Vous êtes sur la page 1sur 11

Current Biology 24, 984987, May 5, 2014 2014 Elsevier Ltd All rights reserved http://dx.doi.org/10.1016/j.cub.2014.03.

010

Report
Leaf Mimicry in a Climbing Plant
Protects against Herbivory

Ernesto Gianoli1,2,* and Fernando Carrasco-Urra2 size and shape and are composed of three leaflets that are
1Departamento de Biologa, Universidad de La Serena, pulvinated and therefore may change their orientation. Field
Casilla 554, La Serena, Chile observations indicate that B. trifoliolata often mimics the
2Departamento de Botanica, Universidad de Concepcion, leaves of its supporting trees in terms of size, shape, color,
Casilla 160-C, Concepcion, Chile orientation, and vein conspicuousness, among other features
(Figure 1). This phenomenon includes the display of a mucro-
nate leaf apex (a small spine at the leaf tip) when twining
Summary around a tree with such mucronate leaves (Figure 1); the
botanical description of B. trifoliolata does not include this
Mimicry refers to adaptive similarity between a mimic organ- feature [14]. Unlike earlier mimicry reports, leaf mimicry by
ism and a model. Mimicry in animals is rather common, this climbing plant is confined not to a single species but
whereas documented cases in plants are rare, and the asso- to several host trees. Moreover, when traversing different
ciated benefits are seldom elucidated [1, 2]. We show the hosts, the same individual vine changes its leaf morphology
occurrence of leaf mimicry in a climbing plant endemic to a accordingly (Figure S1 available online). To quantify this phe-
temperate rainforest. The woody vine Boquila trifoliolata nomenon, we compared 11 leaf traits from both B. trifoliolata
mimics the leaves of its supporting trees in terms of size, individuals and the tree species with which they were associ-
shape, color, orientation, petiole length, and/or tip spini- ated in a mature forest (45 vine individuals associated with
ness. Moreover, sequential leaf mimicry occurs when a sin- 12 host tree species). We further evaluated whether leaf mim-
gle individual vine is associated with different tree species. icry by this vine was related to herbivore avoidance, in analogy
Leaves of unsupported vines differed from leaves of climb- to cryptic behavior against predators in animals.
ing plants closely associated with tree foliage but did not The statistical analysis (a mixed generalized linear model
differ from those of vines climbing onto leafless trunks. [GLM] with observations of tree leaf phenotype nested in spe-
Consistent with an herbivory-avoidance hypothesis, leaf cies, which was a random factor) showed a significant asso-
herbivory on unsupported vines was greater than that on ciation between the leaf phenotype of B. trifoliolata and that
vines climbing on trees but was greatest on vines climbing of the supporting trees in 9 of the 11 leaf traits measured,
onto leafless trunks. Thus, B. trifoliolata gains protection including leaf and leaflet angle, leaf area and perimeter, leaflet
against herbivory not merely by climbing and thus avoiding petiole length, and leaf color (Table 1). These patterns can
ground herbivores [3] but also by climbing onto trees whose hardly be explained by covariation of leaf phenotype with light
leaves are mimicked. Unlike earlier cases of plant mimicry availability because (1) the light environment of sampling sites
or crypsis, in which the plant roughly resembles a back- was rather homogeneous (4%8% light availability), and (2) the
ground or color pattern [47] or mimics a single host [8, 9], host tree species, with contrasting leaf phenotypes, are not
B. trifoliolata is able to mimic several hosts. segregated across the light gradient [15]. Furthermore, leaves
of prostrate individuals of B. trifoliolata (i.e., those vines
Results and Discussion growing on the ground) did not differ from those of vines
that were climbing onto leafless stems or trunks (multivariate
Mimicry cases in plants are not common, and their adaptive analysis of variance [MANOVA]; Table 2) but did differ from
value is rarely reported [1, 2]. The most known example of those climbing onto leafed individuals of the analyzed tree
mimicry in plants occurs in Australian mistletoes, a group of species (7 of 8 species, MANOVA; Table 2; Figure S2). There-
hemiparasitic plants whose leaves mimic those of their fore, when there is no leaf to mimic, climbing plants are not
respective host tree species [8]. The associated benefits or different from plants growing unsupported, which show the
ecological agents involved in this case of leaf mimicry are standard leaf phenotype of the species. We also verified
not clearly discerned [9]. Floral mimicry in which pollinators that individuals growing on bare tree trunks did differ from
are attracted and deceived [10, 11] has also been reported those growing on leafed tree hosts (6 of 8 species, MANOVA;
(mainly describing the resemblance between two species). data not shown; Figure S2).
Other examples of mimicry or crypsis in plants include leaf We found some field evidence supporting the hypothesis
variegation, which is a whitish mottling that resembles leaf that leaf mimicry in climbing individuals of B. trifoliolata is
damage by mining larvae and may deter herbivores that avoid related to herbivore avoidance. First, following the premise
feeding or ovipositing on previously attacked leaves [4], that indistinguishable phenotypes should lead to similar levels
succulent Lithops plants that resemble stones in arid regions of leaf damage [9], we found in paired comparisons that
of Southern Africa [7], and leaves [6] or bracts [5] that may herbivory did not differ between climbing vines and the sup-
make a plant cryptic against a leaf litter background. Even porting host trees (t138 = 21.712, p = 0.09; mean 6 SE of an
though evidence of mimicry in plants has accumulated herbivory index: vines 1.91 6 0.04 and trees 2.01 6 0.04).
recently, it remains a rather contentious issue [1]. Second, leaf herbivory was significantly higher in creeping,
The climbing plant Boquila trifoliolata (Lardizabalaceae) unsupported individuals than in those climbing on trees
is endemic to the temperate rainforest of southern South (Figure 2). Third, leaf herbivory on vine individuals climbing
America [12]. Leaves of this twining vine are very variable in onto leafless supportson which there is no leaf model to
mimicwas higher than leaf herbivory on unsupported
individuals (Figure 2). Given that leafless stems conferred no
*Correspondence: egianoli@userena.cl protection, these results suggest that B. trifoliolata gains
Leaf Mimicry in a Climbing Plant
985

Figure 1. Leaf Mimicry in the Climbing Plant Boquila trifoliolata


Pictures of the twining vine B. trifoliolata co-occurring with woody species in the temperate rainforest of southern Chile, where leaf mimicry in terms of size,
color, and/or shape is evident. White arrows point to the vine (V) and to the host tree (T). Leaf length of the tree species is shown in parentheses [13]; this may
help to estimate leaf size variation in the vine.
(A) Myrceugenia planipes (3.58 cm).
(B) Rhaphithamnus spinosus (12 cm).
(C) Eucryphia cordifolia (57 cm). Notably smaller leaves of B. trifoliolata appear to the left of the focus leaf.
(D) Mitraria coccinea (a woody vine; 1.53.5 cm). Both here and in (F), the serrated leaf margin of the model cannot be mimicked, but the vine shows one or
two indents.
(E) Aextoxicon punctatum (59 cm).
(F) Aristotelia chilensis (38 cm).
(G) Rhaphithamnus spinosus (12 cm).
(H) Luma apiculata (12.5 cm). The inset shows more clearly how B. trifoliolata has a spiny tip, like the supporting treelet and unlike all the other pictures (and
the botanical description) of this vine. See also Figure S1 for pictures showing different leaves of the same individual of B. trifoliolata mimicking different host
trees.

protection against herbivores not merely by climbing and tropical rainforests. Likewise, when vegetative bracts that
thus avoiding ground herbivores or by benefiting from associ- resemble leaf litter and cover reproductive stems in an
ational resistance, as has been shown for other twining vines Appalachian herb were experimentally removed, herbivory
[3, 16, 17], but also by climbing onto host trees whose leaves increased and fruit set decreased [5]. In the case of Australian
are mimicked. For crypsis to be a viable defensive strategy mistletoes that mimic their hosts, protection against arboreal
in plants, herbivores to be avoided should be visually orien- herbivores (possums) had been hypothesized [8], but no
tated [2]. The main herbivores in the study site are small benefit in terms of reduced herbivory has been found [9].
gastropods, weevils, and leaf beetles [18], whose foraging Nonetheless, mistletoe species mimicking host leaves have
behavior relies on both visual and olfactory cues [1922]. higher nitrogen levels (a proxy for palatability to herbivores)
A case of leaf mimicry in which its defensive role has been than their host trees, whereas nonmimicking mistletoes
documented is that of leaf variegation. Thus, variegated leaves show lower leaf nitrogen content than their hosts [24]. This
were less attacked by herbivores than plain green leaves were suggests that mistletoe mimicry might have arisen as an
in both a subcanopy liana [4] and an understory herb [23] from adaptive response against herbivory risk [9, 24].
Current Biology Vol 24 No 9
986

Table 1. Relationships among Leaf Traits of the Climbing Plant Boquila


trifoliolata and Those of Its Supporting Trees

Leaf Traits Adjusted r2 p


Maximum width 0.49 <0.001
Maximum length 0.54 <0.001
Area 0.63 <0.001
Perimeter 0.60 <0.001
Area:perimeter 0.07 0.272
Thickness 0.15 0.130
Leaf petiole length 0.58 <0.001
Leaflet petiole length 0.25 0.037
Leaf angle 0.23 0.045
Leaflet angle 0.62 <0.001
Color 0.32 0.012
Results of mixed GLM analyses are shown. The predictor variable was tree
leaf phenotype (individual observations) nested in species, which was a Figure 2. Herbivory on Leaves of Boquila trifoliolata in Different Scenarios
random factor. Adjusted r2 corresponds to the whole model, with degrees Herbivory index (mean 6 SE) of leaves of Boquila individuals that were either
of freedom (df) = 12, 32. In all cases, the multiple r coefficients were >0. creeping unsupported on the forest understory (no support), climbing on
n = 45 Boquila individuals in 12 host tree species. supporting trees that had their leaves close to those of the vines (host trees),
or climbing onto leafless supports (leafless stems). Different letters above
bars indicate significant differences between groups (p < 0.05, Tukey
honestly significant difference test, following a one-way ANOVA in which
Unlike the Australian mistletoes, which are hemiparasites
F2,102 = 42.01; p < 0.001).
with physiological connections to the host tree [9],
B. trifoliolata is merely a climbing plant whose stems are in
contact with the trunks or branches of host trees. In fact, the model is not directly in contact with the vine, we might hy-
we have observed that leaf mimicry may occur even when pothesize that host plant volatiles trigger specific phenotypic
there is no contact at all between the vine and the mimicked changes in neighboring vine leaves. Volatile organic com-
tree. The Australian mistletoes typically show specific asso- pounds have been shown to elicit specific responses in neigh-
ciations with only one or two host trees [8, 25], thus allowing boring plants, including induction of secondary metabolites,
a long-term evolutionary process of mimicry, and show little increased expression of defense-related genes, and, overall,
phenotypic variation when found on hosts other than their spe- numerous changes of the plant transcriptome [2830]. How-
cific hosts [8, 26]. In contrast, mimicry by B. trifoliolata involves ever, research on airborne plant-plant signaling has not re-
phenotypic plasticity because the same individual is able to ported specific morphological changes in leaves, and nothing
resemble several host trees simultaneously. Furthermore, un- is known about the involvement of volatiles of the model in leaf
like other reported cases of mimicry or crypsis in plants [47], mimicry via induced reprogramming of genes in the mimic.
in which the mimetic plant roughly resembles an undetermined An alternative hypothesis, but perhaps less plausible, would
background or color pattern, B. trifoliolata is able to mimic consider horizontal gene transfer between plants, a phenome-
several hosts. The imitation of several distinct models by a non that is increasingly being reported [3136]. These cases
single species has been deemed particularly advantageous include both single and multiple transfer events per species,
in light of frequency-dependent selection [27], but evidence which are hypothesized to be mediated by a vector or result
of such mimetic polymorphisms is rare and limited to butterfly from plant-plant parasitism or natural grafts [3136]. The plas-
species [27]. Moreover, these phenomena are explained not ticity in leaf mimicry in B. trifoliolata could involve horizontal
by plastic responses but by population divergence. gene transfer on an ecological timescale and might be medi-
We currently lack a mechanistic explanation for this unique ated by airborne microorganisms. The latter speculation is
phenomenon. Because leaf mimicry is observed even when based on the fact that mimicry is observed with respect to
the foliage to which the vine is nearest, irrespective of whether
this foliage belongs to the host tree that the vine has climbed.
Table 2. Comparison of the Leaf Phenotype of Prostrate Vines versus Further research on leaf mimicry by B. trifoliolata might lead
Vines Climbing on Leafless Trunks and on Tree Hosts to the identification of the host tree volatiles or vector-medi-
Prostrate Boquila versus
ated gene transfers that trigger differential gene expression
Boquila Climbing on: Wilks l F df p in this singular climbing plant.
Leafless trunks 0.712 1.469 11, 40 0.182
Aristotelia chilensis 0.107 15.879 11, 21 <0.001 Supplemental Information
Amomyrtus luma 0.393 3.087 11, 22 0.012
Azara lanceolata 0.534 1.268 11, 16 0.324 Supplemental Information includes Supplemental Experimental Procedures
Aextoxicon punctatum 0.158 8.209 11, 17 <0.001 and two figures and can be found with this article online at http://dx.doi.org/
Caldcluvia paniculata 0.141 8.875 11, 16 <0.001 10.1016/j.cub.2014.03.010.
Eucryphia cordifolia 0.157 9.264 11, 19 <0.001
Myrceugenia planipes 0.219 7.139 11, 22 <0.001 Acknowledgments
Rhaphithamnus spinosus 0.364 3.172 11, 20 0.012
This study was supported by FONDECYT (Chilean Agency of Research
Results of paired MANOVAs in which the leaf phenotype (all leaf traits taken Fund) grant 1100585 to E.G. We are grateful to A. Saldana, C. Salgado-
together) of prostrate vines was compared against the leaf phenotype of Luarte, O. Godoy, M. Cisternas, and D. Gianoli for their help in the field
vines climbing on leafless trunks and against the leaf phenotype of vines and their contribution of pictures. We thank K.C. Burns and three anony-
climbing on tree hosts (see Table 1 for trait list). See also Figure S2 for a scat- mous reviewers for thoughtful comments that improved the manuscript
terplot of canonical scores for the leaf phenotype of the three vine groups and CONAF (Chilean Agency of Protected Areas) for granting us permits
included in this table. to work in Puyehue National Park.
Leaf Mimicry in a Climbing Plant
987

Received: December 12, 2013 25. Downey, P.O. (1998). An inventory of host species for each aerial mistletoe
Revised: January 22, 2014 species (Loranthaceae and Viscaceae) in Australia. Cunninghamia 5,
Accepted: March 5, 2014 685720.
Published: April 24, 2014 26. Blick, R.A.J., Burns, K.C., and Moles, A.T. (2012). Predicting network to-
pology of mistletoehost interactions: do mistletoes really mimic their
References hosts? Oikos 121, 761771.
27. Joron, M., and Mallet, J.L.B. (1998). Diversity in mimicry: paradox or
1. Schaefer, H.M., and Ruxton, G.D. (2009). Deception in plants: mimicry paradigm? Trends Ecol. Evol. 13, 461466.
or perceptual exploitation? Trends Ecol. Evol. 24, 676685. 28. Godard, K.-A., White, R., and Bohlmann, J. (2008). Monoterpene-
2. Burns, K.C. (2010). Is crypsis a common defensive strategy in plants? induced molecular responses in Arabidopsis thaliana. Phytochemistry
Speculation on signal deception in the New Zealand flora. Plant 69, 18381849.
Signal. Behav. 5, 913. 29. Heil, M., and Karban, R. (2010). Explaining evolution of plant communi-
3. Gianoli, E., and Molina-Montenegro, M.A. (2005). Leaf damage induces cation by airborne signals. Trends Ecol. Evol. 25, 137144.
twining in a climbing plant. New Phytol. 167, 385389. 30. Pearse, I.S., and Karban, R. (2013). Do plant-plant signals mediate
4. Smith, A.P. (1986). Ecology of a leaf color polymorphism in a tropical herbivory consistently in multiple taxa and ecological contexts?
forest species: habitat segregation and herbivory. Oecologia 69, J. Plant Interact. 8, 203206.
283287. 31. Richardson, A.O., and Palmer, J.D. (2007). Horizontal gene transfer in
plants. J. Exp. Bot. 58, 19.
5. Klooster, M.R., Clark, D.L., and Culley, T.M. (2009). Cryptic bracts facil-
32. Diao, X., Freeling, M., and Lisch, D. (2006). Horizontal transfer of a plant
itate herbivore avoidance in the mycoheterotrophic plant Monotropsis
transposon. PLoS Biol. 4, e5.
odorata (Ericaceae). Am. J. Bot. 96, 21972205.
33. Bock, R. (2010). The give-and-take of DNA: horizontal gene transfer in
6. Fadzly, N., and Burns, K.C. (2010). Hiding from the ghost of herbivory
plants. Trends Plant Sci. 15, 1122.
past: evidence for crypsis in an insular tree species. Int. J. Plant Sci.
34. Gao, C., Ren, X., Mason, A.S., Liu, H., Xiao, M., Li, J., and Fu, D. (2014).
171, 828833.
Horizontal gene transfer in plants. Funct. Integr. Genomics 14, 2329.
7. Kellner, A., Ritz, C.M., Schlittenhardt, P., and Hellwig, F.H. (2011).
35. Xi, Z., Wang, Y., Bradley, R.K., Sugumaran, M., Marx, C.J., Rest, J.S.,
Genetic differentiation in the genus Lithops L. (Ruschioideae,
and Davis, C.C. (2013). Massive mitochondrial gene transfer in a para-
Aizoaceae) reveals a high level of convergent evolution and reflects
sitic flowering plant clade. PLoS Genet. 9, e1003265.
geographic distribution. Plant Biol (Stuttg) 13, 368380.
36. Stegemann, S., Keuthe, M., Greiner, S., and Bock, R. (2012). Horizontal
8. Barlow, B.A., and Wiens, D. (1977). Host-parasite resemblance in
transfer of chloroplast genomes between plant species. Proc. Natl.
Australian mistletoes: the case for cryptic mimicry. Evolution 31, 6984.
Acad. Sci. USA 109, 24342438.
9. Canyon, D.V., and Hill, C.J. (1997). Mistletoe host-resemblance: a study
of herbivory, nitrogen and moisture in two Australian mistletoes and
their host trees. Aust. J. Ecol. 22, 395403.
10. Gaskett, A.C. (2012). Floral shape mimicry and variation in sexually
deceptive orchids with a shared pollinator. Biol. J. Linn. Soc. Lond.
106, 469481.
11. Benitez-Vieyra, S., Hempel de Ibarra, N., Wertlen, A.M., and Cocucci,
A.A. (2007). How to look like a mallow: evidence of floral mimicry be-
tween Turneraceae and Malvaceae. Proc. Biol. Sci. 274, 22392248.
12. Gianoli, E., Saldana, A., Jimenez-Castillo, M., and Valladares, F. (2010).
Distribution and abundance of vines along the light gradient in a south-
ern temperate rainforest. J. Veg. Sci. 21, 6673.
13. Munoz-Schick, M. (1980). Flora del Parque Nacional Puyehue (Santiago:
Editorial Universitaria).
14. Ruiz, E. (2003). Lardizabalaceae. In Flora de Chile, Volume 2,
Marticorena, C., and Rodrguez, R., eds. (Concepcion: Universidad de
Concepcion), pp. 2427.
15. Lusk, C.H., Chazdon, R.L., and Hofmann, G. (2006). A bounded null
model explains juvenile tree community structure along light availability
gradients in a temperate rain forest. Oikos 112, 131137.
16. Gonzalez-Teuber, M., and Gianoli, E. (2008). Damage and shade
enhance climbing and promote associational resistance in a climbing
plant. J. Ecol. 96, 122126.
17. Sasal, Y., and Suarez, M.L. (2011). Insect herbivory in climberhost
shrubs associations: benefit or detriment? Austral Ecol. 36, 814820.
18. Salgado-Luarte, C., and Gianoli, E. (2012). Herbivores modify selection
on plant functional traits in a temperate rainforest understory. Am. Nat.
180, E42E53.
19. Rollo, C.D., and Wellington, W.G. (1981). Environmental orientation by
terrestrial Mollusca with particular reference to homing behaviour.
Can. J. Zool. 59, 225239.
20. Hamilton, P.V., and Winter, M.A. (1982). Behavioural responses to visual
stimuli by the snail Littorina irrorata. Anim. Behav. 30, 752760.
21. Bell, W.J. (1990). Searching behaviour patterns in insects. Annu. Rev.
Entomol. 35, 447467.
22. Stenberg, J.A., and Ericson, L. (2007). Visual cues override olfactory
cues in the host-finding process of the monophagous leaf beetle
Altica engstroemi. Entomol. Exp. Appl. 125, 8188.
23. Soltau, U., Dotterl, S., and Liede-Schumann, S. (2009). Leaf variegation
in Caladium steudneriifolium (Araceae): a case of mimicry? Evol. Ecol.
23, 503512.
24. Ehleringer, J.R., Ulmann, I., Lange, O.L., Farquhar, G.D., Cowan, I.R.,
and Schulze, E.-D. (1986). Mistletoes: a hypothesis concerning morpho-
logical and chemical avoidance of herbivory. Oecologia 70, 234237.
Current Biology, Volume 24

Supplemental Information

Leaf Mimicry in a Climbing Plant

Protects against Herbivory


Ernesto Gianoli and Fernando Carrasco-Urra
Supplemental Information

Supplemental Data

Figure S1, Related to Figure 1

Figure S1. Each row includes pictures of a case in which the same individual of the twining vine

Boquila trifoliolata changes its phenotype while in contact with different supporting trees or

shrubs. A red arrow points to the vine and a blue arrow points to the host species. A) left: this is

the standard appearance of the vines leaf when there are no heterospecific leaves nearby;

center: the color, veins and toughness of leaves from Rhaphithamnus spinosus are mimicked;

right: Boquila trifoliolata leaves resemble the size and shape of leaves from Aextoxicon

punctatum, and are ca. 10-fold larger than those shown in the left panel. B) the vines leaves

resemble the thin and light-green leaves of Fuchsia magellanica (left) and then the thick and

dark-green leaves of Myrceugenia planipes (right).


Figure S2, Related to Table 2

Figure S2. Scatterplot of canonical scores for the leaf phenotype of the three vine groups

included in Table 2: prostrate (no support green triangles), climbing on host tree species (host

trees black circles) and climbing on bare trunks (leafless stems red circles). Discriminate

functions (roots) resulted from combinations of the leaf traits shown in Table 1. It is shown that

leaves of prostrate vines did not differ from leaves of vines that were climbing onto leafless

trunks, but did differ from those climbing onto leafed host trees.

Host trees
4 Leafless stems
No support

2
ROOT 2

-2

-4

-4 -2 0 2 4

ROOT 1
Supplemental Experimental Procedures

Study species and study site

The woody vine Boquila trifoliolata, a monotypic genus from the Lardizabalaceae family, is

endemic to the temperate rainforest of southern South America [S1]. It is a monoecious plant,

with low growth rate (E. Gianoli, personal observations), and is distributed along the entire light

gradient in the forest [S1]. This twining vine has slender, slightly hairy stems when young, and its

leaves are composed of three leaflets, which are pulvinated and therefore may change their

orientation [S2]. The central leaflet is larger than the lateral ones, and leaflets show significant

variation in size (10-100 mm) and shape. Woody stems of old individuals (3-5 cm ) are used as

ropes by native people [S2].

Field observations and measurements were conducted in a temperate rainforest

located within Puyehue National Park (40 39S, 72 11W; 350400 m a.s.l.), southern

Chile. Mean annual precipitation is 2800 mm and mean annual temperature 9.8 C [S3]. The

old-growth lowland forest is composed of evergreen tree species and woody vines [S4,S5],

and 43% of microsites occur at about 5% canopy openness [S6]. Soils are well-drained and

have low nutrient availability [S7]. The main herbivores are small slugs and snails, weevils

and leaf beetles, and are evenly distributed across the light gradient [S8,S9].

Leaf phenotype

We searched for host trees with climbing B. trifoliolata individuals in random walks across the

mature forest within an area of ca. 2 ha. The light environment of sampling sites was rather

homogeneous, with 48% full sunlight (photosynthetically active radiation, mol m-2 s-1,

measured with a LI-COR LI-250 light meter) compared with a large gap nearby. A total of 45

host trees from 12 species with climbing individuals of B. trifoliolata were identified and

sampled: Aextoxicon punctatum (3), Amomyrtus luma (8), Aristotelia chilensis (7), Azara
lanceolata (2), Caldcluvia paniculata (2), Embothrium coccineum (1), Eucryphia cordifolia (5),

Laureliopsis philippiana (1), Luma apiculata (1), Myrceugenia planipes (8), Pseudopanax

laetevirens (1), Rhaphithamnus spinosus (6). In each of these 45 vine-tree associations we

selected the two closest leaves and measured 11 leaf traits in B. trifoliolata and the supporting

tree (height: 80-160 cm). We measured in situ, with the aid of a ruler, a protractor, and a digital

caliper (Mitutoyo; 0.01-mm resolution), the following leaf traits: angle (with the horizon),

thickness, petiole length, leaflet petiole length and leaflet angle (when trees had no leaflets, we

recorded their leaf traits instead). We also rated leaf color by choosing the most similar tone in a

color palette of 216 different tones. Maximum width, maximum length, area, perimeter and

area/perimeter were quantified after image analysis (SigmaScan) of digital pictures. Separate

mixed GLMs between vine phenotype (dependent variable) and host-tree leaf phenotype were

conducted for each leaf trait. Observations of tree leaf phenotype were nested in species, which

was considered a random factor (Table 1). Significant relationships indicate that phenotypes

covary, i.e., that there is leaf mimicry.

In another sampling event, we measured the same set of leaf traits as described above in

B. trifoliolata individuals that were either: i) creeping unsupported on the forest understory (n =

30), ii) climbing on host trees (the 8 more common tree species found in the previous sampling)

(n = 45), or iii) climbing onto leafless supports (bare tree trunks) (n = 30). These plants were

located across the same mature forest as before. To avoid sampling twice the same genet, all

plants were at least 10 m apart. We measured one leaf per individual, choosing intact (or the

least damaged) leaves. We used Multivariate Analysis of Variance (MANOVA), pooling all leaf

traits, in order to compare the leaf phenotype among the three groups (Table 2). We expected that

leaves of prostrate plants should not differ from those of vines climbing onto leafless supports (no
leaf to mimic), but should differ from those of vines associated to the mimicked trees. A

discriminant function plot was used to visualize the results (Figure S2).

Herbivory

We estimated the magnitude of herbivory in two events. In the first event, we made paired

comparisons between individual vines and their companion host trees (n = 70). In the second

event, we sampled the same plants from the three groups where leaf multivariate phenotype was

determined (see preceding section). Each of three leaves per plant was assigned to one of the

following categories of damage, based on visual inspection of leaf area removed: 1, no damage;

2, less than 25% damage; 3, from 25% to 50% damage; 4, from 50% to 75% damage; and 5,

damage above 75%. The score of all leaves was used to calculate an individual herbivory index,

HI = nC15 N-1; where C is the category of damage, n is the number of leaves in the Cth

category, and N is the number of leaves sampled (three in this case) [S8]. In the first event,

herbivory was compared between groups with a paired two-tail t-test. In the second event,

differences among groups where evaluated with a one-way ANOVA followed by a Tukey test

(Figure 2). All statistical analyses in the study were conducted using Statistica (StatSoft
Supplemental References

S1. Gianoli, E., Saldaa, A., Jimnez-Castillo, M., and Valladares, F. (2010). Distribution and

abundance of vines along the light gradient in a southern temperate rainforest. J. Veg. Sci.

21, 66-73.

S2. Ruiz, E. (2003). Lardizabalaceae. In Flora de Chile Vol. 2 (2), C. Marticorena and R.

Rodrguez, eds (Concepcin: Univ. Concepcin), pp. 24-27.

S3. Dorsch, K. (2003). Hydrogeologische Untersuchungen der Geothermalfelder von Puyehue

und Cordn Caulle, Chile. PhD thesis, Ludwig-Maximilians-Universitt, Mnchen,

Germany.

S4. Lusk, C. H. (2002). Leaf area accumulation helps juvenile evergreen trees tolerate shade in a

temperate rainforest. Oecologia 132, 188-196.

S5. Carrasco-Urra, F., and Gianoli, E. (2009). Abundance of climbing plants in a southern

temperate rainforest: host-tree characteristics or light availability? J. Veg. Sci. 20, 1155-

1162.

S6. Lusk, C.H., Chazdon, R.L., and Hofmann, G. (2006). A bounded null model explains

juvenile tree community structure along light availability gradients in a temperate rain

forest. Oikos 112, 131-137.

S7. Saldaa, A., and Lusk, C.H. (2003). Influencia de las especies del dosel en la disponibilidad

de recursos y regeneracin avanzada en un bosque templado lluvioso del sur de Chile. Rev.

Chil. Hist. Nat. 76, 639650.

S8. Salgado-Luarte, C., and Gianoli, E. (2010). Herbivory on temperate rainforest seedlings in

sun and shade: resistance, tolerance and habitat distribution. PLoS ONE 5(7), e11460.

S9. Salgado-Luarte, C., and Gianoli, E. (2012). Herbivores modify selection on plant functional

traits in a temperate rainforest understory. Am. Nat. 180, E42-E53.