Journal of Insect Conservation 8: 263–274, 2004.
2004 Kluwer Academic Publishers. Printed in the Netherlands.

263

-1

Exotic pine plantations and indigenous Lepidoptera in Australia David R. Britton* and Timothy R. New Department of Zoology, La Trobe University, Bundoora, Victoria, Australia 3083; *Author for Correspondence. Present address: D.R. Britton Australian Museum, 6 College Street, Sydney, New South Wales, Australia 2010 (e-mail: [email protected]; phone: 61-2-93206221) Received 5 April 2004; accepted in revised form 19 May 2004

Key words: Chlenias, Exotic species, Geometridae, Pinus radiata, Polyphagy

Abstract Temperate regions of Australia have extensive commercial plantations of Monterey pine, Pinus radiata D. Don. Replacement of indigenous forests by P. radiata is likely to have significant effects on assemblages of native Lepidoptera, and has been considered a major threat to native fauna through displacing native species. However, many species of Lepidoptera, including ennomine geometrid moths in the genus Chlenias Guene`e, have adopted P. radiata as a larval host. Comparisons were made of oviposition preferences and nutritional ecology of Chlenias auctaria Guene´e on P. radiata and two native host plants [Acacia mearnsii De Wild. (Mimosaceae) and Eucalyptus obliqua L’He´rit (Myrtaceae)]. Females showed no significant oviposition preference for any of the three hosts. Growth of sixth instar larvae was significantly less on P. radiata than on the native hosts. Pupal weights were significantly lower, suggesting that the fitness of moths reared on P. radiata will be significantly reduced. The life history strategy of C. auctaria, which includes dispersal of first instar larvae by ballooning, may predispose this species to feed on poor quality hosts, and this may be common to other polyphagous species found feeding on P. radiata in southern hemisphere plantations. The impact of P. radiata may have a long term effect on lepidopteran communities, beyond the simple replacement of indigenous host plants leading to extirpations of feeding specialists.

Introduction The genus Pinus L. (Pinaceae) occurs predominantly in the northern hemisphere, but Pinus radiata D. Don. (Monterey Pine, Radiata Pine) is planted extensively throughout temperate areas of the southern hemisphere as a commercial softwood. Ohmart (1980) estimated that there were over 1.8 million hectares of plantation P. radiata worldwide, and by 1993, 4 million hectares had been established (Center for International Trade in Forest Products 1993). In Australia alone a conservative estimate of over 600,000 ha of P. radiata (of a total area of

over 870,000 ha given over to softwood production: Resource Assessment Commission 1992) has been superseded by an estimate of closer to one million hectares (CSIRO 1996). Most Australian land planted to pines was originally wet sclerophyll forest dominated by Eucalyptus spp. (Myrtaceae), although some had previously been cleared for pastoral purposes. Annual rainfall in these areas normally exceeds 750 mm, with most rainfall occurring in winter. Other southern hemisphere regions also have very large areas of pines. New Zealand had over 900,000 ha by 1984, with a yearly increment of about 40,000 ha (Sutton 1984). Chile and South Africa also have large

264 areas of softwood production, predominantly using P. radiata. Exotic P. radiata plantations are utilised by local insects as food, shelter and habitat. The limited economic value of the native Californian P. radiata has meant that its associated insects are poorly known (Ohmart 1980, 1981, 1982; Ohmart and Voigt 1981). In contrast, burgeoning economic importance in the southern hemisphere has stimulated documentation of insects on P. radiata in Australia, New Zealand and South Africa. As examples, Minko (1961, 1962, 1965) made a detailed study of an Australian plantation, and recorded 26 species of moths, and an additional 32 other herbivorous insects. Rawlings (1953, 1960) and Spiller and Wise (1982) recorded 28 species of moths from New Zealand plantations, and 84 other herbivorous insects. The presence of native herbivorous insects attacking exotic species such as P. radiata has caused considerable concern (Rawlings 1953, 1960), particularly as large areas of monoculture may increase risk of developing pests. Several species of native Lepidoptera have significantly damaged radiata pine in Australia. These include ennomine geometrids [Chlenias spp. (Tindale 1929; Madden and Bashford 1977a, b) and Parathemis lyciaria (Guene´e) (D. King pers. comm.)], Lichenaula sp. (Oecophoridae) (Neumann 1976), Epiphyas postvittana (Walker) (Tortricidae), Hyalarcta huebneri Westwood (Psychidae) and Teia anartoides Walker (Lymantriidae) (Minko 1961, 1962). All are polyphagous species which have added Pinus to their already considerable host range. None of these species has persisted as pests, and in nearly all cases the outbreaks have been associated with silvicultural practices and climatic conditions which have promoted poor tree health and provided suitable conditions for pest species (e.g. Neumann 1976; Madden and Bashford 1977a, b). Nevertheless, their association with pines reflects an unusual host range expansion and the effects of this on native moth assemblages may be far-reaching. Concerns have been raised in regards to displacement and disruption of native fauna by the planting of large areas of radiata pine, particularly when native hardwood forests have been cleared for this purpose. Little is known about the quality of radiata pine as a resource for herbivorous insects. This paper addresses the question of whether P. radiata is a suitable nutri-

tional resource for Chlenias auctaria Guene´e compared to two native host plants and, thus, whether it might either increase or decrease the wellbeing of native moth populations in relation to long term native hosts. The oviposition preference of female moths is also examined, and the implications of polyphagy and the potential impact of P. radiata on native moth species are discussed. Effects of this predominant exotic host on native moths have ramifications for their future wellbeing, as well as for the wider assemblages they represent. Most attention has been paid to displacement or extirpation of ecologically specialised insect herbivores by such exotic plants as a direct consequence of loss of natural habitat, and the capability of some native species to exploit such hosts has largely been ignored in considering their conservation. The study contributes to the wider question of how changed floral balance and host plant quality may influence native moth assemblages.

Moths Considerable confusion occurs in the literature in regard to the specific identity of Chlenias feeding on pines in Australia. For example, Tindale (1929) described C. pini from individuals reared on radiata pine in South Australia, and also identified an additional species (C. pachymela Lower) from collections in the same area. Examination of the genitalia of C. pini, C. pachymela and comparison with other species of Chlenias indicates that these two species are synonymous with C. auctaria Guenee (pers.comm. P.B. McQuillan). Madden and Bashford (1977a, b) encountered problems assigning a specific identity to the species reared from pine in Tasmania. Specimens collected in this study were identified by P.B. McQuillan (University of Tasmania) as C. zonaea, C. auctaria, and C. banksiaria and an undescribed species. McQuillan and Edwards (1996) treated C. auctaria as a synonym of C. banksiaria, but larvae reared from the moths identified as C. auctaria and C. banksiaria in this study clearly represent two species. Pending revision of this genus, for the purposes of this paper the species used for the consumption experiments is C. auctaria, whilst C. auctaria and C. banksiaria were used together for the oviposition trial.

265 The source moths for our trials were lighttrapped in May and June 1993 at two localities in Victoria, Australia. Most came from Kinglake West (Mount Robertson plantations, Melbourne Water Wallaby Creek plantations), 3729¢E, 14514¢S, 600 m above sea level, with others from Blackwood (Lerderderg State Forest) 3729¢E, 14419¢S, 530 m above sea level. Mature P. radiata in the plantations where trapping occurred were about 20–30 years old. At Kinglake the vegetation consisted of P. radiata stands interspersed with areas of sclerophyll forest dominated by Eucalyptus obliqua L’He´rit, with an understorey of Hakea sericea Schrad. & J.Wendl. (Proteaceae), Acacia mucronata Willd. ex H. Wendl., Acacia dealbata Link, A. mearnsii De Wild., A. melanoxylon R.Br. in Ait. (Mimosaceae), Cassinia aculeata (Labill.) R.Br. (Asteraceae), and Pteridium esculentum (Forst.f.) Nakai (Dennstaedtiaceae). Gully areas also included large shrubs of Bursaria spinosa (Pittosporaceae). Some cleared regions are present within the plantations; some of these have regrowth from a previously harvested pine plantation, or are dense thickets of A. dealbata and E. obliqua. Most light-trapping at Kinglake West was done on roadsides within P. radiata plantations. At Blackwood most former vegetation was removed during the goldmining era (1850–1870). Several introduced conifer species are present in small numbers including P. radiata. Remaining native vegetation in the region is similar to that at Kinglake West, with A. melanoxylon sub-dominant to Eucalyptus spp.

Oviposition trials Methods Adult females of Chlenias spp. (both C. auctaria and C. banksiaria, as above) were kept separately in labelled vials and weighed (Sartorius MC1 Analytic AC 2105 electronic top-loading balance, accuracy ±0.0001 g) the day after capture. Moths were then transferred to oviposition cages. Outside dimensions of these aluminium framed, mesh sided cages were 145 · 80 · 80 cm. Trials were conducted in an open paddock in Bundoora, Victoria where moths were subject to weather conditions similar to those at Kinglake West. Each cage contained four different plants in 20 or 25 cm

diameter pots. All plants were similar in height; when trials were initiated the height of plants was noted as a hierarchy from one (tallest) to four (shortest). The initial experimental design included four species: Acacia mearnsii, Eucalyptus obliqua, Pinus radiata (all known hosts) and Allocasuarina verticillata (Lam.) L. Johnson (Casuarinaceae) (not recorded as a host). There was no evidence that adult moths took liquid nourishment, but a weak honey-water mixture was supplied in the cages. Thirty moths were placed in the cages between 1500 and 1700 h the day after their capture. Cages were checked for eggs each day between 1500 and 1700 h and the following parameters scored: (1) Oviposition preference. Whether the eggs were laid on a particular plant species, or on the cage surface. (2) Number of egg clusters. Number of clusters may indicate the suitability of certain hosts. (3) Number of eggs. A measure of the reproductive output of the individual moth. This was related to the initial weight and size as indicated by the length of a single forewing (wing length). (4) Position of egg clusters. This was scored to see if other factors unrelated to the plant hosts may have affected oviposition behaviour. Position in the cage was related to compass directions (e.g. North-east corner, North-west corner, South-east corner, South-west corner), to the height of the cage (Top or bottom) and to the height of the plants. (5) Longevity. Moths which spent more time in cages had greater opportunity to lay more clutches and more eggs than short-lived moths. Longevity was measured from the introduction of the female into the cage until the moth was dead. Analyses of significant differences in oviposition due to physical characteristics of the female moth and the caged environments were conducted using ANOVA and ANCOVA tests on the statistics package JMP Statistical Visualization for the Macintosh, Version 2.0, 1989. Results Twenty of the 30 females laid a mean of 2.8 egg clusters per moth, with a mean total of 111.6 eggs laid within the first 8 days of capture. The mean

266 adult female longevity in cages was 10.1 days, but no eggs were laid after 9 days in the cages. Initial weight of each female correlated with wing length, total number of eggs laid and the largest cluster laid. The total number of clusters laid is related to the wing length of each female, and shows a weakly positive, but non-significant correlation with initial weight and the longevity of each female. The initial cluster represents approximately half of the total number of eggs laid by each moth, with successive clusters decreasing in size in a loglinear fashion. Chlenias spp. oviposited on the cage netting in preference to the four plant species, with only five moths laying on P. radiata, two on A. mearnsiii and two on A. verticillata. No eggs were laid on E. obliqua, or on flat smooth surfaces such as the aluminium and plastic cage frames, plant pots or the floor of the cage. P. radiata had the greatest number of eggs laid on foliage of a plant in the experiment, but the low overall numbers of females laying on plants prevented any definite conclusions in relation to host plant preference. Positional effects were noted in the experiment, with moths tending to lay eggs in the top half of the cage and on the tops of plants. Most eggs were laid in the south-west, south, south-east and north-east sides of the cages.

Rearing of larvae Larvae of C. auctaria used for food consumption and assimilation experiments were reared from eggs obtained as above. Eggs laid by females matured after a 3 week period, by which time larvae were visible within the eggs. Eggs from different females were mixed together to minimise any effect of individual cohorts within each treatment. Approximately equal numbers of eggs were placed in each cage with the food plant and a quantity of paper towelling. McFarland (1973) recommended paper towelling as a way of trapping the excess silk produced by first instar larvae, which can sometimes trap and kill many larvae. To facilitate eclosion mature eggs were moistened thoroughly with distilled water from an atomiser. Once eggs were well-developed (larvae were visible through the chorion) they generally hatched within two days of this application of water (at 20 C). First instar larvae are small (<3 mm long), active and strongly phototactic.

This behaviour is probably related to larval dispersal by ‘ballooning’. Larvae were kept in perspex and gauze-sided cages (size 440 mm wide · 505 mm high · 340 mm deep) with a removable perspex lid. Cages were placed on white plastic tote boxes to hold trees. The modular cage design allowed for cages to be stacked, so that potted trees up to 1.3 m high could be accommodated with two cages and a deep tote box. Two- or three-year old nurserypurchased trees of Acacia mearnsii, Pinus radiata and Eucalyptus obliqua were used as host plants. All plants received two light doses of a liquid fertiliser prior to use as rearing and experimental host plants. Additional plant material for caterpillars reared on A. mearnsii and E. obliqua was obtained from local trees on campus at times when insufficient pot plants were available, although only leaves from potted plants were used for the feeding trials. The cages were housed in an insectary with constant temperature (20 C), humidity (80 ± 5% RH) and lighting (two 25 W gro-lamps, 12:12 day/night lighting regime). Plants were changed each week, or more frequently if insufficient foliage for the larvae was present. A small atomiser was used to mist the plants as an additional source of water for larvae and to maintain turgor in the host plants. Paper towelling was used to line the floor of each cage and to cover the top of the plant pots. The towelling on the cage floor soaked up excess water from the pots, and the towelling on the rim provided an alternative route to the trunk of the pot plant for larvae that fell from the host plants. Larvae were allowed to pupate in plastic vials (250 ml, 7 cm diameter). Pupae were transferred to vials containing moist sterile sand until emergence of the adults, which were used to confirm identification.

Consumption, growth and utilisation experiments Methods The relative performance of final (sixth) instar larvae of C. auctaria on three different plant hosts was assessed by determining the amount of plant material ingested, the amount of faeces egested and the gain in body weight over a 4 day experimental period for individual larvae provided with known amounts of food. Juvenile foliage of three

267 host plants were used: A. mearnsii, P. radiata and E. obliqua, after previous attempts (Britton 1994) at rearing C. auctaria and C. banksiaria had shown that larvae will not consume mature leaves. A 4 day feeding period was chosen as sufficient time within which to observe meaningful patterns in the feeding ecology of the larvae. A longer period may have introduced errors as larvae approach the prepupal stage of development. Each larva was reared on its respective experimental host from hatching to avoid any bias towards a particular host plant. Gravimetric methods The gravimetric method described by Waldbauer (1968) and further detailed by Scriber and Slansky (1981) and Slansky and Scriber (1985) was adopted as a method for determining basic consumption parameters. The gravimetric method requires the collection of the following data from the experimental larvae over a fixed feeding period: (1) The weight gain of the larvae (g dry weight/ day). (2) The weight of food eaten by the larvae (g dry weight/day). (3) The weight of faeces produced by the larvae (g dry weight/day). Weights were obtained with a Sartorius electronic balance (above). Plastic vials (250 ml, 7 cm diameter) were used for holding individual larvae and foliage for experimental and control trials. The initial dry weights of larvae and host plant material were estimated from the initial wet weights by using the following protocol. (1) Estimations of initial dry weights of larvae. To minimise variation in initial wet weights recently moulted larvae were used for experimental trials. These larvae were assumed to have empty or nearly empty guts in the immediate post-moult period, although several larvae produced small amounts of frass in the period between removal from the food plants and weighing. Calculation of dry weight estimates relied upon data from dry and wet weight of controls. The initial dry weight of larvae was calculated from a sacrificial cohort of freshly moulted sixth instar larvae (n = 30) which

were weighed to obtain a wet weight, killed by freezing, oven-dried to constant weight (96 h at 60 C) and weighed. These larvae were all reared on Acacia mearnsii; it was initially assumed that the species of host plant consumed had minimal effect on the wet weight/dry weight ratio of larvae immediately after moulting. The linear relationship between wet weight and dry weight of sacrificial larvae was used to estimate the initial dry weight of experimental larvae. The mean wet weight of freshly moulted sacrificial larvae was not significantly different from the mean wet weight observed for freshly moulted experimental larvae on all three plant hosts (Student’s t-test, t(104) = 1.7935, p = 0.0915) or the mean wet weight observed for freshly moulted experimental larvae on P. radiata and E. obliqua (Student’s ttest, t(74) = 0.4510, p = 0.6533). The relationship between wet weight and dry weight of sacrificial larvae was highly significant (ANOVA, F(1,28) = 189.0299, p < 0.001). A straight-line of best fit for wet weight by dry weight gave the equation y = 0.193x 0.0028. This was used to estimate initial dry weight of larvae used in the experimental treatments. Estimated dry weights of the larvae used in experimental treatments ranged from 0.0203 to 0.0438 g, with a mean dry weight of 0.0310 ± 0.0007 g. (2) Estimations of initial dry weights of experimental host plant provided. Control aliquots were used to estimate the initial dry weight of the wet food provided for the experiment. Linear regressions of dry weight against wet weight were plotted and the equation for each regression used to calculate an estimate of the original dry weight of the experimental food. The critical assumption for this estimation was that the foliage provided in the control vial had a similar percentage dry weight to wet weight compared to the experimental foliage. To minimise errors each piece of foliage was bisected with a razor blade, placing one half in the experimental vial and the other in the control vial. The differing foliage forms found in each host plant meant that the bisection of leaves or shoots required a different treatment for each host, as follows: (a) Acacia mearnsii The leaves of A. mearnsii are finely dissected bipinnate structures with six to twelve pairs of pinnae, each pinna with 30–40 pairs of leaflets. The entire leaf was bisected

268 across the mid-rib. It was initially expected that the cross-wise bisection of A. mearnsii control foliage would introduce an error into the calculations due to the difference in moisture content between the distal and proximal halves of each piece of foliage. An attempt was made to control for this by placing equal numbers of distal and proximal halves into the experimental and control vials in a similar fashion to the distribution of large and small halves of leaves for E. obliqua treatments (see below). (b) Pinus radiata The shoots of P. radiata consist of numerous two-needle fascicles arranged spirally around a central stem. Lengths of shoots varied between 3 and 15 cm. Needles in an early stage of growth near the distal portion of the shoot are covered by a scale-like sheath. Entire shoots of P. radiata were bisected length-wise with approximately equal numbers of needles and sheaths on either half. (c) Eucalyptus obliqua The leaves of E. obliqua are asymmetrical, so it is impossible to divide each leaf into equal halves; leaves of E. obliqua were divided along the mid-rib, leaving one half larger than the other. To ensure that there an even distribution of leaf tissue between treatments, half of the larger leaf halves went into the control treatment with the other half going into the experimental treatment. The smaller halves of the leaves were distributed between control and experimental vials in a similar fashion. When the wet weight of control foliage was plotted against dry weight E. obliqua showed some minor variation from the fitted straight-line, whilst A. mearnsii and P. radiata showed only very slight deviations from the fitted straight-line. Larval faeces were collected every 2 days to prevent fungal and bacterial decomposition which may occur over the full period (Waldbauer 1968). Faeces were oven-dried to constant weight over 48 h. Final dry weights of larvae were obtained by freezing the larvae and drying to constant weight over 48 h. Dry weights are often preferred over wet weights for experiments with nutritional ecology due to the errors introduced by water loss from plant material over the experimental period. However, wet weights are of considerable importance for the analysis of consumption rates in relation to growth. Final weights of experimental

foliage were used for calculation, rather than estimated wet weights.

Analysis of growth and assimilation: analysis of covariance Analysis of standard ratio-based nutritional indices does not incorporate the effect of initial larval weight, which may be responsible for significant differences in larval consumption and assimilation of host foliage between hosts (Raubenheimer and Simpson 1992; Horton and Redak 1993). Raubenheimer and Simpson (1992) used initial weight of the insect as a covariate and final weight as a response; significant effects of treatment on the final biomass. Horton and Redak (1993) modified the analysis of Raubenheimer and Simpson (1992) by using consumption as the covariate and growth as the response variable. This has the benefit of providing information on the effect of treatment on growth rather than final weight, and also allows the researcher to determine the relative importance of preingestive effects (e.g. deterrence) and postingestive effects (e.g. antibiosis). Probabilities generated by the analysis were adjusted with a sequential Bonferroni adjustment to avoid potential Type I errors (Rice 1989). Tested parameters were deemed significant if the adjusted probability value was less than 0.05.

Other measurements of growth and assimilation for C. auctaria The effect of host plants on the weight of pupae was measured by weighing sexed pupae 2 days after the moult from prepupa to pupa. Pupae used for the comparison of the three host plant species were obtained from preliminary feeding trials in 1993. Because the sex ratio of larvae surviving to pupation was highly skewed, Wilcoxon rank sums test was used to test the effect of sex on pupal weights and initial wet weights of larvae. Pupae from larvae reared on P. radiata weighed significantly less than pupae reared on the other two hosts (Tukey–Kramer HSD, q* = 2.41, p < 0.05). The mean pupal weight for P. radiata was 0.2878 ± 0.0187 g compared to A. mearnsii (0.3719 ± 0.0163 g) and P. radiata (0.3975 ± 0.0167 g).

269 Some larvae died before pupating. Of these 36% mortality occurred on P. radiata, 30% mortality for larvae on E. obliqua and 13% for larvae on A. mearnsii.

Overall results independent of host plant treatments The wet weights of freshly moulted experimental larvae ranged from 0.1206 to 0.2427 g, with a mean wet weight of 0.1763 ± 0.0036 g. Wet weights of freshly moulted larvae sacrificed for estimation of initial dry weights of experimental larvae ranged from 0.1297 to 0.246 g with a mean wet weight of 0.1888 ± 0.0051 g. The water content of sacrificial larvae ranged from 80.8 to 84.6% by weight, with a mean value of 82.2 ± 0.2%. There was a significant degree of auto-correlation (ANOVA, F(1,28) = 9.5798, p < 0.01) between dry weight and the water content of larvae, but the relationship between wet weight and water content was not significant. Larger larvae tended to have smaller water to dry weight ratios. This was due to the previous digestion of plant material, despite attempts to sacrifice larvae before they had fed. The final wet weight of larvae after the 4 day feeding trial ranged from 0.1790 to 0.5994 g, with a mean of 0.3802 ± 0.0094 g. Final dry weights of larvae ranged from 0.0320 to 0.1275 g, with a mean of 0.0682 ± 0.0022 g. The wet weight of food consumed over 4 days by larvae ranged from 0.4237 to 1.1108 g with a mean of 0.7627 ± 0.0190 g. The dry weight of food consumed ranged from 0.3517 to 0.0900 g with a mean of 0.2044 ± 0.0070 g. The dry weight of frass produced by larvae over four days ranged from 0.2568 to 0.0742 g with a mean of 0.1573 ± 0.0048 g. Pupal wet weights were highly variable, ranging from 0.2502 to 0.6468 g with a mean of 0.3573 ± 0.0114 g. Female pupae had a mean weight of 0.4294 ± 0.0145 g and were significantly heavier than male pupae (0.3152 ± 0.0111 g) (Wilcoxon rank sums test, v2 = 29.62, p < 0.001). The sex ratio of pupae reared on all three hosts was significantly biased, with 36 males and 21 females (v2 test, v2 = 3.95, p < 0.05). Initial wet weights of larvae for this subset of the data ranged from 0.0938 to 0.6422 g with a mean of 0.3165 ± 0.0155 g. The initial wet weights of larvae were significantly different for each sex (Wilcoxon rank

sums test, v2 = 3.9, p < 0.05) with male larvae heavier than female larvae (0.3562 ± 0.0225 g compared to 0.2908 ± 0.0292 g).

Results of analysis of growth, consumption and excretion data analysed using ANOVAs and ANCOVAs The effect of host on the growth of sixth instar larvae of C. auctaria was initially tested by comparing the estimated initial dry weight of larvae with the dry weight of growth for each host with an ANCOVA. There was no significant effect of host on dry weight of growth (Tables 1 and 2). Similarly there was no significant effect of host on wet weight of growth independent of the initial wet weight of the larvae (Tables 3 and 4). The effect of host on the dry weight of food consumed by larvae was tested with estimated initial dry weight of larvae as a covariate. There Table 1. Analysis of covariance for dry weight of growth for sixth instar larvae of C. auctaria on three plant hosts with the estimated initial dry weight of larvae as a covariate Source

df

MS (·1000)

F

p (all ns)

Host Est. initial dry weight larvae Host*est. initial dry weight larvae

2 1

0.1918 0.0775

1.19 0.96

0.3108 0.3306

2

0.2852

1.77

0.1785

Table 2. Estimated parameters from the linear model used for the ANCOVA for the dry weight of larval growth on three plant hosts with the estimated initial dry weight of sixth instar larvae of C. auctaria as a covariate with a t-test for the hypothesis that each parameter is equal to zero Term

Host [1–3] Host [2–3] Est. initial dry weight larvae Host [1–3]*est. initial dry weight larvae Host [2–3]*est. initial dry weight larvae

Estimate of slope

Standard error of estimate

t ratio

p (adjusted significance)

0.0173 0.0067 0.2990

0.0121 0.0128 0.3051

1.44 0.52 0.98

0.1547 (ns) 0.6032 (ns) 0.3306 (ns)

0.0851

0.3754

0.23

0.8213 (ns)

0.7268

0.4014

1.81

0.0746 (ns)

Host 1 = Acacia mearnsii, Host 2 = Pinus radiata, Host 3 = Eucalyptus obliqua.

270 Table 3. Analysis of covariance for wet weight of growth for sixth instar larvae of C. auctaria on three plant hosts with the initial wet weight of larvae as a covariate Source

df

MS (·1000)

F

p (all ns)

Host Initial wet weight larvae Host*initial wet weight larvae

2 1

1.4434 1.3751

0.43 0.88

0.6496 0.3663

2

3.8274

1.15

0.3223

Table 4. Estimated parameters from the linear model used for the ANCOVA for the wet weight of larval growth on three plant hosts with the initial wet weight of sixth instar larvae of C. auctaria as a covariate with a t-test for the hypothesis that each parameter is equal to zero Term

Host [1–3] Host [2–3] Initial wet weight larvae Host [1–3]*initial wet weight larvae Host [2–3]*initial wet weight larvae

Estimate Standard t ratio p (adjusted of slope error of significance) estimate 0.0416 0.0379 0.2429

0.0597 0.0635 0.2671

0.70 0.60 0.91

0.4887 (ns) 0.5529 (ns) 0.3663 (ns)

0.0307

0.3286

0.09

0.9259 (ns)

0.5311

0.3514

1.51

0.13354 (ns)

Host 1 = Acacia mearnsii, Host 2 = Pinus radiata, Host 3 = Eucalyptus obliqua.

Table 5. Analysis of covariance for dry weight of food consumed by sixth instar larvae of C. auctaria on three plant hosts with the estimated initial dry weight of larvae as a covariate Source

df

MS (·1000)

F

p

Host Est. initial dry weight larvae Host*est. initial dry weight larvae

2 1

0.4339 6.5373

0.26 7.75

0.7741 (ns) 0.0071 (sig)

2

0.8028

0.48

0.6236 (ns)

Table 6. Estimated parameters from the linear model used for the ANCOVA for the dry weight of food consumed on three plant hosts with the estimated initial dry weight of sixth instar larvae of C. auctaria as a covariate with a t-test for the hypothesis that each parameter is equal to zero Term

Host [1–3] Host [2–3] Est. initial dry weight larvae Host [1–3]*est. initial dry weight larvae Host [2–3]*est. initial dry weight larvae

Estimate of Standard t ratio p (adjusted slope error of significance) estimate 0.0096 0.0288 2.7803

0.0405 0.0418 0.9989

0.24 0.69 2.78

0.8134 (ns) 0.4934 (ns) 0.0071 (sig)

1.1502

1.2531

0.92

0.3621 (ns)

0.6341

1.3073

0.49

0.6293 (ns)

Host 1 = Acacia mearnsii, Host 2 = Pinus radiata, Host 3 = Eucalyptus obliqua.

Table 7. Analysis of covariance for wet weight of food consumed by sixth instar larvae of C. auctaria on three plant hosts with the initial wet weight of larvae as a covariate

Figure 1. Dry weight growth of last instar larvae of C. auctaria in response to dry weight of plant material consumed for three host plants.

was a significant effect of estimated initial dry weight of the larvae on the dry weight of food consumed, with larger larvae consuming more dry weight of food (Figure 1, Tables 5 and 6). When

Source

df

MS (·1000)

F

p (all ns)

Host Initial wet weight larvae Host*Initial wet weight larvae

2 1

7.6957 17.9771

0.30 1.39

0.7435 0.2424

2

2.7720

0.11

0.8985

the wet weight of food consumed was tested for the effect of host with initial wet weight of larvae as a covariate there was no significant effect of initial wet weight or host (Tables 7 and 8). An ANCOVA testing for the effect of host on the dry weight growth of sixth instar larvae of C. auctaria with dry weight consumption as a covariate was significant for the effect of dry weight

271 Table 8. Estimated parameters from the linear model used for the ANCOVA for the wet weight of food consumed on three plant hosts with the initial wet weight of sixth instar larvae of C. auctaria as a covariate with a t-test for the hypothesis that each parameter is equal to zero Term

Estimate Standard t ratio p (adjusted of slope error of significance) estimate

Host [1–3] Host [2–3] Initial wet weight larvae Host [1–3]*initial wet weight larvae Host [2–3]*initial wet weight larvae

0.1157 0.0633 0.8784

0.1665 0.1770 0.7449

0.69 0.36 1.18

0.4894 (ns) 0.7217 (ns) 0.2424 (ns)

0.4241

0.9164

0.46

0.6450 (ns)

0.0527

0.9801

0.05

0.9572 (ns)

Host 1 = Acacia mearnsii, Host 2 = Pinus radiata, Host 3 = Eucalyptus obliqua. Table 9. Analysis of covariance for dry weight of growth for sixth instar larvae of C. auctaria on three plant hosts with the dry weight of food consumed as a covariate Source

Df

Host Dry weight of food consumed Host*dry weight of food consumed

MS (·1000)

F

2 1

0.3112 2.8142

5.29 47.83

0.0075 (sig) 0.0000 (sig)

2

0.4402

7.48

0.0012 (sig)

p

Table 10. Estimated parameters from the linear model used for the ANCOVA for the dry weight of growth for sixth instar larvae of C. auctaria on three plant hosts with the dry weight of food consumed as a covariate with a t-test for the hypothesis that each parameter is equal to zero Term

Host [1–3] Host [2–3] Dry weight of food consumed Host [1–3]*dry weight of food consumed Host [2–3]*dry weight of food consumed

Estimate of slope

Standard error of estimate

P. radiata has a negative effect on dry weight growth of the larvae independent of dry weight consumption; larvae on P. radiata lose dry weight in respect to the amount of dry weight consumed. When this test was repeated with wet weights the wet weight of food consumed has a significant effect on the wet weight growth of larvae, but there was no effect of host on the wet weight growth (Tables 11 and 12). Figure 1 shows dry weight growth of larvae plotted against the dry weight of food consumed with straight-line fits for each host. The plots for A. mearnsii and E. obliqua show similar slopes and ranges, with dry weight growth increasing as more dry weight of food is consumed. Dry weight growth of larvae on P. radiata does not increase with the dry weight of food consumed (ANCOVA); there is a significant negative effect of P. radiata on the growth of larvae. The range of consumption values for P. radiata is less than half of that for the other plant hosts.

t ratio

p (adjusted significance)

0.0016 0.0171 0.1746

0.0048 0.0070 0.0253

0.22 2.39 6.92

0.8235 (ns) 0.0196 (ns) 0.0000 (sig)

0.0266

0.0326

0.82

0.4167 (ns)

0.1397

0.0428

3.27

0.0017 (sig)

Host 1 = Acacia mearnsii, Host 2 = Pinus radiata, Host 3 = Eucalyptus obliqua.

consumption and for the interaction between host and dry weight consumption (Tables 9 and 10). The parameter estimates (Table 10) indicate that

Table 11. Analysis of covariance for wet weight of growth for sixth instar larvae of C. auctaria on three plant hosts with the wet weight of food consumed as a covariate Source

df

MS (·1000)

F

p

Host Wet weight of food consumed Host*wet weight of food consumed

2 1

1.9921 135.5044

1.83 124.68

0.1678 (ns) 0.0000 (sig)

2

0.2336

0.21

0.8072 (ns)

Table 12. Estimated parameters from the linear model used for the ANCOVA for wet weight growth of sixth instar larvae of C. auctaria on the three plant hosts with wet weight of food consumed as a covariate with a t-test for the hypothesis that each parameter is equal to zero Term

Host [1–3] Host [2–3] Wet weight of food consumed Host [1–3]*wet weight of food consumed Host [2–3]*wet weight of food consumed

Estimate of slope

Standard error of estimate

t ratio

p (adjusted significance)

0.0474 0.0549 0.3045

0.0273 0.0324 0.0273

1.73 1.69 11.17

0.0873 (ns) 0.0952 (ns) 0.0000 (sig)

0.0228

0.0352

0.65

0.5195 (ns)

0.0224

0.0449

0.50

0.6199 (ns)

Host 1 = Acacia mearnsii, Host 2 = Pinus radiata, Host 3 = Eucalyptus obliqua.

272

Figure 2. Wet weight growth of last instar larvae of C. auctaria in response to wet weight of plant material consumed for three host plants.

Figure 2 shows wet weight growth plotted against the wet weight of food consumed with straight-line fits for each host. The ANCOVA indicated a significant effect of wet weight consumption; as wet weight consumed increases there is a corresponding increase in wet weight growth. The wet weight of food consumed on P. radiata produces similar changes in the wet weight growth of larvae compared to A. mearnsii and E. obliqua, but the range of values are slightly lower than those for A. mearnsii and E. obliqua. This analysis indicates that in feeding trials C. auctaria consumes a similar wet weight of P. radiata at a similar rate to the other two hosts, but when dry weights are analysed, the larvae are consuming less nutrients over the experimental period with a corresponding lack of dry weight increase.

Figure 3. Comparison of the mean percentage water contents of the three plant hosts and the larvae reared on those host plants.

post-trial larvae were significantly different to each other (Tukey–Kramer HSD, q* = 2.39, p < 0.05). Figure 3 is a graph of mean percentage water contents for host plants and post-trial larvae. The patterns seen for mean percentage water contents of host plants were mirrored by the mean percentage water content of larvae.

Water content of host plants and larvae Discussion Mean percentage water contents of host plants ranged from a mean of 60.73 ± 0.58% (A. mearnsii) to 74.99 ± 0.57% (P. radiata). Percentage water content of larvae after the four day feeding trial ranged from 89.51 to 78.73% with an overall mean of 82.43 ± 0.24%. Mean percentage water contents for host plants were significantly different to each other (Tukey–Kramer HSD, q* = 1.98, p < 0.05). Mean percentage water contents of post-trial larvae for each host plant ranged from 80.87 ± 0.21% (A. mearnsii) to 84.79 ± 0.25% (P. radiata). Mean percentage water contents for

Oviposition trials with the two species of Chlenias indicated that, under caged conditions, female moths do not seem to show strong preferences for particular host plants. The height and shape of oviposition sites seemed to elicit a definite response, with the higher parts of plants and the tops of cages preferred. Flat smooth surfaces were avoided, with needles or bipinnate foliage of P. radiata, A. mearnsii and A. verticillata preferred. Nutritional trials using last instar larvae show that P. radiata is a substandard host compared to

273 native food plants such as E. obliqua and A. mearnsii. Despite the frequent records of Chlenias spp. feeding on P. radiata, this host may have a negative effect on populations of these moths by leading to lower fecundity of adult females and increased mortality of late instar larvae and pupae compared to those feeding on native hosts. Why is P. radiata, a substandard host, still utilised in the field by C. auctaria? Oviposition in these moths may be driven more by flight activity and proximity of suitable plant surfaces than by active choice of a particular host species. Females mate immediately after emergence from the pupa (Madden and Bashford 1977a), and lay soon afterwards. Females are unlikely to disperse very far from their emergence site, and may lay eggs on any suitable substrate as determined by height and surface texture. The active dispersal of first instar larvae observed in the laboratory may be a major method of dispersal for this moth species, one which is almost as critical as the distance that the female moth flies from her emergence site. First instar larvae then commence feeding on whichever plant they are fortuitously on after the 1–2 day dispersal period. Polyphagous moths using this method of dispersal may decrease risks from predators and parasites by increasing their chances of locating ‘enemy-free’ spaces in the habitat but, as these trials indicate, they may increase mortality and decrease fecundity by increased proportions of resident populations being ‘forced’ to develop on substandard hosts. The potential impact of exotic pine plantations on native moth communities in Australia may thus be more subtle than simple habitat displacement. Species lists for moths on Australian pine plantations (Britton 1994) initially suggest that many species survive and breed in plantations, but the species lists for Australia and other southern hemisphere countries may only contain those species which use dispersal strategies which do not rely on females choosing the most appropriate host plants for oviposition. These species may also be part of a longer term decline rather than simply an increasing list of species which are hostswitching onto Pinus spp. Further detailed collections of Lepidoptera on P. radiata in Australia and in other locations are required to determine whether this impact is occurring. The pool of species available for such investigation is considerable, although many are con-

siderably scarcer than Chlenias. Moths recorded feeding on P. radiata by 1993 (Britton 1994) include representatives of 21 families, three of which (Argyresthidae, Gelechiidae, Sesiidae) feed only on Northern American tree populations. Although possibly a sampling artefact rather than a relatively complete list, the 70 species then recorded from Australian P. radiata considerably exceed the diversity reported from other countries (New Zealand 31, sub-Saharan Africa 26, South America 13). Allowing for taxonomic uncertainty such as that manifest in Chlenias, about 15 Australian Geometridae have adopted pines as hosts. Seven species of the largely endemic family Anthelidae have also done so, as a more isolated evolutionary lineage. However, it is important to emphasise that the apparent security suggested by the relatively high moth richness and abundance on such exotic hosts does not mask the loss of numerous other species from such habitats which have resulted from removal of native vegetation over large areas of the country.

Acknowledgements The work presented here represents part of the M.Sc. thesis of D.R.B. with Dr T.R. New as the primary supervisor. D.R.B. would like to acknowledge the support of the Department, in particular that provided by the late Professor I.W.B. Thornton as co-supervisor and Head of Department. Analysis of the consumption and assimilation data was greatly improved by discussions with Dr D.F. Hochuli, Dept. of Biological Sciences, University of Sydney. Assistance in identifying Chlenias species was provided by Dr P.B. McQuillan, Dept. of Geography and Environmental Studies, University of Tasmania and Mr E.L. Martin, Dept. of Mines and Primary Industries, Hobart, Tasmania.

References Britton D.R. 1994. Nutritional Ecology of Chlenias spp. on an Introduced Conifer Pinus radiata D. Don: The Effects of Choosing a Non-Native Host Plant in Lepidoptera. M.Sc. Thesis, La Trobe University, Bundoora, Victoria, Australia. Center for International Trade in Forest Products 1993. Radiata Pine: A Competitive Force in Asian Markets. Fact Sheet No. 7, April 1993.

274 CSIRO 1996. Report of the Division of Forestry 1992–1995. CSIRO Forestry and Forest Products, Yarralumla, ACT, Australia. Horton D.R. and Redak R.A. 1993. Further comments on analysis of covariance in insect dietary studies. Entomol. Exp. Appl. 69: 263–275. JMPR 1989. Statistical Visualization for the Macintosh. Version 2.0. McFarland N. 1973. Some observations on the eggs of moths and certain aspects of first instar larval behaviour. J. Res. Lepid. 12: 199–208. McQuillan P.B. and Edwards E.D. 1996. Geometridae. In: Nielsen E.S., Edwards E.D. and Rangsi T.V. (eds), Checklist of the Lepidoptera of Australia. Monographs on Australian Lepidoptera. CSIRO Publishing, Melbourne, Australia, pp. 200–228. Madden J.L. and Bashford R. 1977a. The life history and behaviour of Chlenias sp., a geometrid defoliator of Radiata Pine in Tasmania. J. Aust. Entomol. Soc. 16: 371–378. Madden J.L. and Bashford R. 1977b. Population biology of Chlenias sp., a geometrid defoliator of Pinus radiata in Tasmania. J. Aust. Entomol. Soc. 16: 379–388. Minko G. 1961. The insects of Pinus radiata plantations in NorthEastern Victoria. Bull. Forests Commn. Victoria 13: 1–14. Minko G. 1962. Insects observed attacking Pinus radiata D.Don in nurseries. Forest Tech. Pap. Forests Commn. Victoria 10: 34–47. Minko G. 1965. Injurious insects affecting Pinus radiata (D. Don) in Ovens Plantation. unpublished thesis for Diploma of Forestry (Victoria). Neumann F.G. 1976. Biology, epidemiology and control of a species of Lichenaula Meyrick (Lepidoptera: Xyloryctidae) in Radiata Pine. Vic. Forest Tech. Pap. 25: 15–25. Ohmart C.P. 1980. Insect pests of Pinus radiata plantations: present and possible future problems. Aust. Forest 43: 226–232. Ohmart C.P. 1981. An annotated list of insects associated with Pinus radiata D.Don in California. Divl Rep. Div. Forest Res, CSIRO.

Ohmart C.P. 1982. Destructive insects of native and planted Pinus radiata in California, and their relevance to Australian forestry. Aust. Forest Res. 12: 151–161. Ohmart C.P. and Voigt W.G. 1981. Arthropod communities in the crowns of the natural and planted stands of Pinus radiata (Monterey Pine) in California. Can. Entomol. 113: 673– 684. Raubenheimer D. and Simpson S.J. 1992. Analysis of covariance: an alternative to nutritional indices. Entomol. Exp. Appl. 62: 221–231. Rawlings G.B. 1953. Insects of Pinus radiata forests in New Zealand. Forest Res. Notes Wellington 1(8): 1–19. Rawlings G.B. 1960. Fungi and Insects associated with Pinus radiata. F.A.O. For. Forest Prod. Stud. 14: 250–320. Resource Assessment Commission. 1992. A survey of Australia’s forest resource, Forest and Timber Inquiry. Australia. Rice W.R. 1989. Analysing tables of statistical tests. Evolution 43: 223–225. Scriber J.M. and Slansky F. Jr., 1981. The nutritional ecology of immature insects. Annu Rev. Entomol. 26: 183–211. Slansky F.Jr., and Scriber J.M. 1985. Food consumption and utilization. In: Kerkut G.A. and Gilbert L.I. (eds), Comprehensive Insect Physiology and Pharmacology. Digestion, Nutrition, Excretion. Pergamon Press, Oxford, England, pp. 87–163. Spiller D.M. and Wise K.A. 1982. A catalogue (1860–1960) of New Zealand insects and their host plants. In: Dale P.S. and Maddison P.A. (eds), DSIR Bulletin. New Zealand Department of Scientific and Industrial Research. Sutton W.R.J. 1984. New Zealand Experience With Radiata Pine. Macmillan HR Lectureship in Forestry, Vancouver, British Columbia, Canada. Tindale N.B. 1929. Species of Chlenias attacking pines (Lepidoptera, Family Boarmiidae). Rec. S. Aust. Mus. 4: 43–48. Waldbauer G.P. 1968. The consumption and utilization of food by insects. Adv. Insect Physiol. 5: 229–288.