Search

Premise of the Study: The capsular-fruited genus Metrosideros (Myrtaceae) is one of the most widely distributed flowering plant genera in the Pacific but is extinct in Australia today. The center of geographic origin for the genus and the reason for and timing of its extinction in Australia remain uncertain. We identify fossil Metrosideros fruits from the newly discovered Golden Fleece fossil flora in the Oligo-Miocene of Tasmania, Australia, shedding further light on these problems. Methods: Standard paleopalynological techniques were used to date the fossil-bearing sediments. Scanning electron microscopy and an auto-montage camera system were used to take high-resolution images of fossil and extant fruits taken from herbarium specimens. Fossils are identified using a nearestliving-relative approach. Key Results: The fossil-bearing sediments are palynostratigraphically dated as being Proteacidites tuberculatus Zone Equivalent (ca. 33-16 Ma) in age and provide a confident Oligo Miocene age for the macrofossils. Two new fossil species of Metrosideros are described and are here named Metrosideros dawsonii sp. nov. and Metrosideros wrightii sp. nov. Conclusions: These newly described fossil species of Metrosideros provide a second record of the genus in the Cenozoic of Australia, placing them in the late Early Oligocene to late Early Miocene. It is now apparent not only that Metrosideros was present in Australia, where the genus is now extinct, but that at least several Metrosideros species were present during the Cenozoic. These fossils further strengthen the case for an Australian origin of the genus.

The risk posed to New Zealand native flora by the recently-established pest of Australian origin Uraba lugens Walker (Lepidoptera: Nolidae) (gum leaf skeletonizer) was assessed. Weed biological control host range testing methods were applied to identify those New Zealand plant species potentially at risk. Native plants tested were primarily in the Myrtaceae, the family that contains all the Australian hosts of U. lugens. 2 Experimental methods included no-choice larval feeding assays and field cage and laboratory oviposition trials. Difficulty in ascertaining reliable oviposition preference data from cage oviposition trials limited the confidence with which the invader's field host range could be predicted. 3 Field surveys of plants attacked by U. lugens in the infested area supported the initial predictions but only some of the at-risk native Myrtaceae were present in the area. 4 The risk to native New Zealand plants is presented in terms of two mechanisms: development of self-sustaining populations and temporary spill-over of solitary larvae. Development of self-sustaining populations of the pest within native forests is deemed to be highly improbable. 5 Temporary spill-over impacts are most likely in urban areas within mixed species plantings or boundaries between native/exotic forests and coinciding with the mobile larval life stages. Spill-over impacts from U. lugens have only been recorded to date on Metrosideros excelsa.

A new species of fly, Fergusonina metrosiderosi Taylor sp.n. (Diptera: Fergusoninidae) [and its associated nematode, Fergusobia pohutukawa Davies sp.n. (Tylenchida: Neotylenchidae)], is described from unilocular basal shoot-bud galls on Pohutukawa (New Zealand Christmas Tree), Metrosideros excelsa Banks ex Gaertn. (Myrtaceae). The fly is characterized by the short, sclerotized postabdomen of the female, distinctly grey wings, reduction in length of the reclinate orbital, postocellar and vertical setae, undifferentiation of the ocellar setae, supernumerary setation of the genae, and three pairs of scutellar setae rather than two. The dorsal shield is restricted to a few raised sclerotized spicules on the first to third thoracic segments of the larva and epidermis of the puparium. The nematode is characterized by the combinations of arcuate shape, small ‘a’ ratio, swollen cuticle, small cephalic region with elevated circum-oral area with a flat tip, coarse annulation and relatively slender conical tail of the parthenogenetic female, and coarse annulation, long bursa and angular spicule with short, broad, offset manubrium of the male. Fergusonina metrosiderosi sp.n. and Fergusobia pohutukawa sp.n. represent the first record of the fly-nematode association from the host genus Metrosideros, and the first record from New Zealand. Aspects of the biology and biogeography of the fly-nematode association are discussed.

Nutrient resource availability and host-plant foliar pubescence both influence arthropod food webs, but multifactor studies are needed to under- stand their interdependence and relative importance. Arthropods were sampled by clipping foliage from Metrosideros polymorpha (Myrtaceae) trees of pubescent, glabrous, and intermediate leaf forms on fertilised and unfertilised plots. 2. Fertilisation decreased leaf mass per area (LMA) but did not change the relative mass of pubescence within leaf morphological classes. 3. Fertilisation increased densities of individuals in four taxonomic orders, densities of individuals and species of all trophic levels, and the biomass of Collembola and Homoptera. Herbivore relative diversity (Shannon H 0 ) also increased with fertilisation, but detritivore diversity declined due to increasing dominance of Salina celebensis (Schaeffer) (Collembola). 4. Detritivore density, driven again by S. celebensis, increased with decreasing leaf pubescence, but Heteroptera and Acari were most abundant on the inter- mediate pubescence class, and Psocoptera density and biomass increased with increasing pubescence. Trophic-level species density did not change with leaf morphological class, but relative diversity of all arthropods and of detritivores increased with increasing pubescence. 5. Both resource availability and leaf pubescence affected Metrosideros arthro- pod communities. However, the pervasive positive influence of fertilisation did not translate to compositional shifts, and there were no interactions with leaf morphological class. In contrast, the effects of leaf pubescence on arthropod density, biomass, and diversity were more restricted taxonomically, and non- parametric MANOVA and redundancy analyses demonstrated significant differen- tiation in community composition on the pubescent morphology.

Carnivore (top-down) and resource (bottom-up) influences in food webs are strong and pervasive, but few studies have investigated their interactive effects in species-rich terrestrial ecosystems. This study focused on arthropods associated with the dominant tree species, Metrosideros polymorpha (Myrtaceae), in Hawaiian forests. Severe soil nutrient limitation on a 120-yr-old lava flow was removed by fertilization and combined with bird predator exclusion cages in a randomized block design. Arthropod densities were measured from clipped foliage at the outset and conclusion of a 33-mo experiment, and their biomass was estimated using regression equations. Metrosideros foliar nitrogen, tree growth, and biomass increased directly in response to fertilization but did not change with bird exclusion. Fertilization increased detritivore densities but not biomass, and both density and biomass of herbivores, while bird exclusion increased both density and biomass of carnivores. Fertilization also increased spider density and biomass, but bird exclusion increased spider numbers (15 species) only in high resource plots. Overall, trophic level biomass responses were less pronounced than density because smaller bodied individuals responded more to enrichment. Bottom-up factors controlled basal trophic levels, and detritivores comprised the largest fraction of arthropod density and biomass. Conversely, top-down impacts were apparent but variable, limited to higher order consumers, and did not cascade to the level of primary producers. These experimental results were consistent with the view that complex forest ecosystems are structured on a bottom-up template.

Aim The majority of Hawaii's plants are derived from the south and west of the Pacific rim, suggesting that at least some of the founders have crossed the climatic discontinuity at the equator. However, the pathways of these dispersal events are not well understood. We sought to elucidate such a pathway for the plant genus Metrosideros Banks ex Gaertn., which appears to have colonized Hawaii from New Zealand. Location We have previously analysed the internal transcribed spacer (ITS) sequence of nrDNA for thirty-six taxa of Metrosideros subgenus Metrosideros distributed across Melanesia and Polynesia. Further we collected two taxa from the Austral and Marquesas Islands. Methods Phylogenetic analysis was undertaken using both the ITS and external transcribed spacer (ETS) regions of nrDNA for twenty-three taxa from Polynesia and East Melanesia. Results The increased resolution achieved by combining ETS data with that of ITS allowed discrimination of two subclades in eastern Oceania. One subclade groups all the Hawaiian taxa together with M. collina (J. R. & G. Forst.) A. Gray from the Marquesas Islands. Main conclusions From an origin in New Zealand, the woody angiosperm Metrosideros is likely to have crossed the equator to Hawaii using a staging point in the remote Marquesas Islands. Colonization of Hawaii by the wind-borne seed of these plants has been achieved against the flow of the prevailing Trade winds north of the equator. A possible climatic mechanism for this dispersal is described.

Premise of the study: Primers were developed to amplify low-copy nuclear genes in Hawaiian Metrosideros (Myrtaceae). Methods and Results: Data from a pooled 454 Titanium run of the partial transcriptomes of four Metrosideros taxa were used to identify the loci of interest. Ten exon-primed intron-crossing (EPIC) markers were amplified and sequenced directly with success in Metrosideros, as well as in a representative selection of Myrtaceae, including Syzygium, Psidium, and Melaleuca for most of the markers. The loci amplified ranged between 500 and 1100 bp, and up to 117 polymorphic sites were observed within an individual gene alignment. Two introns contained microsatellites in some of the species. Conclusions: These novel primer pairs should be useful for phylogenetic analysis and population genetics of a broad range of Myrtaceae, particularly the diverse fleshy-fruited tribes Syzygieae and Myrteae.

Cladistic analyses are presented of matK sequence data as well as a nonmolecular database for an identical set of exemplar species chosen to represent the core genera or groups of genera in Myrtaceae. Eleven robust clades are recognized on the molecular data. Polyphyly of the previously recognized Metrosideros and Leptospermum alliances is confirmed, and several smaller informal taxonomic groupings are recognized from among the members of the former alliance, i.e., the Tristania, Tristaniopsis, Metrosideros, and Lophostemon groups. The nonmolecular analysis provides only limited resolution of relationships. A degree of congruence exists between the two analyses in that two separate fleshy-fruited clades, the Acmena and Myrtoid groups, are identified, as are the Eucalypt and Tristania groups, and Psiloxylon and Heteropyxis are the first lineages to diverge in both analyses. A combined analysis recognized all 11 clades that received strong support from the molecular data. A high level of homoplasy is revealed in many of the nonmolecular characters when they are examined against the combined estimate of phylogeny.

New Zealand pohutukawa (Metrosideros excelsa), a member of the Myrtaceae, is a large, mass-flowering tree endemic to northern New Zealand coastlines. Mainland populations have been reduced to fragmented stands, and the original suite of bird pollinators has been largely replaced by introduced species. The native pollinator fauna on several offshore islands is largely intact and includes three species of the New Zealand honeyeaters (Meliphagidae) and native, solitary bees. We estimated multilocus outcrossing rates for three mainland and two island populations and found that they were among the lowest in the Myrtaceae (t(m) = 0.22-0.53). The shift in pollinators had no measurable effect on the mating system. Mass-flowering facilitates geitonogamous selfing, and inbreeding depression in seedling height was detectable at 6 mo of growth. F(s) [Wright's (1965) Fixation Index] was consistently higher than F(m) in all populations, indicating that selection may eliminate selfed offspring from populations prior to achieving reproductive maturity. Results suggest that increased selfing in mainland populations due to pollinator changes is not responsible for current patterns of poor regeneration of this species.

Metrosideros bartlettii (Myrtaceae) is a distinctive and extremely rare tree, endemic to New Zealand, first discovered in 1975. Prior to this study, a total of 19 adult individuals of the species had been reported; these are located in three small forest remnants in the far north of the North Island of New Zealand. Here we describe a total of 31 adult M. bartlettii at the three sites, including 12 individuals newly discovered by us. We analyse the genetic diversity of the species, using microsatellites to examine the chloroplast genome and amplified fragment length polymorphisms (AFLPs) to monitor nuclear variation. The results clearly demonstrate that M. bartlettii is a unique species, distinct from its two closest relatives M. robusta and M. excelsa. Analysis of genetic diversity within the 31 remaining individuals of M. bartlettii showed an average heterozygosity (H) of 0.18 and a proportion of polymorphic genes (P) of 0.44. Population structure, as shown by 286 AFLP loci, varied between the three geographical sites; the site with fewest individuals, containing two trees, showed some separation from the populations at the other two locations. These two latter sites, by contrast, had highly overlapping AFLP population diversity profiles. The implications of these results for conservation of the species are discussed.

We recorded the times of flowering, fruiting, and leafing for ten native canopy and subcanopy trees and shrubs (monthly from December 1994 through December 1997) in a montane cloud forest with relatively aseasonal rainfall on the island of Maui, Hawaii. These species represented the great majority of individual woody plants at the site. Flowers and fruits were available in the community year-round; however, all species exhibited annual patterns of flowering, and four species showed annual patterns of fruiting while the rest fruited in supra-annual patterns. Many species had protracted flowering or fruiting peaks, and some bore small numbers of flowers or fruit year-round. Most species flowered in a monthly peak mainly between May and August, corresponding to the period of greatest solar irradiance and marginally higher temperatures. Fruit ripening followed at varying intervals. In contrast, the heaviest flowering occurred between November and March, resulting from bloom of the dominant tree, Metrosideros polymorpha. At the highest elevations, Metrosideros flowering was heaviest during September, but peak flowering of lower elevation trees occurred in late fall and winter. Two forms of this species differed in their temporal and spatial patterns of flowering. For M. polymorpha var. polymorpha and var. incana, bloom peaked annually between November and January; however, for M. polymorpha var. glaberrima, flowering peaked from April through July, with an earlier secondary peak in January.

We developed 15 polymorphic microsatellite loci from the Hawaiian tree Metrosideros polymorpha. These loci were screened against two varieties from several populations and from 23 individuals from one mid-elevation population on Hawaii Island. Loci were variable with the number of alleles per locus ranging from three to 24. Polymorphic information content ranged from 0.222 to 0.941, and observed heterozygosity ranged from 0.261 to 0.955.

Summary Dicranopteris linearis (Gleicheniaceae), a native fern common throughout the Old World tropics and Polynesia, forms dense thickets > 3m deep over large areas of open-canopy, oligotrophic, wet Hawaiian rainforests. Our objectives were to identify leaf- and whole plant-level traits that are key to its success and to determine its community-and ecosystem-level consequences in primary successional sites. Along an elevational gradient from 90 to 1660m, mean maximum net assimilation rates of Dicranopteris ranged from 2.9 to 5.0 μmol m−2 s−1, compared with 3.6–9.5 μmol m−1 s−1 in the codominant tree Metrosideros polymorpha. Gas-exchange characteristics did not explain Dicranopteris' success, nor its trends in production. However, indeterminate, clonal growth form, shallow rhizomes, marcescent leaves with low decomposability, and a mat-forming capacity enabled Dicranopteris to colonize sites and to maintain dominance via high effective leaf area, despte its low biomass. Phosphorus use efficiency, which reached 24 kg g−1, was exceptionally high, allowing colonization of phosphorus-poor sites. Dicranopteris contributed up to 74% of above-ground net primary productivity in a site where it contained only 14% of live biomass. It accounted for up to 57% and 47% of total nitrogen and phosphorus uptake by plants, respectively, where it contained only 24% and 30% of plant nitrogen and phosphorus. Dicranopteris leaves are short-lived and slow to decompose; thus, fixed carbon is transferred quickly to soil detrital pools where it contributes to aggrading soil organic matter pools and may exacerbate oligotrophic conditions, thereby strongly influencing soil genesis and ecosystem development. The fern therefore influences forest-floor light regimes and directs later community development. An exclusion experiment demonstrated that Dicranopteris competed with Metrosideros, but lack of revegetation in 40% of the exclusion area after 39 months showed that Dicranopteris also colonized microenvironments unavailable to its endemic codominants. Dicranopteris may play an important role in resisting invasions of exotic species into Hawaiian rainforests.

Non-native perennial grasses form 30% of the live understory biomass in seasonally dry, submontane forests in Hawaii Volcanoes National Park, yet their effects on native species are unknown. We removed these grasses from plots of 20 m × 20 m in 1991 and maintained removal and control areas over the next three years. Two fast growing shrub species, Dodonaea viscosa and Osteomeles anthylidifolia, increased in size significantly more in removal areas than in controls. Individuals of the most abundant shrub species, Styphelia tameiameia showed no net growth response to grass removal. They did, however, change their architecture: many branches along the mid and upper sections of the main trunk died and a proliferation of new leaves and shoots occurred in the lower 40 cm of trunk. Basal diameter increase was very small in Metrosideros polymorpha, the dominant tree species in these sites. All species except Styphelia had significantly increased leaf tissue nitrogen in removal plots by 18 months after removal when compared to shrubs in control areas suggesting that removal plot shrubs had greater access to soil nitrogen. Available soil-N pools, which were generally higher in the removal plots, support this interpretation. Light levels near the soil surface were also higher where grasses were removed than where they were present which may have contributed to increased shrub growth. By contrast, soil moisture was consistently lower where grasses were removed than where they were still present. Shrub tissue carbon isotope values were consistent with the interpretation that shrubs in removal plots had less rather than more water available to them. Hence, the increased growth observed in removal plot shrubs could not be due to release from moisture competition. Lastly, our results showed that seedlings of all woody species except Metrosideros were significantly more abundant in removal plots at both one and three years after removal and initially high sapling mortality was balanced by high recruitment into the sapling class. We believe that over time this will result in increased densities of native shrubs if grasses are kept out. With the presence of grasses, shrub growth in these woodlands is reduced and biomass is shifting towards grasses.

An earlier study suggested that soils on the windward slope of Mount Haleakala are excessively waterlogged at low altitudes but become better drained upslope. We analyzed altitudinal changes in soil N turn over, vegetation mass, and foliar chemical composition of the dominant canopy species, Metrosideros polymorpha (Myrtaceae), on this slope. The basal areas and DBH2× height (an index for volume) of woody species ≥2 m in height increased with altitude (as opposed to a general trend elsewhere), peaked at 1800 m elev., and abruptly declined above this altitude. Foliar N and P concentrations of Metrosideros showed a similar altitudinal trend with DBH2× height values. Foliar Fe and Al concentrations steadily increased downslope, but were low again at 450 m, the lowest altitude, contrary to expectation based on waterlogging. Toxic forms of Fe and Al may retard root penetration in waterlogged soils and thus the uptake of these elements. The net N mineralization rate for topsoils incubated in situ was 78.6 μg/g/20d at 1800 m elev., while small amounts of N were immobilized at 450, 1200 and 2200 m elev. The net N mineralization rates of the topsoils, which were air-dried to their permanent wilting points and incubated at 26°C, were greater than those of the wet samples at 26°C, only for the lowest altitude. Conversely, the net N mineralization rates of the topsoils which were kept wet and incubated at the higher temperature of 26°C were greater (P < 0.1) than those of the in situ wet samples only for 1800 m elev. The rates were the same for 450 and 2200 m elev. where the soils were waterlogged and very dry, respectively; thus, N turn over was more strongly limited by moisture than by temperature in these cases. Nutrient availability improved with altitude up to 1800 m elev. in association with the moisture gradient, and appeared to be responsible for the observed pattern in vegetation mass.

Hawaiian forests are subject to the effects of periodic hurricane conditions. Hurricane Iniki struck the island of Kauai, Hawaii on September 11, 1992 with winds exceeding 200 km/h and caused defoliation, felling of trees by snapping and uprooting, and standing tree mortality due excessive limb and leaf loss. The purpose of this study was to evaluate if measured wood mechanical characteristics could be correlated with stem failure of trees under windstorm conditions. A field survey indicated that post-hurricane stem condition (snapped, uprooted, or standing) differed among five common canopy species and was significantly correlated with stem apparent elastic modulus (relative flexibility). Species that tended to snap had significantly higher apparent elastic moduli than those that remained standing or were uprooted. Wood density and stem diameter were not significantly related to stem failure mode. Native trees had a higher percentage per species of standing individuals but also had increased uprooting. Nonnative tree species were more often snapped and fewer were standing after the hurricane. The higher incidence of stem failure for introduced canopy trees may increase the spread of alien understory species following wind disturbance events. These relationships provide a simple means to predict relative differences in stem failure due to high wind conditions and should be considered in planning reforestation efforts on the Hawaiian Islands.

Throughout the 1960s and 1970s there was a rapid decline and canopy dieback in the Metrosideros polymorpha dominated rain forest of Hawai'i. An analysis of air photo sets from 1954, 1965, and 1972, covering the windward slopes of Mauna Kea and Mauna Loa, gave support for an alien disease hypothesis. A total demise of the native forest was predicted for the early 1990s. This prediction as well as the disease hypothesis proved to be wrong. Various searches for a single climatic cause also failed to explain the dieback. The spatial dynamics of the dieback phenomenon were newly analyzed with an additional air photo set from 1977 and by using GIS with spatial statistics. Two juxtaposed and climatically similar landscape matrix samples of ca. 200 km2, one each on Mauna Loa and Mauna Kea, were subjected to an analysis of landform heterogeneity and superimposed dieback patterns. The Mauna Loa matrix displays up to 15 000 yr old lava flows, while the Mauna Kea matrix displays up to 250 000 yr old substrates. Initiation of dieback occurred simultaneously on both mountains and was highly correlated with poorly-drained sites. The progression of dieback, however, followed a gradient of decreasing soil moisture, which often terminated at clearly recognizable substrate boundaries in the Mauna Loa matrix and moved over well-drained hill sites in the Mauna Kea matrix. Metrosideros dieback spread across the entire spectrum of volcanic substrates and habitat moisture regimes and developed from a smaller into a larger patch mosaic. By 1977, ca. 50 % of the forest area in both sample matrices had gone into dieback. Thereafter, the dieback came to a halt. The domino-type collapse, which frequently came to a halt at volcanic substrate boundaries, indicates that stands in better drained sites were also predisposed to die. Stands on adjoining substrates often survived. Substrates with dieback stands displayed no other obvious vigor-reducing stresses. The canopy trees on such substrates may have a common history, such as a major disturbance (including dieback) that synchronized stand development in the past. Subsequent weather disturbances and other abiotic/endogenous stresses associated with stand maturation, such as nutrient limitations and stand-level senescence, may reinforce a rhythmic synchrony over several generations of canopy cohorts.

We determined the consequences of systematic changes in nutrient availability during long-term soil development by measuring foliar nutrient concentrations. Sun leaves of the dominant tree Metrosideros polymorpha and of eight other species were sampled in Hawaiian rain forests developed on substrates that were 0.3 x 10 3 , 2.1 x 10 3 , 5 x 10 3 , 20 x 10 3 , 150 x 10 3 , 1400 x 10 3 , and 4100 x 10 3 yr old. Elevation, annual precipitation, parent material, and dominant species were nearly constant along this gradient. Foliar N and P concentrations in Metrosideros were lowest in the youngest site (0.72% and 0.052% for N and P, respectively), increased to a maximum on 20 X 10 3 and 150 x 10 3 -yr-old substrates (1.45% and 0.108%), and then declined close to the initial concentrations in the oldest site (0.86% and 0.061%); N:P ratios in foliage varied relatively little across the sites. Most other species followed a similar pattern of variation. On a per unit leaf area basis, foliar N and P contents in Metrosideros also peaked on intermediate-aged substrates. Foliar nutrient concentrations in Metrosideros sun leaves were determined across a parallel but wetter substrate age gradient. The pattern of variation was similar on both gradients, but the magnitude of variation was smaller on the wetter sequence of sites. Overall, the pattern of variation in foliar nutrients with substrate age is consistent with conceptual models for the dynamics of soil nutrient availability during long-term soil development, and with measurements of soil properties along this sequence.

We determined controls on litter decomposition and nutrient release for the widespread native tree Metrosideros polymorpha in 11 sites arrayed on gradients of elevation, precipitation, and substrate age on Hawaiian lava flows. The effects of site characteristics were evaluated using three common substrates (Metrosideros leaf litter from one of the sites, wood dowels, and filter paper) decomposed in each of the sites, and the inherent decomposability of tissue (substrate quality) was evaluated using Metrosideros leaf litter from each of the sites decomposed in a common site. Site characteristics were responsible for most of the variation in rates of decomposition in the range of sites and substrates examined. Common substrates decomposed much more rapidly in warm, low elevation sites; apparent Q10 values, calculated on the basis of variation in mean annual temperature with elevation on individual lava flows, ranged from 4 to 11. Litter decomposed slowly in the dry sites, but leaf litter produced in the dry sites decomposed more than twice as rapidly as litter from wet sites when both were measured in the same site. The higher substrate quality of litter from dry sites could be due to trade—offs among nutrient—use efficiency, water—use efficiency, and carbon gain by water—limited Metrosideros. We used these results to test a revision of the CENTURY soil organic matter model that had been designed to simulate the decomposition of surface litter. Simulations accurately matched the pattern but underestimated the magnitude of among—site differences in the decomposition of common substrates in a range of sites. Analyses of both field and simulation results suggested that the decomposition of Metrosideros leaf litter could be limited by nitrogen availability.

Myricafjaya, an introduced nitrogen-fixing tree, is rapidly invading volcan- ically disturbed forests in Hawaii Volcanoes National Park on the island of Hawai'i. We investigated the direct effects of Myrica on the establishment and growth of the native tree, Metrosideros polymorpha, in a forest that is recovering from burial under a 10-200 cm deep layer of volcanic cinder. The number of Metrosideros seedlings found under Metro- sideros trees, in the open, and under Myrica trees was 0.12/M2, 0.01/m2, and zero, re- spectively. Myrica litter inhibited Metrosideros germination in field experiments; germi- nation occurred only when Myrica litter was removed. In open sites, the addition of nitrogen-rich Myrica soils increased dry mass accumulation of transplanted Metrosideros seedlings. Shade from Myrica canopies (or shade cloth) in- creased survivorship and height growth but not dry mass accumulation of Metrosideros seedlings. Isolated Metrosideros trees increased diameter growth in response to nitrogen fertilizer but not in response to the presence of adjacent Myrica trees, despite previous findings that Myrica trees elevate available nitrogen in the soil. Myrica basal diameter and height growth were consistently greater than that of Metrosideros in all size classes. Although Myrica does not readily invade closed, late-successional Metrosideros forests, on young, volcanically disturbed soils it is rapidly establishing dense, monospecific stands under which Metrosi- deros does not regenerate.

The ratios of carbohydrate sink to source in the shoot systems of mature and juvenile Metrosideros polymorpha trees were compared, testing a prediction of the pipe model theory that this ratio is proportional to height, thus explaining the synchrony of forest decline in even-height stands. Volumes of sapwood (S) and inner bark (IB), and leaf mass (L) were measured by destructive sampling in 5 mature and 12 juvenile trees. S:L and IB:L of mature tree shoot systems were higher than of juvenile trees, strongly supporting the hypothesis that mature trees have a lower surplus of carbohydrate production and are less resistant to many forms of stress. The specific leaf burden (S:L and IB:L jointly) within the mature tree sample varied by a factor of about 2, compared with a variation in height by a factor of about 1.3 5, suggesting that a relatively strong environmental stress is needed to trigger the synchronous death of all or most trees in a stand. The pipe model, with its prediction of constant Huber value (stem cross-sectional area per unit supported leaf mass), does not accurately represent the distribution of either the sapwood or the inner bark in M. polymorpha. The sapwood Huber values varied significantly among the organ systems of individual trees and between life states. Within each organ system the sapwood cross-sectional area in larger stems was strongly correlated with supported leaf mass, while the inner-bark cross-sectional area was better correlated with the cross-sectional area of the stem. Sapwood Huber values were similar in twigs and small branches of juvenile and mature trees, but were much lower in the larger stems of mature trees than at any point in the juvenile trees. The pattern of variation of the inner-bark Huber values among organ systems was similar in juvenile and mature trees. A reduction of the sapwood Huber value in the large stems of mature trees associated with heartwood development decreased the rate of increase of S:L with size. IB: L increased very little in the juvenile state, but it increased rapidly in the mature state when leaf mass failed to increase in linear proportion with total tree volume. The volume of inner bark was found to be about one-half as great as the volume of sapwood in mature trees, but is likely to be responsible for a disproportionate part of total aboveground respiration.

Because of their isolation and geographical position, and in contrast to the multi-species tree canopies of tropical rain forests on the continents, the Hawaiian Islands have only two native dominant canopy species in their rain forests, Acacia koa and Metrosideros polymorpha. The wetter forest ecosystems are dominated by only the latter. In 1905, a dieback of lowland tropical Metrosideros rain forest was observed over a 35 km stretch on the lower east slope of Haleakalā Mountain on Maui Island. This was dubbed ‘The Maui Forest Trouble.’ Although the synchronous decline of so many trees was initially believed to be caused by an epidemic disease, a decade of research yielded no pathogen. The conclusion was that the Hawaiian flora consisted primarily of colonizer species that were unable to continue growing on aging soils. Although this made ecological sense at that time, it was a rather limited and thereby unfortunate conclusion. Further research has shown that the Maui Forest Trouble was a ‘bog-f...

Large numbers of Metrosideros polymorpha trees have died in the montane rain forest on the Island of Hawai'i, but previous research has failed to identify a principal cause. This paper describes an experiment that tests the hypothesis that nutrient deficiency is the principal cause of tree death and stand—level dieback. Treatments were fertilizing, stand thinning, and a combination of the two. No significant change in the mortality rate was caused by the treatments; however, stem diameter growth of surviving trees was found to be nutrient limited at all sites. We conclude that nutrient deficiency is not the principal cause of most Metrosideros dieback, but it may be a contributing factor. See full-text article at JSTOR

This paper tests three predictions about resource selection in competing species de- veloped from theoretical studies of Rosenzweig and others. These predictions are distinct from those predicted by simple, single-species optimal-foraging models. Under single-species models, a species is expected to take only its preferred resource when that resource is abundant and to generalize by also taking poorer resources when its preferred resource is scarce. Rosenzweig's models are of two species which can exploit and compete for two habitat patches. (Habitat patches are defined in a very general way.) There are two competition-based cases which can lead to resource partitioning. In the distinct preference case, one species prefers one resource; its competitor, the other. In the shared preference case, both species share the same preference, but one species, the dominant, will exclude, often behaviorally, the other species, the subordinate, from the mutually preferred resource. In con- trast to single-species models, Rosenzweig's models predict three occasions where resource scarcity will be accompanied by specialization: (1) In the shared-preference case, the dominant may be unable to exploit the poorer patch. If it can, it is likely to drive the subordinate to extinction, leaving a one- species system. (2) The subordinate may specialize on clearly inferior resources because the dominant forces it from the better resources. (3) In the distinct-preference case, the scarcity of both resources leads to specialization of both species because interspecific competition (causing specialization) is greater than intraspecific competition (causing generalization). These are the three predictions tested in this paper. During February through June 1979 we recorded the feeding activities of the three commonest endemic nectivores in areas in and adjacent to the Hawaii Volcanoes National Park on the island of Hawaii. The birds fed principally on the flowers of two trees: Metrosideros collina and Sophora chrysophylla. During the study, Metrosideros went from few flowers per tree to peak bloom, and Sophora did the reverse. Both distinct- and shared-preference cases were noted. The former involved Himatione sanguinea on Metrosideros and Loxops virens on Sophora. The latter involved Vestiaria coccinea (the dominant) and subordinates Himatione and Loxops, with the preferred resources being trees with the highest numbers of flowers. Vestiaria only exploited trees with high numbers of flowers and may have had no alternative. Its competitors were forced onto poorer resources by its presence. These results support the theory's first two predictions. The data are also at least consistent with the third prediction: when flowers were scarce, Metrosideros flowers were exploited only by Himatione, and Sophora only by Loxops. However, these results are also explicable in terms of single-species foraging models. The restriction of the behaviorally dominant species to high-quality resources and the temporal and spatial scarcity of such resources suggest that dominant species are more likely to be threatened by human activities than are subordinate species that are capable of using poorer resources. The endemic nectivores that have become extinct or are currently endangered in Hawaii, have all been relatively large, showy, and among the more behaviorally aggressive species.

Twenty 100-m2 plots in a forest dominated by tree fern (Cibotium spp.) and ohia (Metrosideros polymorpha) on the island of Hawaii were subjected to different degrees of canopy removal. Seedlings growing in the control plots received an average of 10% full irradiance, whereas those growing in completely cleared plots received an average of 43% full irradiance. Invasion of Metrosideros tree seedlings into the completely cleared plots averaged more than three times that in the undisturbed forest. These open-born seedlings grew in height by an average of 3.5 cm/yr, 80% more than comparably sized seedlings that had started life in the shade and were then subjected to higher irradiances. The average height increment of all shade-born seedlings was 2.5 cm/yr in the control plots and 6.6 cm/yr in open plots; in terms of relative growth rates, this represents an increase from 0.19 to 0.47 cm -cm- yr-1. Height growth was optimal (10.5 cm/yr) at microsites receiving 55- 60% relative irradiance. During the course of monitoring the growth and irradiance of 1130 seedlings that were tagged before canopy disturbance, it was found that these shade-born seedlings suffered mortality at a rate of 57%/yr when growing at 5 50% indicate that many of these Metrosideros seedlings cannot be readily conditioned to open habitats. It therefore appears that shade-born seedlings of this variety of Metrosideros are intermediate in shade tolerance. Although seedlings may survive in the shade for a year or two, dense canopies of tree ferns or older Metrosideros trees are able to suppress their growth effectively, so that the Metrosideros population may have to depend on canopy gaps or canopy diebacks to maintain itself in mature rain forests. Some individuals were nevertheless able to acclimate successfully to very high light intensities, and this same species is a successful colonizer of barren volcanic deposits, suggesting that Metrosideros polymorpha exhibits broad acclimation ability and/or a wide range of genetic variability.

Generic concepts developed slowly in the Myrtaceae. At the beginning of the 19th century few species were known, and these were traditionally distributed among ten or twelve genera. Willdenow, for example (Sp. P1. 2: 935. 1800), knew among the capsular-fruited Myrtaceae 12 species of Leptospermum, 14 of Metrosideros, and 12 of Eucalyptus. The Myrtaceae with baccate fruits he referred to Eugenia (30 species), Psidium (8 species), Myrtus (28 species), Calyptranthes (6 species) and Plinia (1 species). The American Myrtaceae, except for the monotypic Chilean genus Tepualia, belong to the fleshy-fruited group that since the time of De Candolle has been recognized as a tribe, Myrteae. Outstanding contributions to the knowledge of the species and genera of the American Myrteae were made by De Candolle in 1828 and by Berg in 1855-62, but since Berg's time no one person has had an opportunity to become familiar with the group as a whole. The number of described species is very large, especially in tropical South America, and available herbarium material until recent years consisted chiefly of the classical specimens scattered through European herbaria. Within the last quarter century, however, a considerable amount of new herbarium material has been collected, and a number of regional studies of the American Myrtaceae have been published. The species of the North American continent, of the West Indies, of the Guianas, of the central Andes, of Chile and of the Uruguayan-south Brazilian region have become reasonably well known, and it now seems profitable to assess the work of Berg in the light of modern work on the family, and particularly to evaluate his generic concepts and the concepts of later workers. Until the time of A. P. De Candolle, most workers accepted without much question the generic limits that had been set forth by Linnaeus, who assigned most of the Myrtaceae known to him to one of three genera (cf. Syst. Nat. ed. 13. 2: 336. 1770): Eugenia: Calyx and corolla 4-merous; drupe 1-locular, 1-seeded. Psidium: Calyx and corolla 5-merous; berry 1-locular, many-seeded. Myrtus: Calyx 5-fid; corolla "sub 5-petala"; berry 3-locular, 1-seeded.

THE VASCULAR PLANTS of New Zealand are of special interest because of the location, arrangement and geologic isolation of these islands. A large proportion of its plants are endemic and hybridism is common. This paper summarizes a leaf survey of 38 species of woody dicotyledons, all native to North Island and representing 28 genera. The collection is a general sampling, without regard to individual characteristics or classification, except that for purposes of comparison, five species of the genus Metrosideros were used, and two species each for seven additional genera; the others are represented by one species each. Extreme xeromorphs, succulents and juvenile leaves were avoided. Most of these leaves were collected in the vicinity of Wellington, with a few from the Auckland area. When available, only sun leaves were taken and usually at heights within reach from the ground. All these species are evergreen and the leaves were presumably mature because they were collected in August which is late in the winter season for that region.2 At the time of collection this material was intended for general study and as a convenience the leaves were killed and preserved in 4 per cent formalin. Later median lateral areas from the blades were cut into small rectangles, treated for ten days with 15 per cent hydrofluoric acid and

A B S T R A C T Hawaiian Metrosideios is a highly polymorphic complex that is distributed throughout the six major islands of the Hawaiian Island chain from tropical climates near sea level to altitudes up to 8,509 ft with frost. Its populations extend continuously over many areas with average annual precipitation ranging from 30 to 450 in. and with diverse edaphic and topographical features. The Hawaiian forms are probably all derived from one or a very small number of ancestral introductions that arrived within the last 20 million years by longdistance dispersal. Seeds collected from diverse altitudinal sites on the islands of Hawaii and Maui and grown under uniform greenhouse conditions at Honolulu show evidence of ecotypic differentiation along altitudinal gradients. Seedlings from these islands separated by 50 miles of ocean show parallelism in their altitudinal differentiation in plant height and leaf size, but with strongly overlapping variation from site to site. IN THE HAWAIIAN Islands the most prevalent native tree is a polymorphic taxon with ill-defined taxonomic boundaries, referred to here as the