An identifying feature of Quercus agrifolia Ne ́e (Fagaceae) is the presence of hair tufts on lower leaf surfaces. In other plant species, hair tufts act as domatia for arthropods such as mites, which in turn feed on leaf fungi or small herbivores and possibly benefit plant health. However, this mutualistic relationship remains untested in Q. agrifolia. In this study two primary questions were addressed within a natural stand of Q. agrifolia in San Luis Obispo, CA: 1) Do hair tufts act as domatia for mites? and 2) Does the removal of hair tufts impact mite abundance, herbivory or fungal pathogens on leaves? In an observational study of 377 leaves from 20 trees, we found a significant association between the presence of hair tufts and the presence of mites. When we experimentally removed hair tufts, we found a significant reduction in mites, yet there was no impact on leaf herbivory or necrosis. We conclude leaf hair tufts on Q. agrifolia serve as domatia for mites, but we found no evidence that mites reduce herbivory or fungal pathogens. Thus, while mites likely benefit from housing provided by hair tufts on Q. agrifolia, it is unclear that the tree benefits from the mites, i.e., whether this is a mutually beneficial relationship.

This study took place in a natural stand of Quercus agrifolia in “Poly Canyon” at California Polytechnic University, San Luis Obispo, California (latitude 35.3057, longitude -120.6581). To address our first study question, we conducted an observational study on 22 February 2019. Twenty teams, each of 2-4 students, were assigned a random GPS coordinate within the study population and sampled leaves from the nearest tree. Random GPS coordinates were obtained by creating a roughly 9-acre polygon around oak woodland habitat in Poly Canyon using ArcGIS software (ESRI 2018) and randomly sampling coordinates from within the polygon. Thirty leaves were haphazardly sampled, each from a unique branch within the sub-canopy.  For each leaf, the petiole was cut using a razor blade to avoid flinging any mites off the leaf and leaves were placed individually in a plastic bag. Bags were carefully transported back to the lab in plastic trays so as to cause minimal disturbance to any mites.

In the lab, each team of students recorded data on 2-3 leaves from each of the twenty sample trees. This was done in order to avoid confounding the effects of student group and sample tree. For each leaf, the number of leaf vein axils with hair tufts was recorded.  Hair tufts were defined as clusters of three or more hairs.  For example, in Fig. 2, leaf vein axils 3 and 6 contain hair tufts.  To determine the association between hair tufts and mites, a single axil on each leaf was observed for the presence or absence of mites.  All observations were made with a dissecting microscope at 10-40X magnification. Mite presence included larvae, nymph, or adult mites (eggs were excluded due to uncertainty in their identification). In order to sample approximately the same number of ‘no hair tuft” as “hair tuft” leaf axils for each tree, each leaf was pre-assigned to either the “no hair tuft” or “hair tuft” category.  If “no hair tuft” category was assigned, an axil without a hair tuft was selected for observation, always beginning with the second leaf vein axil from the bottom and to the right of the mid vein (Fig. 2, axil 2), and moving counterclockwise until an axil without hair tufts was located (in Fig. 2 this would be axil 2). Upon locating a suitable axil, the absence or presence of mites was recorded.  The same protocol was used for sampling leaves pre-assigned to the “hair tuft” category (in Fig. 2 this would be axil 3).

Next, we performed a manipulative experiment to determine whether removal of hair tufts impacted the abundance of mites and leaf health. This experiment took place between 25 February through 10 May 2019, using a paired study design on 20 haphazardly sampled trees. Within each tree, 10 branches were selected that met the following criteria: branches contained at least 5 fully expanded leaves, were close enough to the ground to allow easy access by researchers, and were not in contact with any neighboring branches (to reduce the opportunity for mites to move from untreated to treated branches). Branches were then randomly assigned as removal or control treatment. On the ‘removal treatment’ branches, all hair tufts located in leaf vein axils on the abaxial surface were removed by gently scraping the leaf surface with a sharp tool (e.g. a scalpel). This was performed on every leaf within the branch. On ‘control treatment’ branches, abaxial leaf surfaces were scraped immediately adjacent to but just outside all leaf vein axils with hair tufts, thus leaving the hair tufts intact. The control treatment accounts for any effect of leaf scraping, for example, potential damage to the epidermis of the leaf.

Treatments were left untouched for 4-6 weeks, at which point we returned to assess mite abundance and leaf health. On each branch, we removed three leaves, always sampling the 3rd, 4th, and 5th fully expanded leaf from the apical meristem. The three leaves were placed in a plastic bag, transported back to the laboratory, and assessed within 24 hours of collection. Using a dissecting scope, on each leaf we recorded the number of each leaf vein axils with: hair tufts, mites, herbivory, necrosis due to fungal or bacterial pathogens, and fly larvae/larvae scars (Fig. 1). Herbivory was defined as leaf areas with visible physical damage, while necrosis was defined as leaf areas with visible discoloration or decay.  Fly larvae/larvae scars were not included in downstream analyses as only 27 of 450 leaves had any presence of larvae/larvae scars.