Forest health surveillance in New Zealand was initiated in 1956 after insect attack triggered concern about the susceptibility of exotic forests. Surveillance techniques and scientific support developed as the types and areas of greatest risks were identified from operation reviews. The detection of newly established forest insects and fungi has become the top priority and surveillance has extended to port environs. A Forest Disease Contingency Plan formalises actions when new introduced insects and fungi are detected, and a computerised forest health database has been established for records going back 25 years. A "user-pay" system to cover the cost of surveillance and scientific back-up was intitiated in 1987. A Forest Health Advisory Committee involves industry in planning and advising the Government on matters pertaining to forest health.

Regular surveys of port environs and forest areas are justified to detect new introductions of harmful insects or fungi. Early detection allows timely eradication or control action, so minimising losses of forest value. Historical records show an average 4.6 new introductions each year, and timely response to all of these will yield a maximum national benefit, excluding costs of detection, of $8.95 million per annum. Effective forest health survey methods, in order of decreasing cost efficiency, include aerial survey, drive-through roadside survey, and random point sampling. Used in combination, at a range of regional survey intensities appropriate to risk and cost, they yield a maximum national net benefit of $7.33 million at survey levels which give 95% detection of new introductions. This analysis justifies increasing current levels of survey towards 95% detection, while refining the methods and assumptions used so as to increase confidence in defining the point of maximum net benefit.

Since 1973, poplars in New Zealand have been attacked by five new pathogens: Melampsora larici-populina Kleb. (1973), Melampsora medusae Thum. (1973), Marssonina brunnea (Ell. & Ev.) Magn. (1976), Marssonina castagnei (Desm. & Mont.) Magn. (1985), and Xanthomonas campestris pv. populi de Kam (1985). The Melampsora rusts and Marssonina species have defoliated susceptible cultivars and killed mature trees whereas X. campestris pv. populi has caused little damage. During the same period tree and shrub willows have been defoliated by two newly introduced rusts, Melampsora coleosporioides Dietel (1978) and Melampsora sp. (1985) respectively. With the exception of Marssonina brunnea which was introduced into New Zealand on poplar seed, and X. campestris pv. populi whose time and mode of entry are unknown, the pathogens all entered New Zealand via trans-Tasman wind currents from Australia. The devastating effects of poplar rust and Marssonina castagnei on poplar plantings clearly demonstrate the dangers posed by extensive monoclonal propagation. To avoid future disease epidemics the genetic base of poplars and willows has been broadened considerably and the planting of genetically diverse clonal mixtures undertaken.

The role of clothing and baggage of visitors arriving from Australia at Wellington International Airport in introducing plant pathogenic fungi was evaluated. In the 1980 survey pathogenic fungi comprised 17% of the total spores collected and included Ustilago spp., Puccinia spp., and Drechslera spp. The most frequently encountered spore types in the 1982 survey were Alternaria and Pithomyces (64% of samples), rust urediniospores (57%), Drechslera and Epicoccum (51%), Cladosporium (49%), and smut teliospores (45%). Approximately 10% of spore types were viable. Passengers originating from farms carried a significantly greater number of spore types and more rust urediniospores than those from other areas. The likelihood of air passengers introducing new diseases into New Zealand is considered to be high.

Evaluating the economic impact of an introduced pest is most useful if it can be carried out before the pest arrives, so that contingency plans for eradicating it can be developed. This has been done by MAFQual (the quality control section of the Ministry of Agriculture and Fisheries) for critical quarantine pests such as Tephritidae (fruit flies). Critical quarantine pests are defined as those which would cause immediate loss of valuable export market access. For quarantine pests (i.e., those serious plant pests which do not occur in New Zealand, but which would not necessarily cause loss of export market access) it is difficult to develop contingency plans in advance. For a quarantine pest, therefore, the decision on whether to attempt to eradicate, contain, or accept the pest must be made after establishment. A cost-benefit analysis must be undertaken comparing all of the costs of an eradication with the benefits of avoiding the damages caused by the introduced pest. Provided the benefits outweigh the costs, eradication can be accepted as a viable option. These procedures were followed before removal of the mite Bryobia lagodechiana Reck from a rose garden in Auckland.

Intensive import/export quarantine operations were commenced in 1948 by the Forest Service with costs being met by the taxpayers. The first Regulations came in 1956 and contained a simple schedule of fees. These were extended in 1966 but still covered only sawn timber, box snooks, and timber in round form. In 1986 the Government decided that, under its "user-pays" policy, net funding for forestry services (including quarantine) would be reduced to zero over 5 years. New fee schedules were introduced in 1987 by the Ministry of Forestry. There was client resistance to paying for regulatory quarantine inspections. Some "national good" content was recognised by the Government. The impact of user-pays may initially have had some detrimental effect on import inspection effectiveness.

Reviews have recently been conducted independently by Australia and New Zealand into various aspects of their plant quarantine policy and practice. Quarantine policy trends in the two countries are similar, being shaped largely by the same pressures and demands, both external (e.g., international moves towards trade liberalisation) and internal (e.g., financial stringencies). A key issue that has emerged from these reviews, and the most controversial, is the assessment of "acceptable risk" by means of bio-economic analysis. Such an approach must be extremely conservative when considering the interests of industries such as forestry because of the disproportionate impact the introduction of exotic pests and pathogens may have on forests and forest products, the special difficulties of early detection and control of such organisms, and the paucity of data on which to base risk assessment. Other issues of relevance to forestry are the establishment of databases on pests and diseases, pre-clearance, area freedom, privatisation of quarantine premises, "multiskilling" for border inspections, community consultation in policy formulation, and the promotion of public awareness of quarantine.

A short account of the complex of factors involved in rural dieback which has been reported from all states, with the best example being in the New England tablelands of northern New South Wales.

Identification of fruit-bodies of wood-decay fungi (mainly corticioid and polypore fungi) uses both macroscopic and microscopic characters, and the type of associated wood rot. Examination of the hyphal system of the fruit-body is particularly important. Identification of fungal cultures derived from decayed wood, in the absence of fruit-bodies of the fungus, is hindered by a lack of literature for Australasian species. Use of the Buller phenomenon may assist cultural identification. Herbaria and culture collections of wood-decay fungi are key resources for identification.

Indigenous forests in Australia are managed primarily for timber production, water production, and conservation. Soil infestation by Phytophthora spp. has the potential to affect all of these. Infestation is widespread but is perceived to affect timber production only in the mixed species eucalypt forests in Victoria and the Eucalyptus marginata Donn ex Sm. (jarrah) forest in Western Australia. It is known to affect water production in Western Australia. As Phytophthora spp. have a very wide host range, site infestation leads to vegetation changes which reduce the conservation value of indigenous forests and similar plant communities in southern Australia.

Catastrophic tree mortality is a natural process in beech (Nothofagus spp.) forest ecology. Managers must accept that periodically large areas of forest will die, accompanied by outbreaks of insects and disease. In accepting the inevitability of these changes, managers also need to recognise an educational responsibility to forest users in offering an explanation of the processes occurring and their probable consequences. It is well known that small disturbances in beech forest usually initiate change over much greater areas, especially in mature or overmature stands. Forest managers can take action to reduce the impact of new tracks, camp sites, and picnic areas and retain healthy forest cover. Pinhole borers (Platypus spp.) are the primary agents in extending damage, and action to reduce populations and prevent their build-up can be very effective. Cleared trees and logs should be cut into short lengths to dry and become unsuitable for attack. Residual trees should not be stressed by root or stem damage or have spoil heaped on their roots. Trees that subsequently do become attacked should be felled and also cut into short lengths before brood emerge, i.e., within 2 years of attack. Care in planning tracks and camp sites and in the treatment of adjacent trees can go a long way towards preserving the essential character of the area. Conversely, the development of high populations of Platypus spp. can lead to prolonged and extensive stand decline.

Economically, regrowth dieback is the most serious of Tasmania's eucalypt crown dieback diseases. Uncertainty as to the cause of this disease constrains the management of dieback-affected forests. Intensive surveys after droughts may be warranted to check the area and severity of dieback. Decisions could then be made about the need to adjust yield predictions or harvesting schedules. Silvicultural treatments that increase the rate of wood production could be used to help reduce future crop losses due to the disease.

Chalara australis Walker & Kile is a lethal primary pathogen of Nothofagus cunninghamii (Hook.) Oerst. (myrtle) in Tasmanian rainforests. It is a major cause of gaps in myrtledominated stands but mortality is often most severe in rainforest subject to human disturbance. Possible strategies could be developed to minimise disease in areas of disturbed rainforest.

The prime objective of forest health surveillance, as defined in this paper, is the early detection of introduced insects and pathogens in order to minimise damage, enhance the possibility of eradication, and limit the cost of control. Since the introduction of systematic surveillance in New Zealand the emphasis has been on protecting plantation forests and little has been done in indigenous forests. The Department of Conservation has not yet taken steps to address this apparent imbalance. The cost of a comprehensive surveillance programme for New Zealand's indigenous forests is estimated at $4.5 million, compared with current costs of $2.4 million for control of fire, $2.3 million for weeds, and $5.8 million for wild animals. There are differences between plantation and protected indigenous forests which might influence susceptibility to, or seriousness of, introduced insects or pathogens, and various factors will need to be considered in developing a forest health surveillance strategy.

Three types of factors influence the dieback of forest stands — factors that predispose stands, trigger factors that initiate dieback, and factors that contribute to further decline. All known examples of dieback in New Zealand Nothofagus spp., Metrosideros spp., and beech/hardwood forests can be explained using this three-factor framework.

Worldwide we could be said to be in the middle of a "dieback decade". The term is relatively new but evidence is accumulating which indicates that periodic recurrence of many dieback epidemics, with intervals of decades or centuries, is not uncommon. Canopy dieback, starting with gradual defoliation and twig death, occurs over a wide spectrum of forest types in indigenous forests of the Pacific region. Stress in some form or other is commonly recognised as a contributory factor, e.g., drought, excess water, insect infestation, competition, and combinations of these. Gross divergence from "normal extremes" is often the critical feature and the decline set in motion may be irreversible, even if deterioration takes several years. Primary foliage pathogens appear to be relatively innocuous in our zone. Primary insect infestation of leaves and twigs does cause concern although manifestation may be irregular. Analogous, but more devastating, is browsing by possums. Amongst other effects, possums have changed the dominance pattern of rata-kamahi in several regions of New Zealand. A group of less spectacular but nevertheless widespread lethal influences includes damping-off and root-rot fungi, browsing mammals, predators of seed, and feral pigs, as transporters of soil-borne pathogens. Between them, these factors can change the whole course of regeneration. Various physical aspects of the environment have direct and indirect impacts on forest health: gross disturbance by storm, roadworks, logging; changes in hydrology of a site (e.g., podocarp dieback); passive introduction of a pathogen (e.g., Phytophthora) increase in inoculum food base (e.g., Armillaria), soil compaction; increased insolation causing a rise in soil temperature with consequent influence on disease severity (Phytophthora).

A short account of the pests and diseases of plantations in Spain (the most important European concentration of P. radiata) likely to be of interest to growers elsewhere in the world.

Of the 36 species of paropsine chrysomelids occurring naturally on eucalypts in Tasmania, five species have been observed attacking the introduced eucalypt, E. nitens (Deane et Maiden) Maiden, which is the main eucalypt species used in commercial plantations in Tasmania. The chrysomelids are Paropsis porosa Erichson on seedlings; Chrysophtharta agricola (Chapuis) on juvenile foliage; and C. bimaculata (Olivier), P. delittlei Selman, and P. charybdis Stal on adult foliage. To date C. bimaculata has been the most significant pest causing extensive and severe defoliation on several occasions to E. nitens plantations in northern Tasmania. A preliminary appraisal has been made of the significance of attack experienced to date for plantation growth rates. With current planting rates of E. nitens at about 3500 ha per annum and likely to increase in the future, integrated control and resistance breeding programmes are needed to keep these pests in check.

Pruning has been adopted in managing eucalypt plantations for sawlog and veneer log production in New Zealand; however, internal decay can gain entry through pruned branch stubs. Longitudinal movement of decay along the pith was observed in 15 destructively sampled Eucalyptus regnans F. Mueller, as was radial movement towards the pith, but no decay was detected outwards into wood laid down after pruning. Eucalypt pruning guidelines with respect to branch diameter at time of pruning were supported. Three core zones are distinguishable in eucalypts — the decay core, the defect core, and the brittle heart core. In E. regnans, brittle heart rather than internal decay or pruned branch stubs will probably determine the ultimate diameter of the central core. The silvicultural trade-offs involved in the timing and frequency of pruning mean that future research will need to focus on occlusion of pruned branch stubs, patterns of spread of internal decay, and a quick and simple means of detecting brittle heart in standing trees.

Both Sirex noctilio Fabricius and Ips grandicollis Eichh. have been involved in serious outbreaks since their respective establishments. Some progress has been made with biological control but success in the long term may depend on attention to key silvicultural aspects of stand management.

The Armillaria root disease caused by A. novae-zelandiae (Stevenson) Herink and A. limonea (Stevenson) Boesewinkel continues to affect plantation forestry in New Zealand. Although the land area being converted from indigenous forest to pine plantation is declining, limited evidence demonstrates that stands already planted on such sites may harbour nonlethal, chronic infection throughout the rotation, with a consequent reduction in growth yield. There are indications that chronic infection may persist in certain second-rotation stands, although the total area of forest infected by Armillaria species is unknown. Effective control and management of this disease must be based on a sound knowledge of its development during the course of the rotation.

Since 1966 copper fungicides have been aerially applied to control Dothistroma needle blight within the Pinus radiata D. Don stands of Kinleith Forest. The total cost of this programme has been $18,400,000 (1988 dollars). Current control costs are substantially less than they were when spraying began because of improvements in application techniques and reduction in spray volume. Spray records for 40 stands (10 300 ha total) showed that the average spray frequency per hectare per rotation was 5.45 (range 2.10-10.30). Yield information was available for 10 of the 40 stands. There was no correlation between expected yield and spraying frequency for these areas. A theoretical Dothistroma-resistant breed of P. radiata was evaluated in terms of spraying cost. Savings of 56% are estimated if this breed was established within Kinleith Forest in place of existing breeds.

The fungal toxin dothistromin has the furobenzofuran moiety in common with aflatoxin Bl, a known mutagen, human hepatotoxin, and potential human carcinogen. Environmental contamination with dothistromin has been assessed in field studies by sampling forest air, water in the catchment area, and clothing and skin of forestry workers. The possible adverse health effects of such contamination have been assessed retrospectively in two independent epidemiological studies. Dothistromin has also been tested for mutagenicity in a wide variety of in vitro bio-assays, most of which have been positive, and for the purpose of a more complete risk assessment, in a mouse in vivo mutagenicity assay. Dothistromin appeared to be just as genotoxic as aflatoxin Bl in the mouse bone marrow/peripheral red blood cell assay. The toxin was also detected in the urine of these animals. This work and all the earlier findings with dothistromin have implications for the health and well-being of forestry and other workers.

Although plantations are not yet a large part of South-east Asian forestry, they are becoming increasingly important as wood supplies from natural forests are depleted. Insect and fungal pathogens can have severely detrimental effects on tree plantations and usually where this has happened foresters have responded by changing tree species or siting. Awareness of the potential dangers of forest pathogens is highest in areas which already have a plantation crop industry, such as rubber or cocoa. -Author

Although plantations are not yet a large part of South-east Asian forestry, they are becoming increasingly important as wood supplies from natural forests are depleted. Insect and fungal pathogens can have severely detrimental effects on tree plantations and usually where this has happened foresters have responded by changing tree species or siting. Awareness of the potential dangers of forest pathogens is highest in areas which already have a plantation crop industry, such as rubber or cocoa.

In reviewing the current mechanisms for maintaining health in plantation forests, this paper suggests that the health of New Zealand's radiata pine (Pinus radiata D. Don) forests has declined over the last 30 years and can be expected to continue to decline. There is a need to monitor forest health, largely for the predictive capability this will provide. It is also necessary to have alternative replacement species for the time when radiata pine may become uneconomic. It is proposed that in the next decade the attitude to forest health in the industry should be somewhat modified from its present "specialist' status, and instead forest health should be regarded as a normal economic component of silvicultural decision-making. Forest staff should be more involved in health maintenance and forest health data should be incorporated in decision-making models.

A set of growth and yield models for intensively managed Pinus radiata D. Don plantations in Tasmania have been constructed so that temporary inventory plot data can be used as the starting point for the simulation of stand dynamics. Stand-level increment models have been derived for mean dominant height, basal area, mortality, and volume. A combined parameter prediction and recovery method was used to forecast the parameters of the Weibull distribution which was used as a model of the diameter distribution. A model for the simulation of thinning has also been constructed.

Where possible the models have been cast as generalised linear models and a quasi-likelihood approach was adopted in the modelling of the random component of each model, resulting in an iteratively reweighted least squares procedure for the estimation of model parameters. A Poisson-like variance function was used to model the error variance of forecasts of stand variables - mean dominant height, basal area, and volume. Binomial or binomial-like variance functions were used for the mortality, diameter distribution, and thinning models.

A tree volume model based on Schumacher's form and a compatible, segmented, and variable-form stem taper model were found to be the best models for predicting volume, volume and diameter to a given height, and volume to a top diameter limit for Pinus radiata D. Don sample trees. Schumacher's tree volume model was modified by incorporating DBH to total height ratio (R) and mean annual DBH increment (DBH/age) as extra predictor variables. The variable-form stem taper model incorporated DBH, R, and DBH/age as extra predictor variables to the average-form model. The model predicts increasing butt swell with increasing DBH and increasing upper-stem taper with increasing DBH/age. Compatibility enforced algebraically gave more precise and less biased predictions of volume in the butt section compared to the empirically-compatible version of the taper model. The model suffered from bias when predicting diameter near the tip but bias was not a problem in volume prediction. A random coefficients regression procedure was used to generalise the taper model from an average-form to a variable-form model.

Data on 357 logs from 25 standardised timber grade studies (including Pinus radiata D. Don, Cupressus macrocarpa Hartweg, and Cupressus lusitanica Miller) were used to derive methods of classifying pruned sawlog samples and rating sawmill recoveries in pruned log conversion. Results from SEESAW simulations were used to set upper benchmarks for total conversion and recovery in clears grades. Six pruned log classes, ranging from "unpruned" through to excellent, were established. Sample means from the 25 studies were represented in five of these classes demonstrating the wide range of pruned log quality currently available in New Zealand. However, over half the sample means were grouped within the third lowest class and a further quarter were in the class below that, which confirms well-pruned logs are still scarce.

Large differences in mill conversions and clears grades recovery were also identified. This was demonstrated by comparing results from five sawmills with the simulation-established benchmarks. On similar log types, clears grades recovery from the best-performed mill came within 4.5 percentage points of the benchmark while the worst mill was down 18.4 percentage points. Such diversity in results demonstrated why the quality of pruned logs should be assessed on their potential rather than on grade recovery by any particular mill.

Methods presented for evaluating pruned sawlogs are based on accurate and detailed individual log measurements. The measurements provide data both for the calculation of Pruned Log Index (PLI), which is an absolute measure of pruned log quality, and for sawing simulations. The prime objective in pruned sawlog evaluations is to determine accurately the potential of logs to produce clears grade timber. Sawing simulations, using sample log data, provided better estimates of pruned sawlog potential than could be derived from a timber grade study in a real sawmill. The SEESAW simulator was used to design a standardised sawpattern (STDSP) which maximised recovery in clears grades. SEESAW was then used on a range of logs from the database to provide optimum STD SP results on clears grades recovery. Upper benchmarks for STD SP were set by deriving relationships between PLI and those simulated sawing results.

Twenty pruned Pinus radiata D. Don logs from Waratah in the central North Island were sawn to STD SP at the Timber Industry Training Centre sawmill. Sawing results and data from those logs were used in validations of SEESAW and the upper benchmarks. The Waratah logs were also used to provide an example of how a straightforward evaluation of pruned sawlog samples should be carried out. This included a demonstration of how the benchmark relationships established between PLI and SEESAW results can substitute for the continual use of the simulator and reduce the time taken for data analyses from days to minutes.

Data from 89 real pruned Pinus radiata D. Don logs, representing most of the pruned log types currently being traded in New Zealand, were used in a comparison of sawing systems. Each log was "sawn" three times in the SEESAW simulator to provide results from No Taper Cant, Half-taper Cant, and Half-taper Live sawing.

Total conversion to sawn timber and volumes of clearwood recovered were very similar under all systems but the form in which this timber was produced differed markedly. Half-taper Cant sawing proved the best system for most purposes because it produced the fewest pieces and the most desirable distribution of widths. No Taper Cant sawing gave the second best result and is recommended as a good system for mills which do not have a half-taper sawing capacity. Half-taper Live sawing gave the poorest result with a 43% increase in the number of pieces produced and 47% of clears grades recovered in narrows less than 100 mm wide.

A Pruned Log Index (PLI), derived from measurements of log size and shape and the size of the defect core, has been developed as an absolute measure of pruned sawlog quality. It is based entirely on measurable log variables routinely acquired under either of two established pruned log analysis systems. Being dependent on accurate and detailed measurements, PLI is of higher precision than other existing forms of pruned log classification. Consequently its application should be limited to real logs only (other less precise measures are more appropriate for predicting future log quality).

The index was developed and validated with the assistance of the pruned log sawing simulator SEESAW. Simulation was used in preference to the real sawing results available in order to avoid the confounding influences of sawmill variables. This simulation approach also allowed the performance of trial formulae to be examined under three distinct sawing systems. Initial derivation of formulae was intuitive, based on previous experience, and all formulae included terms for log size, log shape, and defect core size. Trial indices were regressed against conversions to clears grades from SEESAW simulations using two sets of real log data from the SEESAW "library". Set 1 consisted of 74 pruned Pinus radiata D. Don logs, originally acquired through sawing studies, and was used to explore and evaluate trial formulae. Set 2 consisted of 15 logs, originally acquired through log cross-sectional analysis, and was used for validation.

Eight years after planting, root systems were compared for 30 Pinus radiata D. Eton trees excavated from each of five different ripping treatments and an unripped control. Taproots of trees from deep-ripped (100 cm) treatments penetrated to a far greater depth (153 cm) than those from the shallow-ripped treatment (60 cm). Taproots of trees planted on the unripped site penetrated to a maximum depth of 48 cm because of the presence of a hard pan.

A single line rip (deep or shallow) resulted in some orientation of primary lateral roots. This orientation became more pronounced in the secondary lateral roots, which were confined along the line of the rip. Although lateral roots were more evenly distributed in the unripped and cross-ripped treatments, least development was in the north-east quadrant and greatest development in the south-west.

The number of butt-swept trees was not significantly affected by treatment, but ths severity of stem deviation and sinuosity resulting from initial butt sweep was related to treatment. The straighter the taproot and the deeper its penetration, the less stem deviation.

It is suggested that all sites with a shallow hard pan be deep-ripped, using winged rippers, to maximise soil disturbance, and that forests on the Canterbury Plains have rows orientated south-west to north-east.

Growth of Pinus radiata D. Don trees growing on and just off four major skidtrails in Tairua Forest was compared. Those growing on the skidtrails were markedly inferior in height, diameter, and form. This significantly affected the selection of trees for thinning and pruning so that, by age 7.5 years, only 4% of the trees planted on the skidtrails remained to form the final crop compared with 15% of those planted off the trails. Trees left on the trails were 2 cm smaller in diameter and 1 m shorter than those left off the trail. These results suggest that, for soils similar to the clays of Tairua, consideration should be given to either leaving major skidtrails implanted, or carrying out some sort of site amelioration.

Dothistroma needle blight caused by Dothistroma pini Hulbary was assessed on Pinus radiata D. Don in New Zealand in nine progeny trials ranging in age from 2 to 10 years. Resistant families could be identified from all assessments and rankings were consistent over sites and years. Heritabilities were moderately high, specific combining ability (SCA) was very small compared to general combining ability (GCA), and there was no indication of substantial genotype x location interaction. The greatest net gain in volume from resistant progeny will result from placing as much selection emphasis on Dothistroma resistance as on improved growth rate and stem form. Genetic gain expected from seed orchard progeny from performance-tested first-generation parents is calculated from progeny test data to be about 11-12%. The actual reduction in disease with disease-resistant stock may be greater because of the epidemiological effect which occurs when all trees in a stand are resistant.