Forty-five clones of Pinus radiata D.Don were studied for repeatability (broad-sense heritability) of levels of individual monoterpenes in cortical oleoresin, using hedged cuttings that had come from 7- to 9-year-old ortets. In these cuttings the estimated clonal repeatabilities of levels of all monoterpenes were very high (generally >0.85). It appeared that clones could readily be distinguished in 99% of pairwise comparisons. For 22 of the clones 15-year-old ortets were sampled in the same month. Ramet-ortet repeatabilities were very high (>0.85) for three of the main monoterpene components (_b-pinene, carene, and limonene), lower (c. 0.6) for a fourth (_a-pinene), and lower still for some minor components, but higher repeatabilities were clearly attainable through straightforward refinements of the chemical analysis.

Various issues arose from the results of the large Pinus radiata D.Don provenanceprogeny trial in Kaingaroa Forest which included all five natural populations and two local "land-race" controls. Differences between the five natural populations are multidimensional, the apparent affinities between populations depending strongly on the traits considered. Assigning any given tree in the species unequivocally to its appropriate population by means other than pedigree information is difficult, and at present would require data sets that would be hard to obtain. The local land-race control populations, Kaingaroa and Nelson, appear to have obtained about two-thirds and slightly over half of their genes, respectively, from Año Nuevo and the rest from Monterey. The closer resemblance of Kaingaroa than Nelson to Año Nuevo was very consistent, yet the resemblances of the controls to their progenitor populations evidently varied markedly among traits; it is suspected that local "land-race" stocks, while otherwise quite broadly based genetically, may reflect relatively large genetic contributions from a small number of Monterey ancestors in the original introductions.

Interestingly, all five natural populations have shown very little local differentiation. Heritabilities seem similar trait for trait among the natural populations. Genetic correlation structures also appear similar among populations, suggesting that the correlations are essentially pleiotropic and thus constitute some important breeding opportunities, or constraints in the case of adverse correlations.

The scope for further research includes prospects for using molecular biology to complement the biometrical analyses covered in the preceding papers and to contribute to future genetic improvement.

Potential benefits of using native populations to bolster the breeding programme include: specific attributes that certain populations could contribute in the short term, new combinations of attributes in the medium term, and, in the long term, greater responses to directional selection. Among the specific attributes are the higher wood density and lower stem sinuosity of the Guadalupe population, and certain soil tolerances of the Monterey population. For the specific attributes to be worthwhile, their economic importance must in some situations outweigh the adverse features of native-population material, or the adverse features must be overcome either by heterotic effects in crossing with domesticated stocks or by genetic segregation in advanced-generation crosses. New seed orchard and vegetative propagation technologies facilitate incorporation of naturalpopulation material in breeding lines and some production populations.

Natural-population gene resources should be managed, probably as both pure populations and hybrid combinations. A major unsolved problem for management is pollen contamination.

Cortical oleoresin was taken from branches of young (c. 7-year-old) trees of Pinus radiata D.Don provenances for gas-liquid chromatograph analysis of monoterpenes. Two studies were undertaken: Study I involved 50 unrelated trees of each of the Californian mainland populations (Año Nuevo, Monterey, and Cambria), Guadalupe Island, and Kaingaroa and Nelson from New Zealand; Study II involved 50 trees from each of the Guadalupe and Cedros Island populations. Monoterpene levels were expressed in percentages of total monoterpenes.

Study I resolved seven frequent monoterpene peaks plus several other peaks that occurred infrequently, the levels of each peak varying widely both from tree to tree within populations and among population means. Several components showed indications of major gene effects. Multiple discriminant analysis, despite highly non-normal frequency distributions for most fractions, was a powerful tool for demonstrating patterns of population differences. Canonical variate distances were least between Cambria and Monterey, followed by those between Año Nuevo and Monterey and between Cambria and Guadalupe, suggesting some north-south trend. The analysis showed a striking intermediacy of the New Zealand populations between Año Nuevo and Monterey, corroborating other evidence that New Zealand stock had derived entirely from those two natural populations, and indicating around 66% and 52% of Año Nuevo genes in Kaingaroa and Nelson respectively.

Study II resolved 12 monoterpenes, the mean levels of 10 differing clearly between the two island populations. Monoterpene composition of Guadalupe differed substantially between the studies, pointing to a need to standardise sampling conditions closely, and preventing reliable cross-reference of Cedros with mainland and New Zealand populations. However, Cedros appeared not to fit a north-south trend. From Study II emerged the most definite correlation between levels of different monoterpenes, a strong positive association between sabinene and terpinolene.

Oleoresin was sampled from wood in lower boles of trees in a Pinus radiata D.Don provenance-progeny trial in Kaingaroa Forest in the central North Island of New Zealand, and the _b-pinene/(_a-pinene + _b-pinene) ratios were determined by gas-liquid chromatography analysis. In one block 379 trees were sampled, representing 50 windpollinated progenies from each of two Californian populations (Año Nuevo and Monterey) and two New Zealand ones (Kaingaroa and Nelson); results for 32 and 24 of the progenies from Kaingaroa and Nelson respectively were cross-referenced with determinations on their seed parents. In another block 161 trees were sampled, drawn from 45 Nelson progenies which were all cross-referenced with seed-parent values.

The pinenes were the only major monoterpene components, except in one tree with 22.5% sabinene and 5.6% terpinolene. Populations differed markedly in _b-pinene ratio (p < 0.001), Año Nuevo averaging -80% and Monterey -71%, Kaingaroa and Nelson being intermediate with -75% and -73% respectively. Año Nuevo and Kaingaroa were less variable than Monterey and Nelson.

Individual estimates of narrow-sense heritability were imprecise, but clustered around unity. They did not differ clearly between the Californian and New Zealand populations, or according to whether sib-analyses or offspring-parent regressions were used, although the latter gave better precision. The economic significance as a breeding goal, however, seems negligible.

The proportions of _b-pinene appeared to rise appreciably as trees got older and taller, raising a caveat for comparing different plantings.

Wood properties were studied in _≥4 trees x 2 sites x 50 wind-pollinated progenies x 7 populations in a Pinus radiata D.Don provenance-progeny trial in Kaingaroa Forest in the central North Island of New Zealand. Specimens were taken at around 1.4 m height and contained rings 1-5 from the pith. Oven-dry wood density and ring width were determined throughout, and grain spirality in rings 2 and 5, compression wood, pith diameter, and tracheid lengths in subsamples. These data were analysed along with height and diameter data.

Wood density averaged about 10% higher in the slower-growing island populations (Guadalupe and Cedros) than in the Californian mainland ones. Compared with Monterey, Año Nuevo and Cambria averaged around 1.5% and 4% lower respectively, and the two New Zealand populations (Kaingaroa and Nelson) _≥1 % higher. Grain spirality was worst in Año Nuevo, but comparisons among other populations were complicated by site interaction (in contrast to other traits) and population differences in pith-to-bark spirality gradient. Compression wood was most prevalent in the island populations, and least in those from mainland California. Differences between local subdivisions of native populations were generally negligible.

Density showed within-population coefficients of variation (CV) of 6-7% but ring5 spirality showed CVs of around 50% or more. Variances and narrow-sense heritabilities (h²) appeared similar among populations. Ring-5 spirality gave h² generally well oyer 0.5 within sites, but some unexplained family x site interaction lowered across-sites h². Ring-2 spirality, which was studied within only one site, gave lower h² and families varied appreciably in between-ring spirality differences. For density h² was generally _≥0.65, but with possibly less additive genetic variation in Año Nuevo than in Monterey and Cambria. Family x site interactions appeared minimal. Estimates of h² for density in the New Zealand populations, from sib-analysis and offspring-parent relationships respectively, agreed closely, suggesting essentially random mating.

Wood density showed a negative association (both genetic and within-family phenotypic correlations) with ring width, but a positive association with tree height, a situation that tends to be obscured by the strong association between height and diameter.

Intercorrelations between traits and age-age correlations for some traits were studied to around age 10 years in a provenance-progeny trial of Pinus radiata D.Don on two test sites in Kaingaroa Forest in the central North Island of New Zealand, using seedlings and clones. The main seedling component of the experiment was supplemented by clones representing cuttings taken from c. 55 cm height on seedling ortets. All the natural populations, except for the Cedros Island population, plus two New Zealand controls were included in the clonal adjunct.

Estimates of genotypic correlations from the open-pollinated seedling progenies and the clones agreed well. Additionally, the clones afforded direct estimates of non-genetic correlations (among ramets within clones). These results, and the inherent precision of clonal experiments, indicate that juvenile or near-juvenile clones have major advantages for studying such correlations.

Between-trait genetic correlations appeared similar among populations, and results confirmed a pattern for volcanic plateau sites of generally favourable genetic correlations between growth rate, stem straightness, a wide-angled "multinodal" branching habit, and lack of stem malformation. Among these correlations those involving branch cluster frequency were stronger on the warmer, higher-index, test site, and so a strongly multinodal branching habit appears to be more advantageous there. Tree height at around 10 years was more favourably correlated genetically with tree-form traits than was current stem diameter.

Age-age environmental correlations between successive heights, and between stem diameters and successive heights, fitted a coherent and consistent pattern, but the pattern of the corresponding genetic correlations appeared to depend appreciably on both population and site. Early heights seem to be excellent predictors of subsequent stem diameters, and in the better-adapted material stem diameter tended to be more strongly correlated genetically with preceding heights than with current height. Analysis of covariance, adjusting 8-year diameter for 7-year diameter, appeared to give very sensitive detection of incipient changes in genotypic rankings for diameter, but it is not yet clear how best to use this feature for selection. Stem diameter increment was correlated with initial height and diameter as well as with foliage retention (Cyclaneusma resistance) score, but the specific contribution of resistance could not be closely evaluated. Age-age genetic correlations, between around 8 and 11.5 years, for subjectively scored tree-form traits appeared close to unity. Among juvenile morphological traits, retarded leader showed weak genetic correlation with subsequent tree form, and the onset of production of sealed buds almost zero correlation with growth rate.

A clonal adjunct to a large Pinus radiata D.Don provenance-progeny trial involved six populations (all the natural populations except Cedros, plus two "land-race" populations from New Zealand) x 30 wind-pollinated progenies (families) x two clones nested within each of two sites x four ramets per clone. The clones were cuttings taken from previously hedged 3-year-old ortets at c. 55 cm height, and came from a subsample of the seedling families used in the main experiment.

Genetic parameter estimates from the clones were compared among populations and with estimates from the seedlings. The parameters estimated included broad-sense heritabilities (H²). narrow-sense heritabilities (h²), phenotypic variances (addressed largely as coefficients of variation), alternative estimates of genetic variances (total genetic v. additive genetic), and genetic variances between and within families in clonal material. Genetic correlations between performance in seedlings and cuttings of the same families were also studied.

Phenotypic variances appeared similar between the cuttings and parallel genetic samples of seedlings. Genetic correlations between seedling and clonal performances appeared to be generally high (≥0.75). These results accord with genotypic effects being very similar between the propagule classes, with important implications for selection and predicting genetic gains.

Estimates of H² were very similar, trait for trait, among populations. This, despite some smaller coefficients of variation in New Zealand material, suggested that genetic variance structures are similar between populations. Estimates of h², made under the provisional assumption of random (i.e., 100% half-sib) mating, tended to be lower in New Zealand than in native-population material, particularly for cumulative growth traits and some tree-form traits including butt straightness, stem straightness, and branch habit quality, so that ĥ² tended to vary much more between traits than Ĥ². Such ĥ² values, however, tended to resemble Ĥ² more in the native populations than in the New Zealand. This, and markedly higher ratios of between-family: total clonal variance [equation omitted] in cuttings of native populations, strongly indicate much greater departures from random mating in natural stands than in plantations. Although differences between ĥ² (assuming random mating) and Ĥ² in New Zealand material tended to be greater in the traits previously reported as showing relatively high levels of specific combining ability, they suggested a generally higher proportion of non-additive gene effects than did earlier reports. Also, the ratios of [equation omitted] in cuttings were hard to reconcile with likely mating patterns in natural stands, unless non-additive gene effects were less than suggested by comparing ĥ² and Ĥ².

Within-population variances (ô²) and narrow-sense heritabilities (h²) were estimated in a Pinus radiata D.Don provenance-progeny trial. Samples from the five natural populations and two New Zealand "land-race" populations were included, each as 50 wind-pollinated progenies, on two sites in Kaingaroa Forest in the central North Island of New Zealand. Assessments were made to around 8 years throughout and to 11-12 years in a sample. Estimates of heritabilities provisionally assumed random mating to give "apparent" heritabilities, but a basis was developed for revising estimates in the light of information or assumptions on departures from random mating.

Natural populations usually gave higher apparent h² than New Zealand plantation material, especially for growth traits in island material and when the trees were younger. Population differences in apparent h² mainly reflected differences in between-family ô², but the New Zealand material also showed less within-family ô² for growth traits. These higher estimates of heritabilities and variances in natural populations appear to reflect appreciable non-randomness of mating, including significant inbreeding, and presumably associated contributions of non-additive gene effects. Allowing for likely biases arising from population differences in mating patterns, the different populations appeared to have generally similar variances (or coefficients of variation) and heritabilities for individual traits.

Genotype-site interaction appeared unimportant for the two sites, but there was some obscure family-block interaction which tended to erode the overall heritabilities for the less heritable traits, notably early heights. Pooling subpopulations within natural populations was usually immaterial to h² and ô².

The results indicated the following h² values around age 8: forking and current leader dieback, 0-0.05; stem diameter (and stem volume), butt sweep score, and current retarded leader, 0.1; branch habit quality score, 0.1-0.15; height, 0.1-0.2; stem straightness score, 0.2; branching angle score, 0.2; branching frequency score, 0.2-0.25; branch cluster counts, >0.4; sealed bud scores, 0.5. Coefficients of variation declined with age for height but increased for stem diameter, recent increments tended to be less heritable than cumulative values for height but more so for diameter, and various heritabilities appeared to rise with age. Although experimental conditions were not ideal the 7-9 year h² values concur well with those in other reports, particularly in respect of comparative heritabilities among traits. Heritabilities in select material thus appear to have been similar to those in both native populations and the New Zealand land races that haveserved as base populations.

Pinus radiata D.Don provenances (populations) were compared to age 12 years on two contrasting sites in Kaingaroa Forest in the central North Island of New Zealand. Differences between populations were detected in most traits, including growth rate, tree-form variables, other morphological traits, and susceptibility to fungal diseases, but variation in economic traits overlapped strongly among almost all populations.

The Californian mainland populations (Año Nuevo, Monterey, and Cambria) showed very similar growth, although for diameter Monterey had evidently begun to draw ahead and Cambria to fall behind. Trees of the Guadalupe Island population were on average about 10% shorter than Monterey; Cedros Island averaged about 30% shorter, and could not compete. The New Zealand populations (Kaingaroa and Nelson) grew more uniformly and averaged about 10% taller and 20% greater diameter than Monterey; Kaingaroa slightly outgrew Nelson. Resistances to dieback and needle-casts were highest in the New Zealand populations, followed by Año Nuevo and Monterey. Cedros appeared to be the worst adapted to local conditions.

Año Nuevo generally showed the worst tree form in stem crookedness, branching, and forking. Kaingaroa was superior, particularly as the trees got older. Cambria had relatively straight butts and boles, but became otherwise inferior. Guadalupe tended initially to have straight stems and tidy "multinodal" branching, but this superiority declined with age. In Cedros, the competition masked an inherently squat growth habit. Population rankings for branch-cluster frequency changed appreciably with age.

Population-site interaction as such was mostly negligible, the main interactions evidently reflecting specific establishment conditions. Local differentiation within mainland populations was usually minuscule. Variation among natural populations was complex and multidimensional, resemblances between populations differing among traits. The Monterey and Año Nuevo populations appeared the most similar, while the island ones, especially Cedros, appeared quite distinct. The island populations were distinguished by thin unfissured bark and binate fascicles. The New Zealand populations evidently derived from Año Nuevo and Monterey, Kaingaroa showing a clearer preponderance of Ano Nuevo ancestry than Nelson. Possible reasons for the superiority of New Zealand material included a release from the "neighbourhood inbreeding" of natural stands.

A provenance-progeny trial of Pinus radiata was planted in three stages (1964, 1965 and 1967) on two sites in Kaingaroa Forest in the central North Island of New Zealand, to study the quantitative genetic architecture of the species and establish a gene resource. Site A (320 m altitude) was generally sheltered with a soil with a deep mantle of various layers of rhyolitic ash capped with 30 cm of raw, basaltic Tarawera scoria. Site B (525 m altitude) had a soil about 3-4 m deep of various rhyolitic ash layers overlying partly welded ignimbrite and with a slight pan about 25 cm below the surface associated with the Taupo ash layer. The trial contained 50 open-pollinated families (progenies) of each natural population (Año Nuevo, Monterey, Cambria, Guadalupe Island and Cedros Island), representing the full natural range of this species, plus two New Zealand populations (Kaingaroa and Nelson), with 20 seedlings per progeny per site. The seedlings were supplemented in the final planting by four juvenile clones from each of 30 progenies from each population except Cedros. Within each of the six site/year-of-planting blocks, randomization was complete except for an interlocking block feature. This layout, a variation of non-contiguous plots, was designed for 50% systematic diagonal thinnings that would leave a balanced classification. The interlocking block layout showed some significant advantages. Its overall success, however, was limited through a combination of very adverse factors: the range of growth rates among populations; intense assertion of crown dominance and the associated stem mortality which were accentuated by needle-cast diseases; and prevalent stem malformation.

Introduction to special issue: Genetic survey of Pinus radiata in New Zealand.

Review of "Armillaria root disease" by Charles G. Shaw III and Glen Kile.

Review of "The conifer manual, vol. 1" by H. J. Welsh.

Kraft fibre, pulp, and handsheet properties for 11 standard Pinus radiata D. Don wood samples were compared with four corresponding samples of the progeny of Clone 850-55 (C55). The standard samples included 9-, 12-, 15-, and 18-year-old toplog and thinnings, and three slabwood samples. Two toplog and two slabwood C55 wood samples were included in the comparison.

It is generally accepted that the basic density of standard New Zealand P. radiata wood chips is strongly correlated with pulp coarseness and handsheet properties. For the C55-toplog pulps such relationships held for handsheet properties but not for pulp fibre coarseness. For the C55-slabwood pulps the converse occurred since standard slabwood and C55-slabwood pulp coarseness and handsheet properties were similar but chip basic densities were very different.

C55 kraft fibres, at a given pulp coarseness or fibre wall area, are proportionately large (broad) and thin walled compared with standard P. radiata kraft fibres. Thus, C55 fibres can be expected to collapse and bond readily one to another and develop high sheet tensile strengths and densities with minimal refining. The handsheet properties obtained supported such interpretations for the C55-toplog pulps, but not necessarily for the coarse long-fibred C55-slabwood pulps. The kraft pulp and fibre properties of the progeny of Clone 850-55 were so unusual that they evoke the question "Is this clone a genetic misfit?"

Fibre cross-section dimensions correlated with pulp coarseness more strongly for dried and rewetted pulps than for undried (never dried) pulps. Fibre wall substance densities and porosities are apparently more variable in undried pulps than in dried and rewetted pulps.

In common with other plantation softwoods, Pinus radiata D. Don exhibits a characteristic corewood zone which contributes significantly to the overall variation in wood properties. Industries have developed preferences for wood with a high or low proportion of corewood and may experience problems if the desired levels are not achieved. With evolving forest management practices worldwide, the trend is to a greater proportion of corewood and this has implications for industries which consider this type of wood undesirable.

Traditional definitions of corewood are based on qualitative assessments of the number of rings from the pith at which an important property (usually wood density) becomes "mature". Since this is an interpretation of a biological pattern, subject to fluctuation in the absolute level under the influence of site and genetic factors, species differences can be large and there is often little relation to wood product performance criteria. An alternative definition based on wood density goes some way towards a technical description of the absolute wood quality.

Microfibril angle was measured for every growth ring at butt, breast height, and 7,12, 18,23, and 30 m height in five 22-year-old trees of Pinus radiata D. Don. Mean angles varied from 9° to 55° with the highest angles occurring in the corewood of the butt log. Angles showed a curvilinear decline from pith to bark, which was more pronounced at the butt end of the stem. Significant variation was observed among trees and growth rings, with an interaction between rings and trees indicating some variation in pith-tobark trends among trees. Microfibril angle declined rapidly with height within the tree, reaching more or less constant values beyond 7 m height, followed by small increases in the corewood of the top log.

A Corrigendum to this paper is available here: Corrigendum: Within- and between-tree variation in microfibril angle in Pinus radiata

Increment core samples were collected at breast height (1.4 m) from 20-year-old trees of 30 of the top-ranked wind-pollinated progenies from the '268' series of Pinus radiata D.Don plus trees, and their wood properties were examined. The objectives were to establish the extent of any possible shift in average wood properties caused by intensive selection for growth rate and stem straightness, and to estimate genetic parameters (variances, narrow-sense heritabilities, and genetic correlations). Samples were also collected from control trees, representing a "felling select" seedlot, within the same progeny trial.

Properties examined included growth rate, wood density, and resin content (all by five-ring core segments), heartwood percentage and compression wood rating (whole cores), and tracheid length (outerwood only).

Although wood density was variable (coefficient of variation about 0.07), the average for the progenies overall was almost identical to that of the control trees. On the other hand, there were slight indications that, along with the improved growth rate of the progenies, compression wood and heartwood resin content levels increased whereas tracheid length decreased. Heritability estimates for all the actual wood properties were high, ranging from 0.4 upwards, and coefficients of variation for resin content variables and heartwood content were very high (0.4 or higher), suggesting good prospects for genetic manipulation.

In a modification of the ratio method for estimating forest biomass, the relationship between the independent variable and the auxiliary variable is linearised prior to calculating the ratio. This is achieved by estimating an exponent for the auxiliary variable by applying a logarithmic regression to data obtained with a sample. This method is mathematically equivalent to the method of estimation based on logarithmic regression where the bias is corrected by a ratio method. A comparison was made of the simple ratio method and the modified method by using simulated samples from seven populations of Pinus radiata D. Don for which total biomass was known. There was little difference between the two methods when simple random samples or samples with probability proportional to the sum of sectional area (an unbiased method) were taken. With samples taken with probability proportional to size (PPS) the results with the simple ratio method were highly biased, thus counteracting gains in precision. With the modified method bias remained small while precision and accuracy were substantially improved. No consistent improvement over the modified method was obtained when Horvitz-Thompson estimators or the mean of ratios method were used. It is recommended that the modified method of estimation combined with sampling with probability proportional to size be used for future estimations of forest biomass.

If temperature increases occur during the next century as predicted by climatologists, then major changes in New Zealand's natural and plantation forests can be expected. Current relationships between temperature and occurrence of natural forest species suggest major changes in forest pattern with an increase in temperature. Because of widely varying relationships between species and climate, changes will likely occur at a species, rather than at a community level. Initial changes should favour species with wide tolerance to climatic factors, good dispersal capacity, and short generation times.

Shifts in distribution will be influenced by the availability of sites for colonisation, i.e., they will depend on mortality of the current site occupants. Disruption in forest composition is most likely to occur in forest patches where there is little potential for dispersal of new, more suitable species. This could result from a narrow, within-patch temperature range, geographic isolation from other patches which might act as seed sources, or dependence on longer-distance latitudinal rather than shorter-distance altitudinal migration.

The plantation estate across the country can be classified in relation to present-day temperature and rainfall regimes. The mean annual temperature is between 10° and 15°C for 90%, and between 10° and 11°C for 35% of the plantation area. Mean annual rainfall is between 1250 and 2250 mm for 75% of the plantation area. If the worst-case climate change scenario is realised, then 96% of the plantation area will experience mean annual temperatures above 13°C. Twenty-one percent of the plantation area will experience mean annual temperatures above 17°C which is near the top of the optimal temperature

range for growth of Pinus radiata D.Don, and warmer than any current values in the present climate. The effect of this on timber yield and log quality is still uncertain. Recommendations for future research include an extension of the current analysis to a national scale for natural forests, and an urgent need for process studies to investigate the effects of climate variables, particularly temperature and carbon dioxide, on the physiological response of all forest species. This is required as a basis for further modelling of the long-term ecological effects.

A Corrigendum to this paper is available here: Corrigendum for Whitehead, D. et al. 1992: How will New Zealand's forests respond to climate change? Potential changes in response to increasing temperature

Data from 2547 randomly selected consignments of part-container-load cargo imported into New Zealand were analysed to examine the incidence of quarantine interception of material that may affect forests, trees, and wood products in this country. Insect damage was found in 4.1 % of the consignments, bark in 3.7%, insects in 2.7%, and fungi in 0.7%. Stone and slate, and goods in the "unknown" category had the highest rate of interception, as did other cargo packed in pallets, cases, and crates. Chemicals, food, textiles, cartons, and rolls of cloth or carpet backing had a low interception rate. Goods originating in Asia and Europe were more frequently infested than American or Australian cargo.

One-year-old Pinus radiata D.Don seedlings were raised at Rangiora Nursery, Canterbury, and subjected to five storage treatments. Seedlings were lifted and packed into waxed cardboard planting boxes which were stored in transporting/ storage crates, or under tarpaulin tents, or in a cool-store for up to 12 days. The crates and tarpaulin covered boxes were stored either on an exposed site or under heavy shade in an adjacent sheltered position. Seedlings were sampled daily from the two treatments on the exposed site, and after 1, 6, and 12 days from the other treatments. Assessments included stem water potential, root growth potential, and growth performance after 24 months in a field trial.

Seedlings stored on an open exposed site under a tarpaulin tent had unacceptable water potentials by 6 days. Seedlings under heavy shade had significantly better water potentials than seedlings on the open exposed site. Cool-stored seedlings had the best water potentials after 12 days' storage. Similar results were recorded for root growth potential.

Height increment was best with cool-stored seedlings, followed by seedlings stored in crates, and then by seedlings stored under a tarpaulin. Height increment fell significantly with longer storage time, more negative water potential, lower root growth potential, and accumulated storage time above 2°C. The results highlighted the importance of ensuring that seedlings are firstly packed into a crate system and then stored (where cool-stores are not available) under sheltered heavy shade to minimise temperature rises above 2°C within the boxes of seedlings.

The herbicide hexazinone was applied in three trials to evaluate the effects of controlling grass, dolly bush (Cassinia aculeata R. Br), and woody weeds in Pinus radiata D.Don plantations. Control of grass on an ex-pasture site strip-sprayed with hexazinone at 2 kg a.i./ha significantly improved the survival and early height growth of P. radiata. Grass control was essential to the commercial viability of plantations on this site. Low tree survival on unsprayed plots resulted in poor stem form and heavy branching, as well as significantly reduced volume at age 16 years. On a dolly bush site, strip-spraying with hexazinone at 4 kg a.i./ha resulted in reduced weed growth and significantly higher tree volume growth at age 14 years compared with unsprayed or 2 kg treatments. On a series of sites with various woody and herbaceous weeds including eucalypts and acacias, strip-spraying with hexazinone at 3 kg a.i./ha prior to planting gave good control of weeds and provided significantly increased early tree height growth. In this trial, control of woody weeds resulted in a significant increase (about 30%) in volume of P. radiata at age 14 years. This increased volume growth meant that commercial thinning to release final-crop stems could be carried out at least 2 years earlier than on untreated stands.

Economic projections of the growth gains to the end of rotation indicate that the initial cost of herbicide treatment was economically justified on these sites.

Historical use of fertiliser to encourage the productivity of pampas grass (Cortaderiaselloana (Schult.) Asch. et Graeb. and C. jubata (Lem.) Stapf) in agriculture suggested that growth of this aggressive forest weed may be encouraged by fertiliser treatment of Pinus radiata D.Don. In glasshouse trials, pampas showed a positive growth response to (a) ammonium nitrate added to a coastal sand, and (b) sodium dihydrogen phosphate added to a silty clay loam soil. Response (b) was increased when ammonium nitrate was also present. These results were consistent with field observations of increased pampas frequency associated with superphosphate treatment of P. radiata on a clay soil, and with reduced tree growth response to calcium ammonium nitrate treatment on a coastal sand. A glasshouse trial showed that pampas residues resulting from herbicide treatment can have a negative effect on P. radiata productivity. Field evidence suggested that fertiliser nitrogen immobilised by decomposing pampas residues was released gradually and utilised by the trees up to 4 years after treatment. Where pampas growth is uncontrolled, a large proportion of the fertiliser intended to promote P. radiata growth may never reach the trees.