Deer predation on leaf miners via leaf abscission
- First Online:
- Cite this article as:
- Yamazaki, K. & Sugiura, S. Naturwissenschaften (2008) 95: 263. doi:10.1007/s00114-007-0318-z
- 117 Views
The evergreen oak Quercus gilva Blume sheds leaves containing mines of the leaf miner Stigmella sp. (Lepidoptera: Nepticulidae) earlier than leaves with no mines in early spring in Nara, central Japan. The eclosion rates of the leaf miner in abscised and retained leaves were compared in the laboratory to clarify the effects of leaf abscission on leaf miner survival in the absence of deer. The leaf miner eclosed successfully from both fallen leaves and leaves retained on trees. However, sika deer (Cervus nippon centralis Kishida) feed on the fallen mined leaves. Field observations showed that deer consume many fallen leaves under Q. gilva trees, suggesting considerable mortality of leaf miners due to deer predation via leaf abscission. This is a previously unreported relationship between a leaf miner and a mammalian herbivore via leaf abscission.
KeywordsEarly leaf abscissionLeaf minerPlant-mediated herbivore–herbivore interactionSika deerStigmella
Plant-mediated interactions among animals have recently attracted much attention from ecologists (e.g., Strauss and Irwin 2004). In particular, multispecies interactions via induced defenses or volatiles of plants have been studied intensively (Denno et al. 1995; Karban and Baldin 1997; Shiojiri et al. 2002).
Early leaf abscission caused by herbivores is a widespread phenomenon in terrestrial plants (Risley and Crossley 1988). Early leaf abscission exerts adverse effects on some sessile endophagous herbivores, especially leaf miners, because herbivore death is caused by desiccation and deterioration of the abscised leaves (Faeth et al. 1981; Bultman and Faeth 1986; Williams and Whitham 1986; Simberloff and Stiling 1987; Stiling et al. 1991; Preszler and Price 1993; Waddell et al. 2001). However, some leaf miners are not negatively affected by premature leaf abscission because they can complete their growth within the abscised leaves on the ground (Hering 1951; Engelbrecht et al. 1969). The survival of leaf miners in fallen leaves is relatively high when the larvae are mature and the deterioration rate of the abscised leaves is low because of relatively high humidity on the ground (Kahn and Cornell 1983; Pritchard and James 1984; Stiling and Simberloff 1989). Thus, the effects of early leaf abscission on herbivores vary considerably among plant–herbivore systems. However, although early leaf abscission has been examined primarily as an induced defense against herbivores, it has only rarely been studied from the viewpoint of multispecies interactions (but see Kahn and Cornell 1983). Kahn and Cornell (1983) demonstrated that early leaf abscission may be beneficial to some leaf miners because the mines in fallen leaves on the forest floor escape parasitism.
We recently observed that leaves of the oak Quercus gilva Blume (Fagaceae) that were mined by Stigmella sp. (Lepidoptera: Nepticulidae) were abscised early and that sika deer, Cervus nippon centralis Kishida (Mammalia: Cervidae), fed on the fallen leaves in a warm-temperate forest in central Japan. This suggests that multispecies interactions may be mediated by early leaf abscission, unless early leaf abscission is an induced defense of oak trees that results in the death of leaf miners in fallen leaves. Leaf miner larvae may be able to survive in fallen leaves on the ground, but deer may prey on them in the fallen leaves. To determine the effects of leaf abscission and deer leaf predation on leaf miner mortality, we examined whether mined leaves tend to abscise earlier than unmined leaves, whether larvae can survive in the abscised leaves, and to what extent deer prey upon the fallen mined leaves.
We first compared the abscission rates of mined and unmined leaves in the field. Second, we reared leaf miners in the abscised leaves on wet soil in the laboratory to examine the possibility that miners complete their growth within the abscised leaves. Third, we observed the feeding behavior of sika deer in the field to infer the effect of their predation on miners. Fourth, we offered both mined and unmined oak leaves to deer in the field to ascertain whether deer discriminate between them. Finally, we discuss the interactions among oak trees, leaf miners, and deer.
Materials and methods
Leaf miner and host plant biology
Q. gilva sheds leaves of the preceding year mainly in mid-April, and new leaves flush beginning in late April. Adult Stigmella sp. (moths) eclose from May to June and oviposit on new host leaves of current-year shoots. Stigmella sp. overwinters as an early instar larva in the leaf mine. From late March to April, the larva matures, egresses from the mine, moves into the soil, and pupates in a small cocoon. The larva completes its growth within a part of a single host leaf. Thus, Stigmella sp. has a univoltine life cycle (Yamazaki, personal observations).
Study site and sika deer
Field observations and experiments were conducted in a 50 × 100-m area of Nara Park (34°41′ N, 135°50′ E; about 110 m above sea level), Nara Prefecture, central Japan in mid-March of 2003 and 2005. Nara Park and the adjacent Kasugayama Hill are covered by warm-temperate old-growth forest, where evergreen oaks such as Q. gilva, Q. sessilifolia Blume, and Q. acuta Thunb. predominate. Sika deer, which are naturally distributed in this area, have traditionally been preserved as “divine messengers” for over 1,000 years (Fujita 1997). Currently, about 1,200 sika deer inhabit the 660-ha area of Nara Park and its surroundings, and the deer density is very high at about 1.8 individuals per hectare (Fujita 1997).
To compare abscission rates between mined and unmined leaves of Q. gilva, we selected five young trees (7–7.5 m tall) in 2003 and five old trees (12–20 m tall) in 2005 at the study site. From each tree, 50 (young trees; tree nos. 1–5) or 100 leaves (old trees; tree nos. 6–10) were haphazardly sampled at a height of 2–3 m above ground. The leaves were only sampled from 1-year-old shoots because Stigmella sp. mines were found exclusively in these leaves. In addition, we randomly collected freshly fallen leaves of Q. gilva from beneath the crown of each tree (Fig. 1b). Each leaf sample was placed individually in a plastic bag and returned to the laboratory. Each leaf was inspected to determine whether it contained mines, and the number of mines was recorded. In addition, to calculate adult emergence rates from fallen leaves, the number of live Stigmella sp. larvae was counted by observing the leaf while allowing light to pass through it. The frequency of mined and unmined leaves was compared both on and under the tree for each tree using the chi-square test. In addition, mine densities in mined leaves were compared between leaves found on and under the tree for each tree using the Mann–Whitney U test.
To examine whether a miner could complete its growth within an abscised leaf, mined fallen leaves from four young trees were placed on wet Akadama-tsuchi soil (standard horticultural soil used in Japan) in 250-ml plastic cups (30–40 leaves per cup) and reared in the laboratory until early July. Several minute holes were punched in the lids of the cups for ventilation. Tap water was sprayed on the soil every 3–4 days to prevent desiccation. In addition, five twigs bearing leaves with many mines of Stigmella sp. were cut from a young Q. gilva tree and returned to the laboratory, and the proximal ends were soaked in a flask of tap water. Egressed mature larvae were placed in a 500-ml plastic cup containing wet Akadama-tsuchi soil and reared in the laboratory, allowing us to estimate the adult eclosion rate from non-abscised leaves. In addition to estimating the adult eclosion rate, these rearing experiments could detect parasitism by parasitoids.
To infer the effect of predation by sika deer on the leaf miners, we observed the feeding activity of deer at the study site. We located a herd of sika deer, comprised of about 20 individuals, on 17 March 2003. We observed 10 adult female deer for 1 min each between 14:30 and 16:00; their food choices were recorded. In addition, between 11:00 and 13:00 on 18 March 2005, we observed four females in a herd consisting of 10 adult deer for 5 min each and recorded their food choices.
If deer discriminate between mined and unmined leaves and prefer to feed on unmined leaves, then mined leaves may predominate on the ground. To ascertain whether this was the case, we offered a pair of mined and unmined Q. gilva leaves to two adult males, ten adult females, and three young deer (unknown sex) on 17 March 2003. The upper surface of the leaves was presented to the deer because the underside is covered with dense trichomes that might prevent deer from perceiving the mines. The feeding response of each deer was recorded.
Percentage of mined and unmined leaves on and under Q. gilva trees
Proportion (%) of leaves mined by Stigmella sp. and without mines that were retained on and abscised from Q. gilva trees
Retained leaves (%)
Abscised leaves (%)
Number of mines (means±SE) of Stigmella sp. per mined leaf in leaves that were retained on and abscised from Q. gilva trees
1.2 ± 0.1
5.9 ± 0.8
1.5 ± 0.3
7.7 ± 1.2
1.9 ± 0.4
5.9 ± 1.2
3.3 ± 0.4
11.6 ± 1.2
4.3 ± 0.7
13.8 ± 0.5
2.2 ± 0.2
3.5 ± 0.1
2.5 ± 0.3
9.1 ± 0.3
2.9 ± 0.3
7.5 ± 0.5
3.4 ± 0.4
11.5 ± 0.4
3.0 ± 0.4
7.1 ± 0.3
Adult eclosion rates from abscised and retained leaves
Eclosion rates of adult Stigmella sp. from abscised (nos. 1–4) and retained leaves of Q. gilva
Eclosion rate (%)
No. of larvae examined
Feeding activity of sika deer
All sika deer observed fed on fallen leaves of Q. gilva (Fig. 1c). The number of fallen Q. gilva leaves eaten within 1 min was 8.6 ± 1.9 leaves (mean±SE; range = 1–22; n = 10) in 2003. In 2005, the number of fallen leaves eaten within 5 min was 87.0 ± 16.5 leaves (range = 54–121: n = 4). Deer also consumed grasses and water but spent most of their feeding time eating fallen Q. gilva leaves.
Feeding preference of sika deer for mined or unmined leaves
One of two adult males, seven of ten adult females, and all three young deer fed on both mined and unmined leaves of Q. gilva. In total, 11 of 15 deer (73.3%) fed on both types of leaf. These deer rushed directly to us and ate additional mined and unmined leaves. The other four deer glanced and sniffed at the leaves but ultimately rejected both mined and unmined leaves.
We examined early leaf abscission of oak trees, the effects of leaf abscission on leaf miner survival, and the feeding behavior of deer to understand leaf miner–deer interactions via oak trees. Abscised Q. gilva leaves contained more Stigmella sp. mines than leaves that were retained on the trees (Tables 1 and 2). Although retained leaves were collected from only 2–3 m above the ground, and mine density and abscission tendency may vary within a tree crown (cf. Hespenheide 1991), a high proportion (74.3–100%) of abscised leaves were mined. Because the leaves of shoots >1-year-old (which bore no miners but could not be distinguished from 1-year-old leaves) were included in samples of abscised leaves from the ground, the tendency of Q. gilva to abscise mined leaves was somewhat underestimated. In addition, many >1-year-old leaves, which were not included in samples of retained leaves, were still in the tree crowns. Because sika deer did not discriminate between mined and unmined leaves when feeding, our results indicate that mined leaves tend to be abscised earlier than unmined leaves and that leaves with high densities of mines fall to the ground more frequently. The possibility that adult leaf miners oviposit preferentially on leaves that tend to abscise early is minimal because of the time lag between the oviposition period and the season of early leaf abscission; a previous study supports this hypothesis (Preszler and Price 1993). Therefore, the presence of mines in leaves may promote early leaf abscission. In addition, early leaf abscission may occur in a spatially density-dependent fashion. Density-dependent leaf abscission has been reported for several species (Simberloff and Stiling 1987; Stiling et al. 1987; James and Pritchard 1988).
Does early leaf abscission negatively affect leaf miners? We found that miners could successfully eclose at similar rates from abscised and retained leaves (Table 3). This high eclosion rate of Stigmella sp. from abscised leaves may be characteristic of the family Nepticulidae; for example, the larvae of Nepticula argyropeza can complete development by maintaining ‘green islands’ on fallen senescent leaves by secreting cytokinins (Engelbrecht et al. 1969). We found that fallen mined leaves remained green for more than 1 month in plastic cups in the laboratory. Therefore, early leaf abscission per se may not negatively affect Stigmella sp. on Q. gilva.
Escape from parasitoids is considered an advantage for larvae in abscised leaves on the ground (Kahn and Cornell 1983). However, we observed no parasitoid emergence from abscised or retained leaves, indicating that parasitoids did not affect early leaf abscission. Epigaeic predators such as ants and ground beetles may also prey on these moth larvae, although this type of predator appears relatively inactive because of low temperatures in early spring.
Our field observations suggest that sika deer feed primarily on fallen Q. gilva leaves in early spring. From late autumn to early spring, sika deer use fallen leaves as their main food source because of a lack of other food (Asada and Ochiai 1996; Torii et al. 2000). The deer feed intermittently while migrating short distances during the daytime (Tatsuzawa and Fujita 2001). One young Q. gilva tree was surrounded by 35–155 fallen leaves (Tables 1 and 2) and one deer consumed eight to nine leaves per minute; thus, a herd of deer (20 individuals) could potentially consume all of these leaves within 1 min, although deer might miss leaves fallen at concealed sites such as under dead leaves.
Sika deer preferred to feed on Q. gilva leaves irrespective of the presence of mines. Usually, because fresh Q. gilva leaves are beyond the reach of deer, they can only eat Q. gilva leaves during the leaf-fall season. Early abscission of mined Q. gilva leaves, i.e., before that of unmined Q. gilva leaves and leaves of other broad-leaved species, may enhance the predation rate by deer. Because it took 1–3 weeks for Stigmella sp. larvae in abscised leaves to complete development and exit the leaves, most larvae in fallen leaves appear to be preyed upon by deer.
Conceivably, the miners may induce early leaf abscission, after which deer accidentally prey on miners in the fallen leaves, thereby contributing to leaf miner mortality. This inadvertent, indirect effect of consumption of fallen leaves by deer may potentially result in reduced leaf miner densities on oak. Hence, we have documented a previously unreported relationship between a leaf miner and a mammalian herbivore via a plant’s response to the herbivore (i.e., leaf abscission).
However, we revealed only the behavioral aspects of the interactions among oaks, leaf miners, and deer. The effects of the interactions on population regulation remain to be explored. We predict the following effects. First, the leaf miner may have little negative effect on oak populations. The negative effects of leaf miners on host plants are not generally considered to be serious (Hespenheide 1991). Because the Stigmella leaf miner density was relatively high in this system (Fig. 1, a and b; Tables 1 and 2), and leaf abscission induced by the leaf miner possibly occurred earlier than nutrient translocation, judging from the fresh green color of the fallen leaves (Prioetti 1998; Fig. 1a and b), the leaf miner may cause a slight decline in oak growth but have little effect on oak populations. Second, deer feeding on oak leaves may have a negative effect on oak populations. Because the deer browse on seedlings and saplings of various plants, including oaks (Shimoda et al. 1994), this negative effect on seedlings may counteract the positive effect of decreasing leaf miner densities in fallen leaves. Third, deer predation on leaf miners in fallen leaves may not considerably decrease the leaf miner population because the study area still supports a relatively high density of leaf miners (Fig. 1, a and b; Tables 1 and 2), despite relatively high deer density. Recently, deer densities have reached overcrowding levels in several areas of the world, and the concomitant enhancement of deer browsing prevents the regeneration of plant communities (McShea et al. 1997; Maeji et al. 1999). The increase in deer feeding activity may decrease leaf miner density in this system and may potentially cause changes in communities elsewhere via unexpected interactions with deer.
We thank Prof. M. Kato of Kyoto University for valuable advice during the course of this research and Dr. T. Czeschlik, Dr. E.F. Connor and two anonymous reviewers for constructive comments on the manuscript. The observations and experiments comply with the current laws of Japan.