1371
American Journal of Botany 88(8): 1371?1389. 2001.
HYDROPHOBIC TRICHOME LAYERS AND EPICUTICULAR
WAX POWDERS IN BROMELIACEAE1
SIMON PIERCE,2,3 KATE MAXWELL,2 HOWARD GRIFFITHS,2 AND
KLAUS WINTER
Smithsonian Tropical Research Institute, Apartado 2072, Balboa, Panama City, Republic of Panama
The distinctive foliar trichome of Bromeliaceae has promoted the evolution of an epiphytic habit in certain taxa by allowing the
shoot to assume a significant role in the uptake of water and mineral nutrients. Despite the profound ecophysiological and taxonomic
importance of this epidermal structure, the functions of nonabsorbent trichomes in remaining Bromeliaceae are not fully understood.
The hypothesis that light reflection from these trichome layers provides photoprotection was not supported by spectroradiometry and
fluorimetry in the present study; the mean reflectance of visible light from trichome layers did not exceed 6.4% on the adaxial surfaces
of species representing a range of ecophysiological types nor was significant photoprotection provided by their presence. Several
reports suggesting water repellency in some terrestrial Bromeliaceae were investigated. Scanning electron microscopy (SEM) and a
new technique?fluorographic dimensional imaging (FDI)?were used to assess the interaction between aqueous droplets and the leaf
surfaces of 86 species from 25 genera. In the majority of cases a dense layer of overlapping, stellate or peltate trichomes held water
off the leaf epidermis proper. In the case of hydrophobic tank-forming tillandsioideae, a powdery epicuticular wax layer provided
water repellency. The irregular architecture of these indumenta resulted in relatively little contact with water droplets. Most mesic
terrestrial Pitcairnioideae examined either possessed glabrous leaf blades or hydrophobic layers of confluent trichomes on the abaxial
surface. Thus, the present study indicates that an important ancestral function of the foliar trichome in Bromeliaceae was water
repellency. The ecophysiological consequences of hydrophobia are discussed.
Key words: Bromeliaceae; epicuticular wax; fluorographic dimensional imaging; SEM; trichomes; water repellency.
Bromeliaceae are flowering plants that are popular in hor-
ticulture and also of great ecological importance in the Neo-
tropics, occupying a diverse range of habitats. One of the first
attempts to classify bromeliad diversity in an ecological con-
text was made by Pittendrigh (1948), who elaborated on the
observation of Tietze (1906) that life form and the function of
leaf hairs was reflected in the taxonomic relationships of gen-
era. Pittendrigh?s scheme was further expanded by Benzing
(2000) into the five ecophysiological types summarized in Ta-
ble 1.
Leaf hairs or foliar trichomes (i.e., unicellular or multicel-
lular structures arising from the epidermal tissues; Bell, 1991)
are almost ubiquitous in Bromeliaceae (Benzing, 1976) and
are perhaps the most distinguishing vegetative feature of the
family. It is well documented that the peltate trichomes be-
longing to species with Type 3, 4, and 5 life forms support
epiphytism by endowing the shoot with the capacity to aug-
ment or replace the absorptive functions of roots (Schimper,
1888; Billings, 1904; Mez, 1904; Benzing, 1970, 1976; Benz-
ing and Burt, 1970; Benzing et al., 1976; Nyman et al., 1987;
Smith, 1989; see Benzing [1980] for a detailed discussion of
their mode of action). The trichomes of terrestrial Type 1 and
1 Manuscript received 27 June 2000; revision accepted 1 February 2001.
The authors thank Harry E. Luther and Bruce K. Holst for assistance with
identification, for generously putting the living collections and herbarium of
the Marie Selby Botanic Gardens at our disposal, and, along with Darren M.
Crayn, for providing comments on an early version of the manuscript; David
H. Benzing for invaluable criticism during the review process; Jason R. Grant
for aid with systematic issues; Jorge E. Aranda for collecting in Fortuna,
Panama; Jorge Ceballos for invaluable technical support with the scanning
electron microscope; and Richard Gottsberger and Aurelio Virgo for general
assistance. We gratefully acknowledge support by a grant from the Andrew
W. Mellon Foundation through the Smithsonian Tropical Research Institute.
2 University of Cambridge, Department of Plant Sciences, Downing Street,
Cambridge, CB2 3EA, UK.
3 Author for reprint requests.
many Type 2 bromeliads are incapable of this function (Benz-
ing et al., 1976; Lu?ttge et al., 1986). Trichome function has
therefore played a pivotal role in the adaptive radiation of
Bromeliaceae via the operation of these different ecophysio-
logical strategies.
However, the function(s) of the trichomes of Type 1 bro-
meliads remains enigmatic. Molecular phylogenetics indicates
that the genera Ayensua and Brocchinia are basal to the rest
of the family (Terry, Brown, and Olmstead, 1997; Horres et
al., 2000; Crayn, Winter, and Smith, unpublished data). Al-
though direct fossil evidence is negligible, mesic Type 1 Pit-
cairnioideae (e.g., Ayensua, some Brocchinia, Fosterella, Pit-
cairnia) are also considered to exhibit a primitive life form
(i.e., ecophysiologically they most closely resemble a hypo-
thetical ancestor of the family). This assessment is based not
only on subfamilial characteristics such as the extensive root
system (Tietze, 1906), but also on the presence of less ad-
vanced nonsucculent C3 physiology (see Medina, 1974) and
the simpler structure of the trichome (Benzing, 1980). Indeed,
within the genus Brocchinia advanced Type 4 species possess
absorbing trichomes, while nonimpounding terrestrial species
possess less highly organized trichomes and are more basal
within the genus (N.B. the most primitive of these, B. pris-
matica, possesses stellate trichomes similar to those of Fos-
terella species; Givnish et al., 1997). Thus, foliar trichomes of
mesic Type 1 Pitcairnioideae mediate primitive functions.
Many roles other than water and nutrient absorption have
been ascribed to bromeliad trichomes, but these functions of-
ten only apply to a small number of species (such as the at-
traction of pollinators or seed dispersers in the case of some
Tillandsia and Billbergia species; Benzing, 2000). More gen-
eral hypotheses concerning the function of bromeliad tri-
chomes include obstruction of predators and pathogens (Benz-
ing, 2000), reduction of transpiration (Billings, 1904), and
photoprotection (Benzing and Renfrow, 1971; Lu?ttge et al.,
1372 [Vol. 88AMERICAN JOURNAL OF BOTANY
TABLE 1. Life forms or ecophysiological types of Bromeliaceae (after Benzing, 2000).
Life form Characteristics
1 Terrestrial herbs of subfamily Pitcairnioideae (and many Bromelioideae) that use roots to acquire water and nutrients?the leaf
hairs being nonabsorbent.
2 Terrestrial Bromelioideae with leaf bases that form a rudimentary watertight ??tank?? into which some axillary roots may grow.
3 Terrestrial or epiphytic herbs in subfamily Bromelioideae, the roots of which have reduced importance in water and nutrient
acquisition with the leaf bases forming an extensive water-holding tank?predominantly crassulacean acid metabolism (CAM),
with leaf hairs that have some capacity to take up water and nutrients.
4 Tank-forming epiphytes in subfamily Tillandsioideae and some Brocchinia?predominantly C3 and with high densities of leaf
hairs on the leaf bases that are highly effective at water and nutrient uptake, the roots functioning primarily as holdfasts.
5 Succulent CAM Tillandsioideae that are epiphytic or lithophytic, with leaf hairs taking up water directly over the entire leaf
surface (without a tank) and possessing holdfast roots, if any.
1986). The deterrence of predators and pathogens currently has
no experimental support. Reduction of transpiration is a xe-
romorphic adaptation, and as such, it is unlikely that this
would be an important selection pressure acting on ancestors
in mesic habitats.
In high densities, bromeliad trichomes produce a whitish
leaf surface that reflects light when dry. This has been quan-
tified in Type 4 Tillandsia fasciculata (Benzing and Renfrow,
1971) and semimesic Type 1 Pitcairnia integrifolia (Lu?ttge et
al., 1986) and is highly suggestive of a role in photoprotection.
However, in the more relevant case of Type 1 P. integrifolia,
trichomes are restricted to the abaxial surface of the leaf; had
these trichomes developed primarily to serve a photoprotective
role, then they would be expected to occur at least in equal
densities on the glabrous adaxial surface. Lu?ttge et al. (1986)
note that the edges of the leaves of P. integrifolia roll inwards
to expose the trichomed abaxial surface during the dry season,
perhaps to promote reflectance, and propose this as a form of
regulation of light reflectance. However, this behavior may oc-
cur simply as a consequence of drought in glabrous species
(e.g., Pitcairnia valerii; personal observation), perhaps as a
response to water loss and concomitant shrinkage of water-
storage parenchyma in the hypodermis (see Billings, 1904).
More importantly, the trichomes of P. integrifolia and P. bif-
rons were not found to influence the heating of leaves (Lu?ttge
et al., 1986). Thus, a photoprotective role for trichomes re-
mains without direct supporting evidence; an investigation of
photoinhibition using fluorimetry techniques has yet to be un-
dertaken.
Evidence for a further general hypothesis concerning the
role of the trichome in terrestrial Bromeliaceae is present in
the literature, but has apparently been overlooked. Krauss
(1948?1949) working on Ananas comosus noted that ??the tri-
chomes on the lower surface of the leaf blade proper appear
unwettable. Drops of water placed on this surface do not
spread, but remain unabsorbed for experimental periods of 3
to 6 h.??
Krauss (1948?1949) also went on to observe that, whereas
the absorbent trichomes of Tillandsia usneoides lost their pale
whitish color when wetted (Billings, 1904), those on the ab-
axial surface of A. comosus did not, as a consequence of air
trapped beneath the trichomes. This implies that the trichomes
on the abaxial surface of A. comosus repel water. Also, the
abaxial surfaces of Pitcairnia integrifolia and P. macrochla-
mys leaf blades appear to be unwettable (Benzing, Seemann,
and Renfrow, 1978; Lu?ttge et al., 1986), and in the case of P.
integrifolia, ??water repellent.?? Indeed, Benzing (1970) dis-
covered that after 12 h of exposure the abaxial surface of P.
macrochlamys had absorbed ;3.5 times less zinc65 than the
glabrous cuticle of the adaxial surface, perhaps suggesting that
the trichome layer hindered absorption. Widespread occur-
rence of repellent trichome layers on the abaxial leaf blade
surfaces of mesic Type 1 bromeliads would therefore suggest
that hydrophobia was an important property of the foliar tri-
chome in ancestral Bromeliaceae.
Also relevant to this study are the hydrophobic waxy sur-
faces of Brocchinia reducta and Catopsis berteroniana. Tom-
linson (1969) suggests that in the case of C. berteroniana these
promote the run-off of water from the leaf blades into the tank
and attraction and entrapment of insect prey by these carniv-
orous species have also been suggested (Fish, 1976; Frank and
O?Meara, 1984). These species also share advanced Type 4
life forms, which usually possess hydrophilic trichomes at
least lining the tank. Determinations of the occurrence of hy-
drophobic surfaces in Tillandsioideae and Brocchinia could
shed additional light on the evolution of the Type 4 life form.
The present study employs a novel technique, fluorographic
dimensional imaging (FDI), to assess the interactions between
aqueous droplets and the leaf blade surfaces of 86 ecologically
diverse bromeliad species representing 25 genera and all three
subfamilies. Fluorographic dimensional imaging is used in
conjunction with scanning electron microscopy (SEM) and
spectroradiometry to reveal the mechanism by which certain
trichomes and epicuticular wax powders repel water. Fluorim-
etry is used to investigate the hypothesized role of trichomes
and wax layers in photoprotection. Nomenclature follows that
of Luther and Sieff (1998), with the exception of the recently
rejected genus Pepinia (Taylor and Robinson, 1999), which is
recognized as a subgenus of Pitcairnia (sensu Smith and
Downs, 1974).
MATERIALS AND METHODS
Plant material of Panamanian origin was collected from the wild, with
voucher specimens being held at the main herbarium of the Smithsonian Trop-
ical Research Institute, Panama (herbarium code SCZ) and at the University
of Panama (PMA). Material of Trinidadian origin was obtained from the living
collections of Moorbank Botanic Gardens (Newcastle-upon-Tyne, UK). An-
anas comosus was grown from meristem culture, with original material pro-
vided by the Centre International de Recherche en Agronomie et Development
(Montpellier, France). All other material was obtained from the living collec-
tions at the Marie Selby Botanic Gardens, Sarasota, Florida, USA (accession
numbers available on request).
Repellency was denoted by the depth of aqueous droplets on adaxial and
abaxial leaf blade surfaces. For FDI of aqueous droplets, calibration standards
were prepared using glass coverslips (;2 cm wide), one-half being coated
with a flat film of paraplast wax (Sigma Chemical, St. Louis, Missouri, USA),
and the other half remaining as an exposed glass surface. The thickness of
these wax and glass standards was measured by micrometer, and these stan-
August 2001] 1373PIERCE ET AL.?WATER REPELLENT INDUMENTA OF BROMELIADS
dards were lightly fixed along one edge of a strong glass plate of ;40 3 40
cm.
Leaf discs were cut from intact and surface denuded midleaf portions of
leaf blade (from two-thirds of the way along the blade). In many species
denudation was achieved using sticky tape, although some species such as
Ananas comosus required careful scraping with a scalpel blade. In the case
of apparently glabrous leaves, the procedure of denudation with sticky tape
was conducted for the sake of consistency. Leaf discs from replicate leaves
(where possible from separate individuals) were then fixed in rows onto the
glass plate, with intact and denuded examples of both surfaces presented up-
permost.
Droplets (10-mL each) of 0.05% (mass by volume in distilled H2O) fluo-
rescein sodium solution were quickly pipetted onto the surface of the leaf
discs and calibration standards and left to stand for 40 min in a darkened
room. In these darkened conditions, the leaf discs and standards were then
illuminated with an ultraviolet (UV) transilluminator (Fotodyne, Hartland,
Wisconsin, USA), and the resulting fluorescence from the excited fluoro-
chrome was photographed using a level camera mounted directly above the
leaf discs. Initial tests determined that the following camera settings provided
the greatest depth of field and contrast, with well-exposed fluorescence and a
darkened background: an aperture of f/22, aperture priority (or a 9-sec ex-
posure with a cable release), using ISO 100/DIN 218 color-reversal film (Ko-
dak Elite). The depth of droplets on wax and glass standards was determined
by micrometer immediately after the fluorograph was taken.
After processing, fluorographs were digitally scanned (LS-2000, Nikon,
Shinagawa-Ku, Tokyo, Japan) and the luminosity of fluorescein droplets was
determined using Corel PHOTO-PAINT7 (Corel, Ottawa, Ontario, Canada)
imaging software (selecting each particular region of the image with the ??eye-
dropper?? tool, and recording the luminosity (L) of the ??paint?? color). To
compensate for possible uneven lighting, eight measurements were taken from
each droplet, and the measurements were averaged. Luminosity and depth
data from the glass and wax standards were then regressed (Excel, Microsoft,
Seattle, Washington, USA) to create a calibration equation, from which the
depth of droplets on leaf discs was calculated using respective luminosity
values. This technique allowed rapid, inexpensive, mass screening of samples.
The difference in droplet depth (DD) due to surface features can be summa-
rized by the following equation:
DD 5 i 2 ed (b) d (b) d (b) (1)
where i 5 droplet depth on intact surface, e 5 droplet depth on denuded
surface, d 5 adaxial surface or alternatively b 5 abaxial surface.
In order to examine the effect of water surface tension on the interaction
between trichomes and water, the above FDI technique was also used on the
leaves of Ananas comosus, using droplets (10-mL each) of fluorescein sodium
solution (5 mL of 0.05% fluorescein and 0.5 mL distilled H2O); with further
replicates on which 10-mL droplets of a solution of fluorescein and household
detergent (5 mL of 0.05% fluorescein and 0.5 mL neat detergent) were used.
Reflectance of light by leaves was measured using an LI-1800 portable
spectroradiometer (LI-COR, Lincoln, Nebraska, USA), via an 1800?12s ex-
ternal integrating sphere (LI-COR). Ranges of reflectance values were nor-
malized to 100% using barium sulfate (BaSO4) as a standard; this compound
has an absolute reflectivity of 99.3% in the wavelength range 300?800 nm
(Munsell Color, New Windsor, New York, USA). Measurements were taken
of intact, water-inundated, and denuded leaf surfaces (both adaxial and ab-
axial). Species with water repellent trichome layers were inundated by soaking
in water for 1 h or until a surface film of water could be sustained on their
removal from the water. Once again, in the case of surfaces that appeared to
have no trichomes, the denudation process was carried out with sticky tape
for consistency?s sake. Average reflectance values of photosynthetically active
radiation (PAR) were calculated as a mean across the wavelength range 400?
700 nm. The reflectance conferred by trichomes or wax powders is defined
as the difference in mean reflection between intact and denuded surfaces.
Photoinhibition of photosystem II was investigated using a PAM-2000 por-
table modulated fluorimeter (H. Walz, Effeltrich, Germany). Aechmea dacty-
lina, Ananas comosus cv. Cayenne Lisse, Catopsis micrantha, Pitcairnia in-
tegrifolia, Tillandsia flexuosa, and Werauhia sanguinolenta were maintained
in seminatural conditions in an open-sided greenhouse at the main Smithson-
ian Tropical Research Institute facility in Panama. Excluding the cultivar of
Ananas comosus, these species grow in semi-exposed to exposed microhab-
itats and may experience several hours of direct sunlight each day (Lu?ttge et
al., 1986; personal observations). A treatment of excessive excitation therefore
consisted of transferring plants grown in moderate sunlight (;450 mmol pho-
ton?m22?sec21 at midday) to direct sunlight at midday (PPFD ?1700 mmol
photon?m22?sec21) for 1 h. The degree of photoinhibition was denoted by the
decline in the dark-adapted ratio of variable to maximum chlorophyll fluo-
rescence (Fv/Fm) following this treatment, with intact and denuded surfaces
being compared.
For scanning electron microscopy, the majority of leaf samples were de-
hydrated through an alcohol series, critical point dried (CPD) in CO2, and
then sputter-coated with gold-palladium (Hummer VI-A, Anatech, Springfield,
Virginia, USA) before examination in the scanning electron microscope (Jeol
JSM-5300LV, Jeol, Tokyo, Japan). However, samples of Catopsis were not
dehydrated in this manner, as the solvents used in CPD may destroy the
structure of wax surfaces (Juniper and Jeffree, 1983); samples were placed in
the scanning electron microscope without preparation.
RESULTS
Light reflectance and photoprotection?An intact layer of
dry trichomes increased the reflectance of visible light (400?
700 nm) by an average of 6.4% on the adaxial surface of
Aechmea dactylina, although not significantly on the abaxial
surface (P . 0.05; Figs. 1?4). Reflectance was increased by
5.0 and 3.9% on adaxial and abaxial surfaces, respectively, of
Tillandsia flexuosa (data not shown), 4.9 and 10.6% on adaxial
and abaxial surfaces of Ananas comosus (Figs. 5?8), and
17.8% on the abaxial surface of Pitcairnia integrifolia (but not
on the glabrous adaxial surface; Figs. 9?12). Powdery epicu-
ticular wax increased reflectance of visible light by a mean of
6.3 and 6.6% on adaxial and abaxial surfaces, respectively, of
Catopsis micrantha (Figs. 13?16). Low densities of filmy tri-
chomes were observed via SEM on the adaxial surface of Type
4 Werauhia sanguinolenta, but these did not alter reflectance
(data not shown). The increased reflectance conferred by tri-
chomes or wax was not sufficient for photoprotection, with the
extent of photodamage (as denoted by a percentage decline in
Fv/Fm) exhibited by leaves with intact surfaces equaling that
of leaves denuded of trichomes or wax powders (after expo-
sure to an equivalent and excessive photon dose; Table 2).
When inundated with water, the adaxial surfaces of Aech-
mea dactylina and Ananas comosus (Figs. 3, 7) and both sur-
faces of Tillandsia flexuosa lost the reflectivity conferred by
their trichomes. The trichomes of Pitcairnia integrifolia and
those of the abaxial surface of Ananas comosus retained their
reflectivity when treated in this manner (Figs. 8, 12). A surface
film of water could not be sustained on the leaves of Catopsis
micrantha even after several days of inundation. Indumenta
did not increase the reflectance of infrared light (800 nm) in
most species, except for Catopsis micrantha and Pitcairnia
integrifolia. Reflectance of infrared wavelengths was higher
(40?50%) than the reflectance of visible light in all species
studied.
Leaf blade interactions with water?A typical fluorograph
for a single species (Catopsis micrantha) is shown in Fig. 17.
Fluorographic dimensional imaging determined that droplet
depth had diminished after 40 min on the intact leaf blade
surfaces of Type 5 species when compared with surfaces de-
nuded of trichomes (DD). For example, on leaf blades of Til-
1374 [Vol. 88AMERICAN JOURNAL OF BOTANY
Figs. 1?4. Aechmea dactylina leaf blade surfaces. 1?2. Scanning electron micrographs of the adaxial and abaxial surfaces, respectively. 3?4. Reflectance of
light by the adaxial and abaxial surfaces, respectively. Reflectance data represent the mean 6 1 SE of four replicates.
landsia nana, DDd 5 2732 and DDb 5 2876 mm; confirming
these leaves to be highly hydrophilic.
Droplets exhibited no significant difference in depth be-
tween intact and denuded leaf blade surfaces in most Type 4
species (P # 0.05; Table 3). However, there were some notable
exceptions; for example the hydrophobic abaxial surface of
Vriesea monstrum (DDd 5 214 mm; Table 3) and both hydro-
philic surfaces of Tillandsia elongata (DDd 5 2210 mm and
DDb 5 2190 mm). Many Type 4 taxa possessed hydrophobic
waxy surfaces, e.g., Catopsis micrantha (DDd 5 800 mm and
DDb 5 960 mm), Guzmania macropoda (DDb 5 216 mm), and
Werauhia capitata (DDb 5 350 mm).
Trichomes, but not wax, lent subfamily Bromelioideae a
range of interactions with leaf surface water. This included no
interaction at all (e.g., both surfaces of Type 2 Bromelia pin-
guin; Table 3), hydrophilic surfaces (e.g., Type 3 Aechmea
dactylina, DDd 5 2220 mm and DDb 5 2130 mm; Type 3
A. fendleri, DDd 5 2130 mm and DDb 5 2110 mm), and the
hydrophobic abaxial surfaces of species such as Type 2 An-
anas comosus (DDb 5 160 mm; Fig. 8) and Type 1 Ronnbergia
explodens (DDb 5 100 mm). A number of bromelioid species
possessed both hydrophilic adaxial surfaces and hydrophobic
abaxial surfaces (e.g., Type 3 Aechmea nudicaulis, DDd 5
2267 mm and DDb 5 226 mm; Type 1 Cryptanthus whitmanii,
DDd 5 2205 mm and DDb 5 407 mm; Type 1 Orthophytum
benzingii, DDd 5 2477 mm and DDb 5 474 mm).
Of the mesic Type 1 pitcairnioids, genera such as Fosterella
and Pitcairnia either possessed hydrophobic abaxial surfaces,
due solely to trichome cover (e.g., Pitcairnia integrifolia, DDb
5 230 mm), or were entirely glabrous and noninteractive (e.g.,
Pitcairnia patentiflora), with a small number possessing hy-
drophobic adaxial surfaces (Pitcairnia arcuata, DDd 5 310
mm). The more xeromorphic pitcairnioid genera showed a
range of trichome-mediated interactions with surface water,
including species that possessed both hydrophilic and hydro-
phobic leaf blade surfaces (e.g., Dyckia marnier-lapostollei,
DDd 5 2457 mm and DDb 5 740 mm; Table 3).
Of the 16 species examined from the elfin cloud forest at
Cerro Jefe in central Panama, six possessed water-repellent
leaf surfaces (Table 3). These were either Type 1 species with
repellent trichomes (Pitcairnia arcuata, Ronnbergia explo-
dens) or Type 4 species with relatively upright leaves that used
trichomes (Vriesea monstrum) or epicuticular wax powders
(Catopsis micrantha, Guzmania macropoda, Werauhia capi-
tata) to shed water. A further six were Type 4 species equipped
with hypostomatous and horizontally orientated leaves.
August 2001] 1375PIERCE ET AL.?WATER REPELLENT INDUMENTA OF BROMELIADS
Figs. 5?8. Ananas comosus leaf blade surfaces. 5?6. Scanning electron micrographs of the adaxial and abaxial surfaces, respectively. 7?8. Reflectance of
light by the adaxial and abaxial surfaces, respectively. Reflectance data represent the mean 6 1 SE of four replicates.
The wax powder layer of Catopsis micrantha was less pro-
nounced towards the tip of the leaf blade, where it still pro-
moted beading up of water (Figs. 18, 19). This layer was con-
tinuous over the leaf blade surface (Figs. 13, 15, 20), but was
not present on the adaxial surface of the leaf sheath within the
tank of the plant. This surface is densely covered with peltate
trichomes (Fig. 21). Powdery epicuticular wax was also pre-
sent on hydrophobic surfaces of Alcantarea odorata, Broc-
chinia reducta, and Werauhia capitata (Figs. 22?25).
Surfaces that showed no trichome- or wax-mediated inter-
action with water generally either possessed thin, filmy peltate
trichomes (e.g., the adaxial surface of Vriesea monstrum; Fig.
26) or lacked surface structures (e.g., the adaxial surfaces of
Fosterella petiolata, Pitcairnia corallina, Pitcairnia integri-
folia; Figs. 9, 27, 28). Water repellent trichomed surfaces fea-
tured either high densities of large, overlapping peltate tri-
chomes consisting mainly of extrusive ring cells (i.e., ??ring-
peltate?? trichomes; Figs. 6, 10, 29?31) or low densities of
tangled stellate trichomes forming a discontinuous indumen-
tum (e.g., Pitcairnia arcuata; Fig. 32). Trichomes of Puya laxa
did not significantly interact with water droplets (P # 0.05;
Table 3)?this species possesses two types of trichome, one
being highly modified with an elongate wing that spirals
around itself to form a hair-like structure (Fig. 33).
Low densities of ring-peltate trichomes occurred on the hy-
drophilic surfaces of Aechmea dactylina (Figs. 1, 2, 34, 35).
Individual trichomes were structurally comparable to the tri-
chomes comprising the continuous hydrophobic trichome lay-
ers of Ananas comosus, Fosterella petiolata, Pitcairnia cor-
allina, Ronnbergia explodens, and Vriesea monstrum (Figs. 6,
29?31, 36, 37). None of these species possessed wax powders,
either on the trichomes or elsewhere.
On the hydrophilic adaxial surface and hydrophobic abaxial
surface of Cryptanthus whitmanii the trichomes appeared no
different, although the lower densities on the adaxial surface
revealed the leaf epidermis proper to SEM (Figs. 38, 39).
Aechmea nudicaulis also has low densities of thin, filmy tri-
chomes on the hydrophilic adaxial surface (Fig. 40) and a
typical hydrophobic abaxial surface (Fig. 41). No species in
any subfamily possessed a hydrophobic adaxial surface com-
bined with a hydrophilic abaxial surface. Water repellent epi-
cuticular wax powders or confluent layers of large ring-peltate
trichomes occurred exclusively on surfaces that possessed sto-
mata in the species studied.
1376 [Vol. 88AMERICAN JOURNAL OF BOTANY
Figs. 9?12. Pitcairnia integrifolia leaf blade surfaces. 9?10. Scanning electron micrographs of the adaxial and abaxial surfaces, respectively. 11?12. Re-
flectance of light by the adaxial and abaxial surfaces, respectively. Reflectance data represent the mean 6 1 SE of four replicates.
The addition of detergent to the fluorescein solution used in
FDI resulted in higher wettability of both adaxial and abaxial
surfaces of Ananas comosus, with aqueous droplets (10-mL
volume) spreading to negligible depth (14.8 6 3.2 mm adax-
ially and 17.6 6 5.3 mm abaxially; Table 4) when the surface
tension of the water was reduced in this manner.
DISCUSSION
Light reflectance and photoprotection?The data indicate
that trichomes and epicuticular wax powders do not have a
significant photoprotective function in a range of ecophysio-
logical types (Types 1?4). Trichomes either did not increase
light reflectance from leaf blades (e.g., Werauhia sanguinolen-
ta) or the mean reflectance conferred by trichomes or wax did
not exceed 6.4% on the adaxial surfaces of the species studied
(with up to 17.8% on the abaxial surfaces). This was not suf-
ficient to significantly alter down-regulation of photosystem II
by excess light in these species (Table 2). Indeed, trichomes
and epicuticular wax powders conferring reflectances of be-
tween ;45 and 55% photoprotect certain desert plants (Eh-
leringer and Bjo?rkman, 1976; Robinson, Lovelock, and Os-
mond, 1993). Also, the present study indicated that the reflec-
tance conferred was correlated with the mode of interaction
between surfaces and water. Hydrophobic surfaces did not lose
reflectivity when wet, whereas hydrophilic trichomes did (see
also Billings, 1904; Krauss, 1948?1949; Benzing, Seemann,
and Renfrow, 1978), and higher reflectivities on abaxial sur-
faces were correlated with the presence of dense hydrophobic
indumenta (e.g., Ananas comosus, Pitcairnia integrifolia).
Thus, the data indicate that hydrophobic and dry hydrophilic
trichome layers inherently scatter light, but are unlikely to
have evolved primarily for the purpose of photoprotection in
Bromeliaceae.
The highly unusual, woolly trichomes of Puya laxa (Fig.
33) did not interact with water droplets on the leaf surface
(Table 3). These trichomes probably act as protection against
frost damage as exhibited by a number of Puya species grow-
ing in high altitude habitats (Miller, 1994). As this example
illustrates, distinct taxa produce trichomes that represent a
more specific adaptation to local environmental conditions.
Thus, dense indumenta could yet prove to furnish photopro-
tection in the case of more extreme xerophytes (Type 5 spe-
cies). A thorough investigation of the fluorescence character-
istics of this life form was beyond the scope of the present
study.
August 2001] 1377PIERCE ET AL.?WATER REPELLENT INDUMENTA OF BROMELIADS
Figs. 13?16. Catopsis micrantha leaf blade surfaces. 13?14. Scanning electron micrographs of the adaxial and abaxial surfaces, respectively. 15?16. Re-
flectance of light by the adaxial and abaxial surfaces, respectively. Reflectance data represent the mean 6 1 SE of four replicates.
TABLE 2. Decrease in Fv/Fm (the dark-adapted ratio of variable to maximum chlorophyll fluorescence) of six species after exposure to saturating
light (PPFD ?1700 mmol?m22?s21) for 1 h, with the leaf blade surface either intact or denuded of surface features. Values are means 61 SE of
four replicates. The absence of differences in letters (a) between means of intact and denuded treatments indicates that there were no significant
differences at the P # 0.05 level as determined by Student?s t test. Life forms or ecophysiological types follow Benzing (2000).
Species Life form Surface
Decrease in Fv/Fm (%)
Intact Denuded
Aechmea dactylina 3 Adaxial
Abaxial
30.0 6 7.6 a
26.9 6 5.8 a
37.9 6 8.8 a
30.1 6 5.9 a
Ananas comosus 2 Adaxial
Abaxial
58.5 6 5.6 a
26.3 6 4.7 a
43.8 6 6.5 a
29.2 6 3.4 a
Catopsis micrantha 4 Adaxial
Abaxial
29.4 6 8.7 a
22.5 6 5.0 a
28.2 6 3.2 a
29.5 6 1.7 a
Pitcairnia integrifolia 1 Adaxial
Abaxial
52.7 6 4.9 a
35.6 6 6.3 a
48.4 6 2.7 a
39.0 6 1.4 a
Tillandsia flexuosa 4?5 Adaxial
Abaxial
11.7 6 4.1 a
30.7 6 9.1 a
21.1 6 8.7 a
38.2 6 10.1 a
Werauhia sanguinolenta 4 Adaxial
Abaxial
47.3 6 7.2 a
34.1 6 1.4 a
39.4 6 3.2 a
30.9 6 2.7 a
1378 [Vol. 88AMERICAN JOURNAL OF BOTANY
Fig. 17. A typical fluorograph for a single species (Catopsis micrantha), used to determine the depth of aqueous droplets on leaf disc surfaces (denoting
repellency) via the comparison of fluorescence signatures of fluorescein droplets against calibration droplets of known depth. In this example, epicuticular wax
powder layers from the adaxial and abaxial leaf blade surfaces are either present (intact) or removed (denuded). Mean depth values presented include 6 1 SE,
with significant differences between means (at the P # 0.05 level) of four replicates determined using Fisher?s multiple comparison procedure.
The mechanism of water repellency?Brewer, Smith, and
Vogelmann (1991) noted three kinds of interaction between
water and the trichomes of flowering plants: (1) low trichome
densities that do not influence droplet retention or the location
of surface moisture, (2) low densities of trichomes that induce
surface water to aggregate into patches, and (3) high densities
of trichomes that lift water off the leaf surface. The leaf blade
surfaces of many Type 4 bromeliads exhibit low trichome den-
sities (Benzing, 1980) and did not interact detectably with sur-
face water in the present study (Table 3). Bromeliads that have
low densities of attenuated stellate trichomes, such as Type 1
Pitcairnia arcuata, appear to interact with water as described
by situation 2, loosely aggregating surface droplets. Consistent
with the third, ?lifting,? mechanism of repellency, continuous
layers of powdery wax or ring-peltate trichomes produce an
irregular hydrophobic surface that prevents water from coming
into contact with the epidermis proper. A summary of the prin-
cipal interactions between leaf blade trichome layers and water
within each ecophysiological type is presented in Table 5.
In many families of flowering plants, water droplets bead
up more readily on irregular than uniform surfaces because
the droplet only contacts the tips of projections from the cu-
ticle (Holloway, 1968; Juniper and Jeffree, 1983), obviating
adhesion (Brewer, Smith, and Vogelmann, 1991; Watanabe and
Yamaguchi, 1993). The physics of these surface?water inter-
actions are outlined by Barthlott and Neinhuis (1997). This
hydrophobic mechanism is readily demonstrable in Bromeli-
aceae. For example, a wetted pineapple leaf (Ananas comosus)
will lose the pale coloration of the abaxial surface only if
detergent is first added to the water. Species with absorbent
trichomes, on the other hand, lose this pale coloration and
reflectivity immediately on wetting (Billings, 1904; Benzing
and Renfrow, 1971; Benzing, 1980; Fig. 3). Also, droplets of
water will only spread on the abaxial leaf surface of pineapple
if detergent is added (demonstrated quantitatively in Table 4).
Pineapple leaves soaked overnight in a detergent solution or
100% acetone will regain their repellency if subsequently
rinsed and dried, suggesting a physical rather than chemical
mechanism (personal observations). Additionally, if a pine-
apple leaf is partially dipped into a detergent solution rather
than pure water, then liquid will be drawn or ??wicked?? up out
of the solution along the trichome layer, i.e., once the surface
tension of the water is broken the leaf surface becomes strong-
ly hydrophilic. Thus, the physical properties of water are cen-
tral to the mechanism of repellency. This mechanism also
August 2001] 1379PIERCE ET AL.?WATER REPELLENT INDUMENTA OF BROMELIADS
demonstrates, at least in part, how Type 4 species prevent wa-
ter loss from the tank via capillary action.
Trichomes that characterize hydrophilic and hydrophobic
surfaces usually share the same structure, with trichome den-
sity differing (e.g., the adaxial and abaxial surfaces of Cryp-
tanthus whitmanii and hydrophilic Aechmea dactylina com-
pared with hydrophobic Ronnbergia explodens; Figs. 34?39).
The lower densities of peltate trichomes of Aechmea dactylina
and the adaxial surface of Cryptanthus whitmanii would allow
water to come into contact with the epidermis proper, with the
interaction between the two presumably allowing water to
spread and envelop trichomes. In addition, the adaxial tri-
chomes of Aechmea nudicaulis differ structurally?lacking the
irregular surface characteristic of the hydrophobic abaxial in-
dumentum (Figs. 40, 41). The chemical composition of hy-
drophilic and hydrophobic surfaces in Bromeliaceae has not
been investigated and the degree to which chemical vs. mor-
phological interactions contribute to repellency remains un-
determined. Nevertheless, the physical characteristics of hy-
drophobic trichome layers in Bromeliaceae are typical of wa-
ter-repellent surfaces in other families, and the qualitative tests
above suggest that surface morphology is paramount to the
operation of hydrophobia.
Ecophysiological consequences of a hydrophobic
indumentum?It may be significant that the majority of bro-
meliads are hypostomatous (Tomlinson, 1969; Benzing and
Burt, 1970), with stomata and hydrophobic trichome layers
occurring together. Intriguingly, Barthlott and Neinhuis (1997)
demonstrate that particulate matter will adhere more readily to
water droplets than to hydrophobic leaf surfaces, lending such
leaves a ??self-cleaning?? capability when wetted. In concert
with a possible function as a physical barrier to pathogens
(Benzing, 2000), this self-cleaning effect could remove path-
ogens and prevent the physical blockage of stomata by partic-
ulates. A continuous trichome layer could also deter herbivores
from the softer underside of the leaf, although to date this
protective role is only evident in two species possessing glan-
dular trichomes (see Benzing, 2000).
Benzing, Seemann, and Renfrow (1978) determined that
photosynthetic gas exchange was not inhibited by wetting the
leaf blades of six species on the surfaces of which water did
not spread (including Pitcairnia macrochlamys). Conversely,
the wetted trichomes of Type 5 bromeliads hold films of water
that slow the exchange of gases between the air and the leaf
(Benzing, Seemann, and Renfrow, 1978; Schmitt, Martin, and
Lu?ttge, 1989). Clearly, most Type 5 bromeliads must reconcile
both gas exchange and water acquisition through the same
surface, relying on temporal separation of these two processes
by performing gas exchange when the leaf is dry. In contrast,
Type 1 and Type 2 bromeliads separate the processes of gas
exchange and water acquisition spatially between roots and
leaves and tank-forming species between the leaf sheath and
blade. Thus, these latter life forms do not need to compromise
carbon gain to acquire water. In this respect, wettable tri-
chomes on the leaf blade would not only be an unnecessary
investment but would be disadvantageous in mesic habitats,
whereas repellent trichomes would favor gas exchange, as per-
haps demonstrated by Pitcairnia macrochlamys (Benzing, See-
mann, and Renfrow, 1978).
Sources of water that may moisten the underside of the leaf
may include dew and, perhaps more importantly in cloud for-
ests, wind-borne mist. These factors in conjunction with the
terrestrial lifestyle (i.e., the close proximity of vegetation and/
or the ground surface from which rainwater can splash up-
wards onto the underside of the leaf) may help explain the
evolution of hydrophobic trichome layers in Bromeliaceae. In-
deed, in the family as a whole, rosulate habits typical of genera
such as Fosterella and Cryptanthus tend to have hydrophobic
abaxial surfaces (Table 3). Also, terrestrial Orthophytum ben-
zingii has basal leaves close to the substrate that possess a
repellent trichome layer on the abaxial surface, but on cauline
leaves this layer is less apparent (personal observation).
Trichome evolution?The mechanism of water repellency
outlined above accords with the scheme of trichome structural
evolution detailed by Benzing (1980). In this scheme, the hy-
pothetical ancestral morphology is stellate (the simple fila-
mentous trichomes of some Navia species appear to be de-
rived; Benzing, 1980; Terry, Brown, and Olmstead, 1997).
Low densities of stellate trichomes provide only discontinu-
ous, patchy repellency (e.g., extant Pitcairnia arcuata), in-
creased densities of which would maintain a greater proportion
of the moistened leaf surface dry. Following this proposed
early increase in trichome density, stellate trichomes may then
have undergone an increase in the number of ring cells, be-
coming truly peltate. This would increase the area covered by
each trichome and thereby foster the ??lifting?? mechanism of
repellency (high densities of intermediate stellate/ring-peltate
trichomes occur in Pitcairnia corallina and P. integrifolia
[Figs. 10, 31] and P. macrochlamys; Benzing, Seemann, and
Renfrow, 1978). Additionally, the extrusive ring cells of such
peltate trichomes appear to lend the overall surface an ex-
tremely irregular small-scale texture (e.g., Figs. 29?31).
Hydrophilic trichome layers among extant Bromelioideae
feature lower trichome densities, suggesting a decline in tri-
chome density from ancestors with dense hydrophobic layers.
This perhaps reflects adaptive radiation into less crowded or
relatively xeric niches. Indeed, Type 1 Ronnbergia explodens
has dense hydrophobic trichome layers and grows in the un-
derstory of cloud forest habitats (Figs. 36, 37; Table 3). More
xeromorphic terrestrial species (CAM equipped and succulent)
such as Cryptanthus warasii and C. whitmanii may possess
hydrophilic surfaces characterized by fewer trichomes (Fig.
38; Table 3; unpublished data), as do many Type 3 species
(Aechmea dactylina, A. nudicaulis; Figs. 1, 2, 34, 40; Table
3).
Dense trichome layers in Tillandsioideae are usually hydro-
philic, unlike those of Bromelioideae and Pitcairnioideae. In-
deed, Billings (1904) points out that one of the most unusual
features of Tillandsia usneoides is that ??unlike most similar
appendages of the epidermis, the scales do not hinder the leaf
from becoming wet.?? Dense hydrophilic trichome layers in
Tillandsioideae must possess a difference that can account for
their lack of water repellency. At present, differences in the
chemical composition of these surfaces cannot be ruled out.
However, a striking structural difference between the tri-
chomes of Tillandsioideae and those of the other subfamilies
is apparent, which could also explain the different interaction
with water. From scanning electron micrographs published in
other sources (Benzing, Seemann, and Renfrow, 1978; Benz-
ing, 1980; Adams and Martin, 1986), it is possible to see that
the parts of adjacent tillandsioid trichomes that overlap one
another are the flexible wings, which overlap when flattened
(wet). Thus, when the leaf is dry and the wings are flexed
upwards, underlying epidermis cells are exposed (Benzing,
1380 [Vol. 88AMERICAN JOURNAL OF BOTANY
T
A
B
LE
3.
L
ea
f
bl
ad
e
su
rf
ac
e?
w
at
er
in
te
ra
ct
io
ns
o
f
sp
ec
ie
s
o
f
th
e
fa
m
ily
B
ro
m
el
ia
ce
ae
,d
iv
id
ed
by
su
bf
am
ily
.
A
da
xi
al
a
n
d
a
ba
xi
al
su
rf
ac
es
w
e
re
e
ith
er
in
ta
c
t
(su
rfa
ce
st
ru
ct
ur
es
pr
es
en
t)
o
r
de
nu
de
d
(su
rfa
ce
st
ru
ct
ur
es
re
m
o
v
e
d)
.S
ur
fa
ce
ty
pe
?
hy
dr
op
hi
lic
,
hy
dr
op
ho
bi
c,
o
r
n
o
t
si
gn
ifi
ca
nt
ly
in
te
ra
ct
iv
e
(n
si)
?
is
de
no
te
d
by
th
e
de
pt
h
o
f
a
10
-m
L
dr
op
le
t
o
f
a
qu
eo
us
flu
or
oc
hr
om
e
a
fte
r
a
pe
rio
d
o
f
40
m
in
c
o
m
pa
re
d
be
tw
ee
n
in
ta
ct
a
n
d
de
nu
de
d
su
rf
ac
es
.D
ep
th
v
a
lu
es
a
re
de
riv
ed
fr
om
flu
or
oc
hr
om
e
lu
m
in
os
ity
(u
nd
er
e
x
c
iti
ng
U
V
)c
o
m
pa
re
d
a
ga
in
st
st
an
da
rd
s
o
f
m
e
a
su
re
d
dr
op
le
td
ep
th
(fl
uo
ro
ch
ro
me
o
n
pa
ra
pl
as
tw
a
x
a
n
d
gl
as
s
su
rf
ac
es
).
V
a
lu
es
re
pr
es
en
tm
e
a
n
s
6
1
SE
o
f
fo
ur
re
pl
ic
at
es
(*
de
no
te
s
sp
ec
ie
s
o
f
w
hi
ch
si
x
re
pl
ic
at
es
w
e
re
u
se
d)
.D
iff
er
en
t
le
tte
rs
(a?
b)
re
pr
es
en
t
si
gn
ifi
ca
nt
di
ff
er
en
ce
s
be
tw
ee
n
in
ta
ct
a
n
d
de
nu
de
d
m
e
a
n
s
a
t
th
e
P
#
0.
05
le
ve
l
a
s
de
te
rm
in
ed
by
St
ud
en
t?s
t
te
st
.
T
he
ph
ot
os
yn
th
et
ic
c
a
rb
on
a
ss
im
ila
tio
n
pa
th
w
ay
(C
3
o
r
c
ra
ss
u
la
ce
an
a
c
id
m
e
ta
bo
lis
m
)e
m
pl
oy
ed
by
e
a
c
h
sp
ec
ie
s
w
a
s
de
te
rm
in
ed
by
D
H
1
,
if
n
o
t
pr
ev
io
us
ly
di
sc
lo
se
d
by
M
ar
tin
(?
;1
99
4)
a
n
d
re
fe
re
nc
es
th
er
ei
n,
o
r
de
te
rm
in
ed
v
ia
c
a
rb
on
is
ot
op
e
di
sc
rim
in
at
io
n
by
C
ra
yn
,W
in
te
r,
a
n
d
Sm
ith
(?
;u
n
pu
bl
is
he
d
da
ta
).
Pl
an
tm
a
te
ria
lo
rig
in
at
ed
fr
om
th
e
fo
llo
w
in
g
c
o
u
n
tr
ie
s:
A
R
5
A
rg
en
tin
a,
B
O
5
B
ol
iv
ia
,B
R
5
B
ra
zi
l,
C
O
5
C
ol
om
bi
a,
C
R
5
C
os
ta
R
ic
a,
C
U
5
C
ub
a,
D
O
5
D
om
in
ic
a,
E
C
5
E
cu
ad
or
,
G
T
5
G
ua
te
m
al
a,
G
Y
5
G
uy
an
a,
H
O
5
H
on
du
ra
s,
JA
5
Ja
m
ai
ca
,
M
E
5
M
ex
ic
o,
PA
5
Pa
na
m
a,
PE
5
Pe
ru
,P
G
5
Pa
ra
gu
ay
,
T
R
5
Tr
in
id
ad
,V
E
5
V
e
n
e
z
u
e
la
.L
ife
fo
rm
s
o
r
e
c
o
ph
ys
io
lo
gi
ca
lt
yp
es
fo
llo
w
th
e
c
la
ss
ifi
ca
tio
n
o
f
B
en
zi
ng
(2
00
0)
.
Sp
ec
ie
s
O
rig
in
o
f
m
a
te
ria
l
Li
fe
fo
rm
C
ar
bo
n
pa
th
w
ay
Su
rf
ac
e
D
ep
th
o
f
dr
op
le
t(
m
m
)
Su
rf
ac
e
in
ta
ct
Su
rf
ac
e
de
nu
de
d
Su
rf
ac
e
ty
pe
SU
B
FA
M
IL
Y
PI
TC
A
IR
N
IO
ID
EA
E
B
ro
cc
hi
ni
a
c
f.
a
c
u
m
in
at
a
L
.B
.S
m
ith
V
E,
G
ra
n
Sa
ba
na
,S
ie
rr
a
de
L
em
a.
4
C
3?
a
da
xi
al
28
4.
8
6
34
.6
a
53
1.
2
6
74
.5
b
hy
dr
op
hi
lic
a
ba
xi
al
70
6.
4
6
58
.9
a
63
5.
2
6
40
.9
a
n
si
B
ro
cc
hi
ni
a
gi
lm
ar
tin
ia
e
G
.S
.V
a
ra
da
-
ra
jan
V
E,
L
a
E
sc
al
er
a
(1
00
0
m
a
.s
.l.
).
4
C
3?
a
da
xi
al
a
ba
xi
al
44
6.
4
6
45
.6
a
93
5.
1
6
58
.7
b
47
8.
1
6
18
.7
a
66
3.
6
6
32
.2
a
n
si hy
dr
op
ho
bi
c
(w
ax
)
B
ro
cc
hi
ni
a
c
f.
he
ch
tio
id
es
M
ez
V
E,
K
av
an
ey
en
.
4
C
3?
a
da
xi
al
96
4.
4
6
18
.5
b
67
2.
8
6
31
.9
a
hy
dr
op
ho
bi
c
(w
ax
)
a
ba
xi
al
89
5.
9
6
64
.3
b
70
8.
5
6
19
.4
a
hy
dr
op
ho
bi
c
(w
ax
)
B
ro
cc
hi
ni
a
m
ic
ra
nt
ha
(B
ak
er)
M
ez
G
Y,
K
ai
te
ur
Fa
lls
.
4
C
3?
a
da
xi
al
47
6.
4
6
33
.9
a
51
4.
8
6
56
.9
a
n
si
a
ba
xi
al
46
0.
0
6
26
.8
a
49
2.
9
6
34
.0
a
n
si
B
ro
cc
hi
ni
a
re
du
ct
a
B
ak
er
V
E,
H
ac
ie
nd
a
Sa
nt
a
E
le
na
(3
00
m
a
.s
.l.
).
4
C
3?
a
da
xi
al
a
ba
xi
al
98
2.
2
6
24
.3
b
94
0.
6
6
25
.9
b
51
2.
2
6
90
.2
a
62
5.
2
6
58
.5
a
hy
dr
op
ho
bi
c
(w
ax
)
hy
dr
op
ho
bi
c
(w
ax
)
D
eu
te
ro
co
hn
ia
sc
hr
ei
te
ri
A
.C
as
te
l-
la
no
s
A
R
,s
in
lo
c.
1
C
A
M
?
a
da
xi
al
a
ba
xi
al
45
1.
9
6
49
.7
a
94
7.
8
6
47
.9
b
35
1.
9
6
87
.4
a
69
9.
9
6
34
.3
a
n
si hy
dr
op
ho
bi
c
(tr
ich
om
e)
D
yc
ki
a
m
a
rn
ie
r-
la
po
st
ol
le
iL
.B
.
Sm
ith
B
R
,E
st
.M
in
as
G
er
ai
s,
D
ia
m
an
ti-
a
n
a
.
1
C
A
M
?
a
da
xi
al
a
ba
xi
al
32
.5
6
10
2.
2
a
14
17
.9
6
22
.5
b
48
9.
3
6
54
.4
b
80
9.
1
6
69
.1
a
hy
dr
op
hi
lic
hy
dr
op
ho
bi
c
(tr
ich
om
e)
D
yc
ki
a
m
ic
ro
ca
ly
x
B
ak
er
PG
,D
pt
o.
Pa
ra
gu
ar
i,
C
er
ro
A
ca
ha
y
(4
50
m
a
.s
.l.
).
1
C
A
M
?
a
da
xi
al
a
ba
xi
al
63
1.
8
6
22
.2
a
37
9.
8
6
57
.2
a
67
8.
0
6
39
.0
a
38
8.
2
6
65
.4
a
n
si
n
si
F
os
te
re
lla
a
lb
ic
an
s
(G
ris
eb
ac
h)
L
.B
.
Sm
ith
B
O
,D
pt
o.
L
a
Pa
z,
Pr
ov
.
N
or
Y
u
n
-
ga
s
(8
00
m
a
.s
.l.
).
1
C
3?
a
da
xi
al
a
ba
xi
al
45
4.
5
6
45
.9
a
11
71
.7
6
25
.3
b
49
8.
3
6
64
.7
a
77
7.
5
6
11
0.
2
a
n
si hy
dr
op
ho
bi
c
(tr
ich
om
e)
F
os
te
re
lla
c
a
u
le
sc
en
s
R
au
h
B
O
,s
in
lo
c.
1
C
3?
a
da
xi
al
51
2.
8
6
65
.7
a
45
6.
3
6
47
.9
a
n
si
a
ba
xi
al
95
2.
2
6
14
.9
b
66
5.
1
6
41
.0
a
hy
dr
op
ho
bi
c
(tr
ich
om
e)
F
os
te
re
lla
c
f.
e
la
ta
H
.L
ut
he
r
B
O
,D
pt
o.
L
a
Pa
z,
Pr
ov
.
N
or
Y
u
n
-
ga
s
(8
80
m
a
.s
.l.
).
1
C
3?
a
da
xi
al
a
ba
xi
al
25
0.
6
6
50
.6
a
89
9.
9
6
12
.9
b
36
0.
1
6
70
.2
a
62
8.
1
6
48
.7
a
n
si hy
dr
op
ho
bi
c
(tr
ich
om
e)
F
os
te
re
lla
pe
tio
la
ta
(M
ez
)L
.B
.
Sm
ith
B
O
,D
pt
o.
L
a
Pa
z,
Pu
en
te
V
ill
a,
Pu
en
ta
de
C
or
ip
at
a
(1
20
0
m
a
.s
.l.
).
1
C
3?
a
da
xi
al
a
ba
xi
al
72
8.
2
6
96
.9
a
12
26
.5
6
68
.4
b
66
2.
5
6
81
.1
a
89
7.
9
6
42
.4
a
n
si hy
dr
op
ho
bi
c
(tr
ich
om
e)
F
os
te
re
lla
sc
hi
do
sp
er
m
a
(B
ak
er)
L
.B
.S
m
ith
B
O
,s
in
lo
c.
1
C
3?
a
da
xi
al
a
ba
xi
al
23
2.
8
6
53
.7
a
85
8.
6
6
59
.8
b
40
9.
2
6
71
.4
a
62
3.
4
6
39
.9
a
n
si hy
dr
op
ho
bi
c
(tr
ich
om
e)
F
os
te
re
lla
sp
.n
o
v
.
B
O
,D
pt
o.
L
a
Pa
z,
Pr
ov
.
M
un
ec
as
,
C
on
sa
ta
(1
20
0
m
a
.s
.l.
).
1
?
a
da
xi
al
a
ba
xi
al
50
6.
1
6
44
.1
a
52
3.
3
6
68
.6
a
48
4.
6
6
10
9.
8
a
54
4.
8
6
70
.2
a
n
si
n
si
H
ec
ht
ia
gu
at
em
al
en
si
s
M
ez
H
O
,T
e
gu
ic
ig
al
pa
-C
om
ay
ag
ua
(1
30
0
m
a
.s
.l.
).
1
C
A
M
?
a
da
xi
al
a
ba
xi
al
19
2.
0
6
69
.7
a
13
18
.9
6
34
.7
b
50
4.
7
6
40
.3
b
91
1.
8
6
88
.5
a
hy
dr
op
hi
lic
hy
dr
op
ho
bi
c
(tr
ich
om
e)
P
itc
ai
rn
ia
a
rc
u
a
ta
(A
nd
re?)
A
nd
re?
PA
,P
ro
v.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
1
C
3
a
da
xi
al
a
ba
xi
al
10
30
.3
6
42
.6
b
83
2.
5
6
45
.4
a
71
4.
3
6
60
.2
a
87
5.
5
6
76
.6
a
hy
dr
op
ho
bi
c
(tr
ich
om
e)
n
si
P
itc
ai
rn
ia
a
tr
or
ub
en
s
(B
ee
r)
B
ak
er
C
R
,P
ro
v.
C
ar
ta
go
,L
a
Su
iz
a
(1
00
0
m
a
.s
.l.
).
1
C
3?
a
da
xi
al
a
ba
xi
al
21
1.
8
6
14
.6
a
48
4.
8
6
46
.2
a
27
4.
8
6
46
.7
a
58
1.
4
6
68
.1
a
n
si
n
si
P
itc
ai
rn
ia
c
o
ra
lli
na
L
in
de
n
&
A
nd
re?
V
E,
E
do
.T
a
c
hi
ra
(1
20
0?
15
00
m
a
.s
.l.
).
1
C
3?
a
da
xi
al
a
ba
xi
al
62
8.
1
6
59
.6
a
89
9.
9
6
16
.7
b
58
2.
8
6
10
3.
6
a
73
3.
8
6
27
.1
a
n
si hy
dr
op
ho
bi
c
(tr
ich
om
e)
P
itc
ai
rn
ia
e
c
hi
na
ta
H
oo
ke
r
PE
,s
in
lo
c.
1
C
3?
a
da
xi
al
53
6.
2
6
44
.7
a
60
5.
0
6
36
.1
a
n
si
a
ba
xi
al
98
3.
4
6
8.
6
b
77
7.
0
6
25
.8
a
hy
dr
op
ho
bi
c
(tr
ich
om
e)
P
itc
ai
rn
ia
im
br
ic
at
a
(B
ro
ng
nia
rt)
R
eg
el
V
E,
si
n
lo
c.
1
C
3?
a
da
xi
al
a
ba
xi
al
51
3.
9
6
24
.3
a
42
9.
5
6
40
.0
a
51
0.
3
6
18
.9
a
54
3.
4
6
28
.4
a
n
si
n
si
P
itc
ai
rn
ia
in
te
gr
ifo
lia
K
er
-
G
aw
le
r*
T
R
,P
oi
nt
G
ou
rd
e,
se
a
so
n
a
lly
dr
y.
1
C
3?
a
da
xi
al
25
8.
6
6
25
.2
a
30
6.
9
6
13
.5
a
n
si
a
ba
xi
al
71
8.
9
6
17
.8
b
48
2.
8
6
20
.9
a
hy
dr
op
ho
bi
c
(tr
ich
om
e)
August 2001] 1381PIERCE ET AL.?WATER REPELLENT INDUMENTA OF BROMELIADS
T
A
B
LE
3.
C
on
tin
ue
d.
Sp
ec
ie
s
O
rig
in
o
f
m
a
te
ria
l
Li
fe
fo
rm
C
ar
bo
n
pa
th
w
ay
Su
rf
ac
e
D
ep
th
o
f
dr
op
le
t(
m
m
)
Su
rf
ac
e
in
ta
ct
Su
rf
ac
e
de
nu
de
d
Su
rf
ac
e
ty
pe
P
itc
ai
rn
ia
m
a
id
ifo
lia
(C
.M
or
re
n)
D
ec
ai
sn
e
PE
,D
pt
o.
Sa
n
M
ar
tin
,T
a
ra
po
to
-Y
u
-
rim
ag
ua
s
(1
00
0
m
a
.s
.l.
).
1
C
3?
a
da
xi
al
a
ba
xi
al
44
4.
2
6
12
.7
a
49
5.
6
6
55
.6
a
52
1.
3
6
61
.6
a
59
1.
2
6
51
.8
a
n
si
n
si
P
itc
ai
rn
ia
m
ic
ro
tr
in
en
si
s
R
.W
.
R
ea
d
D
O
(1
25
0
m
a
.s
.l.
).
1
C
3?
a
da
xi
al
29
5.
4
6
74
.5
a
44
1.
6
6
66
.6
a
n
si
a
ba
xi
al
63
0.
8
6
30
.6
a
64
8.
0
6
18
.6
a
n
si
P
itc
ai
rn
ia
pa
lm
oi
de
s
M
ez
&
So
di
ro
E
C
,P
ro
v.
C
ar
ch
i,
C
hi
ca
l(
13
00
m
a
.s
.l.
).
1
C
3?
a
da
xi
al
a
ba
xi
al
49
7.
3
6
44
.6
a
79
4.
8
6
52
.2
a
55
3.
8
6
34
.0
a
79
1.
8
6
39
.8
a
n
si
n
si
P
itc
ai
rn
ia
pa
te
nt
iflo
ra
L
.B
.S
m
ith
Si
n
lo
c.
1
C
3(C
AM
)?
a
da
xi
al
36
3.
3
6
33
.4
a
36
3.
3
6
42
.0
a
n
si
a
ba
xi
al
46
9.
9
6
21
.1
a
51
0.
3
6
25
.5
a
n
si
P
itc
ai
rn
ia
re
c
u
rv
a
ta
(S
ch
eid
we
ile
r)
K
.K
oc
h
M
E
,E
do
.V
e
ra
c
ru
z
,
Pl
ay
a
E
sc
on
di
-
do
.
1
C
3?
a
da
xi
al
a
ba
xi
al
47
6.
4
6
64
.1
a
12
86
.7
6
38
.3
b
54
7.
6
6
34
.0
a
86
5.
1
6
28
.3
a
n
si hy
dr
op
ho
bi
c
(tr
ich
om
e)
P
itc
ai
rn
ia
ri
pa
ri
a
M
ez
E
C
,M
or
on
a-
Sa
nt
ia
go
.
1
C
3?
a
da
xi
al
46
4.
7
6
58
.0
a
49
5.
4
6
53
.6
a
n
si
a
ba
xi
al
48
3.
1
6
34
.8
a
52
0.
0
6
40
.8
a
n
si
P
itc
ai
rn
ia
ru
br
on
ig
ri
flo
ra
R
au
h
PE
,D
pt
o.
Sa
n
M
ar
tin
,T
a
ra
po
to
(1
00
0
m
a
.s
.l.
).
1
C
3?
a
da
xi
al
a
ba
xi
al
52
2.
4
6
22
.3
a
65
8.
1
6
68
.1
a
54
6.
0
6
56
.7
a
71
1.
2
6
62
.7
a
n
si
n
si
P
itc
ai
rn
ia
u
n
du
la
ta
Sc
he
id
w
ei
le
r
M
E
,E
do
.C
hi
ap
as
.
1
C
3?
a
da
xi
al
31
5.
5
6
12
.6
a
39
2.
7
6
23
.2
b
hy
dr
op
hi
lic
a
ba
xi
al
84
4.
7
6
39
.1
b
67
9.
4
6
24
.4
a
hy
dr
op
ho
bi
c
(tr
ich
om
e)
P
itc
ai
rn
ia
u
n
ila
te
ra
lis
L
.B
.S
m
ith
E
C
,P
ro
v.
M
an
ab
i.
1
C
3?
a
da
xi
al
38
2.
7
6
68
.4
a
48
4.
6
6
89
.6
a
n
si
a
ba
xi
al
89
2.
3
6
9.
5
b
71
4.
9
6
38
.0
a
hy
dr
op
ho
bi
c
(tr
ich
om
e)
P
itc
ai
rn
ia
v
a
le
ri
iS
ta
nd
le
y
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
1
C
3?
a
da
xi
al
a
ba
xi
al
64
7.
4
6
70
.6
a
55
5.
8
6
24
.3
a
61
8.
5
6
11
9
a
75
8.
4
6
86
.7
a
n
si
n
si
P
uy
a
c
te
no
rh
yn
ch
a
L
.B
.S
m
ith
B
O
,D
pt
o.
L
a
Pa
z,
Pr
ov
.
L
ar
ec
aja
,
C
er
ro
Im
in
ap
i,
So
ra
ta
(2
68
0
m
a
.s
.l.
).
1
C
A
M
?
a
da
xi
al
a
ba
xi
al
24
0.
0
6
27
.3
a
45
4.
5
6
83
.1
a
31
6.
1
6
40
.3
a
44
6.
7
6
36
.9
a
n
si
n
si
P
uy
a
la
na
ta
K
un
th
Si
n
lo
c.
1
C
A
M
?
a
da
xi
al
14
.7
6
8.
6
a
23
2.
2
6
35
.2
b
hy
dr
op
hi
lic
a
ba
xi
al
70
.7
6
70
.7
a
34
7.
3
6
38
.3
b
hy
dr
op
hi
lic
P
uy
a
la
xa
L
.B
.S
m
ith
O
ld
ho
rt.
pl
an
t.
1
C
A
M
?
a
da
xi
al
56
3.
1
6
10
5.
9
a
48
9.
3
6
15
.5
a
n
si
a
ba
xi
al
53
8.
5
6
82
.4
a
62
4.
6
6
29
.1
a
n
si
SU
B
FA
M
IL
Y
T
IL
LA
N
D
SI
O
ID
EA
E
Al
ca
nt
ar
ea
c
f.
o
do
ra
ta
(L
em
e)
J.
R
.
G
ra
nt
B
R
,s
in
lo
c.
4
C
3(C
AM
)?
a
da
xi
al
a
ba
xi
al
92
6.
1
6
35
.8
b
10
73
.9
6
4.
4
b
63
8.
9
6
52
.9
a
74
7.
7
6
52
.7
a
hy
dr
op
ho
bi
c
(w
ax
)
hy
dr
op
ho
bi
c
(w
ax
)
C
at
op
si
s
m
ic
ra
nt
ha
L
.B
.S
m
ith
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
4
C
3
a
da
xi
al
a
ba
xi
al
18
70
.5
6
14
9.
2
b
19
71
.5
6
32
.8
b
10
72
.5
6
7.
2
a
10
09
.3
6
71
.7
a
hy
dr
op
ho
bi
c
(w
ax
)
hy
dr
op
ho
bi
c
(w
ax
)
C
at
op
si
s
n
iti
da
(H
oo
ke
r)
G
ris
eb
ac
h
PA
,
Pr
ov
.
C
hi
riq
ui
,F
or
tu
na
,l
ow
er
m
o
n
ta
ne
w
e
t
fo
re
st
(1
20
0
m
a
.s
.l.
).
4
C
3
a
da
xi
al
a
ba
xi
al
61
8.
9
6
15
.8
a
64
0.
8
6
19
.3
a
67
3.
7
6
35
.5
a
60
7.
9
6
13
.8
a
n
si
n
si
C
at
op
si
s
n
u
ta
ns
(S
wa
rtz
)G
ris
eb
ac
h
H
O
,P
ro
v.
C
or
te
s,
Sa
n
Pe
dr
o,
Su
la
(1
00
m
a
.s
.l.
).
4
C
3(C
AM
)?
a
da
xi
al
a
ba
xi
al
85
4.
2
6
48
.2
b
42
7.
2
6
47
.7
a
56
9.
5
6
19
.9
a
46
5.
5
6
41
.3
a
hy
dr
op
ho
bi
c
(w
ax
)
n
si
C
at
op
si
s
se
ss
ili
flo
ra
(R
uiz
&
Pa
vo
n)
M
ez
B
O
,D
pt
o.
L
a
Pa
z,
Pr
ov
.
L
ar
ec
aja
,
Ti
pu
an
i-C
ar
an
av
i(
12
50
m
a
.s
.l.
).
4
C
3?
a
da
xi
al
a
ba
xi
al
95
8.
2
6
30
.2
b
54
7.
6
6
8.
9
a
69
0.
0
6
19
.9
a
49
8.
3
6
54
.7
a
hy
dr
op
ho
bi
c
(w
ax
)
n
si
C
at
op
si
s
su
bu
la
ta
L
.B
.S
m
ith
G
T,
bo
ug
ht
in
m
a
rk
et
in
G
ua
te
m
al
a
C
ity
.
4
C
3?
a
da
xi
al
a
ba
xi
al
42
2.
9
6
53
.9
a
82
4.
5
6
7.
5
b
50
3.
2
6
70
.6
a
68
4.
7
6
29
.9
a
n
si hy
dr
op
ho
bi
c
(w
ax
)
G
uz
m
an
ia
c
ir
ci
nn
at
a
R
au
h
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
4
C
3
a
da
xi
al
a
ba
xi
al
84
8.
2
6
70
.9
a
83
1.
3
6
44
.4
a
82
2.
8
6
92
.6
a
91
5.
8
6
88
.8
a
n
si
n
si
G
uz
m
an
ia
c
o
ri
os
ta
ch
ya
(G
ris
eb
ac
h)
M
ez
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
4
C
3
a
da
xi
al
a
ba
xi
al
67
1.
6
6
16
.7
a
68
6.
0
6
29
.9
a
74
3.
9
6
28
.8
a
75
3.
6
6
48
.2
a
n
si
n
si
G
uz
m
an
ia
m
a
c
ro
po
da
L
.B
.S
m
ith
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
4
C
3?
a
da
xi
al
a
ba
xi
al
34
2.
7
6
9.
4
a
73
7.
6
6
11
.2
b
35
3.
9
6
18
.2
a
52
1.
5
6
45
.7
a
n
si hy
dr
op
ho
bi
c
(w
ax
)
G
uz
m
an
ia
m
o
n
o
st
ac
hi
a
(L
.)
R
us
by
e
x
M
ez
.v
a
r.
m
o
n
o
st
ac
hi
a*
PA
,
C
oc
le?
pr
ov
in
ce
,M
at
a
A
ho
ga
do
,
m
o
n
ta
ne
w
e
t
fo
re
st
(1
00
0
m
a
.s
.l.
).
4
C
3-
C
A
M
?
a
da
xi
al
a
ba
xi
al
66
2.
3
6
86
.6
a
89
0.
4
6
11
9.
5
a
81
4.
4
6
68
.1
a
99
1.
8
6
10
7.
8
a
n
si
n
si
G
uz
m
an
ia
m
u
sa
ic
a
(L
ind
en
&
A
n-
dr
e?)
M
ez
v
a
r.
m
u
sa
ic
a
PA
,
Pr
ov
.
C
hi
riq
ui
,F
or
tu
na
,l
ow
er
m
o
n
ta
ne
w
e
t
fo
re
st
(1
20
0
m
a
.s
.l.
).
4
C
3?
a
da
xi
al
a
ba
xi
al
59
7.
0
6
22
.7
a
60
4.
3
6
10
.3
a
70
2.
9
6
75
.5
a
61
8.
9
6
15
.8
a
n
si
n
si
1382 [Vol. 88AMERICAN JOURNAL OF BOTANY
T
A
B
LE
3.
C
on
tin
ue
d.
Sp
ec
ie
s
O
rig
in
o
f
m
a
te
ria
l
Li
fe
fo
rm
C
ar
bo
n
pa
th
w
ay
Su
rf
ac
e
D
ep
th
o
f
dr
op
le
t(
m
m
)
Su
rf
ac
e
in
ta
ct
Su
rf
ac
e
de
nu
de
d
Su
rf
ac
e
ty
pe
G
uz
m
an
ia
m
u
sa
ic
a
(L
ind
en
&
A
n-
dr
e?)
M
ez
v
a
r.
di
sc
ol
or
H
.L
ut
he
r
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
4
C
3
a
da
xi
al
a
ba
xi
al
66
3.
3
6
44
.2
a
55
1.
8
6
36
.1
a
74
1.
8
6
63
.3
a
51
9.
0
6
10
5.
9
a
n
si
n
si
G
uz
m
an
ia
re
tu
sa
L
.B
.S
m
ith
PE
,D
pt
o.
Sa
n
M
ar
tin
,T
a
ra
po
to
-Y
u
-
rim
ag
ua
s
(1
20
0
m
a
.s
.l.
).
4
C
3?
a
da
xi
al
a
ba
xi
al
47
6.
4
6
72
.7
a
95
5.
2
6
21
.0
b
30
8.
4
6
45
.7
a
35
0.
4
6
86
.9
a
n
si hy
dr
op
ho
bi
c
(w
ax
)
R
ac
in
ae
a
sp
ic
ul
os
a
(G
ris
eb
ac
h)
M
.A
.S
pe
nc
er
&
L
.B
.S
m
ith
v
a
r.
sp
ic
ul
os
a
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
4
C
3
a
da
xi
al
a
ba
xi
al
84
2.
8
6
92
.2
a
90
1.
8
6
18
.8
a
99
3.
8
6
57
.1
a
98
3.
5
6
43
.1
a
n
si
n
si
Ti
lla
nd
si
a
c
a
u
lig
er
a
M
ez
PE
,D
pt
o.
Qo
sq
o,
O
lla
nt
ay
ta
m
bo
,
sa
x
ic
ol
ou
s
(2
70
0
m
a
.s
.l.
).
4?
5
C
A
M
?
a
da
xi
al
a
ba
xi
al
5.
7
6
5.
9
a
19
9.
2
6
16
5.
5
a
41
4.
5
6
14
7.
6
b
65
9.
1
6
48
.9
b
hy
dr
op
hi
lic
hy
dr
op
hi
lic
Ti
lla
nd
si
a
e
lo
ng
at
a
K
un
th
v
a
r.
su
b-
im
br
ic
at
a
(B
ak
er)
L
.B
.S
m
ith
PA
,
Pr
ov
.
Pa
na
m
a?,
G
am
bo
a,
lo
w
-
la
nd
se
a
so
n
a
lly
dr
y
fo
re
st
.
4
C
A
M
?
a
da
xi
al
a
ba
xi
al
81
7.
5
6
22
.4
a
76
8.
3
6
31
.3
a
10
28
.0
6
40
.4
b
95
8.
5
6
37
.6
b
hy
dr
op
hi
lic
hy
dr
op
hi
lic
Ti
lla
nd
si
a
fle
xu
os
a
Sw
ar
tz
*
PA
,
Pr
ov
.
Pa
na
m
a?,
Pa
na
m
a
C
ity
,
C
er
ro
A
nc
on
,l
ow
la
nd
u
rb
an
,s
e
a
-
so
n
a
lly
dr
y.
4?
5
C
A
M
?
a
da
xi
al
a
ba
xi
al
34
9.
0
6
79
.2
a
33
4.
1
6
11
1.
3
a
64
1.
4
6
10
.8
b
59
3.
8
6
14
.2
b
hy
dr
op
hi
lic
hy
dr
op
hi
lic
Ti
lla
nd
si
a
n
a
n
a
B
ak
er
PE
,D
pt
o.
Qo
sq
o,
O
lla
nt
ay
ta
m
bo
,
se
a
so
n
a
lly
dr
y,
sa
x
ic
ol
ou
s
o
n
ro
c
k-
fa
ce
s
(2
70
0
m
a
.s
.l)
.
5
C
A
M
?
a
da
xi
al
a
ba
xi
al
0.
0
6
0.
0
a
9.
4
6
12
0.
9
a
73
2.
1
6
36
.4
b
88
5.
4
6
40
.6
b
hy
dr
op
hi
lic
hy
dr
op
hi
lic
Ti
lla
nd
is
ia
st
ri
ct
a
So
la
nd
er
v
a
r.
st
ri
ct
a
T
R
,s
in
lo
c.
5
C
A
M
?
a
da
xi
al
a
ba
xi
al
0.
0
6
0.
0
a
25
.1
6
54
.0
a
32
1.
3
6
49
.7
b
57
9.
4
6
57
.7
b
hy
dr
op
hi
lic
hy
dr
op
hi
lic
Vr
ie
se
a
m
o
n
st
ru
m
(M
ez
)L
.B
.S
m
ith
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
4
C
3
a
da
xi
al
a
ba
xi
al
83
3.
5
6
44
.6
a
93
8.
3
6
17
.1
b
77
1.
0
6
10
5.
3
a
71
4.
8
6
54
.9
a
n
si hy
dr
op
ho
bi
c
(tr
ich
om
e)
W
er
au
hi
a
c
a
pi
ta
ta
(M
ez
&
W
e
rc
kl
e?)
J.
R
.G
ra
nt
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
4
C
3
a
da
xi
al
a
ba
xi
al
83
1.
3
6
46
.7
a
91
5.
5
6
72
.0
b
80
4.
3
6
46
.7
a
56
2.
0
6
65
.9
a
n
si hy
dr
op
ho
bi
c
(w
ax
)
W
er
au
hi
a
gl
ad
io
lifl
or
a
(W
e
n
dl
an
d)
J.
R
.G
ra
nt
M
E
,E
do
.C
hi
ap
as
(1
00
0
m
a
.s
.l.
).
4
C
3?
a
da
xi
al
a
ba
xi
al
26
8.
7
6
87
.5
a
45
1.
6
6
27
.9
a
52
8.
3
6
10
3.
9
a
54
0.
1
6
58
.6
a
n
si
n
si
W
er
au
hi
a
hy
gr
om
et
ri
ca
(A
nd
re?)
J.
R
.
G
ra
nt
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
4
C
3
a
da
xi
al
a
ba
xi
al
88
5.
5
6
51
.7
a
10
39
.5
6
68
.9
a
86
9.
5
6
96
.4
a
87
5.
0
6
21
.9
a
n
si
n
si
W
er
au
hi
a
pa
na
m
ae
ns
is
(E
.G
ro
ss
&
R
au
h)
J.
R
.G
ra
nt
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
4
C
3
a
da
xi
al
a
ba
xi
al
78
6.
8
6
52
.8
a
56
4.
5
6
83
.9
a
82
5.
3
6
49
.2
a
64
1.
0
6
53
.8
a
n
si
n
si
W
er
au
hi
a
sa
n
gu
in
ol
en
ta
(L
ind
en
e
x
C
og
ni
au
x
&
M
ar
ch
al
)J
.R
.G
ra
nt
PA
,
Pr
ov
.
Pa
na
m
a?,
G
am
bo
a,
lo
w
-
la
nd
se
a
so
n
a
lly
dr
y
fo
re
st
,
4
C
3
a
da
xi
al
a
ba
xi
al
91
1.
8
6
45
.7
a
10
20
.8
6
28
.6
a
96
6.
0
6
54
.1
a
10
38
.0
6
34
.0
a
n
si
n
si
W
er
au
hi
a
v
itt
at
a
(M
ez
&
W
e
rc
kl
e?)
J.
R
.G
ra
nt
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
4
C
3
a
da
xi
al
a
ba
xi
al
85
9.
0
6
59
.5
a
81
0.
0
6
94
.9
a
91
8.
8
6
52
.3
a
89
2.
5
6
36
.4
a
n
si
n
si
SU
B
FA
M
IL
Y
B
R
O
M
EL
IO
ID
EA
E
Ae
ch
m
ea
da
ct
yl
in
a
B
ak
er
*
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
3
C
A
M
a
da
xi
al
a
ba
xi
al
49
0.
2
6
56
.7
a
51
5.
5
6
14
.5
a
70
9.
3
6
18
.7
b
64
7.
2
6
22
.2
b
hy
dr
op
hi
lic
hy
dr
op
hi
lic
Ae
ch
m
ea
fas
cia
ta
(L
ind
ley
)B
ak
er
B
R
,s
in
lo
c.
3
C
A
M
?
a
da
xi
al
61
7.
5
6
24
.5
a
60
2.
3
6
38
.1
a
n
si
a
ba
xi
al
79
5.
9
6
41
.7
b
58
7.
2
6
20
.9
a
hy
dr
op
ho
bi
c
(tr
ich
om
e)
Ae
ch
m
ea
fen
dle
ri
A
nd
re?
e
x
M
ez
T
R
,T
e
x
te
l,
tr
an
si
tio
na
lm
o
n
ta
ne
fo
re
st
(7
10
m
a
.s
.l.
)
3
C
A
M
?
a
da
xi
al
a
ba
xi
al
34
1.
2
6
38
.7
a
48
9.
4
6
45
.0
a
47
7.
3
6
20
.4
b
60
1.
3
6
37
.9
b
hy
dr
op
hi
lic
hy
dr
op
hi
lic
Ae
ch
m
ea
m
a
gd
al
en
ae
(A
nd
re?)
A
nd
re?
e
x
B
ak
er
*
PA
,
Pr
ov
.
Pa
na
m
a?,
B
ar
ro
C
ol
or
ad
o
Is
la
nd
,s
ha
de
d
u
n
de
rs
to
ry
.
2
C
A
M
?
a
da
xi
al
a
ba
xi
al
31
0.
5
6
86
.4
a
59
1.
0
6
61
.6
a
52
3.
5
6
54
.3
b
60
6.
7
6
46
.1
a
hy
dr
op
hi
lic
n
si
Ae
ch
m
ea
n
u
di
ca
ul
is
(L
.)
G
ris
eb
ac
h
PA
,
Pr
ov
.
B
oc
as
de
lT
o
ro
,
C
hi
riq
ui
G
ra
nd
e
(3
m
a
.s
.l.
).
3
C
A
M
?
a
da
xi
al
a
ba
xi
al
21
1.
1
6
10
.5
a
74
0.
5
6
77
.3
b
47
8.
1
6
64
.5
b
51
4.
3
6
46
.9
a
hy
dr
op
hi
lic
hy
dr
op
ho
bi
c
(tr
ich
om
e)
Ae
ch
m
ea
v
e
itc
hi
iB
ak
er
*
PA
,
Pr
ov
.
C
hi
riq
ui
,F
or
tu
na
,l
ow
er
m
o
n
ta
ne
w
e
t
fo
re
st
(1
20
0
m
a
.s
.l.
).
1
C
3?
a
da
xi
al
a
ba
xi
al
60
5.
6
6
47
.7
a
10
50
.3
6
11
.9
b
67
1.
4
6
43
.1
a
85
3.
1
6
24
.5
a
n
si hy
dr
op
ho
bi
c
(tr
ich
om
e)
An
an
as
c
o
m
o
su
s
(L
.)
M
er
ril
c
v
.
C
ay
en
ne
L
is
se
*
A
gr
ic
ul
tu
ra
lc
lo
ne
.
2
C
A
M
a
da
xi
al
a
ba
xi
al
61
9.
5
6
36
.5
a
84
2.
2
6
13
.2
b
66
7.
7
6
11
.0
a
68
5.
3
6
12
.4
a
n
si hy
dr
op
ho
bi
c
(tr
ich
om
e)
B
ill
be
rg
ia
m
a
c
ro
le
pi
s
L
.B
.S
m
ith
PA
,
Pr
ov
.
Pa
na
m
a?,
G
am
bo
a,
lo
w
-
la
nd
se
a
so
n
a
lly
dr
y
fo
re
st
.
2?
3
C
A
M
a
da
xi
al
a
ba
xi
al
43
8.
3
6
10
8.
2
a
11
30
.3
6
16
.6
b
49
8.
8
6
22
.9
a
34
1.
8
6
11
6.
1
a
n
si hy
dr
op
ho
bi
c
(tr
ich
om
e)
August 2001] 1383PIERCE ET AL.?WATER REPELLENT INDUMENTA OF BROMELIADS
T
A
B
LE
3.
C
on
tin
ue
d.
Sp
ec
ie
s
O
rig
in
o
f
m
a
te
ria
l
Li
fe
fo
rm
C
ar
bo
n
pa
th
w
ay
Su
rf
ac
e
D
ep
th
o
f
dr
op
le
t(
m
m
)
Su
rf
ac
e
in
ta
ct
Su
rf
ac
e
de
nu
de
d
Su
rf
ac
e
ty
pe
B
ill
be
rg
ia
ro
be
rt
-r
ea
di
iE
.G
ro
ss
&
R
au
h
PE
,D
pt
o.
M
ad
re
de
D
io
s,
H
ac
ie
nd
a
Sa
lv
ac
io?
n
,
n
e
a
r
Qu
inc
em
il,
e
pi
-
ph
yt
ic
in
fo
re
st
.
3
C
A
M
?
a
da
xi
al
a
ba
xi
al
25
6.
6
6
11
6.
6
a
10
09
.3
6
43
.9
b
68
1.
2
6
51
.7
b
58
9.
5
6
11
4.
5
a
hy
dr
op
hi
lic
hy
dr
op
ho
bi
c
(tr
ich
om
e)
B
ill
be
rg
ia
ro
se
a
H
or
tu
s
e
x
B
ee
r
V
E,
si
n
lo
c.
(D
ist
rib
uti
on
:V
E?
TR
).
3
C
A
M
?
a
da
xi
al
23
3.
3
6
74
.9
a
54
1.
8
6
13
.4
b
hy
dr
op
hi
lic
a
ba
xi
al
57
8.
1
6
28
.9
a
61
1.
4
6
17
.8
a
n
si
B
ill
be
rg
ia
st
en
op
et
al
a
H
ar
m
s
E
C
,P
ro
v.
N
ap
o,
Tu
la
g.
3
C
A
M
?
a
da
xi
al
38
6.
1
6
22
3.
7
a
35
8.
5
6
12
5.
4
a
n
si
a
ba
xi
al
78
6.
3
6
31
.6
b
47
9.
2
6
52
.5
a
hy
dr
op
ho
bi
c
(tr
ich
om
e)
B
ro
m
el
ia
pi
ng
ui
n
L
.*
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
A
zu
l,
tr
op
ic
al
w
e
t
fo
re
st
(6
91
m
a
.s
.l.
).
2
C
A
M
?
a
da
xi
al
a
ba
xi
al
72
5.
8
6
43
.6
a
76
1.
3
6
60
.2
a
78
8.
5
6
22
.5
a
79
5.
2
6
53
.0
a
n
si
n
si
C
an
is
tr
um
se
id
el
ia
nu
m
W
e
be
r
B
R
,e
x
ho
rt
.
3
C
A
M
?
a
da
xi
al
16
9.
8
6
62
.9
a
42
6.
9
6
51
.2
b
hy
dr
op
hi
lic
a
ba
xi
al
75
.9
6
54
.7
a
53
5.
8
6
29
.8
b
hy
dr
op
hi
lic
C
ry
pt
an
th
us
c
f.
br
om
el
io
id
es
O
tto
&
D
ie
tri
ch
B
R
,E
st
.R
io
de
Ja
ne
iro
,B
ar
ra
de
Ti
juc
a,
de
ns
e
fo
re
st
o
n
hi
lls
id
e,
c
la
y
a
n
d
le
af
-li
tte
r
su
bs
tra
te
(3
0
m
a
.s
.l.
).
1
C
A
M
?
a
da
xi
al
a
ba
xi
al
0.
0
6
0.
0
a
64
6.
1
6
58
.6
a
38
8.
8
6
49
.3
b
76
1.
1
6
35
.1
a
hy
dr
op
hi
lic
n
si
C
ry
pt
an
th
us
di
an
ae
L
em
e
B
R
,E
st
.E
sp
iri
to
Sa
nt
o,
Pr
es
id
en
te
K
en
ne
dy
,
Pr
ai
a
de
M
ar
ob
a.
1
C
A
M
?
a
da
xi
al
a
ba
xi
al
18
.6
6
15
.5
a
51
3.
0
6
33
.3
b
13
4.
7
6
18
.9
b
34
3.
4
6
57
.4
a
hy
dr
op
hi
lic
hy
dr
op
ho
bi
c
(tr
ich
om
e)
C
ry
pt
an
th
us
gl
az
io
vi
iM
ez
B
R
,E
st
.M
in
as
G
er
ai
s,
C
ar
ac
a
(1
00
0?
12
00
m
a
.s
.l.
).
1
C
3?
a
da
xi
al
a
ba
xi
al
36
0.
4
6
50
.3
a
76
3.
2
6
47
.6
a
37
8.
5
6
51
.9
a
81
7.
5
6
30
.9
a
n
si
n
si
C
ry
pt
an
th
us
w
a
ra
si
iE
.P
er
ei
ra
B
R
,E
do
.M
in
as
G
er
ai
s,
v
ic
.D
ia
-
m
a
n
tin
a.
1
C
A
M
?
a
da
xi
al
a
ba
xi
al
0.
0
6
0.
0
a
99
8.
5
6
37
.3
b
63
1.
9
6
65
.8
b
78
1.
3
6
11
.7
a
hy
dr
op
hi
lic
hy
dr
op
ho
bi
c
(tr
ich
om
e)
C
ry
pt
an
th
us
w
hi
tm
an
ii
L
em
e
B
R
,E
st
.E
sp
iri
to
Sa
nt
o,
D
om
in
go
s
M
ar
tin
s.
1
?
a
da
xi
al
a
ba
xi
al
17
7.
6
6
61
.2
a
78
2.
9
6
17
.9
b
38
2.
7
6
44
.5
b
37
5.
2
6
49
.9
a
hy
dr
op
hi
lic
hy
dr
op
ho
bi
c
(tr
ich
om
e)
H
oh
en
be
rg
ia
pe
nd
ul
afl
or
a
(A
.R
ic
h-
a
rd
)M
ez
W
e
st
In
di
es
,s
in
lo
c.
3
C
A
M
?
a
da
xi
al
a
ba
xi
al
36
3.
1
6
99
.4
a
99
4.
4
6
41
.3
a
77
0.
2
6
83
.6
b
88
2.
3
6
49
.6
a
hy
dr
op
hi
lic
n
si
N
eo
re
ge
lia
c
ru
e
n
ta
(R
.G
ra
ha
m
)
L
.B
.S
m
ith
*
B
R
,E
st
.R
io
de
Ja
ne
iro
,n
e
a
r
se
a
le
ve
l.
3
C
A
M
?
a
da
xi
al
a
ba
xi
al
56
2.
0
6
44
.9
a
60
8.
4
6
36
.8
a
54
3.
9
6
47
.7
a
58
2.
2
6
37
.9
a
n
si
n
si
O
rt
ho
ph
yt
um
be
nz
in
gi
iL
em
e
&
H
.
L
ut
he
r
B
R
,E
st
.M
in
as
G
er
ai
s,
lit
ho
ph
yt
e,
pa
rti
al
sh
ad
e
(4
50
m
a
.s
.l.
).
1
C
3(C
AM
)?
a
da
xi
al
a
ba
xi
al
0.
0
6
0.
0
a
10
79
.7
6
97
.6
b
47
7.
0
6
38
.1
b
60
6.
1
6
76
.5
a
hy
dr
op
hi
lic
hy
dr
op
ho
bi
c
(tr
ich
om
e)
O
rt
ho
ph
yt
um
gu
rk
en
ii
H
ut
ch
is
on
B
R
,s
in
lo
c.
1
C
A
M
?
a
da
xi
al
47
.7
6
47
.7
a
28
6.
6
6
27
.0
b
hy
dr
op
hi
lic
a
ba
xi
al
13
8.
9
6
11
1.
9
a
43
8.
9
6
38
.5
b
hy
dr
op
hi
lic
O
rt
ho
ph
yt
um
m
a
ga
lh
ae
si
iL
.B
.
Sm
ith
B
R
,E
st
.B
ah
ia
.
1
C
A
M
?
a
da
xi
al
a
ba
xi
al
0.
0
6
0.
0
a
0.
0
6
0.
0
a
28
6.
8
6
94
.2
b
26
5.
3
6
87
.8
b
hy
dr
op
hi
lic
hy
dr
op
hi
lic
Qu
es
ne
lia
bl
an
da
(S
ch
ott
e
x
B
ee
r)
M
ez
B
R
,E
st
.R
io
de
Ja
ne
iro
.
3
C
A
M
?
a
da
xi
al
a
ba
xi
al
33
0.
1
6
39
.9
a
42
0.
8
6
47
.5
a
52
3.
7
6
10
.3
b
62
6.
5
6
9.
1
b
hy
dr
op
hi
lic
hy
dr
op
hi
lic
Qu
es
ne
lia
m
a
rm
o
ra
ta
(L
em
air
e)
R
.W
.
R
ea
d
c
v
.
Ti
m
Pl
ow
m
an
B
R
,E
st
.R
io
de
Ja
ne
iro
,R
io
B
on
ito
.
3
C
A
M
?
a
da
xi
al
a
ba
xi
al
50
3.
4
6
76
.5
a
72
7.
6
6
42
.9
a
55
8.
6
6
48
.9
a
64
8.
3
6
37
.1
a
n
si
n
si
R
on
nb
er
gi
a
e
x
pl
od
en
s
L
.B
.S
m
ith
PA
,
Pr
ov
.
Pa
na
m
a?,
C
er
ro
Je
fe
,e
lfi
n
c
lo
ud
fo
re
st
(1
00
7
m
a
.s
.l.
).
1
C
3
a
da
xi
al
a
ba
xi
al
27
5.
7
6
51
.4
a
66
6.
8
6
13
.6
b
42
0.
95
6
41
.0
a
56
6.
2
6
36
.7
a
n
si hy
dr
op
ho
bi
c
(tr
ich
om
e)
1384 [Vol. 88AMERICAN JOURNAL OF BOTANY
Figs. 18?25. Epicuticular wax powder layers of leaf blade surfaces of bromeliads. 18. Catopsis micrantha, photograph of water droplets on adaxial surface
of leaf blade. 19. Catopsis micrantha, photograph of epicuticular wax powder layer on abaxial surface of leaf sheath. 20. Catopsis micrantha, scanning electron
micrograph (SEM) of trichome embedded in wax layer (unprepared specimen). 21. Catopsis micrantha, SEM of trichomes on the wax-free adaxial leaf sheath
surface (prepared specimen). 22. Werauhia capitata, SEM of trichome on abaxial surface. 23. Werauhia capitata, SEM of abaxial surface. 24. Alcantarea
odorata, SEM of adaxial surface, 25. Brocchinia reducta, SEM of adaxial surface.
August 2001] 1385PIERCE ET AL.?WATER REPELLENT INDUMENTA OF BROMELIADS
Figs. 26?31. Scanning electron micrographs of bromeliad leaf blade surfaces, the adaxial surfaces of which do not interact with water, the abaxial surfaces
hydrophobic. 26?28. Noninteractive adaxial surfaces of Vriesea monstrum, Fosterella petiolata and Pitcairnia corallina, respectively, lacking trichomes or with
filmy trichomes. 29?31. Hydrophobic abaxial surfaces of Vriesea monstrum, Fosterella petiolata, and Pitcairnia corallina, respectively, with well-defined
trichomes in a confluent layer.
1386 [Vol. 88AMERICAN JOURNAL OF BOTANY
Figs. 32?37. Scanning electron micrographs of trichomes from bromeliad leaf blade surfaces, the indumenta of which have different interactions with water.
32. Pitcairnia arcuata, attenuated stellate trichome with radial filaments tangled together, low densities of which form a hydrophobic indumentum. 33. Puya
laxa has two types of trichome, one peltate and the other with a grossly elongate wing that spirals around itself to form a hair-like structure, the indumentum
having no interaction with water. 34. Aechmea dactylina, peltate trichome in a hydrophilic indumentum. 35. Aechmea dactylina, detail of trichome shield. 36.
Ronnbergia explodens, peltate trichome in a hydrophobic indumentum. 37. Ronnbergia explodens, detail of trichome shield.
August 2001] 1387PIERCE ET AL.?WATER REPELLENT INDUMENTA OF BROMELIADS
Figs. 38?41. Scanning electron micrographs of trichomes on hydrophilic and hydrophobic surfaces of the same leaf blade. 38?39. Cryptanthus whitmanii,
hydrophilic adaxial and hydrophobic abaxial surfaces, respectively. 40?41. Aechmea nudicaulis, hydrophilic adaxial and hydrophobic abaxial surfaces, respec-
tively.
TABLE 4. The effect of removal of water surface tension on the leaf
blade trichome-layer?water interactions of Ananas comosus. Re-
pellency was denoted by the depth of a 10-mL droplet of aqueous
fluorochrome after a period of 40 min. The fluorochrome used was
either fluorescein sodium solution (5 mL of 0.05% fluorescein 1
0.5 mL H2O) or a solution of fluorescein and household detergent
(5 mL of 0.05% fluorescein 1 0.5 mL neat detergent). Depth values
are derived from fluorochrome luminosity (under exciting UV
light) compared against standards of measured droplet depth (flu-
orochrome on paraplast wax and glass surfaces). Values represent
means 6 1 SE of six replicates. Different letters (a?c) represent
significant differences between means at the P # 0.05 level as
determined by Tukey?s multiple comparison procedure (ANOVA).
Leaf blade
surface
Depth of aqueous droplet (mm)
Fluorochrome Fluorochrome 1 detergent
Adaxial
Abaxial
559.3 6 81.6 b
1013.1 6 41.7 c
14.8 6 3.2 a
17.6 6 5.3 a
Seemann, and Renfrow, 1978). When the leaf is wetted, sur-
face tension forces acting on the epidermis and/or the under-
side of the trichome wing may permit water to spread. Thus,
dense trichome layers in most Tillandsioideae have different
configurations when wet and dry and will only form a conflu-
ent layer after wetting. The moveable trichome wing of the
Type 5 life form may therefore be regarded as a device allow-
ing the presence of high densities of trichomes while avoiding
repellency.
Indeed, dense layers of peltate trichomes that lack wings in
Tillandsioideae are hydrophobic (e.g., Vriesea monstrum; Fig.
29; Table 3). Also, the immobile trichomes of Type 3 bro-
meliads demonstrate that a moveable wing is not essential for
absorption (Benzing, Givnish, and Bermudes, 1985). The
moveable wing is generally associated with higher trichome
densities and effective water and nutrient absorption by the
leaf surface (Benzing and Burt, 1970).
Epicuticular wax powders?Benzing, Givnish, and Ber-
mudes (1985) suggest that Tillandsioideae and Brocchinia?
both of which include advanced Type 4 tank forms equipped
1388 [Vol. 88AMERICAN JOURNAL OF BOTANY
TABLE 5. Summary of principal leaf blade interactions with water (as determined by fluorographic dimensional imaging) of the different eco-
physiological types of Bromeliaceae. ??Ring-peltate?? trichomes possess a shield composed mainly of ring cells, and ??wing-peltate?? trichomes
possess a shield with a moveable wing. Life forms or ecophysiological types follow the classification of Benzing (2000).
Life form Trichome type Trichome cover Interaction with water Example
1 stellate
stellate/ring
peltate
ring-peltate
discontinuous
continuous
discontinuous
hydrophobic
hydrophobic
hydrophilic
Pitcairnia arcuata
Fosterella petiolata,
Ronnbergia explodens
Cryptanthus whitmanii
2 ring-peltate
ring-peltate
continuous
discontinuous
hydrophobic
hydrophilic
Ananas comosus
Aechmea magdalenae
3 ring-peltate
ring-peltate
continuous
discontinuous
hydrophobic
hydrophilic
Aechmea nudicaulis
Aechmea dactylina
4 ring-peltate
wing-peltate
wing-peltate
continuous
continuous
discontinuous
hydrophobic
hydrophilic
noninteractive
Vriesea monstrum
Tillandsia elongata
Werauhia sanguinolenta
5 wing-peltate continuous hydrophilic Tillandsia nana
with absorbent trichomes?are derived from a common an-
cestor. Indeed, in the present study only Tillandsioideae and
Brocchinia provided examples of species in which epicuticular
wax powders are produced. Waxy Catopsis species have been
shown to use wing-peltate trichomes to take up mineral ions
and amino acids (Benzing et al., 1976; Benzing, 1980; Benz-
ing and Pridgeon, 1983), and this probably also applies to C.
micrantha. Both leaf surfaces bear a powdery layer of epicu-
ticular wax, and this is also one of the few taxa reported to
be amphistomatous (see Tomlinson, 1969; Figs. 13, 14). Thus,
extensive epicuticular wax powders appear to have evolved
only in taxa containing Type 4 life forms, which use trichomes
to acquire water and minerals from tanks.
It is likely that in many Type 4 species a combination of
the horizontal orientation of the leaf and the hypostomatous
condition are sufficient to keep stomata unobstructed by water;
in the present study, predominantly those species that pos-
sessed upright leaves (e.g., Brocchinia reducta, Guzmania ma-
cropoda, Werauhia capitata), and/or stomata on the adaxial
surface (Catopsis micrantha) possessed hydrophobic wax
powders on the leaf blade. Possibly the upright funnelform
habit increases the utility of the tank as an impoundment, and
tank formers face a trade-off between gas exchange and im-
poundment capacity, wax powders being a method of maxi-
mizing both. Reflective epicuticular wax powders have also
been implicated in the attraction and entrapment of insects in
a small number of Type 4 bromeliads?Catopsis berteroniana,
Brocchinia hechtioides, and B. reducta (Fish, 1976; Frank and
O?Meara, 1984; Givnish et al., 1984; Owen, Benzing, and
Thomson, 1988; Owen and Thomson, 1991; Benzing, 2000).
It is possible that a slippery and reflective epicuticular wax
powder helped predispose these lineages to carnivory.
Conclusions?Hydrophobic leaf surfaces of Bromeliaceae
possess a highly irregular microrelief, thereby reducing the
adhesion and spread of water on the leaf blade. Hydrophobic
trichome layers occur on the abaxial leaf blade surfaces of
many mesic Type 1 pitcairnioids and, as these species exhibit
the putative primitive ecological condition, water repellency
appears to have been an important condition in early Brome-
liaceae. The trichomes of Type 4 species are specialized for
the alternative function of water and nutrient absorption from
a water-filled tank, with epicuticular wax powders employed
by some species to shed water from the leaf blades. Hydro-
phobic trichome layers and wax powders could potentially ob-
struct pathogens and particulates, aid in self-cleaning, and/or
maintain gas exchange during wet weather.
LITERATURE CITED
ADAMS, W. W., III, AND C. E. MARTIN. 1986. Morphological changes ac-
companying the transition from juvenile (atmospheric) to adult (tank)
forms in the Mexican epiphyte Tillandsia deppeana (Bromeliaceae).
American Journal of Botany 73: 1207?1214.
BARTHLOTT, W., AND C. NEINHUIS. 1997. Purity of sacred lotus, or escape
from contamination in biological surfaces. Planta 202: 1?8.
BELL, A. D. 1991. Plant form: an illustrated guide to flowering plant mor-
phology. Oxford University Press, Oxford, UK.
BENZING, D. H. 1970. Foliar permeability and the adsorption of minerals and
organic nitrogen by certain tank bromeliads. Botanical Gazette 131: 23?
31.
???. 1976. Bromeliad trichomes: structure, function, and ecological sig-
nificance. Selbyana 1: 330?348.
???. 1980. The biology of Bromeliads. Mad River Press, Eureka, Cali-
fornia, USA.
???. 2000. Bromeliaceae: profile of an adaptive radiation. Cambridge
University Press, Cambridge, UK.
???, AND K. M. BURT. 1970. Foliar permeability among twenty species
of the Bromeliaceae. Bulletin of the Torrey Botanical Club 97: 269?279.
???, T. J. GIVNISH, AND D. BERMUDES. 1985. Absorptive trichomes in
Brocchinia reducta (Bromeliaceae) and their evolutionary and systematic
significance. Systematic Botany 10: 81?91.
???, K. HENDERSON, B. KESSEL, AND J. SULAK. 1976. The absorptive
capacities of bromeliad trichomes. American Journal of Botany 63:
1009?1014.
???, AND A. M. PRIDGEON. 1983. Foliar trichomes of Pleurothallidinae
(Orchidaceae): functional significance. American Journal of Botany 70:
173?180.
???, AND A. RENFROW. 1971. The significance of photosynthetic effi-
ciency to habitat preference and phylogeny among tillandsioid bromeli-
ads. Botanical Gazette 132: 19?30.
???, J. SEEMANN, AND A. RENFROW. 1978. The foliar epidermis in Til-
landsioidae (Bromeliaceae) and its role in habitat selection. American
Journal of Botany 65: 359?365.
BILLINGS, F. W. 1904. A study of Tillandsia usneoides. Botanical Gazette 38:
99?121.
BREWER, C. A., W. K. SMITH, AND T. C. VOGELMANN. 1991. Functional
interaction between leaf trichomes, leaf wettability and the optical prop-
erties of water droplets. Plant, Cell and Environment 14: 955?962.
EHLERINGER, J., AND O. BJO? RKMAN. 1976. Leaf pubescence: effects on ab-
sorptance and photosynthesis in a desert shrub. Science 192: 376?377.
FISH, D. 1976. Structure and composition of the aquatic invertebrate com-
munity inhabiting epiphytic bromeliads in south Florida and the discov-
ery of an insectivorous bromeliad. Ph.D. dissertation, University of Flor-
ida, Gainesville, Florida, USA.
FRANK, J. H., AND G. F. O?MEARA. 1984. The bromeliad Catopsis berter-
August 2001] 1389PIERCE ET AL.?WATER REPELLENT INDUMENTA OF BROMELIADS
oniana traps terrestrial arthropods but harbors Wyeomyia larvae (Diptera:
Culicidae). Florida Entomologist 67: 418?424.
GIVNISH, T. J., E. L. BURKHARDT, R. E. HAPPEL, AND J. D. WEINTRAUB.
1984. Carnivory in the bromeliad Brocchinia reducta, with a cost/benefit
model for the general restriction of carnivorous plants to sunny, moist,
nutrient-poor habitats. American Naturalist 124: 479?497.
???, K. J. SYTSMA, J. F. SMITH, W. J. HAHN, D. H. BENZING, AND E. L.
BURKHARDT. 1997. Molecular evolution and adaptive radiation in Broc-
chinia (Bromeliaceae: Pitcairnioideae) atop tepuis of the Guayana shield.
In T. J. Givnish and K. J. Sytsma, [eds.] Molecular evolution and adap-
tive radiation, 259?311. Cambridge University Press, Cambridge, UK.
HORRES, P., G. ZIZKA, G. KAHL, AND K. WEISING. 2000. Molecular phylo-
genetics of Bromeliaceae: evidence from trnL(UAA) intron sequences of
the chloroplast genome. Plant Biology 2: 306?315.
HOLLOWAY, P. J. 1968. The effects of superficial wax on leaf wettability.
Annals of Applied Biology 63: 145?153.
JUNIPER, B. E., AND C. E. JEFFREE. 1983. Plant surfaces. Arnold, London,
UK.
KRAUSS, B. H. 1948?1949. Anatomy of the vegetative organs of the pine-
apple, Ananas comosus (L.) Merr. Botanical Gazette 110: 159?217, 333?
404, 555?587.
LUTHER, H. E., AND E. SIEFF. 1998. An alphabetical list of bromeliad bino-
mials, 6th ed. Bromeliad Society International, Gainesville, Florida,
USA.
LU? TTGE, U., B. KLAUKE, H. GRIFFITHS, J. A. C. SMITH, AND K. H. STIMMEL.
1986. Comparative ecophysiology of CAM and C3 bromeliads. V. Gas
exchange and leaf structure of the C3 bromeliad Pitcairnia integrifolia.
Plant, Cell and Environment 9: 411?419.
MARTIN, C. E. 1994. Physiological ecology of the Bromeliaceae. Botanical
Review 60: 1?82.
MEZ, C. 1904. Physiologische Bromeliaceen?Studien. I. Die Wasser-O? kon-
omie den extrem atmospha?rischen Tillandsien. Jahrbu?cher fu?r wissen-
schaftliche Botanik 40: 157?229.
MILLER, G. A. 1994. Functional significance of inflorescence pubescence in
tropical alpine species of Puya. In P. W. Rundell, A. P. Smith, and F. C.
Meinzer [eds.], Tropical alpine environments: plant form and function,
195?213. Cambridge University Press, Cambridge, UK.
NYMAN, L. P., J. P. DAVIS, S. J. O?DELL, J. ARDITTI, G. C. STEPHENS, AND
D. H. BENZING. 1987. Active uptake of amino acids by leaves of an
epiphytic vascular plant, Tillandsia paucifolia (Bromeliaceae). Plant
Physiology 83: 681?684.
OWEN, T. P., JR., D. H. BENZING, AND W. W. THOMSON. 1988. Apoplastic
and ultrastructural characterizations of the trichomes from the carnivo-
rous bromeliad Brocchinia reducta. Canadian Journal of Botany 66:
941?948.
???, AND W. W. THOMSON. 1991. Structure and function of a specialized
cell-wall in the trichomes of the carnivorous bromeliad Brocchinia re-
ducta. Canadian Journal of Botany 69: 1700?1706.
PITTENDRIGH, C. S. 1948. The bromeliad?Anopheles?malaria complex in
Trinidad. I. bromeliad flora. Evolution 2: 58?89.
ROBINSON, S. A., C. E. LOVELOCK, AND C. B. OSMOND. 1993. Wax as a
mechanism for photoprotection against photoinhibition?a study of Cot-
yledon orbiculata. Botanica Acta 106: 307?313.
SCHIMPER, A. F. W. 1888. Die epiphytische Vegetation Amerikas. Botanische
Mittheilungen aus den Tropen, Heft 2. Gustav Fischer, Jena, Germany.
SCHMITT, A. K., C. E. MARTIN, AND L. E. LU? TTGE. 1989. Gas exchange and
water vapour uptake in the atmospheric CAM bromeliad Tillandsia re-
curvata L.: the influence of trichomes. Botanica Acta 102: 80?84.
SMITH, J. A. C. 1989. Epiphytic bromeliads. In U. Lu?ttge [ed.], Vascular
plants as epiphytes: evolution and ecophysiology. Ecological Studies 76,
109?134. Springer-Verlag, Berlin, Germany.
SMITH, L. B., AND R. J. DOWNS. 1974. Flora Neotropica. Monograph 14,
Part 1, Pitcairnioideae (Bromeliaceae). Hafner Press, New York, New
York, USA.
???, AND ???. 1977. Flora Neotropica. Monograph 14, Part 2, Til-
landsioideae (Bromeliaceae). Hafner Press, New York, New York, USA.
???, AND ???. 1979. Flora Neotropica. Monograph 14, Part 3, Bro-
melioideae (Bromeliaceae). Hafner Press, New York, New York, USA.
TAYLOR, D. C., AND H. E. ROBINSON. 1999. A rejection of Pepinia (Bro-
meliaceae: Pitcairnioideae) and taxonomic revisions. In J. R. Grant [ed.],
Florilegium Luctatori Botanico Dedicatum: a tribute to Lyman B. Smith.
Harvard Papers in Botany 4: 1?288.
TERRY, R. G., G. K. BROWN, AND R. G. OLMSTEAD. 1997. Examination of
subfamilial phylogeny in Bromeliaceae using comparative sequencing of
the plastid locus NDHF. American Journal of Botany 84: 664?670.
TIETZE, M. 1906. Physiologische Bromeliaceen-Studien II. Die Entwicklung
der wasseraufnehmenden Bromeliaceen-Trichome. Zeitschrift fu?r Natur-
wissenschaften 78: 1?50.
TOMLINSON, P. B. 1969. Anatomy of the monocotyledons. III. Commelinales-
Zingiberales. Oxford University Press, Oxford, UK.
WATANABE, T., AND I. YAMAGUCHI. 1993. The specific adhesional forces of
aqueous droplets on crop leaf surfaces and factors influencing them.
Journal of Pesticide Science 18: 99?107.