Journal of the Royal Society of Western Australia, 80:281-286, 1997 Abundance of arthropods in tree canopies of Banksia woodland on the Swan Coastal Plain R A Tassone1'2 & J D Majer2 ’Entomology Section, Agriculture Western Australia, South Perth WA 6151 2School of Environmental Biology, Curtin University of Technology, Kent Street, Bentley, WA 6845 email: [email protected] Manuscript received January 1997, accepted May 1997. Abstract Canopy arthropods and foliar nutrients were quantified during summer for Banksia menziesii, B. altenuata, B. ilicifolia and Nui/tsia floribunda within a low woodland site at Jandakot Airport on the Swan Coastal Plain, Western Australia. Foliar nutrients were higher for N. floribunda than any of the Banksia species. B. ilicifolia differed from the other two banksias in having lower nitrogen and phosphorus but higher potassium. Invertebrate densities were also higher for N. floribunda than any of the Banksia species, while densities on the three Banksia species were relatively similar, albeit with a few predominantly phytophagous taxa having a higher density on B. ilicifolia. These trends provide evidence that, as in other ecosystems where this has been studied, differing arthropod loads between tree species are related to variation in foliar nutrient levels. Tree usage by predominantly insectivorous birds may reflect the relatively low difference in arthropod loads between the three Banksia species and an uneconomic strategy of seeking out low-density N. floribunda trees. Introduction Western Australian ecosystems, ranging from mallee (Recher et al. 1993) to wandoo woodland (Majer & Recher Invertebrates are an ubiquitous component of most 1988, Majer et al. 1996) to jarrah forests (Majer et al. 1990, terrestrial ecosystems and a key element in the energy 1994). flow and nutrient turnover within a community. At present, natural bushland covers approximately Arboreal invertebrates may function as major herbivores, 237 ha of Jandakot Airport and most of this is in a pollinators, parasites and predators (Majer & Recher relatively undisturbed state (Milewski & Davidge 1981). 1988), and are themselves an important food resource for In 1993, a draft environmental impact statement for vertebrates. The dynamics of bird communities are additional development of aviation facilities, extraction closely linked to the abundance and diversity of of silica sand, and construction of a commercial estate invertebrates, with insectivorous avifauna responding to (Anon 1993) proposed that 96 ha (40.5%). of the airport variations in the abundance of prey (Recher et al. 1991). bushland be conserved. The significance of urban Studies in open eucalypt forests of New South Wales by bushland remnants to vertebrate fauna has been noted Recher & Majer (1994) showed that the abundance and by How & Dell (1994) who have conducted surveys in biomass of invertebrates on eucalypts selected by birds this region. In view of this, it is important to gather were greater than on less preferred species. Similar information on how the area caters for the feeding invertebrate-driven preferences in tree usage by birds requirements of insectivorous vertebrates, such as birds. exist in wandoo (Eucalyptus wandoo Blakely) woodland (Majer & Recher 1988) and jarrah (£. marginata Donn ex Smith) forest (Recher et al. 1996) in Western Australia. Methods In south-western Australia there is a distinct seasonality of climatic patterns and there are also Study site considerable differences in weather patterns between The study was undertaken at Jandakot Airport (32 °6 'S, years. These variations, coupled with soil nutrients, 115 °53 'E), 18 km south of the centre of Perth and 9.5 moisture and the effects of fire, result in the local km east of the coast, in the City of Cockburn. The vegetation having variable levels of invertebrates. Recher vegetation of this region of the Swan Coastal Plain is et al. (1996) described these forests and woodlands as broadly classified as 'Tow woodland" over "heath" and fluctuating environments, which possess significant the airport has mainly mixed woodlands of the temporal and spatial variation in invertebrates. Proteaceae, Banksia menziesii R Br, B. altenuata R Br Another variable influencing invertebrates in tree and B. ilicifolia R Br (Milewski & Davidge 1981). These canopies is vegetation type. This study provides data for banksias can be co-dominant, but vary widely in relative invertebrate densities on tree canopies of Banksia abundance. The study site is located within the woodland at Jandakot Airport on the Swan Coastal Plain. Bassendean Dune system in a swale that retains high soil It complements data on canopy arthropods in other moisture throughout winter. This grey soil is a heavily leached sesquioxide podsol and is acidic (McArthur & © Royal Society of Western Australia 1997 Bettenay 1960). The mean annual rainfall at Jandakot 281 Journal of the Royal Society of Western Australia, 80(4), December 1997 Airport is 802 mm, with most rain falling between May species for subsequent analysis of nutrients. Leaf material and August. was dried at 60 °C for 48 hours and ground to a fine powder prior to analysis. Sampling was carried out within a woodland area measuring 100 m by 130 m. The site was 15 m from the Cotton, funnel-shaped nets with a sampling area of edge of the woodland, and had a 1.5 m sand bank 0.5 m2 were used to collect the chemical knockdown between this edge and an access road. The area was samples. Each net was fitted with a sleeve that held a 100 adjacent to the study site used by the Western Australian ml plastic tube. Within a given tree, five nets were Museum (R A How & J Dell, pers comm) for determining suspended at different heights below the canopy foliage. the significance of remnant urban bushland to reptile and Net positions were selected to equalize the amount of other fauna. We sampled the three Banksia species and foliage (determined by visual inspection) in the column also Nuytsia floribunda (Labill) R Br ex Fenzl directly above the nets. Nets were positioned in the (Loranthaceae), which was scarce and contributed less morning (0800 h), one hour prior to spraying, to allow than 2% to the tree cover. Tree canopy cover was 47%, disturbed invertebrates to return. and averaged 6 - 7 m in height. The dense understorey The canopy above each net was sprayed with was dominated by the 2 m tall shrub, Regelia ciliata synthetic pyrethrin pesticide, synergised with piperonyl Schauer (Myrtaceae), which had a patchy distribution butoxide, using a motorized-knapsack mist-blower. throughout the study site, and also by grass trees, Spraying was done only during dry and calm conditions. Xanthorrhoea preissii Endl (Xanthorrhoeaceae). The Two litres of diluted (0.2%) pesticide was used per tree, ground cover consisted of undershrubs and perennial and trees were left for at least 30 minutes to allow silk- herbs. An additional tree species. Eucalyptus todtiana F attached invertebrates to drop into nets. The canopy was Muell (Myrtaceae), was also present, but was so sparsely then shaken to dislodge remaining invertebrates and distributed that it was not included in this study. specimens were brushed into the collecting tubes and All three Banksia species can be trees up to 10 m in preserved in 70% ethanol prior to sorting and counting height. The leaves of B. menziesii are dull green and very to ordinal level. rigid, 10 - 40 mm wide and 80 - 250 mm long (Marchant The numbers of invertebrates caught could be influ¬ et al 1987). The conspicuous flower heads (held above enced by the amount of foliage above each net. To obtain the foliage) are usually pink or red, occurring mainly in a measure of this, we cut three 1 m long branches (measured February to August. The leaves of B. attenuata are more from growing tip backwards) from each tree species and crowded than B. menziesii and much narrower, being 5 - counted the number of leaves on each branch. Twenty 15 mm wide, 40 - 270 mm long, and strongly serrated. leaves of each species were then removed and their surface The conspicuous flower heads are bright yellow, area measured using a computer-based planimeter. This appearing from September to December. The common enabled a measure of the area of foliage on a 1 m branch name holly-leaved banksia aptly describes B. ilicifolia, to be calculated. with its dark green leaves; they are 20 - 40 mm wide and 30 - 90 mm long, elliptical with an undulating edge, and Data analysis the leaves are slightly serrated with each tooth Following sorting, the number of animals within each possessing a spine. The flower heads are cream, turning taxon was summed for the five nets placed within each deep pink with age. The flowers appear throughout the tree. Mean and standard error of the numbers of each year, but peak in spring. N. floribunda is a tree up to 10 m invertebrate taxon on each tree species were then calcu¬ high, that parasitises the roots of surrounding plants. The lated. Three comparisons were made; (1) between the glabrous leaves are long (40 - 100 mm), narrow (3-8 three Banksia species for each invertebrate taxon; (2) mm) and thick. This tree has masses of orange flowers, between the three Banksia species for all taxa that were which occur from October to January. To touch, the significantly different; and (3) between the four sampled leaves of N. floribunda are much less sclerophyllous than tree species for each invertebrate taxon. Analyses were the leathery Banksia leaves. When bent, B. menziesii and restricted to the common invertebrate taxa, those occur¬ B. attenuata leaves crack and split, unlike B. ilicifolia ring in >65% of the samples. N. floribunda was excluded whose leaves are more supple. from comparisons (1) and (2) because of the low number of trees sampled. Comparison (1) was performed by Sampling univariate analysis of variance (ANOVA) and those The proportional composition of tree species within means which differed significantly were detected using the canopy was assessed above 400 vertical sighting Tukey's pairwise comparisons (P < 0.05). Comparison of points taken at regular distances along transects Banksia species for each invertebrate taxon required natural throughout the study site. log transformations of the data due to the high degree of variability or the presence of zeros. Comparisons (2) and Arthropods were sampled from the canopy of the (3) involved ranking the invertebrate taxa and comparing three Banksia and N. floribunda during early summer these ranks using Kendall's coefficient of concordance. (December 1994). We sampled trees which were not in flower, since we were primarily interested in foliage- Foliar samples from each tree were analysed for total associated arthropods. The number of trees sampled was nitrogen, phosphorus and potassium (see Majer et al. dictated by their availability in the study site. 1992 for methodology). The variation in nutrient levels Consequently, 20 each of B. menziesii and B. attenuata, 10 between the four sampled tree species was analysed by of B. ilicifolia and four of N. floribunda were sampled. The univariate analysis of variance (ANOVA) and those height and crown diameter was recorded for each tree. means which significantly differed were detected using Mature foliage was collected from 10 individuals of each Tukey's pairwise comparisons (P < 0.05). Standard 282 Journal of the Royal Society of Western Australia, 80(4), December 1997 deviations were relatively homogeneous, indicating that explanation for the assayed leaves showing identical pat¬ transformations were unnecessary. terns of significance for nitrogen and phosphorus. B. ilicifolia had significantly lower levels of nitrogen and phos¬ phorus but higher levels of potassium than the two other Results Banksia species. B. menziesii and B. attenuata exhibited no significant difference between their levels of nitrogen, In terms of cover, B. attenuatei was the most dominant phosphorus or potassium. species (70% of projected foliage cover), followed by B. menziesii (20% of projected foliage cover) and B. ilicifolia A total of 7 105 individual arthropods was obtained with 7% of projected foliage cover (Table 1). Table 1 also from the 54 tree canopies sampled. The invertebrate indicates that B. ilicifolia was the smallest tree species count was higher on N. floribunda than on any of the sampled, having up to a 2 m crown diameter. The single Banksia species (collected arthropods averaged 216 from main stem of this species had many branchlets that N. floribunda, 182 from B. ilicifolia, and only 107 - 114 formed a cover of dense foliage around its stem. The from B. menziesii and B. attenuata respectively). The F large crown diameter of N. floribunda was in part ratios for univariate ANOVA comparisons of the three associated with the tendency of this species to have more Banksia species, together with means and standard errors than one main stem. B. attenuata and B. menziesii had for the data, are presented in Table 3. It is evident that similar canopy structure, with the diameter (3.0 - 3.5 m) the densities of only a few invertebrate taxa differ and height above the ground (6.5 - 7.0 m) of their crowns significantly between sampled Banksia species within being similar. The area formed by the crown of N. woodland at Jandakot. These invertebrates are Acarina floribunda (17.3 m2) was twice as large as B. menziesii (7.3 nr) (mites), Blattodea (cockroaches), Homoptera (sucking and B. attenuata (9.5 m2), and approximately six times bugs), Thysanoptera (thrips) and Neuroptera (lacewings) larger than B. ilicifolia (3.2 m2). N. floribunda supported by larvae. In almost all significant differences, numbers far the most leaves per metre of branch, followed by B. were significantly higher on B. ilicifolia than B. menziesii attenuata; B. menzesii branches had the lowest number of and B. attenuata. The exception was for Blattodea, where leaves (Table 1). When leaf area was taken into account, B. ilicifolia and B. attenuata had similar but higher N. floribunda L ranches supported the least foliage, B. densities than B. menziesii. Levels of the various attenuata the most, while the other two banksias were invertebrate taxa generally did not significantly differ intermediate. between B. attenuata or B. menziesii, although Acarina were significantly more abundant on B. menziesii nd, as There were generally significantly higher levels of ni¬ mentioned, Blattodea were significantly more abundant trogen, phosphorus and potassium within the leaves of on B. attenuata. N. floribunda than in the Banksia species (Table 2). Nitro¬ gen and phosphorus are generally mobile nutrients Despite these abundance differences, almost the same within the plant, and their within-plant concentrations variety of taxa were found on the four tree species. Over¬ are usually correlated (Majer et al. 1992). This is a probable all, arthropods from 20 orders of insects (Heteroptera and Table 1 Crown diameter and height (n = 10), the proportional composition of the overstorey, mean leaf area (n = 20), and the number and area of leaves along 1 m branches (n = 3) of sampled tree species in the Banksia woodland study site at Jandakot Airport during December 1994. Values are mean (± se). Crown diameter Tree height Percentage Leaf area Number of leaves Leaf area on Species m m canopy (%) cm2 on 1 m branch 1 m branch (cm2) Banksia attenuata 3.5 ± 0.2 7.0 ± 0.2 69.9 9.4 ± 0.4 979.7 ± 7.7 9256.9 Banksia menziesii 3.0 ± 0.2 6.7 ± 0.2 19.9 25.6 ± 0.6 274.3 ± 5.3 7021.9 Banksia ilicifolia 2.0 ± 0.3 6.4 ± 0.4 6.8 14.5 ± 0.4 490.7 ± 6.7 7124.0 Nuytsia floribunda 4.5 ± 0.4 6.7 ± 0.4 1.7 2.1 ± 0.2 2023.7 ± 15.8 4236.3 Others - - 1.7 - - - Table 2 Nutrient levels (pg g'1) within tree foliage (n = 10) from Banksia woodland at Jandakot Airport during December 1994. The means (± se) of are shown with level of significance (P) between tree species; different superscript letters indicate that means differ significantly using univariate analysis of variance (ANOVA) Total nutrient Species Nutrient comparison B. menziesii B. attenuata B. ilicifolia N. floribunda F Value P Nitrogen 5353.0 ± 6.6b 5540.0 ± 6.4b 3914.0 ± 8.3C 6746.0 ± 9.3a 34.5 < 0.05 Phosphorus 209.5 ± 1.7b 200.5 ± 1.7b 157.5 ± 1.8C 434.5 ± 2.0a 141.2 < 0.05 Potassium 1680.0 ± 6.2b 2065.5 ± 4.0b 2602.5 ± 7.2a 2471.0 ± 7.5a 9.2 < 0.05 283 Journal of the Royal Society of Western Australia, 80(4), December 1997 Table 3 Numbers of invertebrates per tree, sampled by pyrethrin knockdown of Banksia woodland canopy at Jandakot Airport during Decem¬ ber 1994. The number of invertebrates per tree was based on five 0.5 m2 nets within the canopy. Best ranks are assigned to all taxa and also the five taxa where significant differences occurred. Means (± se) for the three Banksia species were compared where possible using a univariate analysis of variance (ANOVA) and those means which differ significantly are indicated by different superscript letters. Taxa with significant differences between banksia species are indicated in bold. Ranks compared using Kendall's Coefficient of concordance. NS = not significant. Species B. menziesii B. attenuata B. ilicifolia N. floribunda Tree comparison Class Taxon n = 20 n = 20 n = 10 n = 4 F Value P Arachnida Pseudoscorpionida 0.2 ± 0.2 0.3 ± 0.2 0.1 ± 0.2 4.8 ± 0.9 - Acarina 23.7 ± l.lb 11.5 ± 0.7C 74.8 ± 2.5a 22.8 ± 2.2 15.4 < 0.05 Araneae 8.0 ± 0.5 12.1 ± 0.6 10.0 ± 0.8 22.0 ± 1.8 NS Crustacea Isopoda 0.1 ± 0.1 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 - Collembola 0.5 ± 0.2 0.2 ± 0.1 1.1 ± 0.5 0.3 ± 0.4 - Insecta Thysanura 0.1 ± 0.1 0.0 ± 0.0 0.0 ± 0.0 3.0 ± 0.6 - Blattodea 0.2 ± 0.1b 1.2 ± 0.3a 1.1 ± 0.3a 1.8 ± 0.6 5.1 < 0.05 Plecoptera 0.1 ± 0.1 0.3 ± 0.2 0.2 ± 0.2 1.0 ± 0.6 - Mantodea 0.2 ± 0.1 0.2 ± 0.1 0.1 ± 0.2 0.5 ± 0.4 - Isoptera 0.5 ± 0.2 0.4 ± 0.2 0.8 ± 0.4 2.5 ± 0.9 - Orthoptera 0.1 ± 0.1 0.0 ± 0.0 0.1 ± 0.2 0.3 ± 0.4 - Pscoptera 3.0 ± 0.4 2.2 ± 0.3 5.6 ± 0.7 54.0 ± 2.3 NS Hemiptera Homoptera 2.7 ± 0.4b 4.4 ± 0.4b 9.1 ± 1.1* 12.5 ± 1.6 4.2 < 0.05 Heteroptera 5.7 ± 0.6 4.7 ± 0.5 8.6 ± 0.8 3.3 ± 1.0 NS Thysanoptera 9.2 ± 0.5b 12.8 ± 0.7b 24.6 ± 1.4* 8.5 ± 1.7 3.3 < 0.05 Neuroptera adults 0.1 ± 0.1 0.5 ± 0.3 0.3 ± 0.3 0.5 ± 0.5 - larvae 1.0 ± 0.3b 1.0 ± 0.3b 2.6 ± 0.4a 7.3 ± 1.6 6.7 < 0.05 Coleoptera adults 12.1 ± 0.6 18.3 ± 0.7 14.5 ± 1.2 37.0 ± 3.6 NS larvae 12.3 ± 0.7 11.3 ± 0.6 6.8 ± 0.6 3.0 ± 1.1 NS Mecoptera 0.5 ± 0.3 0.6 ± 0.3 0.0 ± 0.0 0.0 ± 0.0 - Diptera adults 1.8 ± 0.3 1.8 ± 0.3 3.2 ± 0.6 6.5 ± 1.0 NS larvae 0.0 ± 0.0 0.1 ± 0.1 0.0 ± 0.0 0.0 ± 0.0 - Lepidoptera adults 0.1 ± 0.1 0.0 ± 0.0 0.0 ± 0.0 0.8 ± 0.5 - larvae 1.1 ± 0.3 1.0 ± 0.3 0.2 ± 0.2 0.0 ± 0.0 NS Hymenoptera ants 16.6 ± 0.9 17.9 ± 1.1 7.5 ± 1.2 7.8 ± 1.4 NS others 7.5 ± 0.4 11.9 ± 0.8 10.7 ± 0.8 16.5 ± 1.3 NS Total inverterbrates 106.7 114.4 182.0 216.3 Best rank (All taxa) 4 2 3 1 Arachnida Acarina 2 3 1 _ Insecta Blattodea 3 1.5 1.5 - Hemiptera Homoptera 2.5 2.5 1 - Thysanoptera 2.5 2.5 1 - Neuroptera larvae 2.5 2.5 1 - Best rank (Significant taxa) 3 2 1 - Homoptera counted as one order), arachnids, crustaceans The numbers of invertebrates in each taxon were and collembolans were collected, with 18 sampled on N. ranked across tree species for all taxa and also for those floribunda, 17 on B. attenuatei and B. ilicifolia, and 20 on B. taxa which exhibited significant differences (Table 3). In menziesii. The most abundant invertebrate order on the the first of these rankings, N. floribunda ranked as having four tree species was Acarina, with over 1540 individuals the most abundant invertebrate fauna, followed by B. recovered (22% of total). Along with Acarina, Coleoptera attenuata, B. ilicifolia and finally B. menziesii. This ranking (beetles) and Hymenoptera (ants, wasps, sawflies and was not significant (Kendall's coefficient of concordance; bees) dominated the numbers of invertebrates collected P > 0.05), although it was in partial agreement with the from Jandakot. Of the 20 orders sampled, these three ordering of total invertebrate counts in that N. floribunda composed 61% of the total number of invertebrates. was ranked the highest and B. menziesii the lowest; the 284 Journal of the Royal Society of Western Australia, 80(4), December 1997 other two Banksia sp. were ranked as intermediate. When nearby jarrah forest (Recher et al. 1996). Levels of only those invertebrates with significant differences on invertebrates on N. floribunda were considerably higher Banksia species were ranked (on the basis of their than those normally present on local Eucalyptus species. ordering in the ANOVA analysis), B. ilicifolia had the Arthropod levels on all tree species in the present study most individuals, while B. attenuata was only marginally were higher than on trees in the more arid mallee (Recher ranked higher than B. menziesii. et al. 1993) and wheatbelt woodland sites (Majer & Recher 1988, Majer et al. 1996), suggesting the possible existence of a gradient of canopy arthropod densities Discussion with increasing rainfall. Tree usage by insectivorous birds was recorded in the The higher abundance of invertebrates on the canopy same site during the period when invertebrates were of N. floribunda than on that of the Banksia species sampled through until September 1995 (P Kirkpatrick, supports the conclusions from other studies that School of Environment Biology, Curtin University, invertebrate densities are higher in canopies of tree unpublished data). When the proportion of feeding species with higher foliar nutrient levels (Majer et al observations was assigned to the various tree species, it 1992), although it is not possible to test this statistically was found that values reflected the proportional due to the low sample size. The possibility that higher composition of the overstorey for each tree species. Thus, invertebrate densities can be associated with greater insectivorous birds at Jandakot do not appear to be amounts of foliage can here be eliminated as N. floribunda selecting particular tree species as foraging substrates, was the species with the lowest area of leaves above the but rather use the various species of tree in a random sampling nets. manner. The absence of any apparent selection of tree As with N. floribunda, the densities of invertebrates on species by insectivorous birds is consistent with the the three Banksia species were unrelated to foliage area Banksia species supporting relatively similar invertebrate above the nets. The densities of invertebrates in canopies loads. The absence of any selection for N. floribunda, of the three Banksia species were all relatively similar. despite its higher nutrient loads and invertebrate B. menziesii and B. attenuata had almost identical foliar numbers, is surprising. It is possible that its low density nutrient levels, so it is consistent that the density of within this ecosystem means that birds do not find it invertebrates on these two species was similar. The economical to seek out this tree species. Another position of B. ilicifolia in the invertebrate rankings was possibility is that birds may find it difficult to probe for equivocal, since its foliar nutrient levels were lower in invertebrates amongst the dense foliage of this species. nitrogen and phosphorus, but higher in potassium than These findings suggest that, on a unit area basis, those of the other two banksias. Usually it is the levels of Banksia woodland provides similar amounts of food for nitrogen which correlate with invertebrate levels, canopy-feeding insectivorous birds as does the jarrah particularly herbivores, presumably because this reflects forest in the Darling Ranges. Indications are that, in the protein content of the diet (see references in White contrast to the jarrah forest, insectivorous bird richness is 1969; Morrow 1983). Phosphorus and potassium levels low at Jandakot Airport and that some species are no may also reflect some aspects of the nutritional quality of longer present (How & Dell 1993). Probably the the foliage, so possibly the effects of higher potassium fragmentation of the woodlands of the Swan Coastal cancels out the effects of lower levels of the other two Plain, and the reduction in area of this habitat, has nutrients. reduced the available feeding areas for tree-dependent Certain taxa do exhibit significant differences between birds and this has resulted in the local extinction of some the Banksia species and four of these taxa (Acarina, species. The presence of this array of invertebrates in the Blattodea, Homoptera and Thysanoptera) contain species canopies of Banksia species and N. floribunda presents a which derive their nutrients directly from live or compelling case for maintaining this sizeable remnant of decaying leaves. The fact that they reach greatest native woodland at Jandakot Airport. abundance on B. ilicifolia suggests that some taxa may be responding to the higher levels of potassium in leaves of this species. These trends therefore provide evidence Acknowledgments: We would like to thank K Flecknell and the Federal Airports Corporation for permission to use the Jandakot Airport site and supporting the fact that, as in other ecosystems where also P Kirkpatrick, R P Tassone and E Kostas who assisted with field this has been studied, differing arthropod abundance's work. N Keals and M J de Sousa Majer helped with sorting of arthropod (Recher et al. 1996) and richness levels (Majer et al. 1994) material and K McGregor, J O'Neil and M Costigan of the Department of between tree species may be related to foliar nutrient Ecosystem Management, University of New England, performed the nutrient analyses. P Groom assisted with the leaf area measurements and levels. An alternative possibility is that the higher foliage R How, B Lamont, J Scott and an anonymous referee kindly commented density on B. ilicifolia may create a more favourable on an earlier draft of this paper. microclimate for arthropods. The other studies on canopy arthropods elsewhere in Australia (see Introduction) used identical procedures, References with the exception that 10 nets were used to sample the Anon 1993 Proposed additional developments at Jandakot Air¬ overstorey trees. How do the levels of invertebrates port. Environmental impact statement. Dames & Moore, compare with other trees in Western Australia? If values Perth. are expressed as mean number of invertebrates per net, How R A & Dell J 1993 Vertebrate fauna of the Perth Metropolitan then the Banksia species support similar levels of Region: Consequences of a changed environment. In: Urban arthropods to E. marginata and E. calophylla Lindley Bushland Management (ed M Hipkins). Australian Institute (marri) trees sampled during the same season in the of Urban Studies, Perth, 24-47. 285 Journal of the Royal Society of Western Australia, 80(4), December 1997 How R A & Dell J 1994 The zoogeographic significance of urban Milewski A V & Davidge C 1981 The physical environment, bushland remnants to reptiles in the Perth region, Western floristics and phenology of a Banksia woodland near Perth, Australia. Pacific Conservation Biology 1:132-140. Western Australia. Western Australian Herbarium Research Majer J D & Recher H F 1988 Invertebrate communities on West¬ Notes 5:29-48. ern Australian eucalypts: A comparison of branch clipping Morrow P A 1983 The role of sclerophyllous leaves in determin¬ and chemical knockdown procedures. Australian Journal of ing insect grazing damage. In: Mediterranean-Type Ecosystems Ecology 13:269-278. (eds F J Kruger, D T Michael & J U Jarvis). Springer-Verlag, Majer J D, Recher H F & Ganeshanandam S 1992 Variation in Berlin, 509-524. foliar nutrients in Eucalyptus trees in eastern and Western Recher H F & Majer J D 1994 On the selection of tree species by Australia. Australian Journal of Ecology 17:383-393. Acanthizidae in open-forest near Sydney, New South Wales. Majer J D, Recher H F & Reals N 1996 Branchlet shaking: A Emu 94:1-7. method for sampling tree canopy arthropods under windy Recher H F, Majer J D & Ford H A 1991 Temporal and spatial conditions. Australian Journal of Ecology 21:229-234. variation in the abundance of eucalypt canopy invertebrates: Majer J D, Recher H F, Perriman W S & Achuthan N 1990 Spatial The response of forest birds. Acta XX Congressus variation of invertebrate abundance within the canopies of Intemationalis Omithologicii 20:1568-1575. two Australian eucalypt forests. Studies in Avian Biology Recher H F, Majer J D & Ganesh S 1996 Eucalypts, arthropods 13:65-72. and birds: On the relation between foliar nutrients and species Majer J D, Recher H F & Postle A C 1994 Comparison of arthropod richness. Forest Ecology and Management 85: 177-196. species richness in eastern and western Australian canopies: Recher H F, Majer J D, Gowing G & Sarti N L 1993 Canopy A contribution to the species number debate. Memoirs of the invertebrate communities in woodlands - A comparison of Queensland Museum 36:121-131. morning and afternoon samples by chemical knockdown. Marchant N G, Wheeler J R, Rye B L, Bennett E M, Lander N S Mulga Research Centre Journal 11:27-30. & MacFarlane T D 1987 Flora of the Perth region. Depart¬ White T C R 1969 Food and outbreaks of phytophagous insects, ment of Agriculture, Perth. with special reference to Cardiaspina densitexta Taylor McArthur W M & Bettenay E 1960 The development and distri¬ (Psyllidae, Homoptera) on Eucalyptus fasciculosa F.v.M. bution of the soils of the Swan Coastal Plain. CSIRO (Austra¬ (Myrtaceae) in South Australia. PhD Thesis. University of lia) Soil Publication 16. CSIRO, Canberra. Adelaide, Adelaide. 286