Content uploaded by Hiroaki Sato
Author content
All content in this area was uploaded by Hiroaki Sato on Jul 31, 2017
Content may be subject to copyright.
ORIGINAL ARTICLE
Misaki Iwamoto •Chika Horikawa
Megumi Shikata •Naoko Wasaka
Teiko Kato •Hiroaki Sato
Stinging hairs on the Japanese nettle
Urtica thunbergiana
have
a defensive function against mammalian but not insect herbivores
Received: 3 October 2013 / Accepted: 24 February 2014 / Published online: 13 March 2014
The Ecological Society of Japan 2014
Abstract Thorns and hairs of plants can serve as defenses
against herbivores, although they may not have evolved
under selection by herbivory. Japanese nettles, Urtica
thunbergiana, in Nara Park, Nara Prefecture, Japan,
where sika deer have been protected for 1200 years, bear
many more stinging hairs than those in areas with few or
no deer. Previous studies suggested that such hairy
nettles evolved under natural selection imposed by in-
tense deer browsing, because stinging hairs deterred deer
browsing and because among-population variation in
hair density was associated with deer abundance. To
confirm this hypothesis, we examined (1) whether
stinging hairs affected oviposition and feeding prefer-
ences of herbivorous insects and (2) the degree to which
they deterred deer via laboratory and field experiments
with hairy nettles from Nara Park and with almost-
hairless nettles from another area. A specialist butterfly,
Indian red admiral, showed no oviposition or larval
feeding preferences for either hairy or hairless nettles.
Insect damage levels did not significantly differ between
the two variants. In contrast, deer browsed hairless
nettles more heavily than hairy ones. In hairy nettles,
however, the level of deer browsing was not propor-
tional to stinging-hair density, presumably because the
hairy nettle population had reached a plateau for resis-
tance as a result of long-term strong directional selection
for stinging hairs. These results corroborate the
hypothesis that hairy nettles in Nara Park evolved
through natural selection under intense deer browsing.
Keywords Insect damage ÆEvolution of defensive
traits ÆIndian red admiral ÆIndirect interactions Æ
Larval feeding and oviposition preferences
Introduction
Thorns, spines, prickles (here collectively termed
thorns), and trichomes (called hairs here) on leaves and
along stems can serve as defenses against herbivores
(Hanley et al. 2007). For instance, Milewski et al. (1991)
examined the effects of Acacia seyal thorns on the
browsing behavior of giraffes using thorn removal
experiments in the field and demonstrated that de-
thorned branches suffered more giraffe herbivory than
did intact branches on the same plants. Sletvold et al.
(2010) investigated the interaction between the perennial
herb Arabidopsis lyrata and its specialist herbivore, the
diamondback moth Plutella xylostella; oviposition by
moth females and the level of leaf damage by moth
larvae were negatively correlated with leaf-hair density.
These results indicated that leaf hairs deterred not only
larval feeding but also oviposition.
Nevertheless, thorns and hairs are unlikely to be
effective defenses against both vertebrates and inverte-
brates (Walters 2011). Thorns are more effective against
vertebrates, while hairs are more deterrent to inverte-
brates because of the relative sizes of the plants and
herbivores (Grubb 1992). Thorns deter mammals from
browsing the plant by inflicting painful mouth wounds.
They are, however, ineffective against invertebrates,
such as lepidopteran larvae and aphids, which can
maneuver around them. In contrast, hairs are too small
to wound mammals but can obstruct the free movement
of insects.
Stinging hairs on stems, petioles, and laminae of the
European nettle Urtica dioica are composed of a multi-
cellular pedestal surmounted by an elongate, tapering,
stinging cell (Thurston 1974). Those hairs are thought to
have evolved as a defense against mammalian herbivores
but not invertebrates. This supposition is supported by
the facts that: (1) stinging hairs are silicified and contain
a toxic fluid that causes sharp pain when injected into
human skin (Pollard and Briggs 1984b); (2) stinging
hairs are much denser on leaves in populations that are
M. Iwamoto ÆC. Horikawa ÆM. Shikata ÆN. Wasaka Æ
T. Kato ÆH. Sato (&)
Department of Biological Sciences, Faculty of Science,
Nara Women’s University, Nara 630-8506, Japan
E-mail: scarab@cc.nara-wu.ac.jp
Tel.: +81-742-203937
Fax: +81-742-203937
Ecol Res (2014) 29: 455–462
DOI 10.1007/s11284-014-1137-2
subject to high herbivory pressure by large mammals
than in populations that have lacked browsing pressure
for decades (Pollard 1986; Pollard and Briggs 1982,
1984a; Pullin and Gilbert 1989); (3) mammalian grazers
such as rabbits (Oryctolagus cuniculus) and sheep (Ovis
aries) avoid plants with more stinging hairs (Pollard and
Briggs 1984b); (4) the level of insect damage is not
associated with the density of stinging hairs in the field
(Pollard and Briggs 1984b); and (5) there is no evidence
that stinging hairs deter or interfere with feeding by
invertebrate herbivores such as red admiral butterfly
(Vanessa atalanta), Japanese beetle (Popillia japonica),
grasshopper (Chortophaga viridifasciata), and snail
(Anguispira alternata) (Tuberville et al. 1996).
Like U. dioica, the Japanese nettle U. thunbergiana
Siebold et Zucc. also has stinging hairs. Kato et al.
(2007) examined stinging-hair densities in nettles of
Nara Park, Nara City, central Japan, where many sika
deer (Cervus nippon Temminck) have been protected for
1200 years in an area of 6.6 km
2
. They found that the
park’s nettles had 58–630 times higher stinging-hair
densities on leaves than did nettles in other areas with
few or no sika deer. They also showed that the park’s
native nettles were less frequently browsed by deer than
nettles with few or no stinging hairs that were trans-
planted from an unbrowsed area into the park. Fur-
thermore, they showed that nettles raised from seeds in a
greenhouse retained similar stinging-hair densities of
wild populations from which the seeds originated, sug-
gesting that stinging-hair density was controlled geneti-
cally. Recently, Shikata et al. (2013) investigated the
effects of soil fertility, light intensity, and deer habitat-
use frequency on stinging-hair density in 19 wild nettle
populations, including those of Nara Park. They re-
vealed that stinging-hair density was positively corre-
lated with deer habitat-use frequency and independent
of soil fertility and light intensity. Therefore, we
hypothesize that nettles with high stinging-hair densities
in Nara Park evolved under natural selection imposed
by intense browsing by sika deer.
However, to confirm this hypothesis, we need to solve
at least two problems. First, whether the high density of
stinging hairs on Nara Park nettles affects insect
behaviors such as oviposition and feeding is unclear.
Although stinging hairs on U. dioica were experimen-
tally shown not to interfere with insect feeding (Tuber-
ville et al. 1996), this experiment used nettles that varied
in stinging-hair density by at most threefold. In the
weedy shrub Wigandia urens (Boraginaceae), contrary to
predictions, the level of damage by insects was higher on
bristly leaves with stinging hairs than on smooth leaves
without stinging hairs (Cano-Santana and Oyama 1992).
Thus, dense stinging hairs on Nara Park nettles might
have a defensive function against insects or, conversely,
permit more insect damage. In either case, deer browsing
might induce more stinging hairs, which might indirectly
affect the behavior and/or density of insect herbivores.
This problem is related to the phenomenon of trait-
mediated indirect interactions in ecological communities
(Ohgushi et al. 2012).
Second, the degree to which stinging hairs are effec-
tive against sika deer is uncertain. As mentioned above,
Kato et al. (2007) compared browsing levels on two
extreme variants from different populations, i.e., plants
that were virtually devoid of stinging hairs and those
that were heavily armed. They showed that hairy nettles
suffered much less damage than hairless ones. Unfor-
tunately, they failed to consider within-population var-
iation in stinging-hair density, so whether damage level
decreases consistently with an increase in hair density
remains uncertain.
The objective of the present study was to determine
whether the stinging hairs of U. thunbergiana had a
defensive function against insect herbivores and the de-
gree to which they contributed to resistance against sika
deer. We examined oviposition and larval feeding pref-
erences of a specialist butterfly (Indian red admiral) for
hairy and almost-hairless nettles, compared levels of
damage by insects in the two variants, and analyzed the
relationship between stinging-hair density and the like-
lihood of deer browsing. Finally, from the viewpoint of
trait-mediated indirect interactions among deer, nettle,
and herbivorous insects, we considered the possibility
that an increase in stinging hairs caused by intense deer
browsing would affect herbivorous insects.
Materials and methods
Study organisms
The Japanese stinging nettle Urtica thunbergiana is a
wind-pollinated perennial plant that grows on forest
edges in central and southern Japan (Kitamura and
Murata 1961). Urtica thunbergiana is nitrophilous, as
has been reported for U. dioica (Olsen 1921; Molisch and
Dobat 1979; Kato 2001).
Nettles used in this study originated in Nara Park
(6.6 ha, 3441¢N, 13551¢E, 110 m a.s.l.) and Takatori
Castle Site (3425¢N, 13549¢E, 550–580 m a.s.l.), Nara
Prefecture, central Japan. In Nara Park, several hundred
sika deer have been protected for more than 1200 years,
and nettle leaves have much higher densities of stinging
hairs than those in other locations with no evidence of
sika deer presence (Kato et al. 2007). Takatori Castle Site
is 28.5 km south of Nara Park. The area around the site
was not considered to be inhabited by sika deer, because
there were no signs such as footprints, dung pellets, or
browsing marks during the course of the present study.
The Indian red admiral butterfly, Vanessa indica
Herbst, occurs from India through Indochina and China
to Japan. This butterfly is multivoltine, and its larvae
feed mainly on Boehmeria nivea,B. sylvestris,B. spicata,
B. japonica,U. angustifolia, and U. thunbergiana of the
family Urticaceae in Japan (Teshirogi 1990). In Nara
Park and Takatori Castle Site, Indian red admirals
456
overwinter usually as adults and rarely as larvae. Over-
wintered adult females lay eggs in late March to mid-
April. Eggs hatch in April, and larvae grow into pupae
through five instars.
Oviposition preference in Indian red admiral
Twenty-five well-grown nettles were collected in each of
Nara Park and Takatori Castle Site on 20 November,
2006, and transplanted to a corner of the campus of
Nara Women’s University, which is 500 m from Nara
Park. On the campus, nettles with few stinging hairs on
leaves grow naturally (Shikata et al. 2013). Nettles from
Nara Park and Takatori Castle Site were alternated at
intervals of 1 m in two rows that were 4 m apart. One
row consisted of 24 plants and the other 26. These net-
tles were exposed to oviposition by V. indica females.
The number of butterfly eggs was counted on all leaves
of each nettle on 20 July, 2007. These eggs were laid by
females of the first generation in the year.
The size of each nettle was approximated by the vol-
ume of a cylinder: p·maximum length ·maximum
width ·height on 26 July, 2007. For each plant, two
leaves on each of the second, third, and fourth nodes from
the tip of major shoots were removed, and their densities
of stinging hairs were measured. The upper surface of
each leaf was scanned using an image scanner (EPSON
GT-X770, Seiko Epson Co., Suwa, Japan) with a reso-
lution of 300 dpi. On the digital image, all stinging hairs
on the surface were counted, and leaf area was measured
using the freeware LIA32 for Windows95 ver. 0.376b1
(Yamamoto 1997). Density was calculated by dividing
the number of stinging hairs by the leaf area (cm
2
).
Larval feeding preference in Indian red admiral
Twenty-eight and 23 eggs of V. indica on nettle leaves
were sampled in Nara Park and Takatori Castle Site,
respectively, on 1 April, 2008. Eggs and leaves were
incubated at 24 C under a photoperiod of 14L:10D to
induce hatching. Short-term (24 h) larval preference for
nettles of Nara Park and Takatori Castle Site was as-
sessed by a choice-test. Test arenas were plastic trans-
parent cups (10 cm diam. below, 13 cm diam. above,
10 cm in height) containing one nettle leaf from each
location on the bottom. The areas of those leaves were
nearly equal. A larva within 12 h of hatching was placed
at the midpoint between the two leaves and allowed to
feed ad libitum at 20 C under 14L:10D. Twenty-four
hours later, we recorded which leaf was consumed. Then,
these leaves were removed, and two new leaves were
supplied. This procedure was repeated every day until the
second day of the second instar. Data from the first days
of the first and the second instars were analyzed. When a
larva consumed both leaves nearly equally, data from the
next day were used. When the larva again showed no
preference, that replicate was discarded.
Damage by chewing and leafmining insects
Twenty-five nettles were collected in each of Nara Park
and Takatori Castle Site on 9 and 10 March, 2010. Each
nettle was divided at the rhizome into three clonal parts.
They were planted individually in 1.6-L plastic pots that
were filled with gardening soil (Takii No Baiyoˆ do, Takii
Co, Japan) containing 300 mg/L N, 470 mg/L PO
3
, and
390 mg/L K. They were raised in a greenhouse under
85 % shading and watered periodically. On 18 March,
2011, 80 surviving nettles (40 from each site) were
transplanted individually in 4.3-L plastic pots filled with
the same soil as above. Forty pairs of nettles (one from
each site) were placed at a distance of more than 3 m
from each other on the campus of Nara Women’s
University; the two pots of each pair were 10 cm apart.
For each nettle, the total number of leaves and the
number of leaves consumed by chewing and leafmining
insects (>20 % of area removed) were counted on 31
May and 28 June, 2011. The percentage of consumed
area was estimated by eye. Then, the proportion of
consumed to total leaves was calculated. Common
chewing insects were green weevils (Phyllobius sp.),
mother-of-pearls (Pleuroptya ruralis), and Indian red
admirals, while leafmining insects were agromizyid flies.
The two types of insect damage were combined, because
they had the same impact on the loss of the photosyn-
thetic organ.
The likelihood of deer browsing
Ten of the 80 nettles that were used to assess insect
damage (five from each site) were randomly chosen on 5
July, 2011. For each nettle, the number of leaves was
counted, and four intact leaves were removed from
major shoots. For each leaf, stinging-hair density (cm
2
)
was obtained as described above. Five pairs, each of
which consisted of a nettle from Nara Park and one
from Takatori Castle Site, were placed at least 20 m
apart from each other in Nara Park. Nettles of a pair
were set 3 m apart from one another. They were exposed
to deer browsing under natural conditions for 24 h. To
obtain the likelihood of deer browsing for each nettle,
we counted the number of leaves eaten by deer and di-
vided it by the initial number of leaves. This procedure
was replicated using other nettles on 12–13, 21–22, and
28–29 July and 3–4 August. In total, 30 pairs were
examined. Seven pairs of nettles in which one or both
plants were toppled by browsers were excluded from
subsequent analysis.
Statistical analyses
Before subsequent analyses, stinging-hair density was
log-transformed to improve the fit to a normal distri-
bution (Krebs 1998). The proportions of leaves con-
sumed by insects (>20 % of area removed) and leaves
457
eaten by deer were arcsin-transformed. The data are
presented as mean ± SD before transformation.
The oviposition preference of Indian red admirals for
nettles of Nara Park or Takatori Castle Site was ana-
lyzed by single factor analysis of covariance (ANCO-
VA). Plant size was an independent variable (i.e., the
covariate), the number of eggs per plant was a depen-
dent variable, and nettle population was a fixed factor.
Larval preference for nettle leaves of Nara Park or
Takatori Castle Site was analyzed for the first and sec-
ond instars according to Sokal and Rohlf’s (2012) rep-
licated goodness-of-fit tests. First, homogeneity of the
ratios of larvae choosing leaves of Nara Park and
Takatori Castle Site was tested for each larval popula-
tion using the heterogeneity Gtest. Then, goodness of fit
to a 1:1 ratio was tested for the pooled data by means of
a standard Gtest.
Differences in the levels of insect and browsing
damage between nettles of Nara Park and Takatori
Castle Site were tested using a paired ttest.
We hypothesized that the likelihood of deer browsing
a nettle of Nara Park would be related to the stinging-
hair density on its leaves and the potential intensity of
deer browsing to where and when the nettle was placed.
The potential was expressed by the proportion of leaves
of the Takatori Castle Site nettles eaten by deer, based
on the assumption that nettles of Takatori Castle Site
would be readily consumed by deer because they bore
few stinging hairs. The proportion of leaves eaten by
deer was not significantly associated with stinging-hair
density in nettles of Takatori Castle Site (r=0.390,
df = 21, P= 0.065). The hypothesis was tested via
multiple linear regression analysis, where the dependent
variable was the proportion of leaves eaten by deer in
nettles of Nara Park and the predictor variables were
stinging-hair density and potential browsing intensity.
Statistically significant difference in stinging-hair
density between nettles of Nara Park and Takatori
Castle Site was tested with a mixed model nested ana-
lysis of variance (Sokal and Rohlf 2012), where popu-
lation was a fixed factor, individual was a random factor
nested within population, and leaf was a random factor
nested within individual.
All analyses were performed with SPSS ver 15.0J
(SPSS Inc 2006).
Results
Oviposition preference in Indian red admiral
Forty-seven of 50 transplanted nettles (23 for Nara
Park, 24 for Takatori Castle Site) survived until the
survey date. In total, 27 and 26 eggs of Indian red
admiral were found on nettles of Nara Park and Taka-
tori Castle Site, respectively. About one-third of the eggs
on nettles of Nara Park were laid at the tips of stinging
hairs. ANCOVA revealed no statistical significance in
the interaction between nettle population and nettle size
(Fig. 1; Table 1), and thus the regression slopes of the
number of eggs versus nettle size for the two populations
were considered to be equal. The number of eggs was
significantly associated with nettle size. We did not find a
significant effect of nettle population on the number of
eggs. These results indicated that butterfly females laid
eggs according to nettle size without discriminating the
nettle populations.
Nettles of Nara Park had significantly higher sting-
ing-hair densities than nettles of Takatori Castle Site
(2.4 ± 1.38 cm
2
and 0.04 ± 0.04 cm
2
, respectively,
F
s[1,35]
= 228.9, P< 0.001; data for six nettles of Nara
Park and four nettles of Takatori Castle Site were
accidentally lost).
Larval feeding preference in Indian red admiral
In the first or second instars, there were no significant
differences in the ratios of larvae that chose leaves of
Nara Park or Takatori Castle Site between larvae of the
respective (Gtest for homogeneity; Table 2). The Gtest
for goodness of fit to a 1:1 ratio was applied for the
pooled data. In each instar, the observed distribution did
not significantly deviate from an even distribution
(Gtest for 1:1 ratio; Table 2). These results indicated
that larvae chose leaves without discriminating the nettle
populations. First and second instar larvae were ob-
served to eat leaf tissues without consuming stinging
hairs, which were excised and left in the cup.
Leaf damage by herbivorous insects and sika deer
There was no significant difference in the proportions of
leaves damaged by insects (>20 % of area removed)
between nettles of Nara Park and Takatori Castle Site
on 31 May or 28 June (Fig. 2). In contrast, a significant
difference was found in the proportions of leaves eaten
by deer between the populations (Fig. 3). Deer ate the
0 0.5 1.0
5
4
3
2
1
0
Nara Park
Plant size (106cm3)
Takatori Castle Site
Number of eggs
1.5
2
1
0
Fig. 1 Scatterplot and regression lines of the number of Vanessa
indica eggs per plant versus plant size, which was calculated by
p·maximum length ·maximum width ·height, for nettles orig-
inating from Nara Park and Takatori Castle Site
458
whole shoots of six nettles of Takatori Castle Site, while
they left all or part of shoots of every nettle of Nara
Park. Deer were sometimes observed to jump back and
shake their heads immediately after their mouths con-
tacted nettles of Nara Park. They did not show such
pain responses to nettles of Takatori Castle Site.
Stinging-hair density was significantly higher in net-
tles of Nara Park than in those of Takatori Castle Site
(1.1 ± 0.48 and 0.01 ± 0.01 cm
2
, respectively,
F
s[1,44]
= 115.2, P< 0.001).
Multiple linear regression analysis provided the fol-
lowing prediction equation:
y¼0:310x1þ6:294x2þ3:429
where yis the arcsin-transformed proportion of leaves
eaten by deer for nettles of Nara Park (i.e., the likeli-
hood of deer browsing), x
1
is the arcsin-transformed
proportion of leaves eaten by deer for nettles of Takatori
Table 1 Analysis of covariance table for the data shown in Fig. 1, testing for the effect of nettle size (covariate) and nettle population
(fixed factor) on the number of Vanessa indica eggs laid on nettles (dependent variable)
Source df SS MS F
s
P
For homogeneity of regression slopes
Nettle population ·nettle size 1 0.907 0.907 1.034 0.315
Residual 43 37.741 0.878
For main effect of nettle population
Nettle size 1 16.489 16.489 18.773 <0.001
Nettle population 1 1.233 1.233 1.404 0.242
Residual 44 38.648 0.878
df degree of freedom, SS sum of squares, MS mean square
Table 2 Contingency tables for nettle leaves from Nara Park and Takatori Castle Site which were chosen by 1st or 2nd instar larvae of
Vanessa indica hatched from eggs sampled in Nara Park and Takatori Castle Site when two leaves were offered, followed by the results of
replicated goodness-of-fit tests consisting of a heterogeneity Gtest and a standard Gtest
Source of eggs Nettle leaves chosen Sum
Nara Park Takatori Castle Site
Instar I Nara Park 8 9 17
Takatori Castle Site 9 8 17
Sum 17 17 34
Gtest for homogeneity v2
s= 0.118, df =1,P= 0.732
Gtest for 1:1 ratio v2
s< 0.001, df =1,P= 0.999
Instar II Nara Park 6 11 17
Takatori Castle Site 10 8 18
Sum 16 19 35
Gtest for homogeneity v2
s= 1.458, df =1,P= 0.228
Gtest for 1:1 ratio v2
s= 0.129, df =1,P= 0.720
NP NPTC TC
80
60
40
20
0
Proportion
consumed by insects (%)
31 May 2011 28 June 2011
ts= 0.741
d.f. = 39
P= 0.463
ts= 0.339
d.f. = 39
P= 0.737
100
NP NPTC TC
80
60
40
20
0
t= 0.741 t= 0.339
.
P= 0.737
100
of leaves
Fig. 2 Leaves consumed by herbivorous insects (>20 % of area
removed) as a percentage of the total number of leaves for each of
two nettle populations. Bar + SD, NP Nara Park, TC Takatori
Castle Site
100
80
60
40
20
0
ts= 143.1
d.f. = 22
P < 0.001
eaten by sika deer (%)
NP TC
100
80
60
40
20
0
Proportion of leaves
Fig. 3 Leaves eaten by sika deer as a percentage of the total
number of leaves in each of two nettle populations. Bar + SD, NP
Nara Park, TC Takatori Castle Site
459
Castle Site (i.e., the potential intensity of deer browsing
at that place and time), and x
2
is the log-transformed
stinging-hair density for nettles of Nara Park. The
coefficient of multiple determination was 0.356 and the
F
s[2,20]
was 5.525 (P= 0.015). The standardized partial
regression coefficient of x
1
was significantly >0 (Fig. 4
a), whereas that of x
2
was not significantly different from
0 (Fig. 4b). These results indicated that the likelihood of
deer browsing on nettles of Nara Park varied with po-
tential browsing intensity and, noticeably, was inde-
pendent of stinging-hair density.
Discussion
Our previous studies supported the hypothesis that a
much higher density of stinging hairs on nettles of Nara
Park had evolved as a defense against sika deer (Kato
et al. 2007; Shikata et al. 2013). To confirm this
hypothesis, we needed to address at least two problems:
whether stinging hairs serve as defenses against insect
herbivores and the degree to which stinging hairs deter
sika deer. By comparing (hairy) nettles of Nara Park and
(almost hairless) nettles of Takatori Castle Site, we
found that (1) stinging hairs affected neither the ovipo-
sition preference nor larval feeding preference for nettles
of Indian red admiral, (2) stinging hairs did not reduce
leaf damage by insect herbivores, and (3) although the
likelihood of deer browsing was much lower in hairy
nettles than in hairless nettles, it was not correlated with
stinging-hair density in hairy nettles. These findings
suggest that stinging hairs scarcely serve as defenses
against insect herbivores and that higher stinging-hair
densities are not necessarily more effective in defending
the plants from deer browsing. The former corroborates
our hypothesis but the latter raises doubt about it.
No significant associations between insect damage
and stinging-hair density were reported in the European
nettle U. dioica by Pollard and Briggs (1984b)orTub-
erville et al. (1996). As those authors suggested, insect
herbivores are probably small enough to walk around,
climb over, or eat around stinging hairs. In fact, they
observed that larvae of the red admiral V. atalanta ate
leaf tissues without ingesting stinging hairs. We observed
this behavior in first and second instar larvae of the
Indian red admiral V. indica. Furthermore, females of V.
indica often attached eggs at the tips of stinging hairs as
well as on the leaf epidermis, implying that stinging hairs
did not influence egg laying. Thus, stinging hairs, even at
high densities, have little effect in reducing insect dam-
age.
One might assume that Indian red admirals at the
oviposition preference test site (the campus of Nara
Women’s University) were adapted to lay eggs on nettles
of Nara Park, because the two sites are only 500 m
apart. We cannot categorically deny the possibility, but
it is unlikely because nettles with almost hairless leaves
grow naturally on the campus (Shikata et al. 2013).
Thus, the butterflies may not be adapted to the hairy
nettles of Nara Park and should therefore prefer to
oviposit on the hairless nettles of Takatori Castle Site.
However, our results suggested that butterflies laid eggs
on both variants equally. Thus, Indian red admirals do
not appear to be adapted to either hairy or hairless
nettles for oviposition. This conclusion may apply to
other herbivorous insects as well.
Stinging hairs on leaves of U. thunbergiana effectively
deter deer browsing but not insect oviposition or her-
bivory. Dense stinging hairs were found along the stems
as well as on leaves in nettles of Nara Park (data not
shown), preventing the whole plant from being con-
sumed by the ungulate. These results confirm the
hypothesis that nettles with high stinging-hair densities
in Nara Park evolved under natural selection imposed
by intense deer browsing, as previous studies suggested
(Pollard and Briggs 1982,1984a,b, Pullin and Gilbert
1989, Tuberville et al. 1996, Kato et al. 2007, Shikata
et al. 2013). If the hypothesis is correct, we could predict
that nettles with higher stinging-hair densities would
suffer less damage by deer. However, the level of deer
browsing was not proportional to stinging-hair density
(Fig. 4b). This result is inconsistent with that of Pollard
80
60
40
20
0100080604020
b1’= 0.489
d.f . = 20
P = 0.015
2.001.60.80.4
100
1.2
Likelihood of deer browsing
nettles of Nara Park (%)
Potential intensity of deer browsing (%) Stinging-hair density in nettles
of Nara Park (cm–2
)
b2’= 0.244
d.f . = 20
P = 0.201
ab
80
60
40
20
0100080604020
1
2.001.60.80.4
100
1.2
-
–
2’
Fig. 4 Scatterplot of the
likelihood of deer browsing
nettles of Nara Park (i.e., the
proportion of leaves eaten by
deer) against the potential
intensity of deer browsing (i.e.,
the proportion of leaves eaten
by deer for nettles of Takatori
Castle Site) (a) and against the
stinging-hair density for nettles
of Nara Park (b). b
1
¢,b
2
¢
standardized partial regression
coefficients calculated using
transformed variables
460
and Briggs (1984b), who demonstrated that browsing by
rabbits or sheep was negatively correlated with stinging-
hair density in U. dioica.
The inconsistency between the two studies could be
explained not by the fact that different mammals were
involved but by the difference in stinging-hair density.
The range of hair densities in the present study
(0.29–1.98 cm
2)
was much narrower than that in Pol-
lard and Briggs (1984b) (0–101.2 cm
2
). Statistically, the
lack of correlation between stinging-hair density and
deer browsing level could arise stochastically from the
narrow range of stinging-hair densities.
From an evolutionary viewpoint, we can propose a
more likely explanation. Pollard and Briggs (1984b)
conducted their experiment with plants from several
populations to compare the levels of mammalian her-
bivory across a wide range of hair densities. Thus, they
examined the likelihood of mammalian browsing among
different populations, while we examined it within a
single (Nara Park) population, which has been exposed
to intense browsing pressure for 1200 years. Our results
showed that nettles of Nara Park suffered much less deer
browsing than those of Takatori Castle Site and that the
damage was independent of stinging-hair density within
the Nara Park population. These findings imply that the
resistance of U. thunbergiana to deer browsing would
increase with stinging-hair density until a plateau was
reached. In other words, different densities of stinging
hairs would have the same resistance beyond a threshold
value.
As a result of long-term strong directional selection
for hairy plants, the Nara Park population may not
only be at a plateau for resistance but also at a selection
limit, where the population fails to respond to selection
(Falconer and Mackey 1996). A cessation of response to
selection is not necessarily caused by loss of genetic
variance. Theoretical and empirical studies have indi-
cated that genetic variance can be maintained at a
selection limit (Mousseau and Roff 1987; Falconer and
Mackey 1996). Furthermore, even if genetic variance is
lost at a selection limit, phenotypic variance may be
found; often it increases (Falconer and Mackey 1996).
In fact, the range of stinging-hair densities in the Nara
Park population was wider than that in the Takatori
Castle Site (Fig. 4b). Whether or not the Nara Park
population is at a selection limit deserves to be exam-
ined.
Our finding that stinging hairs had little effect on
insect attack suggested that there was no indirect effect
of sika deer on herbivorous insects through an increase
in stinging-hair density. However, this conclusion does
not mean that there is no indirect relationship between
sika deer and herbivorous insects. Intense browsing by
sika deer could change not only stinging-hair density but
also chemical characteristics, such as water and nitrogen
contents (Shoonhoven et al. 2005; Barrett and Stiling
2007). If so, chemical alteration of leaves might affect the
growth performance of herbivorous insects. Our finding
disagrees with a result of Cano-Santana and Oyama
(1992). They found that Wigandia urens plants with
stinging hairs suffered more damage by herbivorous in-
sects than plants with no stinging hairs. They explained
this result with the fact that bristly plants had higher
contents of water, nitrogen, and phosphorus in their
leaves. In U. thunbergiana, nutritional traits may not
affect insect herbivory. To explore these possibilities, we
must study the nutritional traits of hairy and hairless
nettles and the responses of herbivorous insects to
nutritional variation.
In conclusion, we revealed that stinging hairs effec-
tively deter sika deer from browsing U. thunbergiana and
have little effect in reducing insect attack. These findings
support the hypothesis that nettles with high stinging-
hair density in Nara Park evolved through natural
selection under intense deer browsing. Contrary to our
prediction, there was no significant relationship between
intra-population variation in stinging-hair density and
the likelihood that deer browsed the plants, presumably
because the population was at a plateau for resistance
due to long-term directional selection for hairy plants.
Acknowledgments We dedicate this paper to the memory of
Emeritus Professor Ei’ichi Shibata, who always encouraged our
studies. This study was supported financially in part by a Grant-in-
Aid for Scientific Research from the Ministry of Education, Cul-
ture, Sports, Science and Technology of Japan ((C) No.22570019).
References
Barrett MA, Stiling P (2007) Relationships among Key deer, insect
herbivores, and plant quality. Ecol Res 22:268–273
Cano-Santana Z, Oyama K (1992) Variation in leaf trichomes and
nutrients of Wigandia urens (Hydrophyllaceae) and its impli-
cations for herbivory. Oecologia 92:405–409
Falconer DS, Mackay TFC (1996) An introduction to quantitative
genetics, 4th edn. Longman, Essex
Grubb PJ (1992) A positive distrust in simplicity-lessons from plant
defenses and from competition among plants and among ani-
mals. J Ecol 80:585–610
Hanley M, Lamont BB, Fairbanks MM, Rafferty CM (2007) Plant
structural traits and their role in anti-herbivore defence. Per-
spect Plant Ecol Evol Syst 8:157–178
Kato T (2001) Nettles in Nara Park. Nara Bot 23:11–14 (in Japa-
nese)
Kato T, Ishida K, Sato H (2007) The evolution of nettle resistance
to heavy deer browsing. Ecol Res 23:339–345
Kitamura S, Murata G (1961) Colored illustrations of herbaceous
plants of Japan, vol. II (Choripetalae) (in Japanese). Hoikusha,
Osaka
Krebs C (1998) Ecological methodology, 2nd edn. Harper-Collins,
New York
Milewski AV, Young TP, Madden D (1991) Thorns as induced
defenses: experimental evidence. Oecologia 86:70–75
Molisch H, Dobat K (1979) Botanische Versuche und Beobacht-
ungen mit einfachen Mitteln: ein Experimentierbuch fu
¨r Schu-
len und Hochschulen, 5th edn. Gustav Fischer, Stuttgart
Mousseau TA, Roff DA (1987) Natural selection and the herita-
bility of fitness components. Heredity 59:181–197
Ohgushi T, Schmitz O, Holt RD (eds) (2012) Trait-mediated
indirect interactions: ecological and evolutionary perspectives.
Cambridge University Press, Cambridge
Olsen C (1921) The ecology of Urtica dioica. J Ecol 9:1–8
Pollard AJ (1986) Variation in Cnidoscolus texanus in relation to
herbivory. Oecologia 70:411–413
461
Pollard AJ, Briggs D (1982) Genecological studies of Urtica dioica
L, I. The nature of intraspecific variation in U. dioica. New
Phytol 92:453–470
Pollard AJ, Briggs D (1984a) Genecological studies of Urtica dioica
L, II. Patterns of variation at Wicken Fen, Cambridgeshire,
England. New Phytol 96:483–499
Pollard AJ, Briggs D (1984b) Genecological studies of Urtica dioica
L, III. Stinging hairs and plant-herbivore interactions. New
Phytol 97:505–522
Pullin AS, Gilbert JE (1989) The stinging nettle, Urtica dioica, in-
creases trichome density after herbivore and mechanical dam-
age. Oikos 54:275–280
Shikata M, Kato T, Shibata E, Sato H (2013) Among-population
variation in resistance traits of a nettle and its relationship with
deer habitat use frequency. Ecol Res 28:207–216
Shoonhoven LM, van Loon JJA, Dicke M (2005) Insect-plant
biology, 2nd edn. Oxford University Press, Oxford
Sletvold N, Huttunen P, Handley R, Ka
¨rkka
¨inen K, A
˚gren J (2010)
Cost of trichome production and resistance to a specialist insect
herbivore in Arabidopsis lyrata. Evol Ecol 24:1307–1319
Sokal R, Rohlf JF (2012) Biometry, 4th edn. W H Freeman, New
York
SPSS Inc (2006) SPSS base 15.0 syntax reference guide. SPSS Inc,
Chicago
Teshirogi M (1990) An illustrated book of the Japanese Nymp-
halidae (in Japanese). Tokai University Press, Tokyo
Thurston EL (1974) Morphology, fine structure, and ontogeny of
the stinging emergence of Urtica dioica. Am J Bot 61:809–817
Tuberville TD, Dudley PG, Pollard AJ (1996) Responses of
invertebrate herbivores to stinging trichomes of Urtica dioica
and Laportea canadensis. Oikos 75:83–88
Walters DR (2011) Plant defense: warding off attack by pathogens,
herbivores, and parasitic plants. Wiley-Blackwell, Oxford
Yamamoto K (1997) LIA32 for Windows95 ver.0373. http://www.
agr.nagoya-u.ac.jp/shinkan/LIA32/index.html
462
