Download Endozoochory by beetles: a novel seed dispersal

Annals of Botany 107: 629 –637, 2011
doi:10.1093/aob/mcr013, available online at www.aob.oxfordjournals.org
Endozoochory by beetles: a novel seed dispersal mechanism
Clara de Vega1,*, Montserrat Arista2, Pedro L. Ortiz2, Carlos M. Herrera1 and Salvador Talavera2
1
Estación Biológica de Doñana, Consejo Superior de Investigaciones Cientı́ficas (CSIC), Avenida de Américo Vespucio s/n,
41092 Sevilla, Spain and 2Departamento de Biologı́a Vegetal y Ecologı́a, Facultad de Biologı́a, Universidad de Sevilla,
Apdo-1095, 41080 Sevilla, Spain
* For correspondence. E-mail [email protected]
Received: 17 October 2010 Returned for revision: 30 November 2010 Accepted: 10 December 2010 Published electronically: 7 February 2011
† Background and Aims Due in part to biophysical sized-related constraints, insects unlike vertebrates are seldom
expected to act as primary seed dispersers via ingestion of fruits and seeds (endozoochory). The Mediterranean
parasitic plant Cytinus hypocistis, however, possesses some characteristics that may facilitate endozoochory by
beetles. By combining a long-term field study with experimental manipulation, we tested whether
C. hypocistis seeds are endozoochorously dispersed by beetles.
† Methods Field studies were carried out over 4 years on six populations in southern Spain. We recorded the rate
of natural fruit consumption by beetles, the extent of beetle movement, beetle behaviour and the relative importance of C. hypocistis fruits in beetle diet.
† Key Results The tenebrionid beetle Pimelia costata was an important disperser of C. hypocistis seeds, consuming up to 17.5 % of fruits per population. Forty-six per cent of beetles captured in the field consumed
C. hypocistis fruits, with up to 31 seeds found in individual beetle frass. An assessment of seeds following
passage through the gut of beetles indicated that seeds remained intact and viable and that the proportion of
viable seeds from beetle frass was not significantly different from that of seeds collected directly from fruits.
† Conclusions A novel plant –animal interaction is revealed; endozoochory by beetles may facilitate the dispersal
of viable seeds after passage through the gut away from the parent plant to potentially favourable underground
sites offering a high probability of germination and establishment success. Such an ecological role has until now
been attributed only to vertebrates. Future studies should consider more widely the putative role of fruit and seed
ingestion by invertebrates as a dispersal mechanism, particularly for those plant species that possess small seeds.
Key words: Beetle, Cytinus hypocistis, Cytinaceae, endozoochory, mutualism, parasitic plant, Pimelia costata,
plant–animal interaction, seed dispersal, seed viability, Tenebrionidae.
IN T RO DU C T IO N
Seed dispersal by animals is a complex mutualistic interaction
involving a great diversity of plant and animal species with
significant ecological and evolutionary consequences for
plant community structure and function (Howe and
Smallwood, 1982; Schupp, 1993; Herrera, 2002).
Vertebrates, mainly birds and mammals, are the main seed dispersers in most plant communities, carrying seeds either
internally (endozoochory, seeds consumed and passed
through the gut) or externally (epizoochory, barbed or sticky
seeds attached to their fur/feathers) (van der Pijl, 1982;
Herrera, 1995; Corlett, 2002). With the exception of ants,
which are recognized as important seed dispersers in a wide
variety of habitats (e.g. Levey and Byrne, 1993; Wolff and
Debussche, 1999; Rico-Gray and Oliveira, 2007), seed dispersal by invertebrates has received comparatively little attention.
However, a number of studies have demonstrated that in
addition to ants, other insects (including beetles) play an
important role in the dispersal of seeds. Of these, dung
beetles (Scarabaeidae: Coleoptera) act as secondary seed dispersers by moving and burying seeds embedded in vertebrate
faeces, reducing the probability of rodent seed predation and
simultaneously increasing seedling establishment (Estrada
and Coates-Estrada, 1991; Andresen and Levey, 2004).
Seed size is thought to limit the interaction between plant and
dispersal agent (Hughes et al., 1994; Westoby et al., 1996), such
that insects are not expected to play the role of primary seed
dispersers via fruit ingestion due to purely biophysical
sized-related constraints. Two apparent exceptions to this
general pattern are the New Zealand large endemic flightless
grasshoppers Hemidenina crassidens (Anostostomatidae) and
Zealandosandrus maculifrons (Stenopelmatidae), commonly
known as Weta, that consume the fleshy fruits of several plant
families and disperse seeds by endozoochory (Burns, 2006;
Duthie et al., 2006; King et al., 2010; but see Wyman et al.,
2010). It seems unlikely, however, that this is the only instance
of endozoochory involving invertebrates. Endozoochorous
seed dispersal by ‘medium-sized’ insects could be feasible whenever fruits and seeds are small enough to attract invertebrates and
be ingested and later defecated intact after passage through
the gut.
Cytinus hypocistis (Cytinaceae) is a Mediterranean perennial parasitic plant, inflorescences of which burst through the
host tissues at ground level (de Vega et al., 2007, 2009).
The berry-like fruits ripen from May to July and contain thousands (approx. 25000) of minute seeds (approx. 200 mm in
length) embedded in juicy pulp (de Vega et al., 2009). In
recent years, a comprehensive study has been carried out on
the ecology, reproductive biology, anatomy and population
# The Author 2011. Published by Oxford University Press on behalf of the Annals of Botany Company. All rights reserved.
For Permissions, please email: [email protected]
630
de Vega et al. — Endozoochory by beetles
genetics of this parasite (de Vega et al., 2007, 2008, 2009) and
during the course of observations conducted around Doñana
National Park (south-west Spain), some species of beetles
were observed eating its ripe fruits (de Vega, 2007). Beetles
are important components of the ground insect fauna of such
areas, and include herbivorous and detritivorous species (de
los Santos et al., 1988; Lobo et al., 1998; Cárdenas and
Hidalgo, 2006). In other parts of Spain, beetles have been
observed consuming the pulp of fruits of several species
(e.g. Traveset and Riera, 2005; de la Bandera and Traveset,
2006), although the size of the seeds involved was too large
for beetles to ingest and disperse them endozoochorously. In
contrast, Cytinus possesses some characteristics that may
facilitate the existence of primary seed dispersal by beetles.
First, the seeds have a rigid, thick coat (Guzowska, 1964;
Bouman and Meijer, 1994) and are small enough to be
ingested and defecated intact; second, fruits are borne at
ground level; and third, fruits ripen in late spring and
summer when many species of beetle are most abundant (de
los Santos et al., 1982, 1988; de Vega, 2007). The aim of
this study was to test whether Cytinus seeds are endozoochorously dispersed by beetles, a phenomenon which would represent a novel seed dispersal system in angiosperms.
Specifically we addresses three main questions. (1) Does
beetle behaviour favour efficient seed dispersal with regard
to the quantity factor (number of seeds dispersed) and
quality factor (viability of seeds after passage through the
gut and movements away from the parent plant to suitable
sites)? (2) Are there spatial and temporal variations in the
importance of beetles for seed dispersal? (3) What is the ecological relevance of this plant – animal mutualism? We
combine a long-term field observation study conducted over
4 years with experimental treatments at six study sites to
examine this hitherto unreported plant – beetle interaction.
M AT E R IA L S A ND M E T HO DS
Study species and study area
Cytinus hypocistis (L.) L. is a root holoparasite with a vegetative body reduced to an endophytic system contained entirely
within the root of their Cistaceae host plants (de Vega et al.,
2007, 2008). In spring, the inflorescences burst through the
host tissues and present around six male flowers at basal positions and a similar number of female flowers distally (de
Vega, 2007). Ants are the main pollinators of C. hypocistis,
and plants exhibit a high fruit-set and seed production under
natural conditions (de Vega et al., 2009). The fruits that
ripen from May to July contain thousands of minute seeds
embedded in pulp (de Vega et al., 2009). Two stages in the
ripe fruits were recognized: ripe juicy fruits (yellow) and
ripe dry fruits (brown). In the study region, the yellow,
fleshy fruits are consumed by wood mice (Apodemus sylvaticus) and European rabbits (Oryctogalus cuniculus), both of
which defecate the seeds with no external sign of damage
(de Vega, 2007). When fruits are brown and dried, the dry
pulp seems to play the role of an elaiosome and seeds are
removed by ants (mainly Aphaenogaster senilis,
Crematogaster auberti and Pheidole pallidula; de Vega,
2007).
The study was conducted over 2002– 2005 in the surroundings of the Doñana National Park (Huelva province, southwest Spain, 37818′ N, 6825′ W). The climate is typically
Mediterranean, with mean annual temperatures ranging from
16.6 to 17.68C, and 692 mm of mean annual precipitation in
the study period. We selected six populations of Cytinus parasitizing three host species of the Cistaceae family: two populations parasitizing Cistus ladanifer L. (hereafter referred to
as Cl1 and Cl2), two populations on Cistus salviifolius L.
(Cs1 and Cs2) and two populations on Halimium halimifolium
(L.) Willk. (Hh1 and Hh2). We selected three host species,
because each of them is parasitized by a different genetic
race of C. hypocistis (de Vega et al., 2008). The studied populations were separated by 0.5 –2.5 km and were located at 80–
90 m a.s.l. Cytinus hypocistis on the host C. ladanifer occurred
on clay soils, and vegetation consisted on mixed woodland of
Pinus pinea L. and Quercus suber L. Populations
of C. salviifolius were on sandy soil and vegetation consisted
of a dense forest of Q. suber with scattered individuals of
P. pinea. Populations of C. hypocistis on H. halimifolium
were on sandy soils and the vegetation consisted of woodland
of P. pinea.
Field experiments
Field observations revealed that two species of beetles fed on
fruits of C. hypocistis, Pimelia costata Waltl. (Tenebrionidae),
an Ibero-Balearic endemic (Bruvo-Madaric et al., 2007), and
Calathus granatensis Vuillefroy (Carabidae), an Iberian
endemic (Serrano et al., 2003). Pimelia costata was observed
during the whole study period, while C. granatensis was much
less abundant, being observed only in two populations (Cs1
and Hh1) and only in 2002. The importance of C. granatensis
in the dispersal of C. hypocistis fruits seems to be quite low
(only 0.06 % of fruits were dispersed by this species) and
hereafter all observations and experiments will refer to
P. costata.
Observations were conducted 3– 4 times per week in infructescences freely exposed to dispersal agents in the populations
Cl1, Cl2, Cs1, Hh1 and Hh2 in 2002 – 2005, and in Cs2 in
2002 – 2004. In each study year, the interaction between
C. hypocistis and P. costata was assessed by recording the
rate of natural fruit consumption, the time spent in completely
eating a fruit, number of individuals feeding on the fruits, their
period of activity and their behaviour. The total number of
fruits recorded was 5357 (Table 1).
To assess the relative importance of Cytinus fruits in the diet
of P. costata, 82 individual beetles were captured in 2003 and
2005 in the populations Cs1, Cs2, Hh1 and Hh2 (no beetles
were observed in the populations Cl1 and Cl2). Beetles were
located by walking through the populations and were captured
in situ by hand. Each beetle was immediately placed in a small
plastic container (8.5 × 5.5 × 4.5 cm) kept in shade and
semi-buried for 6 – 8 h, during which time the beetles defecated. Frass from each box was separated in labelled vials
and inspected under a binocular microscope in the laboratory
to check for the presence/absence of Cytinus seeds. In total,
717 frass samples were examined and, in those containing
seeds (n ¼ 189; 26.4 %) the total number of intact seeds was
counted.
de Vega et al. — Endozoochory by beetles
631
TA B L E 1. Percentage of Cytinus hypocistis fruits that were consumed by Pimelia costata in each population and year, and mean
number of seeds per frass
Percentage of fruits consumed (n)*
Population
Cl1
Cl2
Cs1
Cs2
Hh1
Hh2
2002
2003
2004
2005
0.0 (301)
0.0 (178)
1.6 (128)
13.8 (261)
7.1 (392)
0.0 (378)
0.0 (511)
0.0 (207)
0.8 (246)
7.8 (142)
17.5 (468)
9.1 (560)
0.0 (174)
0.0 (117)
4.2 (189)
8.3 (132)
7.7 (91)
9.6 (135)
0.0 (139)
0.0 (48)
0.0 (210)
–
10.6 (94)
9.0 (256)
Number of seeds/frass (n)†
–
–
3.9 + 0.3
10.5 + 1.8
5.8 + 1.4
10.1 + 0.7
(41)
(16)
(13)
(119)
A total of 5357 fruits were monitored at all populations combined.
* n, number of monitored fruits.
†
Values are mean + s.e.; n, number of frass analysed.
Additionally, the extent of beetle movements was investigated
by capture – mark – recapture in 2002 – 2003. In populations Cs1,
Cs2 and Hh2, 72 individual beetles (in total, all populations
combined) were marked with enamel paint, which was persistent
and non-toxic and efficient in marking other darkling beetles
(e.g. Fallaci et al., 1997). We painted each beetle in the field,
with a unique combination of spots on the elytra, and immediately released the marked beetles in the place where they were
initially captured. The presence of marked beetles was recorded
by 15 walking surveys in each study population from May to
July. A minimum of five different transects were undertaken in
each population per survey, and each transect was 100 m long
and any marked beetle within 2 m on either side of the transect
was captured in order to record their code before immediate
release. The same beetle was only registered once per survey.
Each population was sampled at intervals of 5 d. Beetle
surveys were made in the morning (0700 – 1100 h), at the peak
time of beetle activity.
Feeding experiments in the laboratory
Fifteen P. costata were collected in the field and used in
feeding experiments in which beetles were retained in glass containers containing moist filter paper (20 mm diameter). Two
Cytinus fruits were offered to each beetle, and the containers
were inspected every 2 d for frass. Frass samples were collected
with tweezers and stored in Petri dishes. Eight to 12 frass from
each individual were inspected for seeds under a microscope
and the total number of seeds was counted. Feeding experiments
were conducted in 2005 from June to July.
Germination experiments
Ripe Cytinus fruits were collected in populations Cl1, Cs1
and Hh1 in 2002 – 2005. A minimum of 50 fruits were collected per population and year; fruits from the same population
and year were mixed and stored in paper bags from collection
until further processing. One hundred fruits of each of the
Cistaceae host species C. ladanifer, C. salviifolius and
H. halimifolium were collected in the same populations and
stored in paper bags at room temperature.
In an attempt to promote germination, 18 different seed germination treatments were applied. We investigated the germinability of undispersed seeds of Cytinus hypocistis collected
from fruits, and that of seeds dispersed by beetles, and each
experiment included four replicates. For a detailed description
of experimental designs, see Supplementary Data (available
online). After the treatments, seeds were set to germinate in
a germination chamber (Ibercex Mod.F-1, ASL, SA, Madrid,
Spain) over 7 months.
Additionally, germination assays were carried out in the
field using a modification of the design of Rasmussen and
Whigham (1993). Cytinus seeds were placed into packets constructed from 40 × 60 mm rectangles of stainless steal mesh
that were folded once and clipped into 2 × 2 × 36 mm
plastic slide mounts. The mesh pore size of 50 mm retained
the seeds, but allowed them to experience the physicochemical
conditions of the substrate and contact with soil microorganisms. Approximately 500 Cytinus seeds collected from fruits
or 100 seeds from beetle frass were placed in each packet.
Seed packets were buried in the soil and placed in contact
with Cistaceae host plant roots. A set of three packets of
seeds collected from fruits from the three different
C. hypocistis races, and one packet with seeds from frass,
were placed in contact with each host plant root. Each set of
seed packets was buried under eight plants of each host
species, distributed in three populations (Cl1, Cs1 and Hh1).
In total, we used 72 packets with seeds collected from fruits
and 24 packets with seeds collected from beetle frass. A preliminary harvest of two sets of packets was taken from each
host species after 1 month. After collection, the packets were
gently washed with water and kept moist in sheets of wet
filter paper for 1 d before examination. The slide mount was
opened and the presence or absence of germination was
recorded under a dissecting microscope. As no seed germination was observed, the following day those two packets were
buried in the field, in the same places where they were collected. Half of the packets were retrieved 3 months later,
observed as explained above, and buried in soil after 2 d. All
packets were collected 6 months later. All seeds from frass
used in this experiment were obtained from beetles fed exclusively with C. hypocistis fruits in the laboratory.
Seed damage
Seed mechanical damage due to the mastication or chewing
process is a common phenomenon for many plant species
(Bertness et al., 1987; Overdorff and Strait, 1998; Furedi
632
de Vega et al. — Endozoochory by beetles
and McGraw, 2004; Varela and Bucher, 2006). The presence
of mechanical damage in C. hypocistis seeds was assessed
by visual inspection of seeds from all frass collected in the
field and in the laboratory (n ¼ 847 frass). Frass samples
were placed on Petri dishes and carefully observed under a
stereo microscope to identify broken seeds.
Additionally, we used a combination of techniques
described by de Vega and de Oliveira (2007) that allow us
to observe if seeds have normal embryos (i.e. not deformed)
after consumption by animals. Briefly, seeds were successively
softened with Franklin’s fluid [Franklin (1945; modified by
Kraus and Arduin, 1997), a mixture (1 : 1) of 30 % hydrogen
peroxide and glacial acetic acid] and Jeffrey’s fluid
[Johansen (1940), composed of a mixture (1 : 1) of nitric
acid and 10 % chromic acid]. These fluids caused sufficient
seed coat rupture to allow Herr’s clearing fluid [Herr (1971),
lactic acid (85 %), chloral hydrate, phenol, clove oil and
xylene (2 : 2 : 2 : 2 : 1, by weight)] to clear the seed tissues,
permitting effective observation of the embryo.
Seed viability
Following Copeland and MacDonald (2001), viability is
‘the degree to which a seed is alive, metabolically active,
and possesses enzymes capable of catalyzing metabolic reactions needed for germination and seedling growth’. We used
two different methods to test seed viability in C. hypocistis
seeds. First, a 2,3,5-triphenyl tetrazolium chloride (TTC) staining procedure was used following a modified version of the
biochemical viability test of López Granados and Garcı́a
Torres (1999). The TTC test distinguishes between viable
and non-viable seeds on the basis of their respiration in the
hydrated state (Cottrell, 1947). TTC is a white salt that turns
from colourless to the pink/red formazan in living cells in
the presence of cellular respiration (Copeland and
McDonald, 2001), and, accordingly, a red coloration of a
seed indicates viability, as dead and necrotic seeds remain
uncoloured.
Seeds were sterilized for 30 min in 10 % (v/v) commercial
bleach before being washed three times in distilled water and
then immersed in a 1 % (w/v) solution of TTC (Sigma) in
pH 7 water adjusted with 1 M NaOH. Vials with the seeds
were wrapped in aluminium foil and incubated for 15 d at
30 8C. Finally, to bleach the seeds, they were soaked in
50 % (v/v) commercial bleach for 10 min, which allowed the
internal content to be observed. Seeds completely stained red
or dark pink were scored as viable, and those that remain
unstained as non-viable. In total, 828 seeds from frass and
324 seeds from fruits were observed under the microscope.
The other method involved the use of fluorescein diacetate
(FDA) staining by a modification of the procedure of
Pritchard (1985). FDA enters the living cells where it is hydrolysed by esterase enzymes to give fluorescein, and allows a
rapid determination of seed viability (Rasmussen, 1995).
Seeds were surface sterilized for 5 min in 10 % (v/v) commercial bleach, rinsed three times in distilled water and incubated
in distilled water for 24 h at room temperature. Following incubation, seeds were sandwiched between two microscope slides
and rotated in opposite directions to break and remove the seed
coat, and then mixed 1 : 1 (v/v) with FDA 0.25 % (w/v) in pure
acetone. Seeds were viewed under UV-fluorescence
microscopy (Nikon eclipse 80i), and only embryos showing
a uniformly brilliant yellowish green fluorescence were considered to be viable. In total, 280 seeds from frass and 187
seeds from fruits were subjected to the FDA test and observed
under a microscope.
For both procedures, seed viability of ingested seeds was
compared with those collected directly from fruits from the
same population and year, and with autoclaved seeds that
were treated as control.
Statistical analysis
Differences between populations in the frequency of beetle
frass containing seeds were analysed using log-linear analysis
of frequency tables, and the (log-transformed) number of
seeds per frass was compared among populations by means of
a one-way ANOVA. The frequency of beetles that consumed
C. hypocistis fruits was compared among populations using loglinear analysis of frequency tables, and the (log-transformed)
number of seeds per frass obtained in the laboratory and in the
field was compared by means of one-way ANOVA. Viability
differences between seeds from fruits and frass were compared
using chi-square tests for the TTC and FDA procedures. No statistical test was used for germination experiments as all attempts
to germinate C. hypocistis seeds were unsuccessful. All statistical analyses were performed in Statistica 6.0, version 6 (http://
www.statsoft.com). Throughout the text, results are presented
as mean + s.e.
R E S U LT S
Spatio-temporal variations in beetle distribution and
fruit consumption
Pimelia costata was observed feeding on C. hypocistis fruits in
four of the six study populations (Cs1, Cs2, Hh1 and Hh2)
during the whole study period, but never in populations Cl1
and Cl2.
The frequency of fruit consumption by P. costata differed
between seasons and sites (Table 1). The percentage of fruits
consumed by P. costata (when present) ranged from 0.81 %
in population Cs1 in 2003 to 17.52 % in population Hh1 in
the same year (Table 1). We found significant differences
among populations in the frequency of fruits consumed by
beetles when data for each site were combined across all
study years (x2 3 ¼ 2653.6, P , 0.0001), with populations
Hh1 and Cs1 having the highest and lowest frequency of
fruit consumption by beetles, respectively. Within populations,
we found temporal variations in the frequency of fruits
consumed each year (P , 0.0001 in all cases; Table 1).
Beetle foraging activity
Pimelia costata were diurnal with most fruit consumption
occurring in the morning (0700 – 1100 h) and late evening
(1900 – 2100 h). During the middle of the day, when the temperatures at ground level can reach more than 50 8C, activity
drastically decreased and beetles disappeared by burying
themselves in the sand.
de Vega et al. — Endozoochory by beetles
633
F I G . 1. Details of consumers, fruits and seeds of Cytinus hypocistis. (A) Pimelia costata eating C. hypocistis fruits. (B) C. hypocistis fruit with inset at higher
magnification of the area marked in white showing seeds. (C) P. costata frass containing intact seeds. (D) Intact seed isolated from P. costata frass.
(E) C. hypocistis seed softened and cleared allowing visualization of undifferentiated embryo. Abbreviations: end, endosperm; em, embryo; n, nucellus; sc,
seed coat. Scale bars: (A) ¼ 1 cm; (B) ¼ 0.5 cm; (C) ¼ 1 mm; (D, E) ¼ 0.1 mm.
In contrast to mice and rabbits, which selectively consume
yellow-fleshy fruits with a sticky pulp (de Vega, 2007),
P. costata were observed consuming ripe brown fruits with
dry pulp in which thousands of seeds were embedded
(Fig. 1A, B). Beetles were also observed feeding on whole
ripe fruits, and on fruits partially consumed by wood mice,
and also co-occurred with ants on the infructescence. Beetles
fed on the C. hypocistis fruits or infructescences both alone
or in groups of up to five individuals, and often returned
repeatedly to the same fruit for several days until the dry
pulp with embedded seeds was emptied. Based on observations of marked beetles, P. costata individuals spent a
mean time of 9.5 + 2.9 d (n ¼ 15; range 1 – 29 d) to completely consume a fruit in the field.
Increasing temperatures during the middle of the day forced
Pimelia to leave fruits after pulp ingestion seeking thermal
shelters. Beetles were observed waking away from the
feeding site (sometimes up to 50 m) until they buried themselves in the sand. This also resulted in movement of the
seeds from the immediate vicinity of a parent plant. Despite
the observed movement of beetles, the foraging activity of
P. costata in our study system was spatially restricted to
small areas. Eighty-four per cent of marked individuals were
repeatedly observed foraging in the populations where they
were captured and marked, and we did not record any movement of individuals among populations.
Importance of C. hypocistis to beetles
Cytinus fruits were an important part of beetle diet in late
spring and summer. In the field, 46.3 % of live-trapped
individuals (n ¼ 82) had consumed C. hypocistis fruits
before capture, as revealed by the presence of seeds in their
faeces while still in the traps (Fig. 1C). Each individual we collected in the field defecated from one to 39 frass in the plastic
containers, and C. hypocistis seeds occurred in 26.4 % of frass
samples analysed (n ¼ 717), with the number of seeds per
frass ranging from one to 31 (mean ¼ 8.5 + 0.5, n ¼ 189;
Table 1). We found no significant differences among populations in the frequency of beetles with frass containing
C. hypocistis seeds (x2 4 ¼ 4.81, P ¼ 0.307), but there were
significant differences among populations in the number of
seeds per frass (F ¼ 5.48, d.f. ¼ 3, P ¼ 0.001; Table 1).
When beetles were fed exclusively with fruits of
C. hypocistis in the laboratory, they defecated a higher
number of seeds per frass, ranging from 45 to 294 (mean ¼
110.43 + 4.73, n ¼ 130). As expected, there were significant
differences in the number of seeds in the frass collected in
the field and in the laboratory (F ¼ 953.06, d.f. ¼ 1, P ,
0.0001).
Seed damage
Observation under a binocular microscope revealed that all
seeds recovered from frass, both from the field and from the
laboratory, remained intact (Fig. 1D). Broken seeds were
never detected in frass (n ¼ 847). Microscopic examinations
using the technique described by de Vega and Oliveira
(2007) revealed that defecated seeds had normal embryos
(i.e. not deformed). All seeds examined presented a small
embryo composed of about ten cells surrounded by an endosperm composed of larger cells (Fig. 1E), similar to those of
634
de Vega et al. — Endozoochory by beetles
non-ingested ripe fruits (see Guzowska, 1964; de Vega and de
Oliveira, 2007) (Fig. 1E).
Seed germination
All attempts to germinate ingested and non-ingested
C. hypocistis seeds were unsuccessful, both in the laboratory
and in the field. We did not observe any seed with a protruding
radicle. After 7 months of laboratory experiments, the majority
of the seeds showed signs of fungal contamination. In the same
way, most seeds that were buried in the soil turned dark green
or black and showed fungal contamination, including abundant
greenish cottony mycelium.
Seed viability
The two methods for assessing viability indicated that seeds
recovered after ingestion by beetles remained viable (Fig. 2A,
B). Following staining with TTC, viable embryos of
C. hypocistis appeared dark pink or red (Fig. 2A), whereas
seeds with white or yellowish contents were considered to be
non-viable. Autoclaved control seeds did not respond to TTC
and remained unstained. Samples arising from unconsumed
fruits had a similar proportion of viable seeds as those from
beetle frass (8.03 and 11.84 %, respectively, x21 ¼ 3.52, P ¼
0.061, n ¼ 1142) based on the TTC test.
FDA-stained embryos that fluoresced bright yellowish green
were considered viable, whereas no staining was observed in
non-viable seeds, as was the case with autoclaved control
seeds (Fig. 2B). The proportions of viable seeds from fruits
and frass were 32.1 and 30 %, respectively, and the difference
was not statistically significant (x21 ¼ 0. 23, P ¼ 0.632, n ¼
467). Despite the two chemical staining procedures showing
different levels of seed viability, of more importance is that
in both procedures the percentages of seed viability from
fruits and frass were statistically not significant.
DISCUSSION
This study reveals a new seed dispersal mode involving a parasitic plant, which produces fruits with thousands of minute
seeds, and a beetle species that consumes their fruits and
defecates intact and viable seeds. Our data suggest that this
interaction could be regarded as a mutualism, in that both partners benefit from the association; P. costata obtains energy and
nutrients from the pulp, and C. hypocistis achieves seed movement away from the parent plant.
Comparisons of different populations over several years
revealed, however, that the importance of this interaction
was not constant, with striking spatial and temporal variations
in beetle distribution and fruit consumption. Numerous ecological factors, such as vegetation type, shrub cover, temperature, or soil type, may determine local abundance of beetles
(Burel, 1989; Stapp, 1997; de los Santos et al., 2002).
Pimelia costata shows a clear preference for sandy substrates
(Cárdenas and Hidalgo, 2006), and in our study soil type
seems to be an important factor determining beetle presence.
Plants in four populations in which beetles were common
grow on sandy soils, and P. costata responded to near-lethal
ambient temperatures in these populations (up to 50 8C) by
digging and burying themselves in the sand. However, plants
in two other populations occurred on clays, and the severely
compacted soils in summer months may constrain beetle occupation of such populations due to their inability to dig into the
soil for thermal shelter.
Spatio-temporal variations in P. costata population size have
not been estimated in this study, as our focus here was not on
the population dynamics of the beetle, but rather on the overall
importance of this mutualism for both partners. However, it
seems reasonable to suggest that observed variations in the
proportion of fruits consumed were related in part to interannual population size variation of P. costata, a trend that
occurs in other Pimelia species (Fallaci et al., 1997). It must
be kept on mind, however, that variation in fruit consumption
by beetles also depends on the fluctuation of abundance of
other consumers of C. hypocistis fruit. For example, the
absence of fruit consumption by beetles at one site (Hh2 in
2002) was due to the unusual abundance of wood mice (de
Vega, 2007), which nearly consumed all fruits before they
were available for beetles.
Pimelia costata showed a pattern of diurnal activity similar
to other Pimelia species in the Mediterranean Basin (Fallaci
et al., 1997), avoiding the middle of the day when temperatures are highest. The foraging activity of P. costata seems
F I G . 2. Seeds and isolated embryos of Cytinus hypocistis. (A) Seeds stained with triphenyl tetrazolium chloride. Arrowheads indicate viable seeds stained dark
pink. (B) Isolated embryos of C. hypocistis following exposure to fluorescein diacetate, with an unviable unstained embryo on the left and a stained embryo
fluorescing bright green on the right. Scale bars: (A) ¼ 200 mm, (B) ¼ 100 mm.
de Vega et al. — Endozoochory by beetles
to be spatially limited to small areas, as revealed by our
walking surveys and the absence of recorded movement of
individuals among populations, probably due in part to the
fact that Pimelia species are flightless (Moya et al., 2006).
This small-scale spatial foraging behaviour seems to be
similar to that shown in studies of other Pimelia species, in
which individuals marked in a particular zone were recaptured
in the same zone, and only occasionally in other areas (Fallaci
et al., 1997). Although the number of recaptured beetles was
relatively small in this study, these observations could
provide useful information on the limited mobility of
P. costata and the extent to which C. hypocistis seeds could
be dispersed by this insect. Moreover, small-scale dispersal
seems to be in accordance with small population sizes of
C. hypocistis. Although host plants form very large populations with thousands of individuals, C. hypocistis populations
are reduced, on average less than 100 m in diameter, and with
a small number of plants, regularly fewer than 100 individuals
(de Vega, 2007; de Vega et al., 2008).
Our microscopic observations indicate that seeds were defecated intact, as beetle frass never contained fragments of
broken seeds, and embryos showed no sign of mechanical
damage. Germination is the most commonly used method for
determining seed viability, but for many plant species this is
not an easy task. Specifically, for most non-agronomically
important root holoparasitic plants, for example those belonging
to the families Apodanthaceae, Hydnoraceae, Cytinaceae,
Mitrastemonaceae or Rafflesiaceae, information on germination
requirements is absent or inconclusive (Kuijt, 1969;
Garcı́a-Franco and Rico-Gray, 1997; Hidayati et al., 2000; but
see Bolin et al., 2009). Despite our efforts both in the field and
in the laboratory, we were unable to determine the exact germination requirements for C. hypocistis seeds. However, as an alternative to germination, there are several tests for determining seed
viability (Copeland and MacDonald, 2001). The viability tests
used in this study, TTC and FDA procedures, have been widely
used as accurate methods of determining seed viability of
many autotrophic angiosperms (e.g. Pritchard, 1985; Bell
et al., 1995; Copeland and McDonald, 2001; McDonald and
Kwong, 2005) and parasitic plants (Aigbokhan et al., 1998;
Daws et al., 2008; Thorogood et al., 2009). In this study, both
tests indicated that C. hypocistis seeds remained alive and
viable after passage through the beetle gut, with samples
arising from beetle frass having a similar proportion of viable
seeds as those from unconsumed fruits. This demonstrates that
beetles can act as effective seed dispersers, and may play an
important role in the reproductive cycle of the plant.
Efficient seed dispersal for a parasitic plant, however,
involves not only the transport of seeds but also their placement near a suitable host, as most parasitic plants have
minute seeds with very little reserves that require chemical
signals from hosts to initiate seed germination, and successful
attachment and development (Stewart and Press, 1990; Logan
and Stewart, 1991; Yoder, 1999; Bouwmeester et al., 2003; but
see Meulebrouck et al., 2008; Mescher et al., 2009). The seed
dispersal assemblage of C. hypocistis comprises a very heterogeneous array of animals, differing widely in body size, ecological habitats and mobility, such as rodents, lagomorphs,
differently sized ants and beetles (de Vega, 2007). Few endozoochorous angiosperm species rely on as broad a taxonomic
635
diversity of endozoochorous dispersers, and this variety probably led to a heterogeneous seed shadow.
Despite the fact that the number of fruits consumed by
beetles is less than that for other putative dispersers, their
importance should not be underestimated. Rodents and lagomorphs could appear to be better dispersers than beetles, as
they consume a greater percentage of fruits, and potentially
move seeds longer distances. However, they leave dung at
ground level and not near host plant roots as beetles do.
Moreover, mice and rabbits frequently consume immature
fruits, in contrast to beetles which always consume mature
dried fruits. In this study we observed that beetles made
several visits to each fruit over several days before its pulp
was depleted, and with those constant visits they could act
as significant vectors for seed dispersal by spreading their
frass widely. Moreover, Pimelia costata actively dug into
sand to bury themselves during the hottest part of the day,
increasing the probability of placing seeds near the roots of
the host plant. Consequently, P. costata can be considered to
be an efficient seed disperser not only because it defecates
intact, viable seeds, but also because seeds are carried to
favourable germination microsites. This placement of seeds
close to the host root systems could also facilitate direct
fungus– or mycorrhizal – seed contact, which benefits seed
germination for some parasites (Watanabe, 1942; Ecroyd,
1996; Li and Guan, 2007, 2008), potentially including
C. hypocistis. We have recently reported the existence of a tripartite association in natural conditions among Cytinus, its
host plant and mycorrhizal fungi at an intimate anatomical
level (de Vega et al., 2010). However, it remains unclear
whether a mycorrhizal association is necessary at the germination or establishment stage.
The fact that we found a novel form of endozoochory (by
beetles) suggests that the phenomenon is not confined solely
to vertebrate frugivores and is much more widespread than
the limited examples previously assumed for invertebrates
(Burns, 2006; Duthie et al., 2006; King et al., 2010). Seed dispersal by beetles to below-ground microsites in proximity to
host plant roots may also be vital in ensuring successful establishment of Cytinus seedlings in Mediterranean ecosystems.
Consequently, future studies should consider how fruit and
seed ingestion by invertebrates acts as a dispersal mechanism,
particularly for those plants that possess relatively small seeds,
in order to assess the global significance of this plant – animal
interaction.
S U P P L E M E N TA RY D ATA
Supplementary data are available online at www.aob.oxford
journals.org and give full details of the design of the germination experiments.
ACK NOW LED GE MENTS
We thank R. Vega-Durán, F. de Vega and E. Durán for help with
fieldwork, and Dr F. Sánchez Piñero and Dr V. Ortuño Hernández
for beetle identifications. Special thanks to Dr R. G. Albaladejo,
Dr M. Hanley and three anonymous reviewers for helpful comments on the manuscript. This work was supported by a PhD
grant from the Spanish Ministerio de Educación y Ciencia
636
de Vega et al. — Endozoochory by beetles
(AP2002-0939 to C.dV.) and by funds from the Ministerio de
Educación y Ciencia and Fondo Europeo de Desarrollo
Regional (REN2002-04354-C02-02, CGL 2005-01951 to
M.A.), Consejerı́a de Innovación, Ciencia y Empresa, Junta de
Andalucı́a (EXC/2005/RNM-204 to S.T., P06-RNM-01627 to
C.M.H.), Consejo Superior de Investigaciones Cientı́ficas and
Fondo Social Europeo (JAE-DOC Program to C.dV.) and
Ministerio de Ciencia e Innovación (Juan de la Cierva Program
to C.dV.).
L I T E R AT U R E CI T E D
Aigbokhan EI, Berner DK, Musselman LJ. 1998. Reproductive ability of
hybrids of Striga aspera and Striga hermonthica. Phytopathology 88:
563–567.
Andresen E, Levey DJ. 2004. Effects of dung and seed size on secondary dispersal, seed predation, and seedling establishment of rain forest trees.
Oecologia 139: 45–54.
de la Bandera MC, Traveset A. 2006. Reproductive ecology of Thymelaea
velutina (Thymelaeaceae) – Factors contributing to the maintenance of
heterocarpy. Plant Systematics and Evolution 256: 97–112.
Bell DT, Rokich DP, McChesney CJ, Plummer JA. 1995. Effects of temperature, light and gibberellic acid on the germination of seeds of 43
species native to Western Australia. Journal of Vegetation Science 6:
797–806.
Bertness MD, Wise C, Ellison AM. 1987. Consumer pressure and seed set in
a salt marsh perennial plant community. Oecologia 71: 190–200.
Bolin JF, Maass E, Tennakoon KU, Musselman LJ. 2009. Host-specific germination of the root holoparasite Hydnora triceps (Hydnoraceae). Botany
87: 1250– 1254.
Bouman F, Meijer W. 1994. Comparative structure of ovules and seeds in
Rafflesiaceae. Plant Systematics and Evolution 193: 187–212.
Bouwmeester HJ, Matusova R, Zhongkui S, Beale MH. 2003. Secondary
metabolite signalling in host –parasitic plant interactions. Current
Opinion in Plant Biology 6: 358–364.
Bruvo-Madarić B, Plohl M, Ugarković D. 2007. Wide distribution of related
satellite DNA families within the genus Pimelia (Tenebrionidae).
Genetica 130: 35–42.
Burel F. 1989. Landscape structure effects on carabid beetles spatial patterns
in western France. Landscape Ecology 2: 215–226.
Burns KC. 2006. Weta and the evolution of fleshy fruits in New Zealand. New
Zealand Journal of Ecology 30: 405–406.
Cárdenas A, Hidalgo JM. 2006. Ecological impact assessment of the
Aznalcollar mine toxic spill on edaphic Coleopteran communities in
the Guadiamar river basin (Southern Iberian Peninsula). Biodiversity
and Conservation 15: 361– 383.
Copeland LO, McDonald MB. 2001. Principles of seed science and technology, 4th edn. Boston, MA: Kluwer Academic.
Corlett RT. 2002. Frugivory and seed dispersal in degraded tropical east Asian
landscapes. In: Levey DJ, Silva WR, Galetti M, eds. Seed dispersal and
frugivory: ecology, evolution and conservation. Wallingford, UK: CAB
International, 451– 465.
Cottrell HJ. 1947. Tetrazolium salt as a seed germination indicator. Nature
159: 748.
Daws MI, Pritchard HW, Van Staden J. 2008. Butenolide from plantderived smoke functions as a strigolactone analogue: evidence from parasitic weed seed germination. South African Journal of Botany 74:
116–120.
Duthie C, Gibbs G, Burns KC. 2006. Seed dispersal by weta. Science 311:
1575.
Ecroyd CE. 1996. The ecology of Dactylanthus taylorii and threats to its survival. New Zealand Journal of Ecology 20: 81–100.
Estrada A, Coates-Estrada R. 1991. Howler monkey (Alouatta palliata),
dung beetles (Scarabaeidae) and seed dispersal: ecological interactions
in the tropical rain forest of Los Tuxtlas, Mexico. Journal of Tropical
Ecology 7: 459–474.
Fallaci M, Colombini L, Palesse L, Chelazzi L. 1997. Spatial and temporal
strategies in relation to environmental constraints of four tenebrionids
inhabiting a Mediterranean coastal dune system. Journal of Arid
Environments 37: 45– 64.
Franklin GL. 1945. Preparation of thin sections of synthetic resins and wood–
resin composites, and a new macerating method for wood. Nature 155:
51.
Furedi MA, McGraw JB. 2004. White-tailed deer: dispersers or predators of
American ginseng seeds? American Midland Naturalist 152: 268– 276.
Garcı́a-Franco JG, Rico-Gray V. 1997. Dispersión, viabilidad, germinación
y banco de semillas de Bdallophyton bambusarum (Rafflesiaceae) en la
costa de Veracruz, México. Revista de Biologı́a Tropical 44: 87–94.
Guzowska I. 1964. Reinvestigation of embryo sac development, fertilization
and early embryogeny in Cytinus hypocistis L. Acta Societatis
Botanicorum Poloniae 33: 157– 166.
Herr JM. 1971. A new clearing-squash technique for the study of ovule development in angiosperms. American Journal of Botany 58: 785 –790.
Herrera CM. 1995. Plant–vertebrate seed dispersal systems in the
Mediterranean: ecological, evolutionary and historical determinants.
Annual Review of Ecology and Systematics 26: 705– 727.
Herrera CM. 2002. Seed dispersal by vertebrates. In: Herrera CM, Pellmyr O,
eds. Plant–animal interactions. An evolutionary approach. Oxford:
Blackwell Science, 185– 208.
Hidayati SN, Meijer W, Baskin JM, Walck JL. 2000. A contribution to the
life history of the rare Indonesian holoparasite Rafflesia patma
(Rafflesiaceae). Biotropica 32: 408–414.
Howe HF, Smallwood J. 1982. Ecology of seed dispersal. Annual Review of
Ecology and Systematics 13: 201–228.
Hughes L, Dunlop M, French K et al. 1994. Predicting dispersal spectra: a
minimal set of hypotheses based on plant attributes. Journal of Ecology
82: 933 –950.
Johansen DA. 1940. Plant microtechnique. New York: McGraw Hill.
King P, Milicich L, Burns KC. 2010. Body size determines rates of seed dispersal by giant king crickets. Population Ecology 53: 73–80.
Kraus JE, Arduin M. 1997. Manual básico de métodos em morfologia
vegetal. Rio de Janeiro: EDUR.
Kuijt J. 1969. The biology of parasitic flowering plants. Davis, CA: University
of California Press.
Levey DJ, Byrne MM. 1993. Complex ant–plant interactions: rain forest ants
as secondary dispersers and post-dispersal seed predators. Ecology 74:
1802–1812.
Li AR, Guan KY. 2007. Mycorrhizal and dark septate endophytic fungi of
Pedicularis species from northwest of Yunnan Province, China.
Mycorrhiza 17: 103– 109.
Li AR, Guan KY. 2008. Arbuscular mycorrhizal fungi may serve as another
nutrient strategy for some hemiparasitic species of Pedicularis
(Orobanchaceae). Mycorrhiza 18: 429 –436.
Lobo JM, Sanmartı́n I, Martı́n Piera F. 1998. Diversity and spatial turnover
of dung beetle (Coleoptera: Scarabaeoidea) communities in a protected
area of south Europe (Doñana National Park, Huelva, Spain). Elytron
11: 71– 88.
Logan DC, Stewart GR. 1991. Role of ethylene in the germination of the
hemiparasite Striga hermonthica. Plant Physiology 97: 1435–1438.
López Granados F, Garcı́a Torres L. 1999. Longevity of crenata broomrape
(Orobanche crenata) seed under soil and laboratory conditions. Weed
Science 47: 161–166.
McDonald MB, Kwong FY. 2005. Flower seed: biology and technology.
Wallingford, UK: CABI Publishing.
Mescher MC, Smith J, De Moraes CM. 2009. Host location and selection by
holoparasitic plants. In: František B, ed. Plant-environment interactions.
From sensory plant biology to active plant behavior. Berlin:
Springer-Verlag, 101– 118.
Meulebrouck K, Ameloot E, Van Assche JA, Verheyen K, Hermy M,
Baskin CC. 2008. Germination ecology of the holoparasite Cuscuta
epithymum. Seed Science Research 18: 25–34.
Moya Ó, Contreras-Dı́az HG, Oromı́ P, Juan C. 2006. Using statistical phylogeography to infer population history: case studies on Pimelia darkling
beetles from the Canary Islands. Journal of Arid Environments 66:
477– 497.
Overdorff DJ, Strait SG. 1998 Seed handling by three prosimian primates in
southeastern Madagascar: implications for seed dispersal. American
Journal of Primatology 45: 69– 82.
van der Pijl L. 1982. Principles of dispersal in higher plants. Berlin:
Springer-Verlag.
Pritchard HW. 1985. Determination of orchid seed viability using fluorescein
diacetate. Plant, Cell and Environment 8: 727– 730.
de Vega et al. — Endozoochory by beetles
Rasmussen HN. 1995. Terrestrial orchids from seed to mycotrophic plant.
Cambridge: Cambridge University Press.
Rasmussen HN, Whigham DF. 1993. Seed ecology of dust seeds in situ: a
new study technique and its application in terrestrial orchids. American
Journal of Botany 80: 1374– 1378.
Rico-Gray V, Oliveira PS. 2007. The ecology and evolution of ant– plant
interactions. Chicago: The University of Chicago Press.
de los Santos A, Montes C, Ramı́rez Dı́az L. 1982. Modelos espaciales de
algunas poblaciones de coleópteros terrestres en dos ecosistemas del
bajo Guadalquivir (S.W. España). Mediterránea Serie de Estudios
Biológicos 6: 65–92.
de los Santos A, Montes C, Ramı́rez Dı́az L. 1988. Life histories of some
darkling beetles (Coleoptera: Tenebrionidae) in two mediterranean ecosystems in the lower Guadalquivir (Southwest Spain). Environmental
Entomology 17: 799–814.
de los Santos A, Ferrer FJ, de Nicolás JP. 2002. Habitat selection and assemblage structure of darkling beetles (Col. Tenebrionidae) along environmental gradients on the island of Tenerife (Canary Islands). Journal of
Arid Environments 52: 63– 85.
Schupp EW. 1993. Quantity, quality and the effectiveness of seed dispersal by
animals. Vegetatio 107/108: 15– 29.
Serrano J, Lencina JL, Andujar A. 2003. Distribution patterns of Iberian
Carabidae (Insecta, Coleoptera). Graellsia 59: 129–153.
Stapp P. 1997. Microhabitat use and community structure of darkling beetles
(Coleoptera: Tenebrionidae) in shortgrass prairie: effects of season shrub
and soil type. American Midland Naturalist 137: 298– 311.
Stewart GR, Press MC. 1990. The physiology and biochemistry of parasitic
angiosperms. Annual Review of Plant Physiology and Plant Molecular
Biology 41: 127– 151.
Thorogood CJ, Rumsey FJ, Hiscock SJ. 2009. Seed viability determination
in parasitic broomrapes (Orobanche and Phelipanche) using fluorescein
diacetate staining. Weed Research 49: 461– 468.
Traveset A, Riera N. 2005. Disruption of a plant– lizard seed dispersal system
and its ecological effects on a threatened endemic plant in the Balearic
Islands. Conservation Biology 19: 421– 431.
637
Varela O, Bucher EH. 2006. Passage time, viability, and germination of seeds
ingested by foxes. Journal of Arid Environments 67: 566 –578.
de Vega C. 2007. Reproductive biology of Cytinus hypocistis (L.) L.: host –
parasite interactions. PhD thesis, University of Seville, Spain.
de Vega C, de Oliveira RC. 2007. A new procedure for making observations
of embryo morphology in dust-like seeds with rigid coats. Seed Science
Research 17: 63–67.
de Vega C, Ortiz PL, Arista M, Talavera S. 2007. The endophytic system of
Mediterranean Cytinus (Cytinaceae) developing on five host Cistaceae
species. Annals of Botany 100: 1209–1217.
de Vega C, Berjano R, Arista M, Ortiz PL, Talavera S, Stuessy TF. 2008.
Genetic races associated with the genera and sections of host species in
the holoparasitic plant Cytinus (Cytinaceae) in the Western
Mediterranean basin. New Phytologist 178: 875–887.
de Vega C, Arista M, Ortiz PL, Herrera CM, Talavera S. 2009. The
ant-pollination system of Cytinus hypocistis (Cytinaceae), a
Mediterranean root holoparasite. Annals of Botany 103: 1065– 1075.
de Vega C, Arista M, Ortiz PL, Talavera S. 2010. Anatomical relations
among endophytic holoparasitic angiosperms, autotrophic host plants
and mycorrhizal fungi: a novel tripartite interaction. American Journal
of Botany 97: 730– 737.
Watanabe
K.
1942.
Morphologisch-biologische
studien
uber
Balanophoraceen in Nippon. Journal of Japanese Botany 18: 382 –390.
Westoby M, Leishmann M, Lord J. 1996. Comparative ecology of seed size
and dispersal. Philosophical Transactions of the Royal Society of London,
Series B 351: 1309–1317.
Wolff A, Debussche M. 1999. Ants as seed dispersers in a Mediterranean oldfield succession. Oikos 84: 443 –452.
Wyman TE, Trewick SA, Morgan-Richards M, Noble ADL. 2010.
Mutualism or opportunism? Tree fuchsia (Fuchsia excorticata) and tree
weta (Hemideina) interactions. Austral Ecology doi: 10.1111/
j.1442-9993.2010.02146.x
Yoder JI. 1999. Parasitic plant responses to host plant signals: a model for
subterranean plant– plant interactions. Current Opinion in Plant Biology
2: 65–70.