19 Trees
19.1 Overview
Mature trees, though harder to study than smaller plants, are a key plant growth-form that could have their own characteristic patterns of interactions with herbivores. Therefore, we want to include enough surveys of mature trees in HerbVar’s global sampling to allow us to compare patterns between trees and other growth forms. It is also important to include mature trees because there may be major shifts in tree-herbivore interactions with tree ontogeny, from seedling to sapling and sapling to adult. Here we provide three protocols:
- ‘Tree Seedlings & Saplings’ Protocol
- ‘Mature Trees’ Protocol
- Protocol for selecting leaves for herbivory estimates
Objectives: Provide a protocol for sampling mature trees. Collaborators who do not have a special interest in working with mature trees should restrict their surveys to individuals ≤ 2 m height (seedlings and saplings). That is, survey tree species, but focus on seedlings and saplings. Seedling-sapling surveys won’t be representative of all the individuals in a population of a tree species, but these are key stages in tree ontogeny—perhaps the stages in which herbivory is most influential. We are taking a two pronged approach to including tree species in HerbVar.
19.2 ‘Tree Seedlings & Saplings’ Protocol
Follow the Primary Protocol. This includes (but isn’t limited to) the following data:
- Leaf-level percent herbivory estimates for 10 randomly selected leaves
- Counts of presence/absence of herbivory for up to 60 leaves per plant
- A whole-plant visual estimate of herbivory
- And of course please record the number of galls, mines, and other discrete damage types from sessile herbivores
Please note in the metadata that you surveyed only immature individuals (≤ 2 m) at your site. Such a note can be complemented by recording the height of the individuals in the plantSize columns
19.3 ‘Mature Trees’ Protocol
Follow the Primary Protocol & Site Selection Protocol to pick a site, set up a transect, and randomly select focal trees or quadrats.
Many tree species cover huge geographic areas, making it unreasonable to survey an entire “population” or define a discrete study “site.” If you are working with a widespread species, it is fine to choose a “representative” study area, which actually might just be a small part of a large stand of trees.
Once you have selected a representative study area, you will need to randomly select 30 trees to survey (plus their nearest conspecific neighbors). There are many ways to do this. Here are three methods in somewhat decreasing order of amount of work and rigor:
Follow the Primary Protocol exactly, establishing a transect, selecting 30 points randomly (distance along main transect and distance from main transect), and using a circular quadrat at each point to randomly select 1 individual of the tree species within the quadrat (as in the Primary Protocol). Survey each selected tree and its nearest neighbor.
Follow the Primary Protocol except skip the circular quadrat step, which could need to be prohibitively large in some tree populations: Establish a transect and randomly select 30 points (distance along main transect and distance from main transect). Then select and survey the individual nearest to each random point (plus nearest neighbor).
For trees that are at low density or low abundance, use the Low Density Protocol (Chapter 17) to select trees. For low-abundance plants, we recommend surveying every plant within some area. Take GPS coordinates for each plant. Try to get as close to 60 plants as possible. If you are taking GPS coordinates for each plant, then you do not need to measure distances to nearest neighbors because we can measure spatial relationships using the GPS data.
However you select trees, please make sure to take detailed notes on what you did. Note that some of the trees you select may be seedlings or saplings. We recommend doing whichever individual you randomly select, regardless of its age. This should yield a representative sample of all individuals at the site, across age classes.
Randomly select 30 leaves for quantitative estimates of percent herbivory on each of the 30 leaves.
See Primary Protocol and Damage Estimation Training Document for guidelines on quantifying percent herbivory per leaf.
Please also record the number of galls, mines, and other discrete damage types. Note that mines should be included both in percent damage (because they represent damaged surface area) and as counts.
Randomly select an additional 30 leaves to score for presence/absence of herbivory. Record the number out of 30 with herbivory.
Do not worry about estimating herbivory at the whole-plant scale for mature trees; we will estimate this using the 30 presence/absence leaves and the 30 percent herbivory leaves
19.4 Protocol for selecting leaves for herbivory estimates
Trees… are tall, and we will not be able to reach top branches. We will therefore focus on low branches that can be reached from the ground with a pole pruner, and sample from multiple places around the circumference (see Figure 19.1). We provide some guidelines below, but you should choose an approach that makes sense for you and your species. Remember that we are trying to acquire a random subsample of all leaves on the tree; this means avoiding any preference for/against particular leaves (e.g., young vs old). Ideally, leaves will be sampled in proportion to their frequency on the tree. Here are two alternate methods for selecting random (or at least haphazard) leaves:
The easiest method: if it would work for your trees, is to close your eyes, point at the tree, open your eyes, and take the leaf you were pointing at (“Ian’s nose pointing method” in Chapter 15).
Perhaps the most rigorous but most time-consuming method: is to haphazardly strip several times as many leaves as you need (e.g., >200 leaves). Place leaves individually into a large bag. Mix them. Close your eyes and draw 30 leaves for percent herbivory and 30 leaves for presence/absence of herbivory.
