---

1) Site location 2) Site characteristics 3) Selection of study species and study individuals 4) Plant traits 5) Herbivory 6) Seed predation 7) Herbivore abundance 8) References

---

1) Site location
The primary criteria for site selection was site "naturalness". That is, we wished to locate the sites in areas where the present levels of herbivory were plausibly similar to those that the plant traits we are measuring evolved with. Thus, places like New Zealand, Hawaii and Madagascar, where the assemblage of herbivores has changed dramatically within just a few plant generations, were excluded from the study. Places like Australia, where the extinction of the megafauna occurred tens of thousands of years ago were considered to have had sufficient time to re-equilibrate. Obviously, there is a spectrum of site conditions between these two extremes, and the location of the cut-off is somewhat arbitrary. However, we did our best to choose sites where the disturbance regime and herbivore regime were as close as possible to what the plants evolved with.

We tried to get a good spread of sites across a range of latitudes, to maximise our power to detect latitudinal gradients. We established sites on all the major land-masses, to include multiple suites of species and herbivores, and to ensure that our results were general. We did not restrict our sampling to sites of a particular vegetation type, because we consider the fact that vegetation type changes with latitude to be an ecologically important global pattern, which is likely to influence global patterns in herbivory and seed predation. Thus, we aimed to establish sites in the most abundant natural vegetation type in each area. However, we also aimed to get sites at similar latitudes with contrasting vegetation type, net primary productivity, rainfall or soil nutrients, to increase our power to determine the underlying causes of any latitudinal gradients.

We excluded sites on small islands (defined as land masses smaller than Tasmania), because islands are more likely to have major elements missing from the herbivore fauna, or to have unusual herbivore population densities. We also aimed to exclude sites more than 600m above sea level (though this criterion was relaxed if there were no suitable patches of vegetation at lower elevations), in order to reduce confounding between altitude and latitude.

Exact site locations were determined according to 1. location of prior studies (to facilitate comparison of our data with previous information from the sites), 2. site condition (avoiding close proximity to roads, vegetation edges, and sites of major disturbance), 3. location of canopy access facilities, 4. ease of access, and 5. likelihood of escaping from public interference.

We did not consider it necessary to sample within plots of constant areas (this would be confounded by the latitudinal gradient in plant size anyway). The size of the area sampled was determined by the area needed to encounter 10 individuals of the 5 most abundant species (ie, it was a function of plant size and density).

2) Site characteristics
geographic location
Latitude, longitude and altitude will be recorded using a GPS.

leaf area index
To be assessed using hemispherical photography. Images of the canopy will be taken through a fish-eye lens, under diffuse light conditions (on uniformly overcast days, or close to sunrise or sunset). Images will be taken at 25 locations evenly spread throughout the site, with the camera as close to ground level as possible. We will also use point cover estimates to quantify ground cover.

temperature
Mean annual temperature, mean daily minimum and mean daily maximum temperatures will be taken from the nearest available weather station. In the absence of a suitable station, estimates will be taken from New et al (1999).

precipitation
Mean annual precipitation data will be taken from the nearest available weather station. In the absence of a suitable station, estimates will be taken from New et al (1999).

net primary productivity (NPP)
The geographic location data will be put into BIOME 4 (Kaplan et al 2003), to estimate NPP. BIOME 4 is a coupled biogeography and biogeochemistry model. It uses climate and soils information, linked to an ecophysiologically-based photosynthesis and stomatal behaviour model to simulate NPP for a range of plant functional types.

soil fertility
samples will be taken from five locations within each site. Samples will be an even profile from the top 10cm of soil, excluding the litter layer. Samples will be oven-dried for 2 days at approximately 50 degrees Celsius. All soil nutrient analyses will be done in the same laboratory (in Australia). To import soil, it will be sterilised by gamma-irradiation (this does not affect the nutrient content), at Macquarie University. Soil samples will be crushed (using a puck mill), then analysed for phosphorus (using XRF analysis), carbon and nitrogen (using combustion and mass-spectrometry).
We will do a basic characterisation of the soil profile at one location within each site. We will record details of the soil texture (sandy clay, silt loam etc), consistence (loose, friable, firm, extremely firm), pH, and parent material.

3)  Selection of study species and study individuals
selection of study species
We will sample the 4 most abundant species at each site (by cover), regardless of species identity or clade. Analyses that take phylogeny into account will be performed, but the main questions we are asking are about the traits of the most abundant species at each site, and the factors that shape them. These questions do not require us to sample in an explicit phylogenetic framework.
 
selection of study individuals
We will begin from a central point in each site, and locate at least 5 individuals of each of the 4 most abundant species. Only individuals more than two canopy-diameters apart will be used (ie, we won't sample plants in one dense cluster, because this might lead us to underestimate between-individual variability, and is likely to give non-representative results). Included individuals will be mature, outwardly healthy plants that appear to be reasonably "normal". We may need to include a greater number of very small plants to get sufficient leaves for herbivory measurements.
 
4) Plant traits
GENERAL NOTES
These traits will be measured on the 4 most abundant species at each site.
 
We have followed protocols described by Cornelissen et al (2003) wherever possible. Cornelissen et al's paper is an excellent source of information on the ecological meaning of these traits, and gives much more complete descriptions of the reasons for measuring traits in particular ways, and of the techniques used to deal with tricky cases than I have been able to give here.
 
We will measure "leaf" traits on the photosynthetic organs of the plants, regardless of their true botanical form (e.g. phyllodes, cladodes will be measured as "leaves".) All leaf traits will be measured on recently-produced, yet fully-developed leaves, growing in full sunlight, without obvious signs of pathogen attack or herbivore damage, unless specified otherwise. Fern "leaf" traits will be measured on non-reproductive fronds (ie. fronds without sori). All leaves will be collected in the morning, to maximise their chances of being fully-hydrated.
 
Samples will have to be weighed on different balances in each location. However, balances will be calibrated before use (with standard weights), and we will try to use balances with at least 0.1mg accuracy.
 
projected leaf area (mm²)
We will scan 10 freshly-collected leaves (whole leaves, including petioles) on a flatbed scanner, and analyse the images using Image J software. Leaves will be kept in plastic bags, on moist tissue in a cooler or in a refrigerator until they can be scanned, to avoid shrinkage associated with dehydration. The leaves will be taken from at least 5 individual plants (usually more). We will follow Cornelissen et al's (2003) recommendations for special cases.

leaf fresh mass (g)
We will weigh the same 10 leaves used for leaf area measurement. Leaves will be kept in plastic bags, on moist tissue in a cooler or in a refrigerator until they can be weighed, to avoid weight loss due to dehydration and to limit the amount of weight loss due to respiration. The outside of the leaves will be patted dry with tissue paper before weighing.

leaf dry mass (mg)
Leaves will be placed in a drying oven at approximately 50 degrees Celsius for 2 days. They will be allowed to cool in a desiccator (so they do not absorb moisture from the air as they cool), then weighed.
 
leaf mass per area (LMA; = 1/SLA; mg/mm²)
Calculated by dividing leaf dry mass by leaf area for 10 leaves per species. 
 
leaf dry matter content (LDMC; = 1-leaf water content; mg/g)
Calculated by dividing leaf dry mass by leaf fresh mass for 10 leaves per species.
 
leaf toughness
Leaf toughness will be measured using a leaf fracture toughness tester, built to the same specifications as the machine described by Wright and Cannon (2001). In short, this machine measures the force required to push a razorblade (held at a constant angle) through the leaf lamina, at the widest point of the leaf. The figure to the right is a sample output, showing the force to fracture a leaf of Corymbia gummifera, from Ian Wright. You can clearly see the extra force required to cut through the midrib. The total force to fracture is a sum of the area under the curve.
 
 
Leaf phenology
The number of months per year that the canopy is green. To be taken from local knowledge and/or floras.
 
leaf nitrogen and carbon concentration, and C:N ratio
Leaf carbon and nitrogen concentrations will be measured using a leco C:N analyser.
I would love to find a way to measure P concentration too. I'll work on this when I get back from establishing the sites!

leaf phenolics
Still to be decided!

leaf tannins
perhaps a PEG-binding assay? Still to be decided!

leaf alkaloids
Still to be decided!

presence and type of hairs
Simple presence/absence measure, recorded separately for fully-developed and developing leaves. In the case of hair presence, the type of hairiness will be recorded (e.g. tomentose), and the type of hair will be recorded (e.g. stellate, glandular).
 
presence and location of spines
Simple presence/absence measure. In the case of presence, the location of spines will be recorded.
 
presence of extrafloral nectaries
Simple presence/absence measure.
 
presence of wax or glaucescence
Still to be decided! I might place chopped leaves in a solvent, then allow the solvent to evaporate, and measure the mass of the residue??
 
presence of delayed greening
Simple presence/absence measure.
 
seed mass
25 seeds of each species will be collected, from a minimum of 5 separate plants. These seeds will be oven-dried, at 50 degrees Celsius for a minimum of 2days. After cooling in a desiccator (to avoid excess absorption of water from the air during cooling), seeds will be batch weighed on a microbalance (it is not crucial to get a measure of within-species variation in seed mass, because we know that within-species variation in seed mass is negligible compared to cross-species variation in seed mass (Leishman et al 2000)). "Seeds" will be defined as the unit consisting of the embryo and reserves (cotyledons plus endosperm), the seed coat, and whatever fruit tissues are fused to the seed coat. Thus, in Asteraceae, achenes will be weighed, but in most cases (e.g. Fabaceae, Poaceae) true seeds will be the unit of measurement.
 
diaspore mass
25 diaspores of each species will be collected, from a minimum of 5 seperate plants. These diaspores will be oven-dried, at 50 degrees Celcius for a minimum of 2days. After cooling in a dessicator (to avoid excess absorption of water from the air during cooling), diaspores will be batch weighed on a microbalance. "Diaspores" are the dispersal units (seeds plus any fruit tissues that are routinely dispersed with the seeds).
 
seed defenses:seed reserves
We will dissect diaspores (seed dispersal units) into reserve tissues (endosperm, embryo and cotyledons) and protective tissues (seed coat plus any fruit tissue that routinely surrounds the seeds on dispersal). Both fractions will be oven dried at 50 degrees Celsius for a minimum of 2 days. After cooling in a desiccator (to avoid excess absorption of water from the air during cooling), the tissues will be weighed on a microbalance. In some species, the endosperm is inseparable from the seed coat until germination has commenced. These species will be omitted from analysis. This method is described in Moles et al (2003).
 
seed nutrient concentration
We will collect a minimum of 20 seeds (from at least 5 plants) of each species. These seeds will be oven-dried at 50 degrees Celsius for a minimum of 2 days, cooled in a desiccator, then posted to Australia for nutrient analyses. The seeds will be gamma-irradiated to ensure their sterility. However, this treatment will not affect their nutrient content. The seeds will be finely ground, then analysed on a leco C:N analyser to determine their carbon and nitrogen concentrations (and their C:N ratio). If money and time permit, we will also measure phenolics, tannin and alkaloids on these seed samples (methods as for leaf chemistry).
 
plant height
Measured with rulers, or using an inclinometer and tape.
 
plant growth form
Species will be classified as trees, shrubs (less than 2m at maturity, or multi-stemmed) forbs, grasses or climbers.


5) Herbivory
main study
Herbivory will be assessed four times throughout one year (3 months apart) at each site. This will allow us to quantify seasonality, which is likely to vary systematically along the latitudinal gradient, but does not allow us to quantify inter-annual variation. While we acknowledge that this would be good to do, we simply do not have the time or resources to follow each site through multiple years.

We will measure "leaf" loss on the photosynthetic organs of the plants, regardless of their true botanical form (e.g. phyllodes, cladodes will be measured as "leaves".) We will tag recently-produced, fully-developed leaves, growing in full sunlight, without obvious signs of pathogen attack or herbivore damage.

At each of the four sample times, we will tag 3 "leaves" on each of 4 branches on each of 5 individuals on each of the 4 most abundant species at the site (giving a total of 60 leaves per species per sample time). Any initial damage will be quantified. The leaves will be left for 14 days, then revisited and assessed for leaf area loss. Leaves which have been completely removed (as often results from vertebrate feeding) will be recorded as 100% missing. Leaves which are undamaged will be recorded as 100% intact. Digital images will be taken (against a grid background for scale) of all remaining leaves (those with some damage). These images will be analysed using Image J software. In cases where a high proportion of the leaf lamina has been removed, the proportion of the leaf area lost will be estimated by subtracting the remaining area of the damaged leaf from the average entire leaf area for that species.  In cases where damage is less severe, a more accurate estimate will be obtained by calculating the difference in area between the scanned image of the damaged leaf and the same image edited in Adobe Photoshop to approximate the area of the full, undamaged leaf.  All forms of leaf damage will be recorded (leaf chewing, leaf mining, pathogen attack etc). In grasses, and species that photosynthesise with flattened stems, we will mark known amounts of recently-developed "lamina", and record the percentage of lamina area lost over a two week period (as above).

In this main study, we will not attempt to identify the cause of leaf area loss. It is possible that some leaves will be lost to physical damage to the plant, especially in cases involving large vertebrate herbivores. This study attempts to assess the cost to the plant (in terms of leaf area lost), rather than the benefit to the herbivore (leaf area ingested). Thus, we consider incidental losses (to trampling etc) to be just as important as true herbivory. It is possible that some leaves will be naturally senesced during the sample period. These leaves will be falsely recorded as "lost". However, the fact that we are tagging only the most recently-matured leaves, and revisiting the leaves 2 weeks after the initial tagging in each season should minimise the error due to natural senescence.

Some study species (especially deciduous or ephemeral plants) will not have leaves present at all of the sample periods. Obviously, it will not be necessary to sample these species at times when they do not have leaves. We acknowledge that twig-browsing might be an important source of carbon loss for some species. However, this is beyond the scope of our study.

Study plants will be tagged using flagging tape and/or metal tags held on by wire. Sample branches/stems will be tagged using coloured twist-ties. Sample leaves will be identified with a paint-dot on the stem.

This protocol is a slightly modified form of that used by Moles and Westoby (2000).

quantification of the importance of vertebrate/invertebrate feeding
At a subsample of sites (most likely just the 21 Australian sites), we will establish vertebrate exclosures around 3 branches on each of the 4 most abundant species. The exclosures will be constructed from chicken wire, on a frame of fencing wire, and will be designed to exclude vertebrate herbivores while allowing free access to the invertebrate herbivores. Leaves will be tagged and monitored within these exclosures in the same way as in the main study (described above). This part of the study is designed to find out whether there is a latitudinal gradient in the relative importance of vertebrate vs invertebrate herbivores.

area loss during leaf development
It is well-known that a huge proportion of lifetime herbivory happens during leaf development. We will estimate the percentage of leaf area lost during development using images of leaves that have just completed development. We will also carefully examine the stems for signs of leaf scars among sequences of recently-developed leaves, and will record such scars as 100% herbivory during expansion. Despite this, our measurements will be underestimates, because we will not be able to account for whole-branch loss. This is unavoidable, as full quantification of leaf area loss during expansion would require daily site visits throughout the development period of each species, a time investment that we simply cannot make in this study. It will also not be possible for us to estimate the proportion of the holes that result from tissue removal, vs that due to hole expansion during leaf development.
 
Seed Predation
pre-dispersal seed predation
We will collect 50 recently matured seeds from a minimum of 5 plants of each of the 4 most abundant species (species that do not set sufficient seed during the year of study will be omitted). In species with multi-seeded fruits, seeds will be collected from a minimum of 20 fruits. Seeds will be inspected for damage by pre-dispersal seed predators (using a microscope and dissection where necessary). Seeds will be considered to have been preyed upon when there is clear evidence, i.e. if the seeds show entry/exit holes, or if invertebrates, frass, or fragments of damaged seed coat are present in place of a seed. If all of the seeds in a multi-seeded fruit have been completely destroyed by a seed predator (i.e. are not able to be counted), then the number of seeds preyed upon in that fruit will be taken to be the mean number of seeds set per fruit for that species. Considering any seed with predation damage to be inviable means that we may slightly over-estimate post-dispersal seed predation, as some seeds (especially large seeds) can germinate despite partial consumption. On the other hand, seeds that do not fill will not be counted as preyed upon, although this may sometimes be due to predators feeding on nearby stem or fruit tissue. Pre-dispersal seed predation by taxa that remove the entire fruit from the plant will not be assessed.This sampling scheme follows that used by Moles et al (2003).

post-dispersal seed removal
We will collect 50 recently-matured, apparently sound diaspores from each of the 4 most abundant species, at each sample-season in which the species were setting sufficient fruit (species that do not set sufficient seed during the year of study will be omitted). Eight depots will be established for each species. At each depot, five seeds will be placed into a depression in the ground to reduce the chance that seeds would be blown away by wind. We will use naturally formed depressions wherever possible in order to minimize soil and litter disturbance. Seeds will be set directly on the soil surface, in order to mimic the natural situation as closely as possible. Seeds will be set out complete with whatever protective tissues would normally accompany them after natural seed dispersal. Thus, diaspores (rather than seeds) will be the actual unit of study for post-dispersal survivorship. Structures that are usually lost during seed dispersal (such as the flesh of fleshy-fruited species) will be removed before post-dispersal survivorship trials. At each depot, seeds will be placed within 2cm of a wooden toothpick, and a larger marker (a plastic plant label pushed into the ground, or some flagging tape) will be established 50cm to 1m from the toothpick. Seed depots will be established at regular intervals (interval length dependent on vegetation scale) along transect lines within the vegetation type from which the seeds were collected. Depots for the different species at each site will be arranged haphazardly along the transect lines. Seeds will be set out within two days of collection. Thus, post-dispersal survivorship of the 4 most abundant species will be monitored in the natural environment in which the species occur, at the time of year at which seeds of that species are normally available. We will also set out depots with barley, as standard "seeds", in each site. Trials using standard seed will be run four times during the year, in order to encompass variability in seed predator activity. Trials on local seed will be run only at the times of year at which such seed is normally available to seed predators. The number of diaspores remaining at each depot will be censused 2 weeks after they are set out. Diaspores will be considered to have been removed if they cannot be found within 20 cm of the toothpick, or if they have been damaged to such an extent that germination seems unlikely. This method quantifies post-dispersal seed removal, but not post-dispersal seed predation per se. Most studies of “post-dispersal seed predation” actually measure post-dispersal seed removal in this way. Techniques for following seed fate are available (most commonly used are radio tracking and thread following), but these techniques are too labour intensive for use in a study of this size. This sampling scheme follows that used by Moles et al (2003).

7) Herbivore Abundance
invertebrates
We will use pyrethrum spraying (0.6% pyrethrum-water solution, delivered from a pressurised sprayer) to sample invertebrates from 20*20cm areas of foliage from each of the study species.  We will spray 3 replicate areas on each of the 4 most abundant species in each season. Sampling will be done in the morning, on low-wind days when possible. Invertebrates will be caught in polythene sheets below the sprayed foliage, then transferred to 70% ethanol solution for storage and shipment to Australia. Tracey Adams is sorting the samples to broad feeding guilds (herbivores/predators/omnivores), and the number and mass of individuals of each morphospecies will be quantified.  We will not attempt to identify the invertebrates further than necessary to allocate them to a feeding guild.
 
vertebrates
We do not have the time or resources to quantify vertebrate abundance. However, these data will be taken from previous work at the sites wherever possible.

8) References

Cornelissen, J. H. C., Lavorel, S., Garnier, E., Diaz, S., Buchmann, N., Gurvich, D. E., Reich, P. B., ter Steege, H., Morgan, H. D., van der Heijden, M. G. A., Pausas, J. G. and Poorter, H. 2003. A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. - Australian Journal of Botany 51: 335-380.

Kaplan, J. O., Bigelow, N. H., Prentice, I. C., Harrison, S. P., Bartlein, P. J., Christensen, T. R., Cramer, W., Matveyeva, N. V., McGuire, A. D., Murray, D. F., Razzhivin, V. Y., Smith, B., Walker, D. A., Anderson, P. M., A.A. Andreev, Brubaker, L. B., Edwards, M. E. and Lozhkin, A. V. 2003. Climate change and Arctic ecosystems II: modeling, paleodata-model comparisons, and future projections. - Journal of Geophysical Research 108: 1-17.

Leishman, M. R., Wright, I. J., Moles, A. T. and Westoby, M. 2000. The evolutionary ecology of seed size. - In: Fenner, M. (ed.) Seeds - the ecology of regeneration in plant communities. CAB International, pp. 31-57.

Moles, A. T., Warton, D. I. and Westoby, M. 2003. Do large-seeded species suffer higher levels of pre- or post-dispersal seed predation than small-seeded species? - Ecology 84: 3148-3161.

Moles, A. T. and Westoby, M. 2000. Do small leaves expand faster than large leaves, and do shorter expansion times reduce herbivore damage? - Oikos 90: 517-524.

New, M., Hulme, M. and Jones, P. D. 1999. Representing twentieth century space-time climate variability. Part 1: development of a 1961-90 mean monthly terrestrial climatology. - J. Clim. 12: 829-856.

Wright, I. J. and Cannon, K. 2001. Relationships between leaf lifespan and structural defences in a low-nutrient, sclerophyll flora. - Funct. Ecol. 15: 351-359.

Wright, I. J., Reich, P. B., Westoby, M., Ackerly, D. D., Baruch, Z., Bongers, F., Cavender-Bares, J., Chapin, F. S., Cornelissen, J. H. C., Diemer, M., Flexas, J., Garnier, E., Groom, P. K., Gulias, J., Hikosaka, K., Lamont, B. B., Lee, T., Lee, W., Lusk, C., Midgley, J. J., Navas, M.-L., Niinemets, Ü., Oleksyn, J., Osada, N., Poorter, H., Poot, P., Prior, L., Pyankov, V. I., Roumet, C., Thomas, S. C., Tjoelker, M. G., Veneklaas, E. and Villar, R. 2004. The world-wide leaf economics spectrum. - Nature 428: 821-827 .