Plant communities across three different-aged glacial terraces in a south Westland lowland podocarp forest were classified on the basis of floristic composition from 90 vegetation descriptions. Five forest communities were identified, differing in relative abundance of the main tree species and in species richness. One community was dominant on each terrace. Comparisons with other studies in the same forest suggested that differences in stage of soil development could explain floristic differences. The association of different forest communities with particular terraces suggests that species composition will shift over millennia towards a community dominated by Dacrydium cupressinum Lamb. and Manoao colensoi (Hook.) Molloy. These findings have implications for the sustainable management of south Westland's lowland podocarp forests.

Recent developments in the stand modelling system STANDPAK have allowed the prediction of Mean Log Age, growth rings per small-end diameter (s.e.d.), juvenile wood volume, and the proportion of juvenile wood. This capability was used in a theoretical exercise with Pinus radiata D. Don to examine the influence of a number of site, silvicultural, and genetic factors on Mean Log Age and the proportion of juvenile wood in New Zealand domestic log grades.

Sensitivity analysis of individual factors and/or comparisons between simulations of major forestry management practices indicated that Mean Log Age was affected by felling age, tree stocking, and site quality. The proportion of juvenile wood in merchantable log volumes was found to be affected most by felling age, the number of growth rings containing juvenile wood, and site quality. Silvicultural factors (pruning severity, timing of thinning, and tree stocking) had a relatively small effect. Genetic improvement expressed as a Growth and Form (GF) rating, was indicated in this study to have little effect on the proportion of juvenile wood.

An examination of historic changes in silvicultural regime, felling age, and site quality of P. radiata forests in the central North Island indicated increases in the proportion of juvenile wood resulting from modern practices that may explain the current concerns about wood quality.

Inclusion of log age or number of rings/s.e.d. as additional criteria in the current system of log grading, should better segregate logs by the proportion of juvenile wood and wood density.

During summer 1996/97, a Pinus radiata D. Don plantation (9000 ha) near Whangamata, New Zealand, was sampled to assess how intensity and distribution of brushtail possum (Trichosurus vulpecula Kerr) damage were related to stand maturity. Based on a review of previous surveys, a new damage-scoring system was developed. Using this system, the overall proportion of pine trees greater than 3 years of age with possum damage was found to be very low (3.6% of 1244 trees surveyed) despite possums being present in substantial numbers (mean catch of 28.1 possums per 100 trapnights; range 12.0-54.9). Damage levels varied markedly (0%-30%) between 36 plantation blocks. The majority of possum damage involved needle clipping and damage to lateral branches. All damage types showed negative correlations with tree maturity. Given the low damage levels recorded, the cost-effectiveness of current possum-control efforts in pine plantations of this type is uncertain and would benefit from further economic analysis.

Plants have distinct genomes in their nucleus, chloroplast, and mitochondria. Because chloroplast genome in Pinus species is normally paternally inherited, DNA marker systems capable of detecting polymorphisms in the chloroplast molecule would be useful studying paternity and pollen movement. Short nucleotide repeats analogous to nuclear microsatellites, or simple sequence repeats (SSRs), have been previously identified from DNA sequence analysis of the Pinus thunbergii Pari, chloroplast genome, and polymerase chain reaction (PCR) primers have been designed to amplify of the chloroplast microsatellite loci (cpSSRs). Seven of the 20 cpSSR sites were found to be polymorphic in 51 unrelated P. radiata D.Don trees with between two and four alleles per locus. These seven loci have been combined in two fluorescently labelled multiplexed PCRs to generate chloroplast haplotypes resolved in single electrophoretic lanes. This multiplexed protocol largely automates the collection of chloroplast haplotype data for clonal identification, paternity analysis, pollen genotyping, and studying chloroplast diversity in natural populations of P. radiata and other conifers.

A wild type Agrobacterium tumefaciens strain was used to produce tumours on young Pinus radiata D. Don stems in vitro, and stable expression of the uidA reporter gene in this tumour tissue was confirmed histochemically. The regenerative capacity of detached cotyledons from excised P. radiata zygotic embryos was assessed after different treatments involved in A. tumefaciens infection. Exposure to acetosyringone or the antibiotic timentin did not have a negative effect on tissue health and regeneration, whereas exposure to the antibiotic cefotaxime resulted in poor tissue health. Meristematic activity and shoot elongation from very small portions of cotyledons was demonstrated. A protocol using A. tumefaciens to transfer the uidA reporter gene into adventitious shoot-forming meristems was developed and co-cultivation parameters were improved to maximise transient expression of this reporter gene.

A method was investigated for transforming Pinus radiata D.Don embryogenic tissue based on the aphIV gene from Escherichia coli coding for aminocyclitol phosphotransferase, which confers resistance to the antibiotic hygromycin B. Embryogenic tissue captured from immature P. radiata embryos was transformed using the DuPont Biolistic® particle delivery system. Transformation of embryogenic tissue was confirmed by PCR amplification of the transferred aphIV gene and enzymatic activity of the uidA gene product (_b-glucuronidase) in transgenic material. Transformed embryogenic lines could be recovered at a frequency of approximately 0.14 transformants per bombarded petri dish, an efficiency which is suitable for routine transformation experiments.

Genetic parameters were estimated for growth and form traits in a 7-year-old progeny trial established in 1989 and containing 28 open-pollinated families of Pinus radiata for the Basque Country breeding programme of the species. Observed individual heritabilities were moderate for height (h² = 0.20), stem diameter (h² = 0.22), branch size (h² = 0.14) and branch angle (h² = 0.15) and negligible for stem straightness. Stem diameter, height, and branch angle were positively correlated. Branch size had an adverse genetic correlation with growth traits. Genetic gains under index selection strategies were examined. Results from this small-scale trial suggest that is not possible to achieve simultaneous improvements in both growth and branch size.

Using stochastic simulations, the effects of size of unrelated groups, mating designs, and selection strategies were investigated to address the issue of assuring long-term genetic gains. The parameters analysed were status number (Ns), inbreeding coefficient (F), and genetic gain for two heritabilities (0.05 and 0.2).

Under a fixed-resource scenario, unrelated and non-inbred founders were clustered into variable group sizes (from 12 to 128), with 0.5 to 5.5 crosses per parent. Also considered were phenotypic selection and combined index selection, with and without restrictions on the number of individuals selected per family.

Breeding schemes with small, disconnected groups were slightly more efficient in preserving status number through a large number of generations than breeding schemes with large groups, but medium- to large-size groups showed larger expected gains. Inbreeding in small groups may become so severe as to cause fertility problems and considerably reduce the efficiency of selection for additive gene effects. Hence, using very small groups would probably not provide a sustainable long-term breeding strategy. Nevertheless, small groups may form a critical component of breeding strategies that employ marker-assisted selection, since the maintenance of marker-QTL associations would be facilitated in these small populations.

In two Cupresssus macrocarpa Gordon progeny tests planted at Strathallan (South Island) and Gwavas (North Island), with 76 New Zealand land race families and 76 Californian families, the land race progenies performed better than the Californian progenies. Among Californian populations Point Lobos showed flatter branch angles than Cypress Point.

Most traits were highly variable and the narrow-sense heritabilities for branching (0.5) and straightness(0.3) were high to moderate. Due to micro-site variability, diameter growth and height showed smaller heritabilities.

Genetic and phenotypic parameters for diameter at breast height, bole straightness, branching, pilodyn penetration, and basic wood density were estimated for two open-pollinated progeny tests at age 5 years, including first- and second-generation material of different provenances. Different values for coefficients of relationship (0.25 and 0.5) were applied in the heritability calculations for each generation, which accounted for the expected difference in their pollination behaviour and the observed difference in their variance component estimates.

Strong evidence was found for substantial realised gains from open-pollinated progenies from seedling seed orchards of North Forest Products (NFP) and AMCOR in Australia and a New Zealand Forest Research Institute progeny trial. Differences in diameter growth between the best seed orchard seedlot and the best native race were 0.67 standard deviations. In terms of wood density, the New Zealand progenies were clearly superior to those from Australian seed orchards. Families from AMCOR and NFP were of low to very low density.

Increasingly, in the East Coast region of the North Island of New Zealand stands of indigenous regenerating kanuka (Kunzea ericoides var. ericoides (A.Rich.) J.Thompson) are being felled and replanted with Pinus radiata D.Don. Conversion has occurred predominantly on erosion-prone hillslopes where storm-generated landsliding has been widespread and severe, but data on the relative effectiveness of these two forest species in enhancing slope stability are rare. For kanuka and P. radiata, shear stress-displacement curves for their corresponding potential shear planes were measured at two sites by in situ direct shear tests on soil with and without roots. The contribution of roots to soil strength was first estimated by calculating the difference between the maximum shear stress of the shear stress-displacement curves obtained for soil with and without roots. Results suggested that for individual trees the contribution from the roots to soil strength on a root cross-sectional area per unit shear area basis was independent of species for the two tree species tested. There were, however, significant differences in stand density between these two species. These results were then used to obtain relationships between shear stress and shear displacement at the peak of the shear stress-displacement curve, and between cross-sectional area of roots per unit shear area. Taking the shear strength of the combined soil-root system as the peak value of the shear stress-displacement curve produced from the in situ direct shear tests, a limit equilibrium slope stability analysis method was used to derive the safety factors. A simple model developed using the relationship between the shear strength of the soil-root system, the specific root crosssection area, and slope angle was then used to determine safety factors for typical stand densities of naturally regenerating kanuka for comparison with different P. radiata management regimes at equivalent stages of growth.

The model predicted that safety factors for stands of P. radiata in the first 8 years after establishment would be lower than for equivalent-aged stands of fully-stocked regenerating kanuka under similar conditions. However, after 16 years the safety factor for a stand of kanuka would be lower than that for P. radiata at final stocking densities typical of framing and biomass regimes.

In areas where vegetation plays a major role in soil conservation and erosion control, the model can be used to compare the stability of forested slopes with different species and stand densities. However, the model does not take into account the effect of buttressing by mature tree roots.

Significant areas of indigenous kanuka (Kimzea ericoides var. ericoides (A.Rich.) J.Thompson) growing on the unstable East Coast hill-country north of Gisborne, New Zealand, are under threat of replacement by exotic plantation tree species. A knowledge of the strength and decay properties of kanuka root-wood would increase understanding of the potential impacts of clearfelling kanuka in this region. About 30 trees were felled and their roots sampled at 6-monthly intervals over the following 48 months to determine tensile strength, rate of decay, and elastic properties of kanuka root-wood. Roots were excavated using high-pressure water and tested in tension on an Instron Universal Testing Machine.

Mean live kanuka root-wood strength and modulus of elasticity were 32.5 MPa and 830 MPa, respectively. Mean tensile root-wood strength increased 33% to a maximum of 43.1 MPa 12 months after cutting, and did not decline below live root-wood strength until 24 months after the death of the parent tree. To demonstrate that the root decay pattern of kanuka was not unique, additional information was collected and added to the dataset of a previous study of root-wood strength deterioration in southern rata (Metrosideros umbellata Cav.). Mean live southern rata root-wood strength was 50.8 MPa, which increased 27% over 15 months after cutting to 64.7 MPa. The root-wood strength of both species declined exponentially. The calculated time for kanuka and southern rata to lose half their Iive root-wood strength was 39 and 45 months, respectively.

It was concluded that the strength and elastic behaviour of roots impart resilience to the soil by allowing a greater magnitude and range of slope shear-displacement before failure occurs. Kanuka root-wood has a greater live strength and a lower decay rate than the main exotic plantation species, Pinus radiata D.Don. From a purely root strength/root decay stand-point, kanuka would at least for the first 4 years, provide a clearfelled slope with greater stability than a similar slope clearfelled of P. radiata. Consequently, kanuka can be seen as an effective stabiliser of slopes and thus has a very important role to play in the maintenance of slope stability.

Overturning bending moments were analysed for Pinus radiata D. Don trees which had been winched over at Eyrewell Forest, Canterbury, between 1967 and 1971. Trees were sampled from four different age-classes in three forest compartments. The bending moment applied by the winch and cable system increased rapidly and linearly to a maximum value before decreasing until the trees toppled under their own weight. The bending moment due to the mass of the offset stem plus crown contributed an average of 9% to the total overturning moment at the point of maximum applied moment. Significant positive relationships were found between the maximum resistive bending moment (M_c) offered by the tree and its total height, diameter at breast height (dbh), and stem volume. The greatest proportion of the variance in Mc was explained by a linear relationship involving dbh. The angle of stem deflection at both the maximum resistive bending moment and the point at which the tree toppled under its own weight was significantly and negatively related to tree height. Analyses of covariance found that root plate diameter had a significant effect on M_c while root plate depth did not. The effect of taper was uncertain.

Although nitrogen dynamics during litter decay have been studied extensively, little work has addressed the contribution of needles in harvest residue to nutrient losses. We examined the decay of Pinus radiata D. Don needles on slash and not touching the soil, from compartments clearfelled between 4 months and 4 years previously in Himatangi and Santoft Forests. Fresh foliage was also collected from 25-year-old P. radiata, together with soil F, H, and A horizons. The needles and soil samples were incubated, both separately and combined, in suction vessels and leached regularly with water. Most nitrogen was leached from the 4-month and 4-year needles, and amounted to about 8% the total nitrogen; little nitrogen was leached from fresh foliage. Nitrate-nitrogen was the dominant form of nitrogen leached from the 4-year needles, whereas ammonium nitrogen was dominant from the 4-month and 1 -year needles. More nitrate-nitrogen was produced when the 4-month needles were incubated on top of F, H, or A horizon samples which increases potential for leaching. These data were consistent with field results from lysimeters showing nitrate-nitrogen was leached under 2-year-old windrows and raked soil in Santoft Forest. Net phosphorus mineralisation occurred readily in the older needles and mineral soil, and with time could contribute to the labile pool of phosphorus. Our results suggest that post-harvest losses of nitrogen could increase when large amounts of slash remain in stockpiles. Water-soluble carbon compounds were released from all the foliage samples. Differences in nitrogen losses and forms of nitrogen mineralised are likely to be related to changes in quality of carbon in needles after harvest.

Realised genetic gain in diameter, multinodality, stem straightness, malformation occurrence, and stem cone whorl incidence was assessed using field plots in 20-year-old stands planted with Kaingaroa bulk, Kaingaroa climbing-select, and "850" seed-orchard seedlots. The "850" seed-orchard seedlot stands had a significantly greater d than Kaingaroa bulk, resulting in 4.2% gain in d; however, there was no difference in d between "850" seed-orchard material and Kaingaroa climbing-select. For tree mean internode length (MIL), incidence of malformation, and degree of stem sinuosity, "850" seed-orchard seedlots showed no significant improvement over Kaingaroa climbing select or Kaingaroa bulk. The incidence of stem cones was significantly lower for "850" seed-orchard material than for the other two seedlots. The results of this study may partly differ from those of large-plot genetic gain and progeny trials for several reasons. Firstly, the first 8 to 10 years of seed production from the "850" seed orchards contained seed of poorer quality than later seed production. Secondly, superior "850" seed-orchard seedlots may have been better than those planted in commercial stands.

Analysis of covariance models of diameter and height of 13-year-old Pinus radiata D. Don grown in commercial forests in the central North Island formed the basis for comparing average stand tree diameter (d) and mean top height (MTH) of non-seedorchard and "850" seed-orchard seedlots. Both the d and MTH models used several groupings of seedlots to determine the impact of these first tree selections ond and MTH. The models included several variables to remove other growth-affecting factors. For the modelling of d, altitude and crop stocking were used, and the MTH model used site index. The analyses of covariance for both d and MTH, show no significant improvement overall of "850" seed-orchard seedlots over Kaingaroa climbing select or Kaingaroa (bulk) seedlots at age 13. The Gwavas "850" seed-orchard 25-clone seedlots did show a significant improvement in d over Kaingaroa bulk, but not over Kaingaroa climbing select. The results of this study differed from those of large-plot genetic gain and progeny trials. This may be for several reasons. The first 8 to 10 years of seed production from the "850" seed orchards contained seed of poorer quality than later seed production. The assessment of P. radiata at age 13 may be too early for the detection of divergence of "850" seed-orchard material from non-seed-orchard material to be noticeable in commercial stands. Superior "850" seed-orchard seedlots may have been better than those planted in commercial stands.

Pinus radiata D.Don shows marked variation in the number of branch clusters formed annually, with a corresponding variation in"internode" length (distance between branch clusters). Long internodes are commercially valuable due to greater yield of knotfree lumber in the unpruned part of the stem. Internode length is under strong genetic control and is highly amenable to selection. In a recent sawing study trees of clones selected for long internodes gave much higher lumber value as U.S. shop grades than short-internode trees.

Breeding for internode length began in 1970 with the selection of 104 long-internode plus-trees. Subsequent work included open-pollinated progeny testing of both the first-generation selections and 74 second-generation selections, making and testing 153 crosses involving first- and second-generation selections, and further first-generation selections in 1985. Advanced-generation selections will be made in the control pollinated trials in 1998.

Gains in internode length entail somewhat reduced gains in volume growth, branch size, and stem quality. The Long Internode breed has been separated from the main breeding population. The goal is a tree with one or two branch clusters per year, to be grown on appropriate sites with suitable silviculture. Deployment of improved stock is by using seedlings or cuttings from control-pollinated seed. A clonal forestry option also exists. The next step to advance the Long Internode breed is likely to be forming two small elite sublines. Future breeding will largely be within the two elite sublines, allowing for unrelated crosses between sublines for production purposes.

In 1970 an international exchange of 319 select lots of Pinus radiata D.Don was arranged among several countries. Most lots were open-pollinated progenies, with 31 pair-crosses, 26 polycrosses, and two bulked lots. Field experiments were planted with varying field layouts and variable representation of lots. Trials were assessed mainly at 6-8 years from planting, with one trial measured at 14 years and one remeasured at 15 years. Main emphasis was on stem diameter (dbhob), stem straightness scores, branch habit scores (primarily relating to a short-internode light-branching ideal), and malformation, with varying scales used for tree-form traits.

Pairwise genetic correlations between sites were studied, trait by trait, genotype-site interaction being manifested as departures from perfect correlation (r_g = 1) between sites.

Usable data came from nine sites: from New Zealand; from South Africa; and, within Australia, from New South Wales (3), from South Australia, from Western Australia (2), and from Victoria. Such data came from 47 lots at one site and from 126-238 lots elsewhere.

Within sites, repeatability of lot means (h² _g) was good (generally >0.5) for dbhob, stem straightness score, and branch habit score, but sometimes near zero for malformation.

Estimates of between-site genetic correlations (r_g) for dbhob generally exceeded 0.5, but tended to be lower where the two Western Australian sites were involved and were often still lower for the South African site. With an adjustment for assessment-age differences between sites, however, many r_g values were increased but not radically. Some results could be interpreted, but there was often no obvious coherent pattern. For straightness, r_g generally exceeded 0.6, the main exceptions involving one Western Australian site. For branch habit score rg tended to be still higher, except at the South African site where the scoring criterion was rather different. For malformation, r_g was often very imprecise, but mostly very high and positive. For some sites or pairs of sites inclusion of some native-population lots inflated both h² _g and r_g, especially for dbhob. It appears that international exchanges of select genetic material can be beneficial. However, for boosting short- to medium-term genetic gain, imported material should be chosen on appropriate selection criteria, and quite heavy culling from testing may often be needed.

A Corrigendum to this paper is available here: Corrigendum for Burdon, R.D. et al. 1997: International gene pool experiments in Pinus radiata: patterns of genotype-site interaction

Corrigendum for NZJFS 26(3) 370-379: Establishment practices can improve longer-term growth of Pinus radiata on a dry-land hill forest.

The paper which this Corrigendum refers to is available here: Establishment practices can improve longer-term growth of Pinus radiata on a dry-land hill forest

Light microscopy of Pinus radiata D. Don field test stakes (20 x 20 x 500 mm) exposed in wet acidic (pH4.5) soil for 12-24 months showed predominance of an unusual type of decay characterised by tunnelling attack of wood cell walls. After 2 years, decay was moderate to severe in wood treated to ground contact CCA specifications and also equivalent retentions of creosote, and a number of new generation preservatives. Relative to other New Zealand temperate test sites and also an Australian tropical site, the New Zealand acidic soil test site was very aggressive.

Correlative microscopy, including light microscopy (LM), confocal laser scanning microscopy (CLSM), scanning electron microscopy (SEM), and transmission electron microscopy (TEM), was used to elucidate the micromorphology of this attack. Tunnels of diameter 0.2—5 _mm were present throughout all layers of the cell wall, and their orientation was not related to the cellulose microfibril orientation. They also showed no preference for particular cell wall layers, indicating a lignin-degrading capability. CLSM images showed that living, connecting fungal hyphae were present in the cell lumina and tunnels, and this was also confirmed by other forms of microscopy used. This type of attack was predominant in wood that was highly saturated with water whereas wood that was less moist was attacked predominantly by classical white rot at the same site. Ongoing isolation and incubation studies in conjunction with further microscopy should enable identification of the fungal species involved.

Microscopic examinations of CCA-treated Pinus radiata D.Don timbers in an industrial cooling tower in New Zealand showed bacteria and soft-rot fungi to be primarily responsible for the decay of these timbers. Of these micro-organisms, erosion bacteria appeared to be most widespread, attacking wood cell walls independently as well as in combination with tunnelling bacteria and soft-rot fungi. Tunnelling bacteria attacked wood often with soft-rot fungi, and less commonly with erosion bacteria.

Wood samples were taken from various locations in the cooling tower, including spray line support, panelling above the spray drift eliminators, and from spray drift elimination slats. Examination of these samples indicated that, among the microorganisms which attacked the wood, erosion bacteria were most tolerant of oxygen limiting conditions as bacterial erosion was the only type of decay present in wood constantly saturated with water. The evidence provided of the presence of widespread bacterial attacks in an industrial cooling tower timbers is the basis for recognising that bacteria may play an important role in the deterioration of cooling tower timbers.

The Pestalotiopsis anamorph of Lepteutypapodocarpi (Butin) van der Aa, occurring on Podocarpus totara G. Bennett ex D. Don, Podocarpus acutifolius T. Kirk, and a hybrid of P. nivalis W.J. Hooker and P. hallii T. Kirk, is reported as a new fungal record for New Zealand.

In New Zealand, low needle retention in Pinus radiata D.Don is often associated with infection by a needle-cast fungus, Cyclaneusma minus (Butin) DiCosmo et al., and with a condition known as upper mid-crown yellowing (UMCY). Variability in the expression of these two disorders is known to be controlled by both environmental and genetic factors. In order to estimate the degree of genetic control, adjacent seedlings and clonal trees (six clones selected for vigour in the nursery) were assessed at age 22 years for symptoms of UMCY and of C. minus in stands thinned to three final-crop stocking rates. Estimates of genotypic and non-genetic variance and of broad-sense heritability (ratio of genotypic to phenotypic variance) were obtained in two ways: (a) by subtracting the observed within-clone variance from the phenotypic variance of seedling trees, and (b) by directly estimating clone-to-clone variance. Values obtained by these methods agreed well with each other and with earlier estimates made in a nearby seed orchard. Factors related to method of propagation and physiological ageing are therefore unlikely to inflate clonal variation. Clones may be the more cost-effective indicator for UMCY studies. Genotypic variance in resistance to both disorders is considered to be sufficiently high (at least 64% for UMCY and 44% for C minus) to warrant consideration in silvicultural practices.

Rehabilitation techniques for native forest were investigated at an opencast coal mine site in cut-over beech (Nothofagus) forest on alluvial river terraces at Giles Creek near Reefton, Westland. Annual precipitation at the site was 2900 mm and soils were dominantly Allophanic and Acid Brown Soils. The survival and growth of 11 nursery-raised native woody species, and natural regeneration, were examined in three covering treatments consisting of (1) overburden gravel, (2) mixed-horizon forest soil, and (3) layered-horizon forest soil. Underlying gravels were either not ripped or ripped to a depth of 80 cm.

Survival of bare-root and container-grown plants 4.5 years after planting was better in overburden gravel than in both mixed and layered soil, largely because of poor survival of beech species in soil. Plant height growth in overburden gravel, however, was minimal because of nitrogen deficiency. Height growth after 4.5 years in layered soil was nearly twice that in mixed soil, the better growth in layered soil being due to improved drainage and improved nitrogen nutrition. Poor survival of the beeches in forest soil is attributed to root-rot pathogens. Ripping of underlying overburden gravel had no influence on plant survival or growth in any of the three covering materials.

In the fifth year of the trial, ground cover in the two soil treatments was 38-40%, but cover in overburden gravel never exceeded 1 %. Tall-growing adventive rushes {Juncus sp.) dominated in mixed soil, reflecting poor drainage, whereas native and adventive herb species dominated in layered soil. Few native species, especially woody species, were introduced from the original forest through soil replacement. Research is required to determine optimum timing of earthmoving operations in relation to natural seed dispersal.

Estimates of nitrogen input to chronically nitrogen-deficient sand-dunes were made at Woodhill Forest, near Auckland, using a site recently replanted with Pinus radiata D. Don. The study was aimed at screening three legumes—Maku lotus (Lotuspedunculatus Cav. "Grasslands Maku"), hairy canary clover (Dorycnium hirsutum (L.) Ser.), and everlasting pea (Lathyrus latifolius L.)—as possible nitrogen-fixing replacements for yellow tree lupin (Lupinus arboreus Sims). Seasonal biological nitrogen fixation (BNF) of the legumes was compared using the isotope dilution technique and above-ground dry matter productivity. Sampling was conducted in winter (July) and spring (November) of 1994, and summer (February) and winter (July) of 1995.

Dorycnium hirsutum plots contained the highest amount of legume dry matter and this did not differ significantly between seasons. When woody components were omitted, La. latifolius was the most productive species. Dorycnium hirsutum and La. latifolius derived, on average, 98% and 95% respectively of their annual nitrogen uptake from the atmosphere (%Ndfa). This was not significantly affected by season. Lotus pedunculatus, on the other hand, showed a lower %Ndfa in summer.

Lathyrus latifolius was found to contain more fixed nitrogen (214 kg/ha/year) above ground than Lo. pedunculatus or D. hirsutum (55 and 71 kg/ha/year respectively). In all legumes studied, high rates of nitrogen fixation were observed between winter and spring and BNF was highly correlated with dry matter production.

Overall, the results showed that La. latifolius may be considered as a replacement for lupin in the ecosystem studied because of its persistence, its capacity for high non-woody dry matter production, its dependence on nitrogen derived from the atmosphere, and its high nitrogen fixation rate.

Height and basal area data from a series of eight provenance trials of Pinus radiata D.Don planted in New South Wales were analysed at age 8 years. The trials tested subpopulations from the three mainland California natural populations—four from Año Nuevo, six from Monterey, and three from Cambria—as well as a single mixed-family seedlot from an Australian first-generation unculled seed orchard. The test sites varied greatly in soil fertility and climate; mean tree height at age 8 years ranged from 4.7 to 11.2 m. There were significant differences in growth rate between populations on several sites, and overall there was a large population x site interaction. Monterey was usually the best of the natural populations, but Año Nuevo grew better on the highest productivity sites and Cambria grew well on some low productivity sites. Despite the interaction, the seed orchard seedlot was the best population for growth at all sites. There was no evidence of subpopulation-within-population x site interaction and, in contrast to several previous studies, there were significant differences between subpopulations within some populations.

Infection of Pinus radiata D.Don with Agrobacterium rhizogenes strains A4 and R1601 was demonstrated through greenhouse inoculation of seedlings and verification of opine production in gall tissue. Frequencies of seedlings exhibiting gall formation after epicotyl stem inoculation was 4% for strain A4 and 7% for strain R1601 (compared to 59% for the positive control inoculation with A tumefaciens strain 542). Adventitious shoots of 22 P. radiata clones, micropropagated from seeds of the same bulk seedlot from which the seedlings had been grown, differed significantly in susceptibility to A4, Rl 601, and 542 when inoculated in vitro. In vitro shoots were more susceptible to both A. rhizogenes strains but less susceptible io A. tumefaciens 542 than seedlings inoculated in the greenhouse. The highest frequency of shoots forming a gall and/or hairy roots was observed in a different P. radiata clone for each Agrobacterium strain: 22% for A4, 44% for R1601, 58% for 542. Agrobacterium rhizogenes strain R1601 was superior to A4 for infection of P. radiata shoots grown in vitro. Gall production and hairy roots were observed in 11 % and 1 %, respectively, of P. radiata shoots 12 weeks after in vitro inoculation with A. rhizogenes strain R1601. Hormone autotrophic, bacteria-free cultures were established from stem segments that possessed tissue proliferating from Agrobacterium inoculation wound sites. Hairy roots formed directly from the wound site as early as 8 weeks after inoculation. Though opines were difficult to detect in callus tissues resulting from R1601 inoculations, the likelihood of detection was increased if the sample included roots. Cultures of detached hairy roots maintained on quarter-strength modified LePoivre medium reached an average length of 19.5 cm after 1 year and were shown to continue opine expression throughout that time. A greenhouse rooting experiment provided no evidence that rooting of P. radiata tissue culture shoots was improved by inoculation with A. rhizogenes R1601.

Flowering in all plant species analysed to date is regulated by highly conserved and developmentally regulated genes. Some of them were found to belong to the group of MADS-box genes. In the research described here, the construction of two opposing degenerate primers, targeted to highly conserved regions within the MADS box, allowed the amplification of a 78 bp segment from genomic DNA of Pinus radiata D.Don. By subcloning these PCR (Polymerase Chain Reaction) products into Ml3 and analysing the sequence of individual subclones, three different DNA sequences, each representing a conserved MADS-box region of three independent P. radiata genes were identified. The three MADS-box sequences shared nucleotide and amino acid sequence identity with homeotic genes of Arabidopsis thaliana:—two of them with the stamen-carpelspecific AGAMOUS gene and one with the stamen-petal-specific AP3 gene. The sequences isolated here are currently being used to isolate and characterise full-length MADS-box gene sequences which could provide a useful tool for the manipulation of flowering in P. radiata.