The Jeweled Armor of Tillandsia—Multifaceted or Elongated ...

10
Aliso: A Journal of Systematic and Evolutionary Botany Volume 23 | Issue 1 Article 6 2007 e Jeweled Armor of Tillandsia—Multifaceted or Elongated Trichomes Provide Photoprotection Simon Pierce Università degli Studi dell'Insubria, Varese, Italy Follow this and additional works at: hp://scholarship.claremont.edu/aliso Part of the Botany Commons , Ecology and Evolutionary Biology Commons , and the Plant Biology Commons Recommended Citation Pierce, Simon (2007) "e Jeweled Armor of Tillandsia—Multifaceted or Elongated Trichomes Provide Photoprotection," Aliso: A Journal of Systematic and Evolutionary Botany: Vol. 23: Iss. 1, Article 6. Available at: hp://scholarship.claremont.edu/aliso/vol23/iss1/6

Transcript of The Jeweled Armor of Tillandsia—Multifaceted or Elongated ...

Page 1: The Jeweled Armor of Tillandsia—Multifaceted or Elongated ...

Aliso: A Journal of Systematic and Evolutionary Botany

Volume 23 | Issue 1 Article 6

2007

The Jeweled Armor of Tillandsia—Multifaceted orElongated Trichomes Provide PhotoprotectionSimon PierceUniversità degli Studi dell'Insubria, Varese, Italy

Follow this and additional works at: http://scholarship.claremont.edu/aliso

Part of the Botany Commons, Ecology and Evolutionary Biology Commons, and the PlantBiology Commons

Recommended CitationPierce, Simon (2007) "The Jeweled Armor of Tillandsia—Multifaceted or Elongated Trichomes Provide Photoprotection," Aliso: AJournal of Systematic and Evolutionary Botany: Vol. 23: Iss. 1, Article 6.Available at: http://scholarship.claremont.edu/aliso/vol23/iss1/6

Page 2: The Jeweled Armor of Tillandsia—Multifaceted or Elongated ...

Aliso 23, pp. 44–52! 2007, Rancho Santa Ana Botanic Garden

THE JEWELED ARMOR OF TILLANDSIA–MULTIFACETED OR ELONGATED TRICHOMESPROVIDE PHOTOPROTECTION

SIMON PIERCE

Universita degli Studi dell’Insubria, Dipartimento di Biologia Strutturale e Funzionale,J. H. Dunant 3, I-21100 Varese, Italy

([email protected])

ABSTRACT

Foliar trichomes of gray-leaved Tillandsioideae (Bromeliaceae) are highly reflective, suggesting arole in protecting the leaf against direct sunlight in exposed niches. The performance of photosystemII, as denoted by the chlorophyll fluorescence characteristic Fv /Fm , was determined for seven Tilland-sia species and Vriesea barclayana that were exposed to excessive light, with trichomes either presentor removed. Additionally, trichome structure and interaction with light was recorded using extendeddepth-of-field photomicrography, and reflectance quantified using a novel photographic technique.Trichomes of mesomorphic Type IV life forms (T. cryptantha, T. cyanea) and of the intermediate lifeform V. barclayana conferred reflectance of between 1 and 11%, which did not significantly influenceFv /Fm when exposed to a high light intensity of 1500 !mol m"2 s"1 (photosynthetically active radiation)for one hour. However, the ornate trichomes of atmospheric species increased the reflectivity of theleaf blade by as much as 18–40%, with a positive correlation apparent between reflectance and pho-toprotection. Type V Tillandsia andrieuxii, T. caput-medusae, and T. mitlaensis have attenuated tri-chome wings extending perpendicular to the leaf surface and catching the light (with leaf surfacesappearing gray and fuzzy). This open configuration was observed to facilitate leaf ventilation and thecondensation of water vapor on the cooler underlying cuticle, with liquid water subsequently envel-oping the trichomes, suggesting a trade-off between water acquisition and light reflectance for airplants from xeric habitats. However, Type IV-V T. albida and T. concolor impound water in leaf basesand the flattened, circular, and overlapping trichome wings did not facilitate dew formation on thecuticle. For these plants with white, smooth leaf surfaces, trichomes are multifaceted and provide moreeffective photoprotection by scattering light in the manner of cut gemstones.

Key words: Bromeliaceae, bromeliad, epiphyte, photoinhibition, photoprotection, Tillandsia, trichome.

INTRODUCTION

Tillandsioideae are a subfamily of Bromeliaceae, charac-terized in part by peltate trichomes (plate-like leaf hairs)with a distinctive flexible shield. The majority of Tilland-sioideae grow as epiphytes, which possess trichomes thatplay a central role in their lifestyle by allowing the leaf toabsorb water and dissolved minerals and with roots that actprimarily as holdfasts (the structure and absorptive mecha-nism of the tillandsioid trichome is described in Fig. 1).Aside from this principal role of the tillandsioid trichome inwater and mineral absorption, the dry, upright trichomewings of atmospheric species (i.e., succulent epiphytes thatabsorb water directly over the entire leaf surface rather thanfrom phytotelmata; Benzing 2000) reflect light, which hasbeen proposed as an adaptation to high light intensities(Benzing and Renfrow 1971; Benzing 1976). Mechanisti-cally, light scattering by dry trichomes has been suggestedto provide photoprotection (Benzing 2000); i.e., shieldingproteins in light-harvesting units from excessive light ener-gy. Indeed, atmospheric Tillandsioideae specialize in highlyexposed canopy niches that regularly experience direct sun-light. Trichomes of Tillandsia fasciculata Sw. reflect ap-proximately 25% of incident visible light (Benzing and Ren-frow 1971) and ecotypes of T. caput-medusae from Sonora,Mexico, have inherently more reflective indumenta than eco-types from less arid sites (Dimmit 1985). However, the non-absorptive trichomes that characterize some of the more

primitive terrestrial bromeliads reflect up to 17% of visiblelight as a consequence of a highly irregular water-repellentmicro-relief, but this is not sufficient for photoprotectionagainst direct sunlight (Pierce et al. 2001). Indeed, trichomesconferring reflectance of up to 56% limit photosynthesis forEncelia farinosa A. Gray ex. Torr. (Asteraceae) in the Son-oran Desert of California (Ehleringer et al. 1976), which isclearly of adaptive significance as pubescent leaves are pro-duced seasonally and distinct ecotypes have been discerned(Housman et al. 2002).

Direct measurements of photoinhibition and specificallyof photoprotection by trichomes of gray-leaved Tillandsia L.have not been made previously. The present study investi-gates the hypothesis that highly reflective foliar trichomesphotoprotect the leaf, as evidenced by photosystem II chlo-rophyll fluorescence characteristics. The nomenclature fol-lows Luther (2002) and life forms or ecophysiological typesfollow Benzing (2000), as summarized by Pierce et al.(2001).

MATERIALS AND METHODS

Plant Material and Cultivation

The Tillandsia species and Vriesea barclayana listed inFig. 2 (authorities noted in Table 1) were obtained from BirdRock Tropicals (Carlsbad, California, USA), with the excep-tion of Tillandsia cyanea (Bouquet Florist, Crookes, Shef-

Page 3: The Jeweled Armor of Tillandsia—Multifaceted or Elongated ...

VOLUME 23 45Jeweled Armor of Tillandsia

Fig. 1.—Transverse views of the tillandsioid trichome in dry andwet configurations (not to scale); based on the photomicrographsand descriptions of Benzing (1976, 2000). The ‘‘shield’’ is com-prised of dead central disc cells, ring cells, and wing, and the ‘‘stalk’’is comprised of live dome and foot cells. When dry, the wing standserect, exposing the underlying leaf surface and trichome undersidesto rainwater, thus avoiding the water repellency typical of densetrichome layers (Pierce et al. 2001). Rainwater or dew enters thecells of the trichome shield, which lack cuticle and cell contents. Asthe central disc swells the increased pressure is directed by the dif-ferentially thickened ring cell walls to flatten the wing against thecuticle, promoting water uptake by enhancing capillary movementof water towards the central disc (Benzing 1976). Transverse wallsof the dome cells are uncutinized, allowing water uptake via os-mosis, transport between cells, and ultimately into the leaf. Domecell ultra-structure suggests that mineral nutrients are taken up ac-tively (Benzing 2000). As the leaf surface dries, collapse of thecentral disc seals the dome cells, preventing evaporative water loss(Benzing 1976).

field, UK). Life forms ranging from soft-leaved Type IVspecies with extensive phytotelmata (T. cryptantha, T. cy-anea) to succulent Type V species with extremely limitedphytotelmata (T. andrieuxii, T. caput-medusae, T. mitlaensis)and intermediate life forms (T. albida, T. concolor, V. bar-clayana) were investigated. Plants were cultivated in the UKunder glass (16–29!C, 39–92% relative humidity [RH],16 h photoperiod), with shade cloth limiting light intensityto 20% of ambient light (e.g., on a cloudless day, with fullsunlight of 1600 "mol Quanta m#2 s#1, plants experienced300 "mol m#2 s#1). The commercial nutrient solutionBabyBio (pbi Home and Garden Ltd., Middlesex, UK) wasdiluted ($ 0.4 of recommended strength) to provide 3.2 mMN, 0.3 mM P, and 0.1 mM K, with all leaf surfaces inundateddaily using a hand-nebulizer, for a minimum of two weeksprior to experimentation.

Reflectance of Light by Leaf Surfaces

Reflectance of light by leaf blades of each species wasquantified from photographs, using barium sulfate as a stan-

dard (BaSO4; absolute reflectivity of 99.3% over the wave-length range 300–800 nm; Munsell Color, New Windsor,New York, USA). Barium sulfate powder was made into anaqueous paste (50% w/v) painted on the mid-leaf portion ofthe leaf blade surface, and left to dry.1 The leaf was pho-tographed on tungsten color-reversal film with even lightingprovided by four 20W halogen lamps. Photographs wereslightly underexposed to avoid over saturation of the lighterportions of images. The image was digitized using a DFSPrimeFilm 1800u transparency scanner (Jessops, Leicester,UK). Corel Photo-Paint vers. 9 imaging software (Corel Cor-poration, Ottawa, Ontario, Canada) was used to compare theluminosity of pixels representing the barium sulfate standardand pixels representing adjacent leaf surface proper (use ofCorel software to determine luminosity values is describedby Pierce et al. 2001). Portions of the leaf blade were alsodenuded of trichomes using sticky-back plastic, as describedby Pierce et al. (2001), and thus the reflectance of the tri-chomes was determined by comparing intact and denudedsurfaces.

Chlorophyll Fluorescence

Photoinhibition of photosystem II was investigated usingan OS5-FL portable modulated fluorometer (Opti-Sciences,Tyngsboro, Massachusetts, USA). The degree of photoinhi-bition was determined from the decline in the dark-adaptedratio of variable to maximum chlorophyll fluorescence(Fv /Fm) following exposure for one hour to 1500 "mol m#2

s#1 photosynthetically active radiation (PAR), provided byHPI-T Plus 400W high pressure iodide lamps (KoninklijkePhilips Electronics N.V., Eindhoven, The Netherlands) at anair temperature of 28!C (see Pierce et al. 2001).

Trichome Structure and Density

Leaf blade trichome structure was observed via light mi-croscopy with extended depth-of-field images of leaf surfac-es and trichomes produced using the method of Rolfe andScholes (2002). An Olympus BX50WI microscope (Olym-pus Optical Company, London, UK) with focus controlslaved to a stepper motor (Proscan, Prior, Cambridge, UK)was used to acquire images at different focal planes at 50"m intervals throughout the sample. Samples were illumi-nated from above by a 150W halogen lamp (Schott KL1500Electronic, Schott-Fostec LLC, Auburn, New York, USA)with light delivered via a fiber optic, reflected from the sam-ple and ultimately photographed using a cooled, back-thinned, illuminated charge-coupled device (CCD; Micro-max 880PB, Roper Scientific, Marlow, UK). Image acqui-sition, processing, and display functions were facilitated byImage Pro Plus vers. 4.1 software (Media Cybernetics, SilverSpring, Maryland, USA). Using DeFuzz vers. 1.0.1 extendeddepth of field software (Rolfe 2002), images of 1000 $ 800pixels in size were subdivided into blocks of 102 or 202

1 NB titanium dioxide (TiO2) in toothpaste may be a more con-venient standard than barium sulfate for use in the field or as analternative to a powdered form when crossing international borders.Using the above technique, an absolute reflectivity of 93% (N % 10)was determined for Colgate Total (Colgate-Palmolive [UK] Ltd.,Guildford, UK) in comparison with barium sulfate.

Page 4: The Jeweled Armor of Tillandsia—Multifaceted or Elongated ...

46 ALISOPierce

Fig. 2.—The seven Tillandsia species and Vriesea barclayana (Bromeliaceae) used in the present study. Pairs of coins for scale are: USquarter (left) and GB pound (right). A 30 cm rule provides scale for T. cyanea.

Page 5: The Jeweled Armor of Tillandsia—Multifaceted or Elongated ...

VOLUME 23 47Jeweled Armor of Tillandsia

Tabl

e1.

Inte

ract

ions

betw

een

ligh

t(p

hoto

synt

heti

call

yac

tive

radi

atio

nor

PAR

)an

dfo

liar

indu

men

taof

epip

hyti

cT

illa

ndsi

asp

ecie

san

dV

ries

eaba

rcla

yana

(Bro

mel

iace

ae:

Til

land

-si

oide

ae).

The

degr

eeof

phot

odam

age

toph

otos

yste

mII

,af

ter

expo

sure

tosa

tura

ting

ligh

t(1

500

!m

olm

"2

s"1

PAR

)fo

ron

eho

ur,

isde

note

dby

the

exte

ntof

redu

ctio

nin

Fv/F

m(t

heda

rk-

adap

ted

rati

oof

vari

able

tom

axim

alch

loro

phyl

lfl

uore

scen

ce)

wit

hth

ele

afbl

ade

surf

ace

eith

erin

tact

orde

nude

dof

surf

ace

feat

ures

.V

alue

sre

pres

ent

the

mea

n#

1S

Eof

four

repl

icat

es.

*D

enot

essi

gnifi

cant

diff

eren

ces

inth

ere

duct

ion

ofF

v/F

mbe

twee

nin

tact

and

denu

ded

trea

tmen

tsat

the

P!

0.05

leve

las

dete

rmin

edby

Stu

dent

’st-

test

,**

P!

0.01

,**

*P

!0.

001,

wit

hn.

s.in

dica

ting

‘‘no

sign

ifica

nt’’

diff

eren

ce.

Pro

port

ion

data

wer

ear

csin

etr

ansf

orm

edpr

ior

tost

atis

tica

lan

alys

isan

dst

anda

rder

rors

ofpr

opor

tion

data

calc

ulat

edvi

ath

em

etho

dof

Fow

ler

etal

.(1

998)

.D

eter

min

atio

nof

phot

osyn

thet

icpa

thw

ayis

base

don

tiss

ueca

rbon

isot

ope

com

posi

tion

(Cra

yn,W

inte

r,an

dS

mit

hun

publ

.dat

a),i

nco

njun

ctio

nw

ith

gas

exch

ange

and

titr

atab

leac

idit

yda

tafo

rV

.ba

rcla

yana

(Pie

rce

etal

.20

02b)

.L

ife

form

sor

ecop

hysi

olog

ical

type

sfo

llow

Ben

zing

(200

0).

Spe

cies

Lif

efo

rmP

hoto

synt

heti

cpa

thw

ayS

urfa

ceT

rich

ome

dens

ity

(sta

lks

mm

"2 )

Refl

ecta

nce

from

inta

ctle

afbl

ade

(%)

Refl

ecta

nce

conf

erre

dby

tric

hom

es(%

)

Dec

reas

ein

Fv/F

mfo

llow

ing

high

ligh

tex

posu

re(%

)

Inta

ctsu

rfac

eD

enud

edsu

rfac

eS

igni

fica

nce

T.

albi

daM

ez&

Pur

pus

IV-V

CA

MA

daxi

alA

baxi

al41

#1.

748

#4.

079

#5.

180

#5.

127

#3.

040

#3.

611

#1.

83

#1.

131

#3.

241

#3.

7**

***

T.

andr

ieux

ii(M

ez)

L.

B.

Sm

.V

C3

Ada

xial

43#

1.7

59#

4.4

25#

2.9

10#

1.8

19#

2.5

*A

baxi

al48

#4.

067

#4.

736

#3.

416

#2.

339

#3.

6**

T.

capu

t-m

edus

aeE

.M

orre

nT

.co

ncol

orL

.B

.S

m.

V IV-V

CA

MC

AM

Aba

xial

Ada

xial

Aba

xial

18#

1.2

57#

6.2

58#

3.7

58#

4.4

72#

4.9

76#

5.0

23#

2.8

6#

1.4

18#

2.5

9#

1.8

12#

2.0

4#

1.2

31#

3.2

16#

2.3

17#

2.4

* n.s.

*T

.cr

ypta

ntha

Bak

erIV

CA

MA

daxi

alA

baxi

al27

#1.

225

#0.

662

#4.

563

#4.

69

#1.

811

#1.

939

#3.

625

#2.

931

#3.

233

#3.

3n.

s.n.

s.T

.cy

anea

Lin

den

exK

.K

och

var.

cyan

eaT

.m

itla

ensi

sW

.W

eber

&E

hler

s

IV V

C3-

CA

M

Ada

xial

Aba

xial

Aba

xial

48#

3.6

38#

2.0

22#

0.9

36#

3.4

40#

3.6

83#

5.2

1#

0.5

2#

0.9

27#

3.0

31#

3.2

10#

1.9

11#

1.9

31#

3.2

14#

2.2

30#

3.1

n.s.

n.s.

*V

.ba

rcla

yana

(Bak

er)

L.

B.

Sm

.va

r.ba

rcla

yana

IV-V

C3-

CA

MA

daxi

alA

baxi

al60

#7.

160

#4.

449

#4.

053

#4.

24

#1.

18

#1.

637

#3.

514

#2.

145

#3.

917

#2.

4n.

s.n.

s.

Page 6: The Jeweled Armor of Tillandsia—Multifaceted or Elongated ...

48 ALISOPierce

Fig. 3.—The relationship between percent reflectance of visiblelight by trichomes and photoprotection (the extent of decline in Fv /Fm circumvented by trichomes of intact surfaces, compared to de-nuded surfaces) for abaxial leaf blade surfaces of Tillandsia albida(T. alb.), T. andrieuxii (T. and.), T. caput-medusae (T. c.-m.), T.concolor (T. con.), T. cryptantha (T. cry.), T. cyanea (T. cya.), T.mitlaensis (T. mit.), and Vriesea barclayana (V. bar.).

pixels. Blocks that were optimally focused within each focalseries were identified by standard deviation parameters andthen combined to form a final montage. For investigation oftrichome density, low-magnification images were taken ofadaxial and abaxial surfaces of leaves from four replicateplants. For species with attenuated trichome wings, difficultyin discerning positions of individual trichomes was avoidedby removing the wings with sticky-back plastic and countingthe remaining trichome caps or stalks within a specified area.

Water Vapor Condensation on Leaf Surfaces

Condensation of water vapor on leaf surfaces was inducedby passing a humid air stream over the leaf, viewed simul-taneously by microscopy. Breath from the microscope op-erator (33!C, "99% relative humidity [RH]) was deliveredto the leaf surface via a length of 5 mm bore butyl tubingat a rate of 0.2–0.3 L min#1, as determined by rotometer(GEC-Elliott Ltd., Chelmsford, UK). Ambient laboratoryconditions were 20!C and 48% RH, with leaves equilibratedto these conditions for several hours before observationswere made. Images were captured at intervals of approxi-mately 4 sec (dependent on automatically determined ex-posure times) by Image Pro Plus, with 30 to 50 images com-bined in sequence to produce time-lapse movies of conden-sation on leaf blade surfaces.

RESULTS

Reflectance of visible light from intact leaves of the studyspecies ranged from 36% for Tillandsia cyanea (adaxial sur-face, ecophysiological Type IV) at one extreme to 83% forT. mitlaensis (abaxial surface, Type V; Table 1). The pro-portion of visible light reflected by trichomes only was 1–11% for Type IV T. cryptantha and T. cyanea and the in-termediate Type IV-V life form of V. barclayana. This didnot reduce the magnitude of photodamage following expo-sure to intense light for one hour, as denoted by a declinein dark-adapted Fv /Fm compared between intact and denudedsurfaces (Table 1). Trichomes of the adaxial surface of TypeIV-V T. concolor similarly reflected only 6% of incidentlight, which did not photoprotect the leaf. However, tri-chomes on abaxial surfaces of this species reflected 18%,significantly reducing photodamage (P ! 0.05), as did ab-axial (40%) and adaxial (27%) reflectance from trichomesof Type IV-V T. albida (P ! 0.01 and P ! 0.001, respec-tively; Table 1). High reflectances were recorded for TypeV species, which also significantly photoprotected leaves(i.e., 23–36%; T. andrieuxii, T. caput-medusae, and T. mit-laensis; Table 1). Trichome layers of Type IV-V and TypeV species were variable in densities with respect to Type IVspecies (Table 1).

The reflectance of visible light from the abaxial surfacesof the bromeliads studied was directly proportional to thedegree of photoprotection conferred (denoted by the de-crease in Fv /Fm vs. reflectance; y $ 0.8244, x $ #0.8155,r2 $ 0.9, P $ 0.003; Fig. 3). High reflectivity was not de-pendent on trichome density per se. For example, relativelylow densities of trichomes were associated with high reflec-tivity and photoprotection for T. caput-medusae and T. mit-laensis (18 % 1.2 and 22 % 0.9 trichome stalks mm#2, re-

spectively, cf. 60 % 4.4 stalks mm#2 for low-reflectivity V.barclayana; Table 1).

Photomicrographs of trichome structure and the indumen-ta of abaxial leaf blade surfaces for each species are pre-sented in Fig. 4–16. The trichome wings of T. albida and T.concolor have multifaceted surfaces that reflect light (Fig. 4,8, 12), with the extensive wing of T. albida conferring thegreatest reflectivity and photoprotection of any species stud-ied (Table 1). Such faceted wing surfaces were not apparentfor the other species studied, with reflection from trichomesof T. andrieuxii, T. caput-medusae, and T. mitlaensis occur-ring along the edges of the radial cells that comprise thewing (Fig. 6, 11, 13). These trichome wings are elongatedalong a single axis, inclined away from the leaf surface.

For T. caput-medusae and T. mitlaensis low-trichome den-sities and erect trichome wings expose underlying epidermison which condensation readily occurred and from which tri-chomes became wetted. (Note that no condensation was ob-served on trichome wings; Fig. 13, 14; time-lapse movieimages are available on request). For species with immobile,flattened, plate-like trichomes, it was possible to induce con-densation on the trichome wings (e.g., T. albida; Fig. 15,16), but considerable time was required, even at the extremeair temperature and humidity differences imposed in the pre-sent study, compared to the almost instantaneous conden-sation occurring on the smooth cuticles of T. cryptantha andT. cyanea (data not shown).

DISCUSSION

The trichome wings of gray-leaved Tillandsia spp. reflectenough light to protect photosystem II against intensitiesequivalent to full sunlight. However, high reflectivity resultsfrom different trichome structures and arrangements for eachlife form. The Type V life form of T. caput-medusae has arelatively sparse covering of trichomes, but the individualtrichome shields are large and the wings extend almost per-

Page 7: The Jeweled Armor of Tillandsia—Multifaceted or Elongated ...

VOLUME 23 49Jeweled Armor of Tillandsia

Fig. 4–11.—Photomicrographs of trichome layers on abaxial leaf blade surfaces digitally enhanced to increase apparent depth of field.—4. Tillandsia albida.—5. Vriesea barclayana.—6. T. andrieuxii.—7. T. caput-medusae.—8. T. concolor.—9. T. cryptantha.—10. T. cy-anea.—11. T. mitlaensis.

Page 8: The Jeweled Armor of Tillandsia—Multifaceted or Elongated ...

50 ALISOPierce

Fig. 12.—Photomicrograph of an isolated multifaceted trichomeof Tillandsia albida, digitally enhanced to increase apparent depthof field.

Fig. 13–16.—Condensation on representative abaxial leaf bladesurfaces of Tillandsia species.—13–14. T. caput-medusae.—13.Dry.—14. With condensation forming on the leaf cuticle.—15–16.T. albida.—15. Dry.—16. With condensation forming on the tri-chome wing. Arrows denote the position of example droplets ofliquid water. These images are frames from time-lapse movies pre-sented at The Third International Conference on the ComparativeBiology of the Monocotyledons (available on request).

pendicular to the leaf surface. This photoprotects the leaf,but also allows leaf ventilation and the condensation of watervapor on the underlying cuticle, and thus in its liquid form,the water ultimately wets the trichomes (time-lapse moviesavailable on request). Similar trichome configurations alsoallow condensation on the cuticle of T. mitlaensis and con-densation readily occurs on the smooth surfaces of meso-morphic Type IV T. cyanea and T. cryptantha. (Note thatthese species bear trichomes with small, immobile shieldsthat lay flat against the leaf surface.)

Leaves of epiphytes are cooler than the surrounding airduring the night and part of the day due to evaporative cool-ing (S. Pierce unpubl. data). In contrast, the slender, erecttrichome wings of T. caput-medusae have a higher surfacearea to mass ratio than the succulent leaves and lack cellcontents; trichome temperature will probably track air tem-perature more rapidly than will the leaf temperature. Thus,in nature, dew probably forms on the cooler cuticle ratherthan the trichome wing, as observed here in the laboratory(i.e., the cuticle, not the trichome wing, is the nucleus fordew formation; Fig. 14). At its simplest, attenuation of thetrichome wing can thus be considered a tradeoff betweenphotoprotection and water acquisition from dew formation,potentially of critical importance for Type V species thathave no significant water-impoundment capacity or absorp-tive roots. Indeed, for epiphytes that may endure long, dryseasons in exposed situations without access to soil water(and thus cannot rely on evaporative cooling) alternativemethods of regulating leaf energy balance are of critical im-portance. Trichome layers with a greater reflectance of in-cident visible light also reflect infrared wavelengths (e.g.,Pitcairnia integrifolia Ker Gawl.; Pierce et al. 2001) andreflection of visible light will reduce the need for radiation-

less energy dissipation (i.e., the conversion of excess lightto heat), thus also cooling the leaf. This is evident as leaftemperatures increase in full sunlight when trichomes areremoved from T. circinnata Schltdl. (Benzing 1976).

However, interpretation of attenuated trichomes as an ad-aptation to mediate leaf energy balance may be complicatedby further trichome-mediated phenomena. A number of TypeV species with attenuated trichomes are native to cloud for-ests and foggy deserts (e.g., T. plumosa Baker, T. tectorumE. Morren; D. H. Benzing pers. comm.), including T. an-drieuxii investigated in the present study (Smith and Downs1977). Wet trichomes hold a film of water over the stomataand impede gas exchange (Benzing et al. 1978; Pierce et al.2001, 2002a). However, the present study demonstrated thatinclined, attenuated trichome wings allow ventilation of theunderlying leaf surface; this could facilitate drying in moisthabitats and attenuated trichome wings potentially wick wa-ter away from the stomata. (Note that trichomes are attenu-ated only on the abaxial surfaces of hypostomatic T. an-drieuxii and a moist habitat is reflected in its use of C3 pho-tosynthesis; Table 1.) Additional selection pressures, whichshape the structure and function of trichomes for particularspecies, are detailed by Benzing (2000).

In contrast to Type V bromeliads, the intermediate TypeIV-V life forms of T. albida and T. concolor possess rudi-mentary phytotelmata formed by leaf sheaths. Indeed, diffi-culty in inducing condensation, even with the extreme air/leaf temperature differential generated in the laboratory, sug-gests that in nature dew formation would not occur readily

Page 9: The Jeweled Armor of Tillandsia—Multifaceted or Elongated ...

VOLUME 23 51Jeweled Armor of Tillandsia

on the indumenta of these species. Trichomes of T. albidaform a dense layer with trichomes flattened even when dry(but still exhibiting limited hygroscopic movement) and withthe extensive trichome wing decorated with concave dim-ples. These dimples produce a multifaceted surface that scat-ters incident light, thus providing the greatest reflectivity andphotoprotection of any species investigated in the presentstudy. Multiple concave facets provide additional surfacearea from which light is scattered, thus amplifying reflectionfrom each trichome. Trichome wings are thick enough thatthe probability of reflection is equal for photons of all wave-lengths (i.e., there is no iridescence; Feynman 1985) withreflected light exhibiting the same spectral composition asincident light. Similar trichome ornamentation was also ev-ident for T. concolor, explaining the high reflectivity of thisspecies despite its relatively small trichome wings. Indeed,the eight species investigated here exhibited diverse tri-chome wing architecture. With over 530 species of Tilland-sia currently described (Luther 2002) and many varied lifeforms evident, it is likely that other light-reflecting structuresand photoprotective strategies are also present within the ge-nus. For instance, trichome wings of T. karwinskyana Schult.f. appear to be papillose and also catch the light (Benzing2000).

Trichome layers are typically more prominent on the un-derside of leaf blades of the more primitive bromeliads, pos-sibly to reduce transpiration (Benzing 2000) or, being waterrepellent, to keep stomata clear of excess water and dirt(Pierce et al. 2001). A propensity for larger trichome wingson the underside of the leaf, providing greater photoprotec-tion for erect, more delicate, expanding leaves, may havefavored the relatively upright and in-rolled leaf blades ofmany Type V Bromeliaceae, such as T. caput-medusae andT. mitlaensis. Thus, an erect and in-rolled morphology couldbe considered part of the photoprotective strategy. Tillandsiabulbosa Hook. has retained the in-rolled leaves typical ofthis life form, but not the same trichome structure, with re-flectance lost as an adaptation to darker cloud forest habitats(Benzing 1980, 2000; Pierce et al. 2002a).

While trichomes of species such as T. albida and T. caput-medusae have adaptations to extremely high light, a contin-uum of reflectance and photoprotection is apparent for thetillandsioid trichome. Indeed, light-response analysis of V.barclayana indicates that even trichomes of this speciesmodify instantaneous photosynthesis at moderate light inten-sities (i.e., shade the leaf and decrease both electron trans-port and nonphotochemical chlorophyll fluorescencequenching; S. Pierce unpubl. data), but do not prevent pho-todamage at high light intensities (Table 1). Trichomes, ingeneral, are thus detrimental in terms of light capture forphotosynthesis. However, trichomes throughout the subfam-ily share common structural and functional features essentialto water and nutrient uptake and hence the epiphytic habit.The smaller number of species that have additional structuraladaptations to high light therefore represent the extreme ofadaptive radiation; trichomes have an inherent reflectivitythat has become more pronounced during radiation into ex-posed niches.

To conclude, the trichomes of gray- or white-leaved Til-landsia exhibit adaptations in structure (wing attenuation orfacets, respectively) that reflect enough light to help protect

the photosynthetic apparatus against full sunlight, whereasthe lesser reflectance by trichomes of mesomorphic Type IVlife forms does not and indeed is characteristic of speciesendemic to shaded habitats. Trichome attenuation for TypeV species represents a trade-off between photoprotection andleaf ventilation that allows water acquisition via dew for-mation on the cooler underlying cuticle for species occu-pying exposed xeric niches. For Type IV-V species that arehydrated from phytotelmata and have smooth, white leafblade surfaces, high light intensity appears to have been theoverriding selection pressure acting on trichome wing ar-chitecture; these trichome wings are multifaceted and there-by extremely reflective.

ACKNOWLEDGMENTS

I thank Jason R. Grant for organizing and inviting me tothe Bromeliaceae sessions; Anna Taylor for training in ex-tended depth-of-field photomicrography; Phil Swarbrick fordemonstrating the use of various pieces of equipment andfor comments on an earlier version of the manuscript; and,particularly Steve Rolfe and Julie Scholes for the use of theirimaging equipment. I also thank David H. Benzing for hisconstructive scrutiny during the review process. Attendanceat the conference was supported by a conference grant fromthe Royal Society, London, UK, and an attendance schol-arship from the Rancho Santa Ana Botanic Garden, Clare-mont, California, USA.

LITERATURE CITED

BENZING, D. H. 1976. Bromeliad trichomes: structure, function andecological significance. Selbyana 1: 330–348.

. 1980. The biology of Bromeliads. Mad River Press, Eureka,California, USA. 305 p.

. 2000. Bromeliaceae: profile of an adaptive radiation. Cam-bridge University Press, Cambridge, UK. 690 p.

, AND A. RENFROW. 1971. The significance of photosyntheticefficiency to habitat preference and phylogeny among tillandsioidbromeliads. Bot. Gaz. 132: 19–30.

, J. SEEMAN, AND A. RENFROW. 1978. The foliar epidermis inTillandsioideae (Bromeliaceae) and its role in habitat selection.Amer. J. Bot. 65: 359–365.

DIMMIT, M. A. 1985. Intraspecific variation in Tillandsia: selectingsuperior forms. J. Brom. Soc. 35: 101–103, 130–132.

EHLERINGER, J. L., O BJORKMAN, AND H. A. MOONEY. 1976. Leafpubescence: effects on absorptance and photosynthesis in a desertshrub. Science 192: 376–377.

FEYNMAN, R. P. 1985. QED: the strange theory of light and matter.Princeton University Press, Princeton, New Jersey, USA. 158 p.

FOWLER, J., L. COHEN, AND P. JARVIS. 1998. Practical statistics forfield biology. John Wiley & Sons, Chichester, UK. 249 p.

HOUSMAN, D. C., M. V. PRICE, AND R. A. REDAK. 2002. Architectureof coastal and desert Encelia farinosa (Asteraceae): consequencesof plastic and heritable variation in leaf characters. Amer. J. Bot.89: 1303–1310.

LUTHER, H. E. 2002. An alphabetical list of bromeliad binomials,ed. 8. Bromeliad Society International, Sarasota, Florida, USA.82 p.

PIERCE, S., K. MAXWELL, H. GRIFFITHS, AND K. WINTER. 2001. Hy-drophobic trichome layers and epicuticular wax powders in Bro-meliaceae. Amer. J. Bot. 88: 1371–1389.

, K. WINTER, AND H. GRIFFITHS. 2002a. The role of CAM inhigh rainfall cloud forests: an in situ comparison of photosyntheticpathways in Bromeliaceae. Pl. Cell Environm. 25: 1181–1189.

Page 10: The Jeweled Armor of Tillandsia—Multifaceted or Elongated ...

52 ALISOPierce

, , AND . 2002b. Carbon isotope ratio and theextent of daily CAM use by Bromeliaceae.New Phytol. 156: 75–83.

ROLFE, S. A. 2002. DeFuzz vers. 1.0.1 extended depth of field soft-ware. University of Sheffield, South Yorkshire, UK.

, AND J. D. SCHOLES. 2002. Extended depth-of-focus imaging

of chlorophyll fluorescence from intact leaves. Photosyn. Res. 72:107–115.

SMITH, L. B., AND R. J. DOWNS. 1977. Tillandsioideae (Bromeli-aceae). Fl. Neotrop. Monogr. 14. Hafner Press, New York, USA.1492 p.