Tangential shrinkage of earlywood and latewood of Pinus radiata in wood from logs which had a severe propensity for within-ring internal checking on drying was compared with wood from logs which showed no tendency to form internal checks. The samples were collected from nine 23- to 26-year-old trees of each category, with a total of 132 samples being assessed. These trees were chosen from three forests in the central North Island of New Zealand, denoted as Sites A, B, and C. The average tangential shrinkage from green to 12% moisture content was 3.68% for all the samples, with a standard deviation of 0.82%. Between-site variation for average shrinkage was less than 0.27% with an average value of 3.82% for Site B, 3.57% for Site A, and 3.55% for Site C. From the measured data, it was found that earlywood and the latewood shrink differently. However, for the severely checked rings, the tangential shrinkage of the earlywood layer was higher than for the latewood layer, and for the rings without checks the trend was the opposite.

Eggs and larvae of huhu (Prionoplus reticularis White) were treated at elevated temperatures and varying controlled atmosphere (CA) conditions. At 35^oC in air, more than 9.4 days were needed to achieve 99% mortality (LT99) of huhu larvae. Treatment time decreased with increasing temperature to 3 h at 45^oC in air. Eggs and neonate larvae (mean weight: 1.25 mg) were more susceptible to the heat treatments than were larger larvae (≥100 mg). When treating larger huhu with controlled atmosphere conditions of 100% N₂, 100% CO₂, or a 50% N₂ / 50% CO₂ mixture at 20^oC, less than 36% mortality was observed after 11 days' exposure. However, increasing the treatment temperature to 40^oC during 100% N₂, 100% CO₂, and 50% N₂ / 50% CO₂ treatments resulted in LT99 of 8.3, 6.9, and 7.6 h respectively. The 100% CO₂ controlled atmosphere treatment was most effective. There was no statistical difference in mortality responses to the 100% N₂ or 50% N₂ / 50% CO₂ treatment. All controlled atmosphere treatments at 40^oC were significantly better at controlling larger larvae than was the 40^oC air treatment.

The risk of severe dothistroma needle blight in Pinus radiata D. Don forests depends on many factors, including the genetic diversity of the pathogen population, and so molecular methods were used to estimate the genetic diversity of the population of Dothistroma pini Hulbary in New Zealand. Samples of D. pini were collected from infected P. radiata trees using two sampling regimes. One involved collections from three forests, using a hierarchy of populations approach. The other involved collections from a Forest Research Institute field trial for Dothistroma resistance in which each tree was of known parentage. In addition to these, four D. pini samples isolated in New Zealand during the 1960s, and DNA from a Central American strain of the teleomorph (sexual form) Mycosphaerella pini E. Rostrup apud Monk, were analysed. Using the PCR-based techniques of RAPD (Random Amplification of Polymorphic DNA) and RAMS (Random Amplified Microsatellites), no genetic diversity was detected within the New Zealand samples. These results suggest that a single strain of D. pini was introduced into New Zealand and that this spread throughout the country. The New Zealand strain of D. pini is easily distinguishable from the Central American teleomorph, using both RAPD and RAMS techniques. The current study does not specifically address virulence levels, but this work implies that although the current New Zealand strain of D. pini is unlikely to overcome the current level of P. radiata resistance, the introduction of an overseas strain into New Zealand could present a threat to forest health.

The 17 fungi and the new combination described in this paper have not previously been recorded in New Zealand. The fungi described are: Caulicolous Mitosporic fungi, Coelomycetes: Cryptosporiopsis hoheriae comb. nov., Sarcostroma mahinapuense sp. nov.; Caulicolous Mitosporic fungi, Hyphomycetes: Chalara myrsines sp. nov.; Foliicolous Basidiomycota: Uromycladium alpinum; Foliicolous Ascomycota: Acrospermum sp., Cashiella sticheri sp. nov.; Foliicolous Mitosporic fungi, Coelomycetes: Catenophoropsis eucalypticola, Coleophoma cylindrospora, Cryptosporiopsis eucalypti, Discula sp., Monostichella robergei, Sarcostroma arbuti, Seiridium eucalypti, Septoria typica sp. nov., Vermisporium acutum, V. brevicentrum, V. eucalypti; Foliicolous Mitosporic fungi, Hyphomycetes: Harzia acremonioides. The host plants are: Acacia dealbata, A. mearnsii, A. melanoxylon, Corokia cotoneaster, Cupressus lusitanica, Eucalyptus calophylla, E. delegatensis, E. fastigata, E. ficifolia, E. leucoxylon, E. muelleriana, E. nitens, E. regnans, E. saligna, Hoheria populnea, Lagunaria patersonii, Myrsine chathamica, Pinus radiata, Prumnopitys ferruginea, Pseudopanax crassifolius, Pteris tremula, Sticherus cunninghamii.

The 20 fungi described in this paper have not previously been recorded in New Zealand. The fungi described are: Caulicolous Ascomycota: Cryptodiaporthe salicina, Stictis stellata, Xenomeris abietis; Corticolous Ascomycota: Lachnellula hahniana, L. pseudotsugae, L. subtilissima, Lophium mytilinum, Sarea resinea; Foliicolous Ascomycota: Meloderma desmazierii, Mycosphaerella pittospori, Rosenscheldiella pullulans, R. styracis; Caulicolous Mitosporic fungi, Coelomycetes: Melanconium oblongum; Foliicolous Mitosporic fungi, Coelomycetes: Coma circularis, Diploceras dilophosporum, Hainesia lythri, Readeriella mirabilis, Septoria unedonis, Sigmina thujina, Strasseria geniculata. The host plants are: Arbutus unedo, Callistemon sp., Chamaecyparis lawsoniana, Cyathodes fraseri, Eucalyptus fastigata, E. fraxinoides, E. nitens, E. regnans, Juglans nigra, Kunzea ericoides, Larix sp., Olearia traversii, O. virgata, Pinus radiata, P. strobus, Pittosporum umbellatum, Pseudotsuga menziesii, Salix fragilis, Thuja plicata.

A model has been developed that predicts the amount of carbon contained in the stem, crown, roots, forest floor, and understorey of managed Pinus radiata D. Don stands at any age over a ritation. A key concept underlying the C_change model is that, with current knowledge of growth partitioning, mortality, and decay of tree components, stem volume production can be used to predict carbon content of forest biomass components. The advantage of taking this approach is that data input requirements for predicting forest carbon are minimised, given a system for determining stem volume growth and mortality over time. The Stand Growth module of STANDPAK predicts P. radiata stem volume for each of the major forest-growing regions in New Zealand, based on an extensive network of permanent sample plots (PSP). By linking the Growth Partitioning module with Stand Growth, a minimum set of data inputs is required to calibrate C_change to the region. The utility of this approach was tested by running C_change to make predictions of stem volume and carbon at several sites where stand biomass measurements had been made. These sites covered a range in nitrogen fertility, stocking, stand ages, and climate. Across all studies, actual above-ground stand carbon content (i.e., excluding understorey and forest floor) was highly correlated with that predicted by C_change (r²=0.97, n=25, p < 0.01). Assuming that suitable regional growth models are available for predicting stem volume and that growth relationships are constant across regions, these results give confidence in the use of C_change for prediction of carbon on a stand and regional scale in New Zealand.

The locations and dimensions of branches encased within a tree-stem influence the appearance, strength, and ultimately the end-use of timber. A model which predicts branch development through time would enable the New Zealand forest industry to obtain better estimates of log and timber grades at time of harvest from inventory information collected at any time during the rotation. Choosing the appropriate level of detail with which to model branch development for forest management requires a compromise between understanding growth processes, data collection costs, model accuracy and precision, and suitability for use by forest management. The approach chosen was to model the occurrence of branch clusters within annual shoots on Pinus radiata D. Don, and to model the diameter growth of branches at the point of attachment to the tree stem. This approach has several major advantages. It is based on an understanding of branch morphology and basic growth processes. It allows the utilisation of existing models of annual height development, stem taper, and growth which are based on extensive field measurements. We believe that the model structure is cost-efficient in terms of the data requirements. The increased costs of data collection to develop a more detailed model of the full extent of the crown would not be matched by a similar increase in worth to the forest industry. The locations and dimensions of branches encased within a tree-stem influence the appearance, strength, and ultimately the end-use of timber. A model which predicts branch development through time would enable the New Zealand forest industry to obtain better estimates of log and timber grades at time of harvest from inventory information collected at any time during the rotation. Choosing the appropriate level of detail with which to model branch development for forest management requires a compromise between understanding growth processes, data collection costs, model accuracy and precision, and suitability for use by forest management. The approach chosen was to model the occurrence of branch clusters within annual shoots on Pinus radiata D. Don, and to model the diameter growth of branches at the point of attachment to the tree stem. This approach has several major advantages. It is based on an understanding of branch morphology and basic growth processes. It allows the utilization of existing models of annual height development, stem taper, and growth which are based on extensive field measurements. We believe that the model structure is cost-efficient in terms of the data requirements. The increased costs of data collection to develop a more detailed model of the full extent of the crown would not be matched by a similar increase in worth to the forest industry.<

Use of genetic transformation, as the most widely pursued form of 'genetic engineering', is often seen as entailing various biological risks. Prominent among the perceived risks for forest trees is genetic contamination of wild populations. However, this cannot happen with exotic species with no local relatives; moreover, suppression of reproduction is desired on independent grounds. Potentially the most serious, yet largely unrecognised, risks for plantation forest species arise from side-effects of genetic transformation on fitness in the field. While these risks may involve low-probability events ('disasters'), such disasters could be catastrophic, except under extremely short rotations, calling for risk management. The risks that are associated just with gene insertion are in principle readily controlled. More problematic are the risks associated with side effects of alien gene (transgene) action interfering with crucial functions such as disease resistance. Quantitative analysis of such risks was made assuming a range of arbitrary probabilities associated with individual transgenes and interactions among different transgenes within the same recipient genotype, showing how such risks might increase markedly as more transgenes are inserted. More crucial, though, may be the need to avoid over-dependence on insertion of any particular gene, despite the associated risk being of low probability, because of the worst-case potential for dire economic consequences. This approach, however, may be difficult to reconcile with the desideratum of suppressing all sexual reproductive activity. Field testing of transformants will be crucial, as it will eliminate many transformation-related risks. The elimination, however, will be slow and never quite complete, which argues for long periods of applying risk spread. In New Zealand a gap exists in the regulatory provisions for risk management with transformants. Use of genetic transformation, as the most widely pursued form of `genetic engineering', is often seen as entailing various biological risks. Prominent among the perceived risks for forest trees is genetic contamination of wild populations. However, this cannot happen with exotic species with no local relatives; moreover, suppression of reproduction is desired on independent grounds. Potentially the most serious, yet largely unrecognized, risks for plantation forest species arise from side-effects of genetic transformation on fitness in the field. While these risks may involve low-probability events (`disasters'), such disasters could be catastrophic, except under extremely short rotations, calling for risk management. The risks that are associated just with gene insertion are in principle readily controlled. More problematic are the risks associated with side effects of alien gene (transgene) action interfering with crucial functions such as disease resistance. Quantitative analysis of such risks was made assuming a range of arbitrary probabilities associated with individual transgenes and interactions among different transgenes within the same recipient genotype, showing how such risks might increase markedly as more transgenes are inserted. More crucial, though, may be the need to avoid over-dependence on insertion of any particular gene, despite the associated risk being of low probability, because of the worst-case potential for dire economic consequences. This approach, however, may be difficult to reconcile with the desideratum of suppressing all sexual reproductive activity. Field testing of transformants will be crucial, as it will eliminate many transformation-related risks. The elimination, however, will be slow and never quite complete, which argues for long periods of applying risk spread. In New Zealand a gap exists in the regulatory provisions for risk management with transformants.

In Pinus radiata D. Don seed orchards throughout New Zealand, controlled pollination of isolated cones produced very few full seeds in 1991-92. Standard X-ray analysis of the empty seeds showed that the female gametophyte was present but shrivelled. Excising and soaking these gametophytes and attached nucellar material in a 15% sucrose solution under vacuum for 48 hours allowed them to be dissected or sectioned. Pollination had occurred in all seeds; however, development of the pollen tube and embryo formation differed among the seeds. In 15% of the seeds either pollen grains did not germinate or pollen tube development was arrested immediately after germination. In 35% of the seeds pollen tubes penetrated only part way through the nucellar cap. Thus, for 50% of the seeds, factors leading to abortion could have occurred while the cones were still within the isolation bags. In 50% of the seeds the pollen tube penetrated through the nucellar cap. Development occurred through the proembryo stage to late embryo in 44% of the seeds. Abortion in these seeds occurred well after the removal of the isolation bags. It is not clear how the isolation bags could be associated with the collapse of these ovules some 14 months after pollination.

The genus Pinus evolved about 100 million years ago, spreading from centres in eastern North America and western Europe throughout middle latitudes of the supercontinent Laurasia. Many early subsections of Pinus are recorded from fossil remains of this period, but it is not until the early Tertiary, when the genus was fragmented by changing global climates and continental tectonics into latitudinal refugia, that secondary centres of origin appeared. From one of these areas, now Mexico and Central America, the subsection Oocarpae is thought to have arisen. This subsection includes the California closed-cone pines, P. radiata D. Don, P. muricata D. Don, and P. attenuata Lemmon, which evolved 15-25 million years ago as they migrated northward to California from Central America. Pinus radiata appears to have occupied coastal or near-coastal habitats throughout its history in the California region. A synthesis of recent evidence from microfossils in sediment cores, oxygen-isotope ratios in ice and sediment cores, and re-evaluation of available macro-fossils yields an improved characterisation of Quaternary climatic and vegetation changes in the California coast region. Over the last million years, California climates fluctuated quasi-cyclically, at long-(multi-millennial), medium-(century), and short-(decadal to annual) periodicities in patterns similar to those documented elsewhere in the world. Changing plant fossil assemblages in the California coast region reflect major and minor fluctuations, with Quercus and Compositae dominating floristic communities during warmest and driest periods, and Taxodiaceae/Cupressaceae/Taxaceae taxa dominating coldest periods. Pines, notably P. radiata, expanded in abundance and shifted locations along the coast during climates intermediate to these extremes. Such climates occurred during transitional periods of major and minor cycles, including onset and terminations of major glacial stages, extended interstadials, and shorter cool, mesic intervals within warmer climates. This evidence is used to argue against a long-standing scenario that P. radiata was broadly distributed along the coast throughout the Pleistocene (10 ka - 2 Ma), and that the Holocene climatic optimum (warm, dry period 4-8 ka) triggered contraction into the present fragmented, relictual populations. Instead, it is proposed that the species has maintained a metapopulation strategy throughout its history in the California region, growing in distinct coastal populations that were subject to repeated events of colonisation, coalescence, and local extirpation in response to fluctuating climates. Population differences are thus likely to reflect frequent founder and bottleneck effects, and complex introgression. Appropriate conservation and restoration approaches to P. radiata may be revised given this perspective. Rathern than restricting views about native populations to the five extant locations, coastal sites beyond the current distribution might also be accepted as 'neo-native'. Areas where P. radiata has naturalised along the California coast during historic climate periods similar to the present include many sites known from the Quaternary fossil record. Rather than treating P. radiata in these areas as an undesired exotic, these and other locations could be considered appropriate places in which to encourage conservation populations of P. radiata. This would mimic the natural potential of P. radiata under the generally cooler climate of the late Holocene (relative to earlier millennia) to colonise and expand in abundance, which may be inhibited currently by human development.

Trigonometric volume ratio equations, which predict stem volume to any specified top height limit as a percentage of the total stem volume, were developed for 25 native tree species in southern New South Wales and Victoria. When used together with companion total stem volume equations, they allow direct predictions of volume to any height limit. Validation statistics obtained through a resampling approach showed little local bias over any range of relative height and the 95% confidence limits of prediction error were mostly within 10% over the entire range of relative height for all species. These equations will be useful for forest managers when estimating the volumes of multiple logs from the same stem during integrated logging, and in forest inventory where yield estimates by product categories are required

Composite tree taper equations were fitted to data from sectionally measured, New Zealand-grown Eucalyptus pilularis Sm., and to combined data from E. muelleriana Howitt and E. globoidea Blakely. Although the sample size was small, the equation for E. pilularis fitted well. The combined E. muelleriana, E. globoidea equation was less satisfactory and should be used only as an interim solution.

Quarter-sawing helps to reduce drying degrade in Eucalyptus regnans F. Muell., and large sawlogs with a minimum small-end diameter of 40 cm are recommended in order to use this technique. The growth rate of dominants will determine the time taken to grow trees capable of producing these large logs. A trial with initial stockings of 2500, 1111, and 625 stems/ha, thinned between ages 5 and 11.8 years, with two unthinned treatments, was analysed at age 19 years to determine the effect of stocking on growth of dominant trees. Stocking was not the main influence although diameter of dominant trees tended to be greater in treatments with lower initial and final stockings

Allometric relationships between stem variables and leaf area were determined for plantation Eucalyptus nitens (Deane et Maiden) Maiden and competing Acacia dealbata Link in 8-year-old and 4-year-old stands. Both cross-sectional area and sapwood area were considered at crown break and breast height. Whilst all four stem variables could be used to calculate tree leaf area adequately, proximity of the stem measurement to crown break was found to be more important than the choice of sapwood area or cross-sectional area in determining the closeness of the derived allometric relationship. When the relationship between sapwood area at crown break and leaf area was used, a single equation was suitable for predictive purposes across treatments for each species

Species and provenance trials of eucalypts, planted independently, near Kaikohe in Northland and at Clive in coastal Hawke's Bay were assessed at ages 7 and 11 years, respectively. At Kaikohe, 11 provenances of Eucalyptus nitens (Deane et Maiden) Maiden, six of E. saligna Smith, three of E. grandis Hill ex Maiden, two each of E. globulus Labill. and E. maidenii Labill., and one of E. robusta Smith were included; at Clive, 21 species were involved, including E. globulus, E. maidenii, and E. bicostata Labill., and provenances of E. nitens from central Victoria (Vic) and southern New South Wales (NSW). At Kaikohe three replications of 49-tree square plots were planted of each seedlot at 2.0 × 1.9m spacing, and at Clive the trial consisted of eight replications of 18-tree plots at 2 × 1 m spacing. At a third site at Patoka, north-west of Napier, transect comparisons were made of E. nitens (Vic), E. maidenii, and E. bicostata at age 11 years. At Kaikohe (age 7 years), the best growth was achieved by E. nitens (Vic) and E. nitens (NSW) (equal; dbh 200 mm), followed by E. globulus and E. maidenii (equal; dbh 172 mm). However, the E. nitens (Vic) provenances showed a widespread, unidentified disorder in the form of progressive loss of the lower crown and eventual death of the tree. Their "survival" (after a 50% early thinning) averaged half that of the NSW provenances which had healthy crowns. Eucalyptus globulus showed a similar problem to E. nitens (Vic), but E. maidenii retained good crowns and high survival. At Clive (age 11 years), a periodic high water table affected survival and health of some species more than others. Eucalyptus nitens (Vic) grew well (dbh 200 mm), though survival was poorer than that of the NSW provenance and crown health was much poorer, with the same progressive loss of lower crown and death (also windthrow) as was seen at Kaikohe. Eucalyptus globulus had grown a little better than both E. nitens provenances, but showed similar crown death to E. nitens (Vic). Eucalyptus maidenii had a higher mean dbh (219 mm) than all other species and had maintained a high survival. At Patoka, growth, form, and health of E. nitens (Vic) and E. maidenii were excellent, with E. maidenii a little slower-growing. The good health and much higher wood density of E. maidenii than of E. nitens indicate its potential value for pulpwood, e.g., at Clive (from a disc study at age 10 years) weighted whole-tree basic density of E. maidenii was 582 kg/m³ as against 450 kg/m³ for E. nitens.

Monthly tree heights and diameters were measured for sample stems of Eucalyptus saligna Smith over a period of 3 years, and of E. regnans F. Muell. and E.fastigata Deane & Maiden over a 5-year period. The E. saligna stand was located in a coastal Bay of Plenty forest

Eucalyptus nitens (Deane et Maiden) Maiden, a high-altitude species from the mountains of central Victoria and eastern New South Wales, is the eucalypt most commonly planted for pulpwood in New Zealand. Parents of these two provenances of E. nitens were crossed with E. grandis Maiden of coastal New South Wales origin to create first-generation hybrids that should be adapted to warmer, low-altitude sites in New Zealand and which might be able to be propagated by cuttings. Single-pair crosses were made between eight New Zealand-selected parents from both central Victorian and southern New South Wales provenances of E. nitens, as pollen parents, with eight selected E. grandis female parents growing in a South African seed orchard. Eight seedlings of each hybrid family and of open-pollinated families from each of the E. nitens and E. grandis parents were planted in a single-tree-plot design at each of four sites in New Zealand. One trial in Southland was destroyed by frost, and hybrids performed poorly at a frosty central North Island site. At two, warm, coastal Bay of Plenty sites the hybrids at age 2 years and 8 months had average height and diameter that was about the same as their E. nitens open-pollinated siblings. There were a large proportion of poorly grown, genetically defective individuals in the hybrid families and a smaller proportion of extremely vigorous trees which exceeded the growth of the best-grown individuals of E. nitens. These successful hybrid genotypes could form the basis of clonal forestry deployment of the hybrids, provided vegetative propagation methods can be developed. Some success has already been achieved in propagation trials and further crosses have been made and await planting.

Eucalyptus grandis x E. nitens is currently being trialled in New Zealand for pulp and wood production. F1 hybrids were successfully propagated using rooted cuttings derived from coppice of 2-year-old seedlings. The hybrid behaved more like the E. grandis parent in its ability to coppice and produce rooted cuttings. Starving stock plants of nutrients and applying 2-stage topping, which are necessary for obtaining coppice with E. nitens, gave poorer results for the hybrid. The ideal window in time for coppicing has not been identified and the experimental period for coppicing should be extended from June through to February. Clonal influence was the only critical factor in the rooting of the hybrid cuttings. Rooting percentages ranged from 0 to 100% depending on clone. Sixteen out of the 135 clones (12%) had 70% rooting or better. This figure should increase with further optimization of the factors important to coppicing and rooting.

Forest harvesting can generate large amounts of woody debris, some of which ends up in riparian areas and stream channels where it can pose problems for post-harvest management due to its potential for mobilisation during floods and for affecting stream ecosystem structure and function. We measured the decay state, size, and orientation of Pinus radiata D. Don large woody debris (>10 cm diameter; LWD) in and alongside 13 spring-fed streams in the central North Island where harvesting had occurred between 1 and 19 years previously, partly to assess the time-scales associated with any impacts of harvesting debris. Most LWD in the streams and riparian zones was oriented perpendicular to the flow. Similar orientations for riparian and submerged wood indicated that it had moved little since harvesting and was therefore a stable feature of the channels in which it was deposited. Most instream LWD was lying level on the streambed where it could potentially play an important role in channel scour processes. Diameters of riparian LWD tended to be larger than instream LWD, apparently due to "case-hardening" of logs on the land and erosion of decayed material from the outer surfaces of logs in streams. LWD in streams was less decomposed than that in riparian areas where it became severely decayed within 10-15 years. Logarithmic regression of time since harvest, and decay data for submerged LWD considered to be from harvesting, indicated that pine logs remained intact in these spring-fed streams for over 20 years. The persistence and stability of pine LWD throughout the rotation period mean that LWD is likely to play an important role for stream processes and biota in central North Island pumice-bed streams.

Four native provenances of Pinus radiata D.Don were sampled in a common environment for studying wood density (as chip density), kraft pulp yield at Kappa number 302, 10 pulp fibre properties, and four handsheet properties. Each provenance sample comprised a nominal 30 unrelated trees. Separate wood specimens were taken to represent toplog and slabwood material respectively. Within each of the eight provenance/ wood-zone subclasses the trees were allocated at random to give duplicate bulked chip samples. Use of four refining treatments gave further replication and allowed various handsheet properties to be evaluated in relation to density, tear index, and tensile index. The various properties were similar for the four provenances (Año Nuevo, Monterey, Cambria, and Guadalupe), except in the Guadalupe toplog material. Chip densities for the Guadalupe toplog and slabwood samples were surprisingly similar, and Guadalupe toplog fibres were short, slender, of low coarseness, stiff (high collapse resistance), and thus present in large numbers compared with those of the other toplog pulps. Opportunities for tree breeding could, therefore, involve selecting Guadalupe trees with high density and large numbers of slender tracheids of low coarseness in the corewood. Such selection should substantially improve corewood properties for both solid-wood products and pulping. Verification of the distinctive tracheid properties found in Guadalupe marterial is now needed.

Fast-grown plantation wood (particularly the juvenile zone) can have some limiting performance characteristics (stiffness, strength, stability). Improvements in the overall performance of the lumber products will depend on a knowledge of the most important factors which can be influenced by either silviculture or tree breeding. Some wood technologists believe that the traditional view of wood density as the single most-important factor may be flawed, and that other parameters (e.g., microfibril angle, spiral grain, and compression wood) may have an equal or greater effect. During 1996-98, intensive utilisation studies were carried out on Pinus radiata D.Don material from two sites in the central North Island of New Zealand. Sample trees were selected from managed crops (23 and 28 years old respectively) to cover a range of wood properties. They were chosen on the basis that they represented a wide range of tree characteristics (diameter, branch habit), and subsequently the full range of log and stem properties was measured from wood disc samples and sawn lumber. One of the objectives of the research was to document the relative importance of a number of properties known to influence both mechanical properties and product performance. Some of the studies comprised a series of standard "small clears" tests, designed to untangle the impacts of wood density, ring width, spiral grain, compression wood, and microfibril angle on bending strength (MoR) and stiffness (MoE). Altogether 450 "small clears" samples were selected, covering the range of log and lumber properties existing in 51 stems from 15 clones across the two trials. Each wood sample was designated "juvenile" (within 10 rings from the pith - 304 pieces) or "mature" (129 pieces) depending on the location within the tree stem. The study data clearly showed very significant effects of both density and microfibril angle on clearwood performance of the juvenile wood. In the mature wood, however, density alone was important. Overall, wood density was confirmed as the most influential parameter, affecting all classifications of wood, being easy to measure, and showing high juvenile:mature correlation. This confirms traditional thinking and justifies the past and present efforts to document factors affecting wood density in P. radiata and to improve the average levels through tree breeding.

Continuous vegetation cover preventing erosion of coastal sand dunes is essential for the protection and development of productive forestry, farming, and other activities in their vicinity. Use of symbiotic nitrogen fixation to enhance the nutritional status (and thus the vigour) of sand dune vegetation currently depends on the introduction of nursery-raised plants. Local sand and a standard nursery potting compost were seen as possible alternatives for seedling rooting media, but their relative effects on plant performance after transfer to exposed dune sites were unknown. Three nitrogen-fixing species representing a range of tolerance to dune conditions in the north of the North Island (Acacia sophorae (Labill.) C. Martius, Chamaecytisus palmensis (Christ) Bisby et K. Nicholls, and Lotus uliginosus Schk.) were used as the basis for comparison. A slight overall plant size advantage was associated with potting compost during the first year after transplanting, but this was not maintained during the second and third years. Long-term plant survival was not influenced by seedling rooting medium, although mortality of C. palmensis (the least tolerant species) was twice as great in compost-raised plants as in sand-raised plants during the first 6 months. There was no evidence of any consistent effect of seedling rooting medium on relative nitrogenase activity. It was concluded that choice of medium for raising seedlings in the nursery is likely to have only a minor and transitory effect on the performance of nitrogen-fixing plants in sand-dune revegetation projects and species-screening trials.

Data produced by the "Method for Assessment of Recoverable Volume by Log Type" (MARVL) from the Tikitere Agroforestry Trial were used to compare the economic consequences of a range of final-crop stockings of Pinus radiata D. Don on a typical farm site with understorey grazing. The 93-ha 21-year-old trial included stockings at 400, 200, 100, and 50 stems/ha; intermediate stockings were simulated using the stand growth program STANDPAK as calibrated by the MARVL data. The highest stocking examined (400 stems/ha) yielded the highest revenue per hectare. Whereas it is conceivable that rising premiums for large pruned logs could enable 200 stems/ha to yield an equivalent revenue, premiums would have to almost double before 100 stems/ha could achieve this. In terms of Internal Rate of Return (IRR), the most profitable investment - given that current log specifications do not include many important features of internal wood quality - is clearly to harvest at or before age 21, and to adopt low final-crop stockings (100-200 stems/ha). On the other hand, if the Net Present Value (NPV) approach is used and combined with discount rates that are markedly lower than the IRR (11-14%), then the optimum felling age occurs as late as age 31 and optimum final-crop stockings are as high as 400 stems/ha. The preferred method (IRR or NPV) will depend on the objectives of the decision-maker.

Data produced by the "Method for Assessment of Recoverable Volume by Log Type" (MARVL) from the Tikitere Agroforestry Trial were used to compare volumes of Pinus radiata D. Don by log type for a range of final-crop stockings on a fertile farm site. The 93-ha 21-year-old trial included stockings at 400, 200, 100, and 50 stems/ha; intermediate stockings were simulated using the stand growth program STANDPAK as calibrated by the MARVL data. Higher stockings tended to have a greater total and recoverable volume, unpruned volume, volume of large-diameter unpruned logs, and volume of small-branched logs (nearly 200 m³/ha). At age 21, all stockings over 250 stems/ha generated approximately the same pruned volume, but higher stockings are expected to have greater pruned volumes at older ages. The greatest pulpwood volume occurred at 100 stems/ha, comprising 30% of the total recoverable volume. The volume at Tikitere for the 400 stems/ha stocking was about one-third higher than expected for an equivalent non-farm site, but this difference may decrease, with age.

The relationship between juvenile and final stem sweep is used to aid selection of final-crop trees at time of thinning, and measurement of juvenile sweep at final pruning (approximately age 6 to 8 years) provides the information with which to predict sweep in the final-crop stems of Pinus radiata D. Don stands. Sweep in the pruned butt log, in conjunction with log diameter, length, and taper, affects the level of conversion to sawn timber, and hence, the value of a log. Pith sweep, as a surrogate for juvenile sweep, and final-stem sweep data calculated from cross-sectional analyses or sawing studies of 815 pruned logs, were used to improve the stand average and individual log sweep predictions made in the stand modelling system STANDPAK. The following juvenile / final sweep model was fitted using regression analysis: S_F = aS_P - b(AGE-28) where: S_F is final sweep (mm/m); S_P is pith sweep (mm/m); AGE is time from juvenile sweep measurement to final harvest (years); a, b are regression coefficients. Validation by residual analysis indicated the model slightly under-predicted final sweep and the error in predicting final sweep could be expected to fall within ±5 mm/m. Comparison of the fit of several distribution models to study data, by comparing chi-squared goodness-of-fit deviances, identified the lognormal distribution as a reliable predictor of individual log final sweep. Validation using an independent data set from six sites indicated the error in predicting percentage frequency of logs in individual log classes could be expected to fall within ±20%. An important limitation of this study was the use of pith sweep as a surrogate for juvenile sweep.<

Top-height growth components and site index curves, based on permanent sample plot data for Pinus radiata in New Zealand, were were used in growth models stratified into 8 growth modelling regions. The application of an age adjustment based on seasonal growth patterns helped to reduce variation caused by different measurement dates and the rapid and sustained growth of P. radiata throughout the year. Global and local parameters in a stochastic differential equation version of the Bertalanffy-Richards model were estimated simultaneously by maximum-likelihood. Various parameterizations were tried. The procedure has worked reliably, giving satisfactory predictions and making full use of the available data. Comparison of the models obtained showed significant differences across regions, justifying the initial regionalization and indicating that a broader aggregation would not be advisable.

Soil fertility at the Tikitere Agroforestry Research Area near Rotorua has been monitored since the site was planted with Pinus radiata D. Don in 1973. Measurements made in 1991 and 1992 showed that after 18 years soil pH had declined and that Olsen phosphorus levels increased with increased tree stocking. Soil magnesium also declined at the higher tree stocking rates. We made further measurements on samples collected in 1991 and 1992 to determine reasons for the change in nutrient status. Soil carbon and soil organic phosphorus in the surface soil (0-75 mm depth) decreased with tree stocking, indicating increased net mineralisation of soil organic matter under P. radiata compared with pasture. Exchangeable cations and cation exchange capacity also declined with tree stocking, which is consistent with the loss of exchange sites in soil organic matter. Total soil inorganic phosphorus declined with increased tree stocking. This was expected because of reduced fertiliserphosphorus input at the higher stocking rates. However, phosphorus fractionation showed that bicarbonate-extractable inorganic phosphorus increased and acid-extractable inorganic phosphorus declined with tree stocking. The latter indicated possible decreases in soil apatite phosphorus. These results were consistent with the mobilisation of soil phosphorus under P. radiata by dissolution of fluorapatite (in parent material deposited during the 1886 Tarawera eruption) due to lower soil pH values and mineralisation of organic phosphorus previously accumulated under pasture. Consequently, Olsen phosphorus has increased despite reduced phosphorus fertiliser applications. Microbiological activity (as indicated by microbial biomass and microbial respiration) and phosphatase activities of soil samples collected in 1995 also decreased with tree stocking, and so mineralisation of organic matter was not a result of microbial activity alone. However, the proportion of soil aggregates > 0.5 mm decreased under P. radiata, indicating there was less physical protection of the soil organic matter than under pasture where soil aggregates were maintained by fine roots.

Surveys of port environs and forests to detect new introductions of harmful insects or fungi are carried out in New Zealand by Forest Health Advisers. The efficiency of three survey methods was analysed, using simulated damage. The first two methods involved the use of drive-through or walk-through sampling of plantation forests, and the third method focused on sampling port environs. At the slowest vehicle speed tested (15 km/h), the drive-through forest sampling gave detection efficiencies very similar to walk-through sampling. In the drive-through surveys, 88%, 79%, and 63% of simulated damage was detected at 0 m, 20 m, and 40 m from road edge, with corresponding efficiencies of 97%, 71%, and 47% for the walk-through surveys. Detection levels for the drive-through survey reduced sharply at greater vehicle speeds. One port environs survey detected 49% of all simulated damage, but cumulative detections from repeated surveys gave a lower gain than at first assumed, with two surveys detecting 66% of all symptoms rather than the 75% predicted. Using the new efficiency data, and a revised method of calculating the probability of detecting a randomly located infection centre, detection probability scores for drive-through surveys and walk-through plots were considerably higher than predicted, but the port environs trial showed that re-inspections gave lower probabilities of detection than predicted.

Woody debris volumes and channel bank disturbance were measured in a 100-m section of stream channel, prior to and after harvesting, in 17 streams in pine plantations in five regions of New Zealand. These sites were harvested using four different harvest methods. Volumes of pre-harvest woody debris and woody debris produced during harvest averaged 105 m³/ha, and 147 m³/ha, respectively. Apart from the stream-cleaned sites where virtually all the pre-harvest and harvest woody debris was removed, post-harvest volumes (pre-harvest + harvest) averaged 289 m³/ha and increased three-fold on average over pre-harvest levels. Most of the woody debris in the stream channel was positioned above the stream - 69% of pre-harvest woody debris, 64% of harvest woody debris, and 66% of total post-harvest woody debris. The remainder lay in-stream or on the floodplain. The most significant change in woody debris characteristics after harvest was size distribution. Small woody debris <10 cm in diameter (SWD) increased from 13% of woody debris volumes at pre-harvest to 38% at post-harvest. The number of pieces of large woody debris ≥10 cm in diameter (LWD) increased significantly, and the average length and piece size decreased significantly after harvest. This was due mainly to the removal of the larger merchantable pieces of LWD from the stream channel. Harvest method had the most impact on harvest woody debris volumes in the stream channel, overriding the influence of riparian buffers which ranged in width from 1 to 30 m at four of these sites. Stream-cleaned sites had the lowest harvest woody debris volumes, followed by sites harvested with ground-based systems (15 m³/ha and 48 m³/ha respectively). When yarder systems were used to extract timber back from the stream edge, woody debris volumes averaged 104 m3/ha, whereas hauling across the stream channel resulted in the highest average woody debris volumes of 287 m³/ha. For hauling across the stream channel only, there was a relationship between stand volume and harvest woody debris volumes. Bank collapses accounted for 68% of all pre-harvest channel bank disturbances. Bank scuffing from felling and log extraction during harvest operations was the most common channel bank disturbance after harvest (46%). Harvest method did not show a clear relationship with the degree of channel bank disturbance.

Appropriate nitrogen-fixing plant species are required to replace Lupinus arboreus Sims, which once played an important nutritional role in the managed succession used to reclaim unstable coastal sand in New Zealand. A fungus disease has reduced L. arboreus populations by up to 95%. Six spaced-plant trials at widely-separated locations on the west coast of the North Island identified 18 introduced species which are able to grow and fix nitrogen in the sand dune environment. Trials were conducted in the area immediately behind the foredune which is partially stabilised by sand-binding grasses. In this zone plant growth is limited by nitrogen supply. Unless the activity of browsing animals on sand dunes can be controlled, only unpalatable species or those combining high tissue regeneration capacity with absence of weed potential are likely to be effective and acceptable substitutes for L. arboreus. The relative performance of the 18 species requires further examination under standard trial conditions before a final list can be recommended for sand revegetation purposes.

Whole zygotic embryos of Pinus radiata D.Don were evaluated for survival after storage in liquid nitrogen (-196ÝC). Zygotic embryos were cryoprotected with selected levels of sorbitol and dimethylsulphoxide (DMSO) and were then frozen, employing an isopropanol-filled cryovial container and -80^oC freezer for 1.5 hours prior to immersion in liquid nitrogen. After removal from liquid nitrogen the condition of embryos was assessed and then cotyledons were excised for shoot induction. Up to 61% of cotyledons cryopreserved in a cryoprotectant formula of 0.4M sorbitol and 15% DMSO regenerated shoots. A sample of the regenerated shoots was successfully rooted. Cryopreserved cotyledons from 84% of the genotypes were able to regenerate shoots upon return to in vitro culture as compared to 92% of the unfrozen controls.

There was substantial variation in the timing of lateral long-shoot initiation in a Pinus radiata D. Don clone, associated with ramet size. Grafted meadow orchard ramets aged 2, 3, and 4 years from planting (with respective heights approximately 1, 2, and 4 m) were sampled on 19 February and 31 March. On 19 February, microscopic examination revealed that vegetative and female strobili primordia were differentiating on 4-year-old ramets, but 2-year-old ramets had not initiated any lateral long-shoot primordia. The 3-year-old ramets had vegetative and undifferentiated long-shoot primordia but no female strobili. On 31 March, buds on the 2-year-old ramets contained lateral long-shoot primordia in early stages of differentiation. At this time strobili primordia of the 2-year-old ramets were similar in size and state of bract formation to those of the 4-year-old ramets 6 weeks earlier. Application of GA4/7 on 19 February stimulated only the 2-year-old ramets to form female strobili.

A software tool was developed for the stochastic simulation of multiple traits and uses sequential culling for the different traits. It selects superior progeny and, at present, can calculate genetic gain from the backward selection of parent clones based on progeny performance, crossing those parent clones, and forwards selection among the progeny. Key information input to the program includes the base population size number of parents selected, economic weights for different traits, genetic correlations between traits, mating design and number of crosses, number of progeny planted and selected, and proportion culled within the progeny for each trait. An application of the software, with particular relevance to the Pinus radiata D. Don breeding population in New Zealand, deals with the development of breeds for specific products and for specific site types. Gains obtained at age 8 from different selection emphases were compared, for an unspecialised main population of 300 parents and four specialised breeds of 24 parents each. The parents came from a base population of 2000 progeny-tested parents. The simulation covered forwards selection in the progeny obtained by crossing the selected parents. The breeds were designated Structural Timber, Growth and Form Elite, Long Internode, and Fibre, and reflected different end-product objectives. Six traits were involved - namely dbh, stem straightness score needle retention score, branch cluster frequency score, spiral-grain angle, and wood density. The simulation provided support for differentiating a Structural Timber breed (with a strongly multinodal branching habit, higher wood density, and reduced spiral grain angle). As well, we were able to simulate breeds similar to the existing Long Internode breed and a Growth and Form Elite population. The software can also be used to study the effect of varying population size.

The North American provenances of Pinus radiata D. Don show differences in growth and form when grown in common-garden trials, but little has been reported on differentiation into regional land races as an exotic. Data at age 8 years from two large trials, each with progenies of plus-tree selections from different regions in New Zealand, were analysed for various traits. The first trial ("850") had polycross progenies of 109 parent clones, selected between 1950 and 1967 and belonging to six regional selection groups, and was planted on six sites in 1975. The second ("888") had 329 open-pollinated progenies, from 279 plus-trees selected in 1988 in six distinct regions and from clones in a Chilean seed orchard. This trial was planted on three sites in 1989. In the "888" trial there were significant differences among seed sources for every trait except needle retention at all sites and malformation at one of the sites. The Tarawera source had the highest branch cluster frequency score, needle retention score, malformation score (least malformation), and percentage of acceptable stems. The Southland and Nelson sources had low branch cluster frequency scores. Branch cluster frequency score was the only trait for which significant seed-source effects were found across sites in the "850" trial. The lowest branch cluster frequency scores were for the Nelson and Southland sources. However, there were significant interactions between seed source and site, and at the site with the highest survival and heritabilities there were significant differences between seed sources for diameter, straightness score, branch cluster frequency score, and malformation score. On this site, the highest score for every trait assessed was for the Kaingaroa second-rotation seed source. Two trends emerged from these studies. The first trend was of faster growth and better form for selections made in second-rotation stands in the central North Island. Given that most of the plus-trees selected and progeny tested for the New Zealand breeding population were selected on such sites, this was an encouraging result. The second trend was tor lower branch cluster frequency for selections made in the South Island Possible reasons for significant seed-source effects are better ability to select on certain sites and regions, founder effects, and natural and artificial selection.