Total reflectance of ultraviolet and photosynthetically effective wavelengths was measured for a range of different leaf types. Two approaches were employed. Firstly, reflectance of monochromatic wavebands at 330 and 680 nm was measured for a total of 45 different species covering a wide range of genera. In the second, specific leaf types that displayed different degrees of reflectance were treated to remove hairs and waxes that contributed to their reflectance. Selected waxy and non-waxy leaves were also studied in more detail over the spectral range 270–500 nm. It was found that both pubescence (presence of hairs) and glaucousness (presence of a thick epicuticular wax layer) had marked effects on total reflectance. Pubescent leaves tended to be more effective in reflecting longer wavelengths than ultraviolet radiation. The extent of this effect depended on hair type. Glaucous leaves demonstrated that surface waxes were very effective reflectors of both UV and longer wavelength radiation.
The aim of this study was to assess the relative contributions of leaf hairs and cuticular waxes to the reflectance of potentially damaging UV radiation and of useful photosynthetically active radiation (PAR; 400–700 nm) in a representative range of plant species. Most of the studies of the deleterious effects of UV radiation on plants cited above have concentrated on the UV-B waveband. However, long-term (weeks) field experiments have shown that excessive levels of radiation in the UV-A (320–400 nm) waveband can also be inhibitory to plants. Examples include Brassica napus L. (Holmes 1997), Senecio vulgaris L. (Holmes 1998) and Cynosurus cristatus L. (Cooley et al. 2001). Preliminary spectral studies showed that measurements of reflectance at 330 nm provide a useful indicator of reflectance in both the UV-B and UV-A wavebands with the exception of Eucalyptus cinerea L. (which was the only glasshouse-grown plant studied). We have therefore measured the reflectance of the leaves of a wide range of genera and species at 330 nm, with parallel measurements at the photosynthetically active wavelength of 680 nm. Detailed spectral studies between 270 and 500 nm were also carried out for selected species.
Materials and methods
Cambridge University Botanic Garden (52°12′ N, 0°07′ E) carries an extensive range of plant species, obtained from different altitudes and latitudes. Accordingly it is possible to study plants from environments with widely differing solar UV fluxes. A total of 45 species were selected for examination (Table 1). With the exception of E. cinerea L., which had been grown in a glasshouse, all plants had been grown outdoors. Although the Botanic Garden carries a wider range of species, those chosen for this study were determined in part by the fact the experiments were carried out in spring.
Table 1. A list of the species used in this study showing also Family, geographical area of origin and leaf surface classification
Prunus lusitanica azorica
S. Europe/N. Africa
Eur/C. Asia/N. Afr.
S. Eur/W. Asia
S. Eur/N. Afr/C.Asia
S. & E. Australia
All measurements were made on the upper laminas of unshaded leaves (avoiding mid-veins) that showed no visible sign of disease, or injury. Classification of leaf types (see below) was based on visual examination of the leaf surface. For monochomatic measurements, preliminary experimentation (data not shown) revealed little variation in reflectance measurements from undamaged leaves of the same species. Accordingly, for these measurements, two leaves from each species were measured. Standard errors (SE) of the means of these measurements were typically less than 8%, with maximum SEs of 12·8 and 18·2% for measurements at 680 and 330 nm, respectively. For all other procedures (spectral scans, wax and hair removal), five leaves from each species were used.
All reflectance measurements were made with an Oriel Scientific integrating sphere (‘Ulbricht’; Oriel Scientific, Leatherhead, UK) using standard techniques (e.g. Seyfried 1989). Both 330 and 680 nm monochromatic radiation were provided by a 100 W mercury arc lamp, with appropriate optics, stabilized ( ± 0·5%) by an XH100 power supply (Müller, Moosinning, Germany). After passage through a 60 mm distilled water cuvette (model MHO 60, Müller) to minimize infra-red radiation, the radiation was filtered through a calibrated 10 nm bandpass interference filter (Andover Series, Salem, NH, USA). Second- and third-order harmonic transmittance of longer (and shorter) wavelengths was reduced to less than 10−5 of peak wavelength. A calibrated Spectralon reflectance standard with > 95% reflectance at the relevant wavelengths was used (oriel scientific). For monochromatic wavelengths, reflected radiation was measured with an UV-enhanced silicon photodiode (SD112UV; Macam, Livingston, UK) connected to a LP102X power meter (Macam), calibrated for both spectral response and linearity.
Spectral scans were made in 1 nm steps over the range 270–500 nm with a temperature-controlled double holographic grating Optronics (Orlando, FL, USA) 742 spectroradiometer using polychromatic radiation from the mercury lamp described above. The scans were restricted to this waveband because the photomultiplier is optimized to the UV and blue wavebands; extension of the polychromatic studies to longer wavelengths would have resulted in unacceptable losses in both accuracy and precision. The spectroradiometer was calibrated against both National Physical Laboratory and National Institute of Standards and Technology standards and against other instruments in use across Europe (for details, see Webb 1994).
Classification of leaf hairs and waxes
Although epicuticular coverings can vary greatly between different species (see Discussion), their description has been reduced to four basic types for the purpose of this study (Table 1). These different types can also occur in combination and are:
1Woolly – very long (>1 mm), thin hairs which are typically white as they are dead cells and contain air.
2Hairy – single short or stellate hairs which are normally visible to the naked eye. This category also includes bristly leaves, i.e. those with numerous very short, stiff hairs.
3Hairless – smooth, glabrous, rugose (rough), or merely papillose epidermis. Leaves with few, widely spaced bristles were also included in this category.
4Waxy (glaucous) – an obvious wax covering which can be detected under a light microscope and can be removed by a wax solvent.
Removal of leaf hairs and waxes
Leaf hairs were removed by gently rubbing the leaf surface with very fine emery paper. Subsequent examination of the leaf surface, using a binocular microscope, enabled only those leaves which showed negligible damage to the epidermal surface, but with almost total elimination of hairs, to be selected for reflectance measurements.
Leaf wax was removed by dipping the leaves in chloroform, wiping gently with a camel hair brush and rinsing in fresh chloroform and reflectance was measured immediately thereafter (Barnes et al. 1996). Again, microscopic examination of the leaf surface was used to confirm wax removal and lack of dehydration.
Where appropriate, data were analysed by analysis of variance using SPSS 7·5 (Jandel Corp., San Rafael, CA, USA). Standard errors were computed for all means. All proportional data were transformed prior to analysis using an arcsin square root transformation (Nogués et al. 1998).
Screening a selection of different leaf types
A selection of 45 species was compared in an attempt to determine possible influences of leaf surface morphology (e.g. hairs and waxes) on the reflectance of radiation in the UV waveband at 330 nm. Comparative measurements were also made at 680 nm in order to estimate the relative reflectance of photosynthetically active radiation. The plants, their origins and leaf surface morphology are shown in Table 1. Data from the reflectance measurements are presented in Fig. 1.
The main conclusion that can be drawn from these results is that four significantly different (P < 0·01) patterns of reflectance are observed which can be related to leaf surface morphology. First, waxy leaves had significantly higher reflectances (P < 0·01) in both the UV (330 nm) and photosynthetic (680 nm) wavelengths than any other leaf type studied. Second, woolly leaves had highly variable reflectances at both wavelengths measured; however, a high reflectance at 680 nm did not necessarily result in a high reflectance at 330 nm. Thus three of the woolly-leaved plants had visible reflectances of greater than 13%, yet had the lowest measured UV reflectances (< 2·5%). Third, hairless leaves tended to have low reflectances at both 330 nm (5·3–9·2%) and 680 nm (4·7–8·4%). Fourth, hairy leaves had significantly lower reflectances (P < 0·01) than hairless leaves at 330 nm (3·9–5·4%), but a similar range of reflectances at 680 nm (3·8–7·1%).
Effects of removing leaf hairs and/or wax
Ten of the species which had been screened in the previous experiment and which had contrasting leaf morphologies and reflectances were treated with abrasives or solvents in an attempt to ascertain relationships between leaf surface morphology and reflectance. Following treatment of the selected leaves (Fig. 2a & b) it was observed that femoval of wax from the leaf surfaces resulted in significant reductions (P < 0·01) in reflectance in both waxy and pubescent leaves at 680 nm. Removal of leaf surface waxes caused significant reductions (P < 0·01) in reflectance at 330 nm. In all three waxy-leaved species examined, the reduction in reflectance was proportionally greater at 330 nm than at 680 nm.
Removal of hairs from the leaf surfaces had variable effects on UV reflectance. In two species (Campanula elationides and Kalanchoe tomentosa), hair removal significantly reduced (P < 0·01) reflectance at 330 nm. Hair removal from the leaves of Verbascum dumulosum, however, had no significant effect on UV reflectance. Attempts to remove surface waxes by chloroform treatment from leaves classified as hairless had no significant effect on reflectance at either wavelength (data not shown).
The presence of leaf wax has a marked effect on leaf reflectance at both UV and photosynthetic wavelengths. In contrast, leaf hairs appeared more effective in reflecting 680 nm radiation and did not have a consistent influence on reflectance at 330 nm. Treatment of the hairless leaves with the same solvent as used for the waxy leaves had no significant effect on their reflectance. This latter observation provides valuable circumstantial evidence that the wax removal procedure had no significant effect on the reflective properties of the cuticles of the plants studied.
In order to achieve a greater understanding of the contribution of leaf waxes to leaf reflectance, a series of spectral reflectance measurements between 270 and 500 nm were performed on selected waxy and non-waxy leaves. Species with waxy leaves were compared with closely related species with non-waxy leaves, both before and after treatment to remove wax. Leaf reflectance was substantially greater at all wavelengths scanned in the two waxy-leaved Eucalypts compared with non-waxy Eucalypts (Fig. 3). Treatment to remove wax significantly reduced reflectance at all wavelengths in the waxy Eucalypts, but had negligible effect on the non-waxy Eucalypts (Fig. 3a,c). Wax removal reduced leaf reflectance in the waxy species to a level similar to the non-waxy (Fig. 3b,d).
Figure 4 compares the spectral reflectance of the two succulents, Kalanchoe blosfeldiana and Kalanchoe pumila. Kalanchoe blosfeldiana is non-waxy and washing the leaf surface with chloroform had no effect on reflectance (Fig. 4a). By contrast, removal of waxes from the surface of K. pumila leaves (Fig. 4b) caused a marked reduction in the reflectance at PAR wavelengths (400–500 nm); a similar reduction was noted in the monochromatic studies at 680 nm (Fig. 2). Reflectance of UV-A and longer UV-B wavelengths was also reduced by wax removal. A small but consistent increase in reflectance was always observed below 300 nm after wax removal.
Regression analysis demonstrated that there was no correlation between UV reflectance and latitude of origin (not shown).
Most plant cuticles are covered with waxes to a greater or lesser extent. These waxes have considerable structural diversity and are of fundamental importance to plants in their interactions with the environment (Barthlott et al. 1998). In a comprehensive survey of over 13 000 species, a total of 23 different wax types were classified, including layers, crusts, platelets and projections (Barthlott et al. 1998). Clearly, a classification of such complexity (see also below) would preclude meaningful comparisons of leaf surface type and reflectance in the present study.
Wax development has long been known to be environmentally dependent. Thus, when Eucalpytus risdoni was overwintered in a heated greenhouse, it did not produce wax, whereas those grown outside did (Barber 1955; Penfold & Willis 1961). Such results are difficult to interpret as several factors, including radiation, temperature and water stress are changed. In this study the only greenhouse-grown eucalypt (E. cinerea) produced wax; however, no outdoor plants were available for comparison. As with leaf waxes, leaf hairs, or trichomes, are found on many plants (Esau 1977). Like waxes, leaf hairs have also been classified into numerous categories, including unicellular and multicellular, which range in size or shape from the short, broad protruberances of papillate hairs, through long, thin hairs, to complex stellate scales (Esau 1977). Furthermore the hairs can be alive or dead, calcified, silicated, or neither and the protoplast can vary greatly in size, or be absent (Esau 1977). As with waxes (Gonzalez et al. 1996; Gordon & Percy 1999), the extent to which hairs develop is dependent on both species and environment (Karabourniotis et al. 1992; Liakoura et al. 1997; Filella & Peñuelas 1999). Accordingly, the classification of hairs and waxes that we have employed in this study is necessarily coarse, but is required if meaningful relationships are to be obtained between leaf hair-type and leaf reflectance.
Relatively little work has been published about the UV reflective properties of leaf hairs and waxes. Caldwell et al. (1983) suggested that, for most species, UV-B reflectance from the leaf surface was low (<5%). Whereas this study shows that this is clearly true for the hairless and the less densely pubescent species, leaves with waxy or with densely pubescent surfaces can have considerably greater UV reflectances (Mulroy 1975; Robberecht et al. 1980). The relative effectiveness of epicuticular waxes and hairs can be compared by reference to Figs 1–4. In particular, Figs 3 and 4 show the effectiveness of wax in reflecting UV radiation, where two out of the three species studied showed higher reflectances in the UV than at visible wavelengths. By contrast, almost all of the densely pubescent, or woolly-leaved, species studied had higher reflectances at 680 nm than at 330 nm.
Although it is clear from Figs 1, 3 and 4 that waxes have fairly broad-spectrum reflectances (but see also below), those of leaf hairs are more variable and a high reflectance at visible wavelengths does not necessarily mean a high reflectance in the UV waveband (Fig. 1). Such variation in the reflectances of highly pubescent (woolly) leaves has been reported previously (Robberecht et al. 1980) and can, in some instances, be explained by the UV-absorbing properties of such leaf hairs (Karabourniotis et al. 1992, 1993). That this is not always the case, however, is demonstrated by the observation that removal of hairs from Verbascum dumulosum (Fig. 2), while significantly reducing PAR reflectance, had no effect on that in the UV. Likewise, the significantly lower UV reflectances of hairy leaves, compared to hairless (Fig. 1), could result from the UV absorbing properties of the hairs (Karabourniotis et al. 1992, 1993), or their increased ability to scatter light (Skaltsa et al. 1994). The absorbance of UV radiation by both leaf hairs and waxes increases strongly with wavelengths below about 300 nm and this increase in absorbance is more marked than that with supposed screening pigments (Bornman & Vogelmann 1988). However, too little is known about the relative absorbance of waxes compared with screening pigments in order to state the relative importance of waxes. It is noteworthy that the spectral reflectance characteristics of the waxy species E. cinerea (Fig. 3b), Eucalyptus gunnii (Fig. 3d) and K. pumila (Fig. 4b) in their natural state differ greatly, whereas removal of the surface wax results in reflectance characteristics which are broadly similar. This insinuates substantial differences in the chemical and/or morphological characteristics of the waxes and deserves further study.
The data presented in Figs 1–4 demonstrate that leaf surface waxes are effective reflectors of both UV and visible wavelengths. The wavelength selectivity of this reflection, however, is less pronounced than that of epidermal absorption (Robberecht & Caldwell 1978; Caldwell et al. 1983). It is therefore possible that reflectance is of less importance in protecting plants against the potentially damaging effects of UV-B radiation than absorption (Caldwell et al. 1983). Nevertheless, a number of studies have demonstrated protective roles for leaf hairs and waxes against UV-B. Robberecht et al. (1980) showed that despite the fact the proportions of reflected (5–40%) and absorbed UV-B (60–95%) varied markedly within four species examined, all had transmittances of UV-B into the mesophyll of less than 1%. Moreover, removal of leaf surface waxes (Clark & Lister 1975; Karabourniotis et al. 1992) has been shown to increase the amount of UV penetrating the mesophyll.
It is clear from the preceding discussion therefore that UV protection is not a fundamental function of leaf waxes but rather one of many. It can therefore be concluded that, for the species studied, epicuticular waxes are effective reflectors of UV radiation, whereas leaf hairs, if profuse, tend to be poor reflectors of UV, but can substantially reduce the amount of PAR arriving at the leaf surface. Selection of the trait for epicuticular waxes may provide an effective method for developing crops which would be more tolerant of increased UV-B radiation, but that would not result in plants that had substantially reduced absorption of essential photosynthetically active radiation.
The authors thank E. Sauber and J. Simchen (Botanical Institute, University of Erlangen, Germany) for their help in collecting and analysing leaf material used in this study.
Received 20 April 2001;received inrevised form 26 July 2001;accepted for publication 26 July 2001