ETAPA DE TRANSICIÓN

To better understand the relative role of processes or mechanisms that structure tripartite

cyanolichens in Craigieburn, manipulative techniques using transplant experiments were used. To determine whether stand age or habitat preferences within ages influence lichen growth, a

factorial experiment design with nested samples was used. Replicate factors were stand age, vertical position on the tree and species.

6.2.4.1 Preliminary transplant experiments

Preliminary transplant experiments were conducted to determine suitable methods and to identify which species could be used successfully for transplants. Transplant methods that were tested included, 1) a modified version of Coxson & Stevenson’s (2007) “cage” method that involved constructing a round frame out of plastic tubing and attaching green nylon mesh to the frame, thalli were then attached to the mesh with nylon monofilament string and the frame was tied to the tree trunk, 2) an “attached thallus” method derived from Sillett et al. (2000), and, 3) the “pendant” method of McCune et al. (1996) (Figure 6.2). All transplant methods were compared based on the total number of thalli that survived after 4 months and this information was used to establish the best method and species to be used for the final transplant experiment.

Figure 6.2: Preliminary transplant methods. Photo (a) the “frame”, (b) is the “glue” and (c) is the “pendant” method.

6.2.4.2 Final transplant experiments

The final transplant experiments used an “attached mesh” method that was derived from the frame method used in the preliminary experiments. The “attached mesh” method involved tying thalli onto a piece of mesh that was then directly tied onto the tree trunk (Figure 6.3). A total of 264 thalli were transplanted using two lichen species, 176 P. colensoi and 88 P. faveolata. Sample lichens were collected in April 2008 from lower tree trunks in four different aged stands (25, 40, 140 and 265). Thalli were carefully removed from the tree trunk using a knife and transferred into paper bags that were labeled according to their source location. Thalli were then dried and weighed using the same methods as described above. Additional data recorded for each weighed thallus included signs of health and signs of herbivory. Thallus health was determined visually by assessing overall condition and discoloration of the cortex and assigned categories where 0 = healthy condition no signs of damage, 1 = little damage; partly necrotic or bleached on less than 5% of thallus surface, 2 = moderate damage; 5- 25% of thallus surface area necrotic or bleached, 3= heavily damaged; > 50% of surface damaged. Herbivory was documented and distinguished from other damage by regions of the thallus where the algal layer was removed from the upper cortex. Categories were assigned where 0 = healthy with no visible damage, 1 = little damage; less than 5% of thallus surface, 2 = moderate; partly damaged 5-25% of surface, 3 = heavily damaged; > 50%. After assessment, three thalli were tied onto a plastic piece of mesh (15cm x 20cm) using a braided monofilament string. Within 3-4 days of initial collection thalli were placed back in the forest (Figure 6.3).

6.2.4.3 Field placement

Two meshes were installed in each tree, one for experimental treatment and one for a control. Each mesh was assigned to a treatment of stand age or treatment of vertical height on the tree trunk based on age of stand and vertical position from where the thalli were collected. Control treatments transplants were placed in the same age stand and on trees from which specimens were collected at the lower vertical position. For treatments transplants treatments involved moving thalli to different aged stands from which specimens were collected and placing thalli at a higher vertical position (3.0m) were thalli were not naturally found. Overall, the experiment included 11 mesh replicates for each treatment of stand age (25, 40, 140 and 265) and vertical position, totaling 33 thalli, and 11 controls with the same number of thalli. Overall, each stand age had 22 mesh transplants for a total of 88 mesh transplants across stands, totaling 264 thalli. For each tree age, 11 mesh treatment transplants and 11 control treatment transplants were tied onto the tree trunk with zip ties and reinforced plastic coated mounting wire (Figure 6.3). The area on the trunk was cleared of all epiphytes before the mesh was tied on. This was to minimize competition and to simulate natural conditions, where the thalli would be directly on the bark and receive runoff from above. All transplants were tied onto the south facing aspect of the tree trunk. Older trees were accessed by ladders

In addition, to examine if the abundance of the target species influenced growth the presence of both species of Pseudocyphellaria were documented. Categories were assigned where, 0 = species not present, 1 = species present in area, 2 = species present on tree and 3 = species present on same tree near the transplant.

To record temperature and relative humidity, HOBO Pro Series Temp, RH ©1998 data loggers were used. Data loggers were attached to two trees within each stand age; one data logger was placed at the base of the tree and another 2.5m up the trunk. Maximum, minimum and mean temperature (°C) and relative humidity were recorded at 30 minute intervals for the duration of the experiment. BoxCar Pro version 4.3 was used to download data.

Thalli growth measurements were recorded from May 2008 to May 2009. All mesh transplants were collected in May 2009, reweighed and thallus health was documented using the same methods described above.

Figure 6.3: Photo of “attached mesh” transplants. The photo on the left shows thalli attached to the plastic mesh and the photos on right shows thalli on mesh attached to mountain beech tree trunks.

6.2.5 Statistical analyses

To examine environmental conditions within each stand age, paired t-tests were used to assess differences in monthly temperatures and relative humidity (%) between vertical positions. Differences between temperature, relative humidity and biomass across stand age were tested with one-way ANOVAs (Zar 1999), and Tukey’s HSD was used to evaluate the differences between means. Pseudocyphellaria species were analysed separately.

Changes in thallus biomass were calculated using % weight change = [(weight after (g) - weight before (g)) / weight before (g)] * 100. To improve normality thalli were eliminated from the data set if their weight change was beyond two standard deviations of the mean biomass (Antoine & McCune 2004). Of the 264 thalli, eight P. colensoi and six P. faveolata were removed that had weight change values lower than -19% and two P. colensoi thalli were removed that had values greater than 117%. For each stand age, the mean weight change per mesh was calculated and t- tests were used to determine if there were significant weight change differences between controls, at lower heights on the trunk, and treatments, at upper positions on the trunk. The significance of the t-test was used to determine whether to combine the different vertical positions within trees and tease out the effect of age.

For comparisons of thalli weight change (%) among stand ages, ANOVAs were used with trees as the treatments blocked by treatments variables of stand age. Data from each tree was pooled since t-test results showed mostly no difference in weight change between mesh transplants on lower and upper heights on a tree. Responses were calculated for each species as the mean weight change per tree. Data achieved a normal error distribution.

GLMMs were used to determine the causal factors that could influence a thallus to increase in weight or grow and a thallus to decrease in weight. Thalli that increased and decreased in weight were separated for analyses to examine the affect of each factor. The random effects included an individual thallus nested in mesh, tree, and stand age. The fixed effects comprised either (1) the difference (after-before) in thallus health or herbivory (only for the duration of the transplant period), (2) the category of target species presence in the area or (3) movement of the thallus where 1= control or same age where thallus was sourced, 2 = treatment of thallus movement from a younger to an older stand or 3 = treatment of thallus movement from older to younger stand. Absolute transformations were used to run negative values. P. colensoi positive and negative values and P. faveolata negative values fit negative binomial error distributions where a log-ratio link function was used to relate random and fixed effects. P. faveolata positive values fit a

Poisson error distribution therefore a logarithm link function was used. Separate models were run for the different species and models were built using the Wald test statistic that sequentially adds terms to the fixed model and drops individual terms from the full fixed model until the best fit model is obtained. The statistical programme Genstat version 12.1 was used to analyse all statistical tests and models (Payne et al. 2009).

6.3 Results