New host records are given for Navomorpha lineata in Cyathodes fasciculata, C. juniperina, Nothofagus solandri var solandri and N. truncata. The tunnel structure of N. lineata is described for the host C. fasciculata. Four distinct types of tunnelling were observed and are related to different stages of larval development. Tunnelling causes localised swellings and lesions and is sometimes responsible for breaking of limbs.

Keywords: Coleoptera; Cerambycidae; Navomorpha lineata; larva; wood-borer; tunnelling behaviour; host-plants.

Introduction

Woodboring insects of live hosts pose a problem for those wishing to study their habits. Although their presence may be detected by surface manifestations, these may be difficult to detect and/or insufficient for specific identification. The insect itself usually cannot be extracted without considerable damage to the host species and there is the danger of damaging the insect itself beyond recognition. These characteristics have perhaps contributed to the relative paucity of information on habits and host range of New Zealand's live-wood borers.

With the exception of Oemona hirta (Fabricius) (Coleoptera: Cerambycidae), which has a very wide range of exotic host species (Dye 1950, Harrison 1976), New Zealand's live-wood borers are largely confined to indigenous hosts. O. hirta is a pest of primary importance within exotic hosts (DSIR Entomology Division Internal Annual Report 1979) and has been the subject of some biological study (Dye 1950, for exotic hosts). Of the indigenous plants which are hosts for live-wood borers, only Nothofagus spp. have a strong potential for major commercial exploitation (Chavasse 1978) with the result that Platypus spp. as potentially important pests, have become the subject of detailed study (Litchwark 1978, Milligan 1979). The only other live-wood borers to receive attention are Aenetus virescens (Doubleday), (Alma 1977) Navomorpha lineata (Bain, 1976) and Ochrocydus huttoni Pascoe (Hosking, 1978), which also occur in Nothofagus spp. (Milligan 1974) as well as other hosts.

Although N. lineata is a recognised pest there is relatively little information on its biology. This study of tunnelling in C. fasciculata may contribute to a more thorough understanding of the behaviour of this insect.

Study Area and Methods

N. lineata (Fig. 1) attack in C. fasciculata was studied at lake Pounui Reserve, Southern Wairarapa, in a 60-year-old broadleaf stand dominated by Leptospermum ericoides (Biggs 1977), and also at Wright Street Bush, Wainuiomata, in a mixed-age stand of Nothofagus solandri var solandri and N. truncata.

External and internal tunnel characteristics and associated limb malformations were studied for 40 N. lineata tunnels. Fifteen larvae were extracted and their identity first confirmed by Mr J. Bain (Forest Research Institute) and subsequently by the author.

Fig. 1 Navomorpha lineata adult.

Tunnel Characteristics

The eggs of N. lineata are usually deposited under bud scales of the host plant (Duffy 1963, Bain 1976). On C. fasciculata tunnels of young N. lineata were all found adjacent to cicada scars. This suggests that cicada scars are used by N. lineata as oviposition sites on this host. Bain (1976) described the early tunnelling stage as a downward spiral encircling the limb close to the surface but the tunnelling pattern in C. fasciculata does not take the form of a complete downward spiral. The wound caused by page 3 cicada oviposition appears to prevent or discourage complete encirclement of the limb with the result that the larval tunnels turn back on themselves at the cicada scar. The diameter of limbs affected by this spiral tunnel vary from 5mm to 15mm (diameter measured immediately below spiral tunnel). The extent of the spiral tunnel varies from one or two loops to a network ramifying around the limb as far as the cicada scar and longitudinally for up to about 50mm. This variation in tunnelling results in different degrees of limb deformity.

Fig. 2 Spiral tunnel malformation by Navomorpha lineata on Cyathodes fasciculata illustrating asymmetry of deformity and position in relation to the cicada oviposition scar. Bottom stem diameter, 7mm.

C. fasciculata reacts to the spiral tunnel damage by a swelling of the limb above and below (figs 2, 3A) which stretches the overlying bark, producing a lacework pattern of bark distortion (fig. 2). Broken strands of bark often lend a serrated appearance to the edge of the damage (fig. 3A). The spiral tunnel and limb swelling is more extensive opposite the cicada scar (fig. 2). Callus tissue is sometimes produced on the edges of dead or exposed tissue and this growth may seal over the damage.

Later instars construct a longitudinal tunnel (fig. 3B) and frass is ejected through holes opening to the limb surface (Bain 1976). In C. fasciculata this tunnel may penetrate both up and down the limb from the spiral tunnel but it reaches its greatest length below. The usual length of the longitudinal tunnel was not determined in this study but lengths of up to 380mm were recorded for occupied tunnels. The longitudinal tunnel varies in radial depth from just below the surface of a limb to the centre.

Fig. 3 A (left) Spiral tunnel malformation by Navomorpha lineata on Cyathodes fasciculata. Bottom stem diameter 14mm. B (centre) Damage caused by extensive subsurface feeding around stem of Cyathodes fasciculata. Bark overlying the subsurface feeding area has been removed. Longitudinal tunnel below the Navomorpha lineata larva. Stem diameter 11mm. C (right) Nothogafus solandri var solandri stem malformation by Navomorpha lineata. Extruded frass and shredded wood is visible in this specimen. Bottom stem diameter 10mm.

Sometimes during the longitudinal tunnel construction phase the larva will undergo a temporary change in feeding behaviour. Instead of a well defined longitudinal tunnel the larva excavates tissue over a broader area beneath the bark, sometimes encircling the limb (fig. 3B). The bark and immediately underlying tissue is usually undamaged, therefore allowing continued terminal growth. Unlike the longitudinal tunnel, the cavity formed by the subsurface feeding is packed with frass. Construction of the subsurface cavity may occur close to the spiral tunnel and has been page 5

Fig. 4 Schematic diagram of circular gallery of Navomorpha lineata in Nothofagus solandri var solandri. Longitudinal section.

observed where the longitudinal tunnel extends down a branch into a main stem. The limb reacts by swelling but this is distinguishable from the spiral tunnel by the difference in tunnel diameter associated with the underlying cavity (3-5mm compared with 1-2mm of the spiral tunnel).

Prior to pupation the larva may construct a circular gallery around the limb, weakening it sufficiently to break in the wind (Duffy 1963). This circular gallery is constructed by the larva tunnelling in the radial plane, across the limb, leaving only bark connecting the limb above and below (fig. 3B). At the circular gallery the occupied longitudinal tunnel is blocked either by shredded wood or a mixture of shredded wood and frass. The vacated tunnel is blocked by frass. Only one circular gallery was found in C. fasciculata in this study. Two examples of limb breakage were observed where the longitudinal tunnel was slightly spiralled rather than truly axial at the breakage point.

Discussion

Cyathodes fasciculata is described as a shrub species (i.e. characterised by the lack of a distinct trunk) by Allan (1961), and Poole and Adams (1964) and in the study areas C. fasciculata tends to be less than fully erect with a winding or twisted stem. The occurrence of N. lineata in the shrub C. fasciculata is of no surprise since it is known to attack a variety of trees and shrubs (Bain 1976). It is interesting to note that C. juniperina, a shrub of similar proportions to C. fasciculata (Poole and Adams 1964) which occurs at each study site, showed no evidence of attack. However one N. lineata larva was found in C. juniperina near the study site at Lake Pounui—suggesting that this plant may be a very incidental host.

No published accounts of host responses to N. lineata feeding in other plant species are available for comparison. However during the present study some examples of the tunnelling of N. lineata in Nothofagus solandri var solandri and N. truncata were examined. In these two hosts there is stem or branch swelling (fig. 3C) and varying amounts of localised dead tissue at the site of the spiral tunnel. The swelling was often difficult to distinguish in size and form from the swelling at a cicada scar where N. lineata was absent.

There appear to be differences between larval behaviour of N. lineata in N. solandri var solandri and C. fasciculata. The circular gallery (fig. 4) is page 6 much more common in N. solandri var solandri than in C. fasciculata. In June 1980 at both study sites it was possible to find branches of N. solandri var solandri which were broken off as a result of the circular gallery without any special effort while only one example of this behaviour was found in C. fasciculata between December 1979 and July 1980. In N. solandri var solandri the circular gallery is sometimes constructed at both ends of the pupal chamber, isolating a length of stem or branch about 70mm long. This behaviour was not observed in C. fasciculata.

Lloyd (1949) observed severe damage to the terminal growth of various species of indigenous trees on the edge of clearings and on the forest margin which he attributed to cicada oviposition in both leading and lateral shoots. Among trees affected were Dacrydium cupressinum, Podocarpus totara and Leptospermum scoparium, all of which are known hosts of N. lineata (J. Bain 1980 pers. comm.). An unidentified cerambycid larva was found in terminal growth (probably those affected by cicada scars but this was not directly stated) of L. scoparium and D. cupressinum. Although damage, including resulting windbreak, was attributed to cicada oviposition it is possible that N. lineata was responsible for much of the damage through weakening caused by the spiral tunnel and perhaps the longitudinal tunnel associated with the cicada scars. N. lineata appears to have a definite preference for cicada scars as oviposition sites in C. fasciculata and branches of Douglas fir with cicada scars are particularly susceptible to N. lineata attack although N. lineata is not considered to be necessarily dependant on damage by other insects in all its hosts (Bain 1976). Some wind breakage of leading shoots at sites of N. lineata tunnelling was observed in C. fasciculata but this was not recorded quantitatively.

In C. fasciculata weakening induced by N. lineata larvae may hasten the death of a limb but I have not studied this aspect. I have observed that branches with larval tunnelling sometimes have noticeably fewer leaves and many of them are dead. Presumably the spiral tunnel inhibits phloem transport, while the longitudinal tunnel is often wide enough to occupy much of the cortex and therefore impede xylem transport.

The live-wood boring habit of N. lineata is shared by other cerambycid beetles such as Gastrosarus nigricollis Bates (Duffy 1963, J. Bain 1979 pers. comm.), Oemona hirta (Fabricius), Mesolamia sp., Ochrocydus huttoni and Brounopsis hudsoni Blair (Dugdale 1975). Gastrosarus nigricollis has been recorded in C. fasciculata (J. Bain pers, comm.) and may have similar effects on this plant. The appearance of infested limbs and tunnels as described in this paper are therefore not regarded as diagnostic of N. lineata. Further knowledge of the habits and effects of other cerambycids in the future may allow specific determination from these features. The only other wood borer found in C. fasciculata in the two study areas is Aenetus virescens (Lepidoptera: Hepialidae). Its larval activities can be distinguished from N. lineata by the presence of a silk-frass web where the larval tunnel opens to the surface of the host. No evidence was found of A. virescens attacking C. juniperina.

Specimens of Xanthocryptus novozealandicus (Dalla Torre) (Hymenoptera: Ichneumonidae) pupae were found in tunnels of N. lineata at Wright Street Bush. X. novozealandicus is a larval parasite of at least nine cerambycid species (Valentine 1967) but has not been reported as a parasite of N. lineata. Vacant tunnels of N. lineata are sometimes occupied by the larvae, pupae and adults of Artystona wakefieldi Bates (Coleoptera: Tenebrionidae).

Acknowledgements

I thank Mr W. J. Winstanley, Dr G. W. Gibbs (Victoria University of Wellington), Mr J. Bain (Forest Research Institute, Rotorua) and Mr J. S. Dugdale (D.S.I.R., Entomology Division) for helpful criticism of the manuscript. Mr Bain and Mr Dugdale also provided valuable discussion during the study.