Japanese oak silkmoth feeding preference for and performance on upper-crown and lower-crown leaves

Study organism

Antheraea yamamai is widespread in Japan. It is polyphagous, feeding on oaks and chestnuts of Fagaceae and cherry and apple trees of Rosaceae. The adults fly in September–October, and the females lay their eggs on the twigs of their food trees. The eggs overwinter and hatch in mid- to late April. The larvae feed for approximately 2 months, and pupate in spun leaves (Inoue 1982; N. Teramoto, unpubl. data, 1994–1999). A deciduous oak, Quercus acutissima, is one of the most suitable food plants for larval growth of A. yamamai (Teramoto 2001).

Strains of A. yamamai have been established and maintained for their silk in sericulture research stations. For each of the present experiments, we used larvae hatched from eggs laid by a single female of a strain reared in Shiga Prefecture Agricultural General Center to minimize the effects of genetic variability on the results.

Leaf sampling site

Leaves of Q. acutissima were sampled in a secondary forest of Kasuga Hill in Nara City (380 m a.s.l.; 34°41′N, 135°51′E). The forest is dominated by two deciduous oaks, Q. serrata and Q. acutissima, which are 12–16 m in height. Larvae of A. yamamai are occasionally found on these oaks.

Leaf traits

Three or five trees of Q. acutissima were selected monthly from May to August 2000 (13 May, 24 June, 19 July, and 26 August). We haphazardly sampled 30 undamaged leaves from each of the upper (12–16 m above the ground) and lower (4–6 m above the ground) parts of the individual tree crowns. These leaves were brought to the laboratory in an insulated box filled with ice. For 20 leaves the outline was scanned using an image scanner (GT-4000; Epson, Suwa, Japan), and the area was measured by using free image analysis software (LIA32 for Windows 95, version 0373; Yamamoto 1997). After being weighed, the leaves were dried at 80°C for 48 h to determine their dry weight. The leaf dry mass was then calculated as dry weight per unit area (g/cm²) and the water content (%) as ((wet weight − dry weight)/wet weight) × 100. Afterwards, the leaves were ground together into fine powder. The sample was analyzed two or three times for the nitrogen and carbon contents using an NC analyzer (Sumigraph NC-90 A; Shimazu, Kyoto, Japan).

The other ten leaves were cross-sectioned at the middle of the mid-vein with a razor by hand, and the thickness was measured using a microscope with an ocular micrometer.

Feeding preference of A. yamamai larvae

Eggs deposited by a single A. yamamai female of Shiga SG strain were exposed to an incubator (25°C and 0 h light : 24 h dark (LD 0:24)). After 3 days, the light condition was changed to LD 24:0 for induction of hatching. Twenty-three hatchlings that emerged within 3 days after induction (2–4 July 2000) were used for the experiment. They were individually placed in a transparent plastic box (100 mm × 100 mm × 25 mm) that contained a sheet of dry filter paper on the bottom and moistened sanitary cotton on the inside of the lid, and were kept in an incubator maintained at 23 ± 1°C with a photoperiod of LD 14:10. They were reared on primary young or secondly flushed young leaves of Q. acutissima except during the following trials. On the fourth day after the third moulting (26–28 July) and the fifth day after the fourth moulting (5–10 August), larvae were offered a set of four or six expanded undamaged leaves, half of which were from the upper crown and half of which were from the lower crown. The area of each leaf was measured in advance so that areas of leaves from the upper and lower crowns were almost equal. After 6 h, the area of each leaf was measured again, and the total area consumed by the larva was determined. The dry weight consumed was calculated as area consumed × dry mass, where dry mass was the mean value obtained in the leaf trait investigation mentioned previously.

Leaf consumption and performance on upper-crown and lower-crown leaves

Offspring deposited by a single female of Shiga YG strain were used for the experiment. A total of 160 hatchlings that emerged during 28–30 May 2001 were individually reared on young leaves until the second instar (12–17 June) by using the same method as described above. After being weighed on the first day of the second moulting, they were divided into two groups with the same mean and variance in fresh weight. The larvae of one group were individually reared on expanded undamaged leaves collected from the upper crown, whereas those of the other group were reared on leaves from the lower crown. The larvae were checked daily and were given new undamaged leaves every 1–2 days. Rearing boxes were randomized daily within the incubator to avoid position effects.

Immediately after the fifth molting stage, individual larvae were offered four to six expanded undamaged leaves, the areas of which were measured in advance. After 24 h, the area of each leaf was measured again to determine the total leaf area consumed by the larva. Leaf wet and dry weights consumed by individual larvae were calculated as area consumed × wet or dry mass, where wet and dry masses were the mean values obtained from 15 upper-crown leaves (wet mass: 19.1 × 10⁻³ g/cm²; dry mass: 9.44 × 10⁻³ g/cm²) or 15 lower-crown leaves (wet mass: 15.1 × 10⁻³ g/cm²; dry mass: 6.56 × 10⁻³ g/cm²).

For each larva, we measured fresh weights at apolysis phases of the third, fourth and fifth instars and on the 14th day of the sixth instar, developmental time from the third instar to the start of spinning for pupation, and fresh weight of the pupa on the 16th day after spinning. Pupae were dried at 80°C for 48 h, and their dry weight was measured. They were sexed on the basis of the morphology of the antenna.

The relative growth rate (RGR) of each individual was calculated using the formula (Radford 1967):

where W_P is the dry weight of the pupa, W_L3 is the dry weight of the second moulting larva, and T is the time (days) from third instar to spinning for pupation. W_L3 was represented by a mean value from 11 larvae (i.e. 28.1 × 10⁻³ g).

Statistical analyses

Differences in leaf area, leaf dry mass, thickness and water content between upper-crown and lower-crown leaves were investigated using split-plot analyses of variance (anova). Sampling date and leaf were treated as fixed factors, and tree was nested within sampling date and was considered to be a random factor. In order to meet the assumptions of normality, leaf areas were log_e transformed and water content data were arcsine transformed before analyses.

Differences in nitrogen content and N/C ratio between upper-crown and lower-crown leaves were examined by mixed factorial anova. Sampling date and leaf were treated as fixed factors, and tree was considered as a random factor. Nitrogen content data were arcsine transformed before analysis.

The preference of larvae for upper-crown and lower-crown leaves (variables: leaf dry weight) was examined using mixed factorial anova. Leaf and larval instar were fixed factors, whereas larva was a random factor.

Differences in larval fresh weight between upper-crown and lower-crown leaves were investigated using repeated-measures anova. Individual larvae were subjects, leaf was the between-subject factor, and instar was the within-subject factor.

The statistical analyses were carried out using SPSS version 7.5 (SPSS 1997).

Leaf traits

Budburst occurred in mid-April in both upper and lower crowns. Leaf area was significantly larger in the lower crown than upper crown (Fig. 1A). Leaf dry mass and leaf thickness gradually or somewhat increased until August, being significantly greater in the upper crown than lower crown (Fig. 1B,C). Leaf water content gradually decreased until August, being significantly higher in the lower crown than upper crown (Fig. 1D). Leaf nitrogen content and N/C ratio decreased until late June, afterwards remaining almost constant (Fig. 1E,F). These parameters were slightly but significantly higher in the upper crown than lower crown.

Feeding preference, performance and leaf consumption of larvae

When fourth and fifth instar A. yamamai larvae were simultaneously offered upper-crown and lower-crown leaves, they consumed a significantly larger dry weight of upper-crown leaves (Fig. 2). This indicates that A. yamamai larvae prefer upper-crown leaves to lower-crown leaves.

The percentage mortalities of larvae reared on upper-crown leaves and those on lower-crown leaves were 12.5% (10/80) and 20% (16/80), respectively. No significant difference was found between them (Fisher’s exact probability = 0.284). Those larvae that died during rearing were excluded in the analyses described later here.

Larval fresh weight was somewhat greater in individuals reared on lower-crown leaves than in those reared on upper-crown leaves at each instar for males and females; the difference as a whole was significant for both sexes (Fig. 3). Interactions between leaf and instar were significant, indicating that differences in larval fresh weight between two categories of leaves vary according to instars.

In contrast, pupal fresh and dry weights were larger in individuals reared on upper-crown leaves than in those reared on lower-crown leaves; significant differences were found in male dry pupae and in female fresh and dry pupae (Fig. 4A,B). In males, neither development time nor RGR significantly differed between the two categories of leaves. In females, however, both indices were significantly larger on upper-crown leaves (Fig. 4C,D).

When sixth instar larvae were fed with either upper-crown leaves or lower-crown leaves, both males and females consumed a significantly larger amount (wet and dry weights) of upper-crown leaves (Fig. 5).

Leaf traits in the upper and lower crowns

In Q. acutissima, upper-crown leaves are smaller and thicker, and have higher nitrogen content than lower-crown leaves (Fig. 1). This is consistent with the theory proposed by Mooney and Ehleringer (1997) that leaves developing under conditions of high light intensity, or sun leaves, tend to have a smaller area, greater thickness, and higher nitrogen content compared with those developing under conditions of low light intensity, or shade leaves. Le Corff and Marquis (1999) and Fortin and Mauffette (2002) obtained similar results for deciduous oaks (Q. alba and Q. velutina) and the sugar maple (A. saccharum), respectively. However, Yamasaki and Kikuzawa (2003) found that sun leaves had lower nitrogen content than shade leaves in a beech (Fagus crenata). Thus, we should be cautious about inferring a generalized positive relationship between light intensity and nitrogen content. Unfortunately, several studies have examined the effects of sun and shade leaves on folivore performance without analyzing the nitrogen content of leaves (Futuyma & Saks 1981; Harrison 1987; Watt 1992; Suomela et al. 1995).

Water content is lower in upper-crown leaves than in lower-crown leaves in Q. acutissima (Fig. 1D). Fortin and Mauffette (2002) also found lower water content in upper-crown leaves of the sugar maple. This is probably attributable to higher transpiration rates in upper-crown leaves in greater light intensity (Grace 1997).

Performance and feeding preference

The most commonly used index for folivore performance is the RGR, expressed as amount of growth attained (dry matter) per unit of body weight (dry matter) per day (Schoonhoven et al. 1998). Pupal weight is positively related to adult size (Stamp & Bowers 1990; Scheirs et al. 2002) and varies with sex (Bauce et al. 1994; Pöykkö & Hyvärinen 2003). Thus, folivore performance should be represented by the RGR for each sex based on larval development time and pupal dry weight. As far as we know, no studies have satisfied these conditions to examine the effect of upper-crown and lower-crown leaves or sun and shade leaves on folivore performance. The RGR values obtained in the present study meet these conditions.

The RGR of female Japanese oak silkmoths is significantly higher on upper-crown leaves than lower-crown leaves, whereas male RGR values do not significantly differ between the two categories of leaves (Fig. 4D). Clearly, the higher RGR of females on upper-crown leaves results from not only a larger dry weight of pupae (Fig. 4C), but also a shorter development time (Fig. 4B). Hence, the effects of upper-crown and lower-crown leaves on folivore performance are likely to vary with the sex.

Upper-crown leaves of Q. acutissima have significantly higher nitrogen content and N/C ratios than lower-crown leaves, as mentioned previously, although the differences are rather small. The higher RGR of females on upper-crown leaves is consistent with the theory that folivore performance increases with nitrogen content and N/C ratio of leaves (Schoonhoven et al. 1998; Awmack & Leather 2002). However, this is not the case with males. The different effects of the nitrogen content and N/C ratio of leaves on female and male RGR values might be partly because females need more protein than males to produce eggs (Nation 2002).

The higher RGR of female larvae on the upper-crown leaves is, however, inconsistent with the theory that folivore performance increases with the water content of leaves (Scriber 1977; Schoonhoven et al. 1998; Henriksson et al. 2003), because upper-crown leaves of Q. acutissima contain lower water content than lower-crown leaves (Fig. 1D). This does not necessarily mean that the larval growth of A. yamamai is independent of leaf water content. The present experiment allowed the larvae to supplement leaf water by taking water from moistened sanitary cotton in rearing boxes. Upper-crown leaves may be more suitable for larval growth than lower-crown leaves only when larvae are free from water deficiency.

The present results raise two points that are relevant when we examine folivore performance. First, the performance based on fresh weight may depend on whether it is measured at larval or pupal stadia. Larvae of the Japanese oak silkmoth on lower-crown leaves have a larger fresh weight at larval stadia (Fig. 2), but they, especially females, attain a smaller fresh weight at pupal stadia (Fig. 4A). This inconsistency seems to be linked to excess water intake from lower-crown leaves, which contain a higher water content than upper-crown leaves. Larvae appear to ingest a greater amount of lower-crown leaves (Fig. 5) to compensate for the lower nitrogen content of the foliage, so they have excess water intake. Larvae defecate rather liquid frass when they are ready to spin cocoons (Gardiner 1982; M. Oishi, unpubl. data, 2001). Larvae fed with lower-crown leaves could excrete excess water in the frass, and they would thus become pupae with a smaller fresh mass.

The second point of note is that performance based on pupal weight may depend on whether it is measured by fresh weight or dry weight. Male pupae of the Japanese oak silkmoth had similar fresh weights when reared on upper-crown and lower-crown leaves, but they had a significantly larger dry weight when reared on the former (Fig. 4A,B). Unlike male pupae, female pupae had larger fresh and dry weights when reared on upper-crown leaves. This suggests that female pupae do not have so much excess water as male pupae when they feed on lower-crown leaves, although we cannot offer any reason for this. In any case, considering that larvae on the lower-crown leaves have an excess of water, conversion formulae for calculating dry weights of larvae and pupae from their fresh weights (Stamp & Bowers 1990; Bauce et al. 1994; Henriksson et al. 2003) should be treated with caution.

Leaf thickness, which is related to toughness, usually reduces folivore feeding and performance (Choong 1996; Schoonhoven et al. 1998). Larvae of A. yamamai, however, exhibited a preference for and better performance on upper-crown leaves (2, 4), despite the fact that these leaves are thicker than lower-crown leaves (Fig. 1C). Because fourth and fifth instar larvae were used in the preference experiment, they may have easily overcome the thickness of the upper-crown leaves. In fact, first and second instar larvae of A. yamamai are unable to eat expanded leaves (Teramoto 2001).

Feeding preference of A. yamamai larvae for upper-crown leaves is in accordance with better performance (larger pupal dry weight for males and higher RGR for females) on the foliage. However, the experiment for feeding preference was carried out when the differences in nitrogen content and N/C ratio between the two categories of leaves were rather small (late July and early August). Hence, there is a possibility that larvae do not select leaves by judging the nutritional value. Larvae might use as feeding stimulants physical or chemical characters that have very little relationship to leaf nutritional value.