Rao 1999). Totally six measurements were made for each leaf type. Data of chlorophyll contents between variegated and green leaves were statistically tested with the two-sample t test.

Ten seedlings of both types (c. 10 cm tall) were collected and potted for 2–3 days before measurements. One of the fully expanded and healthy leaves of the top 2nd or 3rd nodes from each individual was selected. Due to the irregular shape and small size of the white area of Blastus, chlorophyll fluorescence was measured from (1) an area of a variegated leaf (v) containing both white (c. 1/2) and green area, (2) a green area of a variegated leaf (vg), and (3) a green area of a green leaf (ng) of a green seedling (n = 10 for each type). The maximum quantum yields of PSII [F v/F m] were measured with a pulse-modulated fluorescence system (FMS1) (Hansatech Instruments Ltd, Norfolk, UK) follow the same protocol as Sheue et al. (2012).


Variegated pattern and micromorphology

Spotted to irregular white areas on the adaxial surface of the variegated leaves (v) of Blastus are located between the primary and the secondary veins (Fig. 1a–c). Irregular white areas are formed by coalescence of white spots. A distinctive multicellular unbranched (mu) trichome is always located at the center of a white spot (Fig. 2a). The abaxial leaf surfaces of both seedling types are often reddish (Fig. 1a–c), although a few of them are green. For the green seedlings, there is no noticeable temporal morphological change in the leaves (ng) beyond the early occurrence of mu trichomes (usually the 1st–4th pairs) (Fig. 1d).

Fig. 2

Leaf optical property and micromorphology of Blastus cochinchinensis (ah) and Sonerila heterostemon (im). bg bladder-like glandular trichome, mg mature green leaf, mu multicellular unbranched trichome, ng green leaf of a green seedling, ss short-stalked glandular trichome, v variegated leaf, vg green area of a variegated leaf, vw white area of a variegated leaf. ac The adaxial surface views of a variegated leaf of Blastus observed with reflected light (a low magnification, b high magnification) and transmitted light (c image matched with b). Each adaxial epidermal cell in the white area shows an irregular polygonal white ring under reflected light, outlining the cell, but this pattern disappears under transmitted light. d, e The adaxial epidermis of Blastus. d A variegated leaf. The epidermal cells of the vw area are polygon-like and flat, with arched to straight anticlinal cell walls. e The ng leaf showing similar epidermal characters to those of the vg area, with irregular cell shapes and repand anticlinal cell walls. Note that more abundant ss trichomes on the ng leaf. fh The abaxial epidermis of Blastus. f The vw area with more short-stalked glandular hairs, but fewer stomata. g The vg area with higher stomatal density than the vw area. h The ng leaf with the highest stomatal density, but fewer ss trichomes. The presence of the bladder-like glandular trichomes is not significantly different between these three areas. ik The adaxial surface views of a variegated leaf of Sonerila observed with reflected light (i low magnification, j high magnification) and transmitted light (k image matched with j). Under transmitted light, the variegation pattern is unrecognizable. l, m The adaxial epidermis of Sonerila. The cell shapes of the vw and vg areas (l), and the mg leaf (m) are similar. The outer paradermal cell walls of the vw area are flat, unlike those of the other two areas, which have protruding cell walls (Scale bars a, i = 2 mm; b, c, j, k = 0.25 mm; dh, l, m = 300 μm)

White areas of the adaxial variegated leaves appear as isolated white spots scattered within an oblong area enclosed by primary and secondary veins, or connected together. In Sonerila, isolated white spots only are found, and the white spots are located also in the areas enclosed by the primary and secondary veins (Fig. 1f).

Variegated leaves of Blastus and Sonerila were observed and compared with reflected and transmitted light (Fig. 2). Under reflected light, each adaxial epidermal cell of the white area (vw) of a Blastus leaf shows a striking thick white ring outlining the cell wall and a thin irregular white ring outlining the interior cell on green background (Fig. 2b). This polygonal pattern disappears under transmitted light, and both white and green areas (vg) show roughly similar green color, but with more white spots on the white area (Fig. 2c).

In Sonerila, the vw area displays irregular white patterns mingled with green color. Like Blastus, the white and green areas of Sonerila appear similar under transmitted light, with slightly more white spots on the white area (Fig. 2i–k). These optical properties of these two taxa, suggest that the leaf chlorophyll contents of the vw and vg areas do not differ greatly. This finding suggests that chlorophyll type is not a significant variegation mechanism in these plants.

Based on the SEM micrographs, the adaxial epidermal cells differ between the vw and vg areas of a v leaf, but the vg area is similar to green leaves in both Blastus and Sonerila (Fig. 2). In Blastus, the adaxial epidermal cells of the vw area are polygonal, with arched and straight anticlinal cell walls. In contrast, in the vg area, and also in the ng leaf, these cells are irregular, with repand anticlinal cell walls (Fig. 2d, e).

A comparison of stomata and trichomes of Blastus shows that these characters are significantly different between leaf types (Figs. 2f–h, 5). The average stomatal density (stomatal no. per unit area viewed with a tabletop microscope at 300×) increases significantly from 22.9, 39.4 to 51.2 in the vw, vg areas and the ng leaf respectively (Fig. 5b). Guard cell size does not differ significantly between vw, vg and ng (Fig. 5c). Three types of trichome, mu trichomes, short-stalked glandular trichomes (ss) and bladder-like glandular trichomes (bg), were found on the leaf surface. The mu trichomes located only in the centers of the vw spots are large and distinctive (Fig. 2a). The distribution of ss trichomes shows strongly contrasting patterns between leaf surfaces and between leaf types (Figs. 2d–h, 5d). Both the vg area and the ng leaf have significantly higher densities of ss trichomes on the adaxial side, while the vw area has significantly higher densities of ss trichomes on the abaxial side. The bg trichomes also occur on both sides of each of the three leaf types, with no statistically significant density variation.

Fig. 3

Comparative leaf structures between the variegated and green leaves of Blastus cochinchinesis (aj), Sonerila heterostemon (km) and K. pulchra (np). C chloroplast, Cd druse crystal, Ch f funnel-shaped chlorenchyma cell, Cr raphide crystal, Cry crystal, Ep ad adaxial epidermal cell, Ep ab abaxial epidermal cell, Hd hypodermal cell, IS intercellular space, mg mature green leaf before flowering, mu multi-seriate unbranched trichome, ng green leaf of a green seedling, P palisade cell, Sp spongy cell, Sp l spongy-like cell, vg green area of a variegated leaf, vw white area of a variegated leaf. ac Fresh leaf sections of Blastus. The vw area is thicker, with three cell-layered colorless spongy-like cells below the adaxial epidermis. Note that the ng leaf is similar to the vg leaf, but with darker green upper mesophyll. df Transverse leaf sections. Note that the funnel-shaped chlorenchyma in the vg area and the ng leaf is tightly connected to the adaxial epidermis. In contrast, the vw area has intercellular spaces (open arrows) between the spongy-like cells (with few chloroplasts indicated by small arrows) and the flat adaxial epidermal cells. g Top view of the adaxial epidermal cells (some cell walls marked) of the vw (top) and vg (bottom) areas observed with a LM provided with an external light. White patches of crystals are observed in the vw area. In contrast, the cells of the vg area show reflections of protruding cell walls (arrows), but no crystals. hj The adaxial epidermal views of a variegated leaf treated with the clearing method. h A comparison of the vw and vg areas, with a dash-line marked the boundary. i A close view of the raphide crystals in the vw area. j The mu trichome filled with druse and raphide crystals. km Transverse leaf sections of a variegated and a mature leaves of Sonerila. The intercellular spaces between the flat adaxial epidermal cells and the funnel-shaped chlorenchyma cells are only found in the vw area (open arrows). The outer cell walls of the adaxial epidermal cells are flat in vw, but lens-like in the vg area and the mg leaf. no Free hand sections of a variegated leaf of Kaempferia. n The vw area has flat outer epidermal cell walls, intercellular space between hypodermal cells and palisade tissue (open arrow) and crystals in the hypodermal cells. o The vg area with lens-like protruding adaxial epidermal cells. p A top view showing that only the vw area has crystals and polygonal light patches surrounding cell walls (Scale bars af, km = 50 μm; g, i = 25 μm; h, j = 100 μm; np = 200 μm)

Fig. 4

The mesophyll chloroplast ultrastructure of variegated leaves of Blastus cochinchinesis (ab) and Sonerila heterostemon (cd). Ch f funnel-shaped chlorenchyma cell, S starch grain, Sp spongy cell, Sp l spongy-like cell, vg green area of a variegated leaf, vw white area of a variegated leaf. a The chloroplast in the colorless spongy-like cells (the upper mesophyll) in the vw area still has very dense thylakoid membranes forming grana and starch grains similar to the vg area and the green leaf of the green seedling. b A close view of chloroplasts in the funnel-shaped chlorenchyma (upper mesophyll) in the vg area showing abundant thylakoid membranes and a starch grain. c A close view of a chloroplast in a funnel-shaped chlorenchyma cell in the vw area of Sonerilla. d A close view of a chloroplast in a spongy cell in the vw area of Sonerilla

Fig. 5

Comparisons of leaf micromorphology, chloroplast size and chlorophyll fluorescence between leaf types of Blastus cochinchinesis. Abbreviations (including lower scripts), ab abaxial side, ad adaxial side, l lower mesophyll, ng green leaf of a green seedling, ss short-stalked glandular trichome, u upper mesophyll, v variegated leaf, vg green area of a variegated leaf, vw white area of a variegated leaf. a Leaf thickness (n = 5). b Stomatal density (n = 50). c The cell size of guard cells (n = 50). d The density of short-stalked glandular trichomes on both surfaces (n = 50). e Comparison of chloroplast size (length) in the upper mesophyll and the lower mesophyll (n = 30). Note that the chloroplasts of the lower mesophyll (spongy tissue) are significantly larger than those in the upper mesophyll in the vw area. They are the largest chloroplasts of spongy mesophyll among all leaf types. f The maximum quantum yields of PSII (F v/F m) (n = 10). Due to the limitation of the small size of the vw area, data were measured from the area containing both the vw and vg areas to represent a variegated leaf as a whole (v), the vg area and the ng leaf. The different letters above histograms indicating the significant difference between groups as determined by Tukey’s HSD Test (P < 0.05); error bars represent ±SE about the mean

In Sonerila, the adaxial epidermal cells in the vw and vg areas are similarly polygonal shaped (Fig. 3l, m). However, the outer paradermal cell walls in the vw area of both taxa are much flatter than those in the vg area and the mature green leaf (mg) of Sonerila. There was no significant difference between types and sizes of the adaxial epidermal cells.

Leaf structure

All young leaves of both seedling types may be red underneath, which is especially common in variegated leaves. Fresh and resin-imbedded leaf sections of Blastus show that the structure of the vw area of a variegated leaf differs strikingly from the vg area and the ng leaf (Fig. 3a–f). The leaf is significantly thicker (P < 0.05) in the vw area (113.6 ± 10.6 μm) than in the vg area (66.2 ± 4.5 μm) and in the ng leaf (63.3 ± 2.1 μm), where the notation in parentheses is mean ± standard error of the mean. The leaf thickness of the vg area and the ng leaf is very thin (4-cell-layered in total), with one layer of funnel-shaped chlorenchyma cells (upper mesophyll) and one layer of spongy chlorenchyma cells (lower mesophyll) between the epidermal layers (Fig. 3b, c, e, f). In the vw area, the upper mesophyll is composed of 3–4 layers of colorless spongy-like cells, while the lower mesophyll has 2–3 layers of spongy cells (Fig. 3a, d). The leaf thickness of the vw area gradually thickens toward the central mu trichomes and gradually thins to the connection with the vg area.

Based on the transverse view of Blastus leaf sections, the paradermal outer cell wall of the adaxial epidermal cells is flat in the vw area, but protruded and lens like (convex) in the vg area and the ng leaf (Fig. 3a–f).

Additional intercellular spaces between the adaxial epidermal cells and the upper mesophyll were found only in the vw area of Blastus (Fig. 3d). In contrast, the adaxial epidermal cells are tightly connected to the funnel-shaped chlorenchyma cells in the vg area and in the ng leaf (Fig. 3e, f).

Druse type crystals are commonly scattered in the mesophyll of both leaf types of Blastus, with no notable difference between these leaf types. However, a strong contrast between leaf types in crystal distribution patterns was found in the adaxial epidermal cells (Fig. 3g, h). In the adaxial epidermal cells of a fresh leaf, raphide crystals appear as bundles of white filaments when observed under reflected light. In contrast, the cells of the vg area show slightly irregular reflections of cell walls.

Raphide type crystals were found in 88% (± 11.5%) of vw adaxial epidermal cells, but in 0.0% (± 0.0%) of both vg and ng cells. These raphides with different size appear as elongated forms located either in the middle or peripheral of an epidermal cell, with a compact aggregation of needle crystals, or as loosely packed randomly orientated needle crystals (Fig. 3i). In addition, the mu trichome in the center of the vw area of Blastus is densely packed with abundant raphide and druse crystals (Fig. 3j).

For Sonerila, there are two major differences between leaf types. The intercellular spaces between the adaxial epidermal cells and funnel-shaped chlorenchyma cells are only present in the vw area. However, significant lens-like outer cell walls of the adaxial epidermal cells appear in the vg area and in the mature green leaf (mg), but not in the vw area (Fig. 3k–m). These results on the adaxial cell wall characters of Blastus and Sonerila from sections accord with the observations from scanning electron microscopy (Fig. 2d, e, l, m).

In the transverse view of the vw area of Kaempferia leaves, the adaxial epidermal cell walls have flatter outer paradermal cell walls and each hypodermal cell has a crystal (druse-like) roughly located in its center (Fig. 3n–p). The crystal appears as a white spot when observed under reflected light. The intercellular spaces between hypodermal cells and the upper mesophyll cells appear as reflected light area in the vw area of a transverse fresh leaf (Fig. 3n). In a top view, each vw adaxial hypodermal cell is surrounded by several small white spots forming a polygonal pattern outlining the cell (Fig. 3p).


A close comparison of the chloroplast size (length) and number per cell in the upper and lower mesophylls between leaf types of Blastus shows that the sparsely scattered chloroplasts in the upper mesophyll (colorless spongy-like cells, Spl) in the vw area (Fig. 3a, d) are significantly smaller and fewer than those in the upper mesophyll (funnel-shaped chlorenchyma cells) of the vg area and the ng leaf (Figs. 3e, f, 5e). However, the chloroplasts of the lower mesophyll are significantly larger than those in the upper mesophyll in the vw area (Figs. 3d, 5e). These chloroplasts in the lower mesophyll of the vw area also are larger than those in the lower mesophyll of the vg area and the ng leaf, but the test of difference is not significant (P > 0.05) (Figs. 3d–f, 5e).

The chloroplast number per upper mesophyll cell (Spl) in the vw area (mean no. 0.57 ± 0.75, counted from sections with 1 μm thickness) is significantly less than that of the vg area (5.58 ± 1.16) and the ng leaf (4.83 ± 0.98) (P < 0.001). However, the chloroplast number per lower mesophyll cell (Sp) in the vw area (3.14 ± 1.23) is higher than that of the vg area (2.25 ± 0.87) (P = 0.046), but similar to that of the ng leaf (2.50 ± 0.54).

Examined with TEM, all chloroplasts in the leaves of Blastus and Sonerila, regardless of their size and location, were found to have extremely dense thylakoid membranes, filled with abundant grana and starch grains (Fig. 4). Unexpectedly, even the chloroplasts of the colorless spongy-like upper mesophyll cells of the vw area of a Blastus leaf are functional, despite their small size.

Chlorophyll content and chlorophyll fluorescence

The chlorophyll a content and the ratio of chlorophyll a/b of Blastus in the variegated leaf (v) are significantly lower than those in the green leaf (ng) (Table 1). The chlorophyll b content and total chlorophyll (a + b) of the variegated leaf are lower than in the ng leaf, but do not reach statistical significance (P > 0.05).

Table 1

Chlorophyll a, chl. b, chl. a/b and total chlorophyll contents in the variegated leaves (v) of variegated seedlings and the green leaves (ng) of green seedlings of Blastus cochinchinensis n = 6

Leaf type

Chl. a

Chl. b

Chl. a/b

Chl. a + b


9.41 ± 1.80a

4.46 ± 0.86a

2.13 ± 0.09a

13.87 ± 2.42a


13.63 ± 1.16b

5.58 ± 0.41a

2.44 ± 0.09b

16.35 ± 1.31a

The uppercase alphabetical letters mean significant difference between two groups (P < 0.05) as determined by the two-sample t test; error bars represent mean ± SE (μg/g, fresh weight)

The maximum quantum yields of PSII, as measured by F v/F m, differ significantly between the vg area and the ng leaf (Fig. 5f). Surprisingly, the vg area shows significantly higher F v/F m (0.834) than the ng leaf. The area of a variegated leaf (including 1/2 white and 1/2 green areas) also maintains high F v/F m (0.827), similar to that of the ng leaf (0.824).


The remarkable foliar variegation of the two taxa of Melastomataceae and Kaempferia comes from multiple combined mechanisms. We now show that the mechanisms leading to the variegated leaves of juvenile Blastus can be classified into the following five types: (1) epidermal type, (2) air space type, (3) upper mesophyll type, (4) chloroplast type, and (5) crystal type (Fig. 6). Several of these types are also shared by Sonerila and Kaempferia.

Fig. 6

Diagrammatic leaf structures of Blastus cochinchinensis showing the combined mechanisms of foliar variegation. bg bladder-like glandular trichome, Cd: druse crystal; Chf: funnel-shaped chlorenchyma cell; Cr raphide crystal, E ab abaxial epidermal cell, E ad adaxial epidermal cell, IS intercellular space, ng green leaf of a green seedling, mu multi-seriate unbranched trichome, S stoma complex, Sp spongy cell, Sp l spongy-like cell, ss short-stalked glandular trichome, vg green area of a variegated leaf, vw white area of a variegated leaf. The leaf structure of the vw area is strikingly different from the vg area and the ng leaf. Five mechanisms were found to be associated with such remarkable variegation creating white areas on a leaf: (1) epidermal type, 1Eab (polygonal cell shape and flat outer cell walls of the adaxial epidermal cells), (2) air space type, 2IS (with additional intercellular spaces between the adaxial epidermal cells and the upper mesophyll cells), (3) upper mesophyll type, 3UM (the upper mesophyll is composed of spongy-like cells, instead of funnel-shaped chlorenchyma), (4) chloroplast type, 4CP (smaller size and lower abundance of chloroplasts in the upper mesophyll, but larger chloroplasts in the lower mesophyll), and (5) crystal type, 5Cr (additional crystals appear in the adaxial epidermal cells). In addition, the vw area has a mu trichome filled with crystals, higher leaf thickness, the lowest stomata density and the highest ss trichome density on the abaxial surface

Epiderminal type

Hara’s (1957) epidermal type refers to differences between the epidermal cells on different parts of the leaf, creating variegation. However, only a single example of the epidermal type (i.e. Oxalis martiana Zucc.) has been reported previously (Hara 1957). Interestingly, the variegated plants studied here all provide new examples of the epidermal type. These three species have convex lens-like epidermal cells in the vg and ng areas, presumably to focus light into the mesophyll (Lee 2009; Vogelmann 1993). However, the outer paradermal cell walls of the adaxial epidermal cells of the vw area of them are flat and ss trichome free in Blastus. Such flat cell walls in the vw area potentially increase light reflectance and scattering relative to the more common convexity of the epidermal cell walls of shade plants (Lee 2009).

Air space type

Hara (1957) noted that air space between the epidermis and the mesophyll can create light areas on a leaf. A detailed study of Begonia showed that this airspace could be above the mesophyll but beneath water storage tissue or beneath the adaxial epidermis (Sheue et al. 2012). Blastus and Sonerila are characterized by air space directly beneath the adaxial epidermis in the vw areas, but not elsewhere. In the vw area of a Kaempferia leaf, additional air spaces occurring between the hypodermis (functioning as water storage tissue) and the upper mesophyll correspond to a report for Begonia (Sheue et al. 2012). This air space enhances reflection of light before it reaches the mesophyll, as explained by Snell’s law, discussed previously for variegation in Begonia (Sheue et al. 2012).

Zhang et al. (2009) showed experimentally that if the air space is replaced with water, the polygonal pattern of the vw area of the Begonia leaf disappears implying that the polygonal pattern in the vw area is derived from the air space type. In the present study, the vw areas of the studied taxa all displayed similar strong polygonal patterns suggesting that polygonal patterns can be used as an indicator of the air space type of foliar variegation. At present, most cases of structural variegation reported in the literature are the air space type, including Aglaonema nitidum (Fooshee and Henny 1990), various taxa of Begonia (Sheue et al. 2012; Zhang et al. 2009), Pulmonaria officinalis L. (Esteban et al. 2008) and Schismatoglottis calyptrata (Tsukaya et al. 2004).

Upper mesophyll type

A close view of the vw area of Blastus revealed starkly different upper mesophyll between the vw (colorless and spongy cell-like, but with functional chloroplasts) and both the vg and ng (green funnel-shaped) areas. We term this mechanism the ‘upper mesophyll type.’ It is defined at the tissue level by differences in the upper mesophyll structure, cell type and arrangement, between different areas of the leaf. Although, we know of no direct empirical evidence, the optical physics of reflection and refraction in differently shaped or arranged cells argue theoretically that a difference in appearance would result, creating variegation.

The upper mesophyll type embraces the ‘variable palisade type’ reported from Arum italicum L. (La Rocca et al. 2011), since the upper mesophyll cells may vary in form between different plant taxa, including palisade cells (for most of the dicots), funnel-shaped chlorenchyma cells (for many shade plants) (Sheue et al. 2007), and spongy-like cells as in Blastus (this study). In A. italicum, one layer of palisade tissue was found in the pale-pigmented area of the leaf, while two layers of palisade tissue were found in the dark-pigmented area (La Rocca et al. 2011).

Chloroplast type

Surprisingly the upper mesophyll cells of the vw area of Blastus look colorless in the fresh section (Fig. 3a) but still have chloroplasts with very dense thylakoid membranes and starch grains (Fig. 4a), which do not differ from the chloroplasts in the vg area and the ng leaf, although they are significantly smaller and fewer in number. In contrast, the spongy mesophyll beneath the vw area has the largest chloroplasts of the spongy mesophyll tissue over all leaf types, although this is not statistically significant (Figs. 3d–f, 5e). The lower spongy tissue with larger chloroplasts in the vw area becomes the major photosynthetic area, and presumably can compensate for potentially reduced photosynthetic performance of the upper mesophyll. This chloroplast arrangement in the vw area apparently differs from that in the vg area and the ng leaf, in which upper mesophylls (funnel-shaped chlorenchyma) with larger chloroplasts are the major photosynthetic area, like most flowering plants. Here we term this mechanism the ‘chloroplast type’, a new mechanism of variegation reported in this study. With this definition, ‘chlorophyll type’ defined by Hara (1957) is a special case of ‘chloroplast type’ when chlorophyll is absent from the plastid. In Blastus, however, ‘chloroplast type’ is caused by small but functional chloroplasts, and few per cell.

Since chloroplasts are located in the upper mesophyll cells, it might be argued that the chloroplast type is included in the upper mesophyll type. However, an effect originating from the gross structure of the tissue (upper mesophyll) is distinct from an effect originating from the organelles (chloroplast) within that tissue. Moreover, one of these effects stems from physical structure and the other from chemical composition. Thus, we retain these two terms as referring to two different mechanisms.

Crystal type

In both Blastus and Kaempferia crystals in the adaxial dermal cells occur exclusively in the vw area. This is the first report of a strong association between the presence of dermal-cell crystals and the vw area of a variegated leaf. Both taxa show crystals in association with variegation, suggesting that this phenomenon may be underreported. Possible difficulties arise in finding crystals in a leaf, including small crystal size (e.g. Blastus), loss during free hand sectioning and difficulty of observation in a semithin section, which may be too thin for intact crystal retention. Thus, applying the clearing method to include a relatively large leaf area with intact leaf structure is helpful to observe crystal characters. However, this technique requires more skill to obtain samples with good quality and a proper staining. Based on a simple test with dilute hydrogen chloride, we noticed that the crystal composition of Kaempferia is different from that of Blastus (calcium oxalate). Further study is needed to identify the chemical composition of the crystal of Kaempferia.

Although the crystal type is not known empirically to produce variegation in the absence of other mechanisms, the fact that both crystals are easily seen as intense white spots from a magnified top view of a leaf (Fig. 3g, p) confirms their optical effect. Second, the high refractive index, n = 1.58, of calcium oxalate crystals suggests high scattering ability (Gal et al. 2012, a study of druses of Carya illinoinensis (Wangenh.) K. Koch, pecan). Here we have shown how these crystals can be associated with variegation, defining the ‘crystal type’ mechanism.

Calcium oxalate crystals are commonly found in a broad range of green plants, from small algae, gymnosperms to angiosperms (Franceschi and Nakata 2005), but are not found in submerged aquatic plants (Kuo-Huang et al. 1994). Several hypotheses regarding crystal function in plants have been proposed, including calcium regulation, plant protection, detoxification (e.g. heavy metals or oxalic acid), ion balance, tissue support, and light gathering and reflection (Franceschi and Nakata 2005). Horner (2012) reported that druse crystals of palisade cells, occurring below the thin pit field regions in the wall interface, may collect and disperse light into the surrounding chloroplasts of deep shade Peperomia (Piperaceae).

However, apart from some obvious examples where crystals serve in calcium regulation, plant protection, and metal detoxification, there is a lack of evidence in support of some of the other proposed functional roles (Franceschi and Nakata 2005). The presence of crystals in the vw area of two studied taxa implies intriguing possible optical and protection effects for further exploration.

Independence of the mechanisms

We have identified five mechanisms that jointly create the patterns in Blastus. If they always co-occurred across the plant kingdom, one might doubt that these mechanisms were independent, but that is not the case. Independence means that a mechanism can be present and cause the phenomenon of interest independently of whether other mechanisms are present.

For the mechanisms here, chloroplast type, air space type, and epidermal type can be argued, based on previous evidence, as independent in that each can occur without the others, and each can create the phenomenon of interest without the others. The upper mesophyll type has not been shown generally to create variegation independently based on empirical information. However, in the form designated “palisade type” by La Rocca et al. (2011), different numbers of layers of palisade cells on different parts of a leaf were shown to create variegation. In Blastus, the upper mesophyll type consists of the difference between spongy-like tissue and funnel-shaped chlorenchyma. Although we know of no direct empirical evidence, independently of other differences, the physics of reflection and refraction in differently shaped cells argue theoretically that a difference in appearance would result.

The crystal type also is not known empirically to produce variegation in the absence of other mechanisms, but again the fact that the crystals are easily seen as intense white spots from a magnified top view of a leaf confirms their optical effect. Thus, the five mechanisms that we list are arguably independent.

Possible effects of combined mechanisms

The vw areas of Blastus and Kaempferia remarkably combine three physical mechanisms of variegation in the dermal tissue: epidermal type, air space type and crystal type. These mechanisms are active in the vicinity of the adaxial epidermis. Such a combination, which to our knowledge has never previously been reported, may potentially enhance light scattering and reflectance to create the intense whites on variegated leaves.

The upper mesophyll type and chloroplast type of Blastus involve different physical structures, including leaf structure, chloroplast size and abundance, but these physical structural differences have substantial effects on the chlorophyll content of the variegated leaf. The colorless spongy cell-like upper mesophyll revealed by the fresh leaf section is the net outcome of the chloroplast size and number differences found in the vw area, similar to the effect of chlorophyll deficiency. In this case, a physical mechanism has chemical consequences.

Past reports in the literature primarily implicate a single mechanism of variegation in any given plant (Fooshee and Henny 1990; Hara 1957; La Rocca et al. 2011; Sheue et al. 2012; Tsukaya et al. 2004; Zhang et al. 2009). However, the foliar variegation of some Aglaonema cultivars stems from two combined mechanisms, chlorophyll and air space types (Sheue, personal observation). Previously, it was suggested that the air space mechanism could be identified by observing the leaf from both sides with the naked eye under reflected light or with a stereoscope under transmitted light (Sheue et al. 2012). However, the presence of multiple mechanisms interferes with this test.

Temporal presence

Blastus and Sonerila display variegated leaves only at the young stage, unlike various Begonia (Sheue et al. 2012; Zhang et al. 2009), which maintain this trait for the lifetime of an individual and so have ‘persistent variegation.’ The variegated seedlings of Blastus usually start producing green leaves exclusively before reaching 15 cm in height. As the mature size is 2.5–3 m, it is clear that variegated plants are just at an early juvenile stage. In contrast, Sonerila produces variegated leaves until just before flowering when two-pairs of unvariegated leaves commonly appear forming an unvariegated background to the flowers. The leaf color patterns of Kaempferia are highly variable: young plant leaves are marked with silver; some mature plants retain these markings but others turn mostly green (National Gardening Association). To distinguish these patterns we use the term ‘early juvenile variegation’ for Blastus and ‘juvenile phase variegation’ for Sonerila. They are non-persistent variegation with ontogenic changes of leaf traits during development (variegation, trichome and leaf anatomy). As the term heteroblasty refers to a plant individual with abrupt and substantial changes in form and function during development (Zotz et al. 2011), both of these types of juvenile variegation are forms of heteroblasty. Other taxa display juvenile variegation including some members of Sonerila (Lee 2010), Smilex and Clematis (Sheue, personal observation), but it is a pity that this trait is often neglected.

Possible ecophysiological implications

The maximum quantum yields of PSII (F v/F m) in the Blastus v leaf are similar to the ng leaf, and these values in the vg area are significantly higher than those of the ng leaf. Such results starkly differ from variegated leaves created artificially using the chlorophyll type mechanism as observed in the variegated leaf of Ficus pumila ‘Sonny’ (Sheue et al. 2012). Although our Blastus results are similar to previous findings for Begonia with air space variegation in which the light areas had the same values of F v/F m as the green areas (Sheue et al. 2012), we should note that the maximum quantum yield of PSII does not fully characterize the photosynthetic performance.

An interesting study on European variegated plants with structural foliar variegation showed that net photosynthetic rates (expressed on a leaf area basis) were similar in the light-green and the dark-green areas of the two cyclamen species. Yet, the net photosynthesis of Cyclamen persicum was higher in the light-green areas, if expressed on a dry mass basis (Konoplyova et al. 2008).

Variegation may lead to only limited loss of photosynthetic performance in Blastus because the vw area is usually much less than 20% of a variegated leaf (Sheue et al. unpublished), is caused primarily by physical mechanisms, and the chlorophyll content of the variegated leaf does not significantly differ from the green leaf. Moreover, the overall effects of multiple combined mechanisms on photosynthesis may be more complicated than in a taxon with a single variegation mechanism.

As we have noted, in Blastus and Sonerila, the occurrence of variegation is heteroblastic. The combined mechanisms in Blastus variegated leaves represent intricate structural modifications that may compensate for and minimize photosynthetic loss, and reflect changing plant needs. A considerable number of hypotheses on the adaptive function of heteroblastic leaves have been proposed, including light capture and carbon gain, nutrient uptake, water use efficiency and resistance to herbivory, and no adaptive significance (Zotz et al. 2011), but evidence is scarce. Although the early juvenile variegation of Blastus and Sonerila may reduce herbivore damage, and the transformation of green leaves before flowering of Sonerila may encourage pollinator visits, field studies to test these ideas are lacking. Since natural variegation is especially common in shade plants (Fooshee and Henny 1990; Givnish 1990; Sheue et al. 2012; Tsukaya et al. 2004) and variegated and non-variegated forms of a taxon often occur in close proximity (Kiew 2005; Sheue et al. 2012; Tsukaya et al. 2004), variegated forms appear to be able to compete successfully with green forms of the same species and persist in local populations. Further empirical studies are warranted to better understand the mechanisms, benefits and costs of foliar variegation.



The authors thank the editor and two anonymous reviewers for valuable comments and suggestions, Dr. Cecilia Koo of the Botanic Conservation Center (KBCC) in Pingtung, Taiwan for providing Kamepferia (K040469) for this study, Huisun Experimental Forest of National Chung Hsing University for providing a permit to collect the plant materials of Blastus; nurseries in Singapore for providing materials of Sonerila, Dr. W. M. Chou (National Chiayi University) for providing a microslicer, and Dr. W. T. Chao (Tunghai University) and Miss P. C. Chaw (National Chung Hsing University) for help with TEM. This study was partially supported by the Ministry of Science and Technology [MOST-97-2126-B-005-002-MY3; MOST 104-2621-B-005-002-MY3], Taiwan.


  1. Arnon DI (1949) Copper enzymes in isolated chloroplasts. Polyphenoloxidase in Beta vulgaris. Plant Physiol 24:1–15CrossRefPubMedPubMedCentralGoogle Scholar
  2. Campitelli BE, Stehlik I, Stinchcombe JR (2008) Leaf variegation is associated with reduced herbivore damage in Hydrophyllum virginianum. Bot 86:306–313CrossRefGoogle Scholar
  3. Cui WH, Guan KY (2013) Diversity of leaf variegation in Chinese Begonia. Plant Divers Resour 35:119–127Google Scholar
  4. Davison GWH, Ng PKL, Ho HC (2008) The Singapore red data book: Threatened plants and animals of Singapore. Nature Society, SingaporeGoogle Scholar
  5. Esteban R, Fernádez-Marín B, Becerril JM, García-Plazaola JI (2008) Photoprotective implications of leaf variegation in E. dens-canis L. and P. officinalis L. J Plant Physiol 165:1255–1263CrossRefPubMedGoogle Scholar
  6. Fooshee W, Henny RJ (1990) Chlorophyll levels and anatomy of variegated and non-variegated areas of Aglaonema nitidum leaves. P Fl St Hortic Soc 103:170–172Google Scholar
  7. Franceschi VR, Nakata PA (2005) Calcium oxalate in plants: formation and function. Annu Rev Plant Biol 56:41–71CrossRefPubMedGoogle Scholar
  8. Gal A, Brumfeld V, Weiner S, Addadi L, Oron D (2012) Certain biominerals in leaves function as light scatterers. Adv Opt Mater 24:OP77–OP83Google Scholar
  9. Givnish T (1990) Leaf mottling: relation to growth form and leaf phenology and possible role as camouflage. Funct Ecol 4:463–474CrossRefGoogle Scholar
  10. Hall DO, Rao KK (1999) Photosynthesis, 6th edn. Cambridge University Press, CambridgeGoogle Scholar
  11. Hara N (1957) Study of variegation leaves, with special reference to those caused by air space. J Jpn Bot 16:86–101Google Scholar
  12. Horner HT (2012) Peperomia leaf cell wall interface between the multiple hypodermis and crystal-containing photosynthetic layer displays unusual pit fields. Ann Bot 109:1307–1316CrossRefPubMedPubMedCentralGoogle Scholar
  13. Hsiao PM (2008) A study on leaf morphology of Melastomataceae in Taiwan. Master Thesis. National Sun Yat-sen UniversityGoogle Scholar
  14. Huang SF, Huang TC (1993) Melastomataceae. In: Bufford DE, Chang CE, Chaw SM, Chen CH, Deng MJ, Hartley TJ, Hoch PC, Hsieh CF, Hsieh TH, Huang SF, Huang TC, Kao MT, Li HJ, Lievens AW, Lin SC (eds) Flora of Taiwan, 2nd edn. vol 3. National Taiwan University Press, Taipei pp 905–928Google Scholar
  15. Kiew R (2005) Begonias of peninsular Malaysia. Natural History Publications, Kota Kinabalu (Borneo)Google Scholar
  16. Konoplyova A, Petropoulou Y, Yiotis C, Psaras GK, Manetas Y (2008) The fine structure and photosynthetic cost of structural leaf variegation. Flora 203:653–662CrossRefGoogle Scholar
  17. Kuo-Huang LL, Sheue CR, Yang YP, Chiang SHT (1994) Calcium oxalate crystals in some aquatic angiosperms of Taiwan. Bot Bull Acad Sinica 34:179–188Google Scholar
  18. La Rocca N, Rascio N, Pupillo P (2011) Variegation in Arum italicum leaves. A structural–functional study. Plant Physiol Bioch 49:1392–1398CrossRefGoogle Scholar
  19. Lee DW (2009) Plant tissue optics: micro- and nanostructures. Proc. SPIE 7401, Biomim Bioinspir, 740104 (August 21, 2009). doi: 10.1117/12.826085
  20. Lee DW (2010) Nature’s palette: the science of plant color. University of Chicago Press, ChicagoGoogle Scholar
  21. National Gardening Association, Accessed 17 September 2016
  22. Pao SH, Chesson P, Peng CI, Yeh WB, Sheue CR (2014) Characteristics and ecological significance of natural foliar variegation from Begonia formosana. The 99th Annual Meeting, 10th–15th August, Ecological Society of America. Sacramento, California, USAGoogle Scholar
  23. Sheue CR, Sarafis V, Kiew R, Liu HY, Salino A, Kuo-Huang LL, Yang YP, Tsai CC, Lin CH, Yong JWH, Ku MSB (2007) Bizonoplast, a unique chloroplast in the epidermal cells of microphylls in the shade plant Selaginella erythropus (Selaginellaceae). Am J Bot 94:1922–1929CrossRefPubMedGoogle Scholar
  24. Sheue CR, Pao SH, Chien LF, Chesson P, Peng CI (2012) Natural foliar variegation without costs? The case of Begonia. Ann Bot 109:1065–1074CrossRefPubMedPubMedCentralGoogle Scholar
  25. Sheue CR, Chang HC, Yang YP, Liu HY, Chesson P, Hsu FH (2014) A new classification of marginal resin ducts improves understanding of hard pine (Pinaceae) diversity in Taiwan. Flora 209:414–425CrossRefGoogle Scholar
  26. Smith AP (1986) Ecology of a leaf color polymorphism in a tropical forest species: habitat segregation and herbivory. Oecologia 69:283–287CrossRefGoogle Scholar
  27. Soltau U, Dötterl S, Liede-Schumann S (2009) Leaf variegation in Caladium steudneriifolium (Araceae): a case of mimicry? Evol Ecol 23:503–512CrossRefGoogle Scholar
  28. Tsukaya H, Okada H, Mohamed M (2004) A novel feature of structural variegation in leaves of the tropical plant Schismatoglottis calyptrata. J Plant Res 117:477–480CrossRefPubMedGoogle Scholar
  29. Vogelmann TC (1993) Plant tissue optics. Annu Rev Plant Biol 44:231–251CrossRefGoogle Scholar
  30. Zhang Y, Hayashi T, Hosokawa M, Yazawa S, Li Y (2009) Metallic lustre and the optical mechanism generated from the leaf surface of Begonia rex Putz. Sci Hortic Amst 121:213–217CrossRefGoogle Scholar
  31. Zotz G, Wilhelm K, Becker A (2011) Heteroblasty—a review. Bot Rev 77:109–151CrossRefGoogle Scholar

Copyright information

© The Botanical Society of Japan and Springer Japan 2016

Authors and Affiliations

  • Yun-Shiuan Chen
    • 1
  • Peter Chesson
    • 1
    • 2
  • Ho-Wei Wu
    • 3
    • 4
  • Shang-Hung Pao
    • 1
  • Jian-Wei Liu
    • 1
  • Lee-Feng Chien
    • 1
  • Jean W. H. Yong
    • 5
  • Chiou-Rong Sheue
    • 1
    • 2
  1. 1.Department of Life Sciences and Center for Global Change BiologyNational Chung Hsing UniversityTaichungTaiwan, Republic of China
  2. 2.Department of Ecology and Evolutionary BiologyUniversity of ArizonaTucsonUSA
  3. 3.Genome and Systems Biology Degree ProgramNational Taiwan UniversityTaipeiTaiwan, Republic of China
  4. 4.Institute of Plant and Microbial BiologyAcademia SinicaTaipeiTaiwan, Republic of China
  5. 5.School of Plant BiologyUniversity of Western AustraliaPerthAustralia

Personalised recommendations