Pilot Study

A small pilot study was undertaken for 2 weeks in March 2019 using 8 artificial nests, made of rope and metal wire, to examine basic questions about the predator behaviour and whether the experiment was suitable for a larger study. Specifically, the pilot study aimed to answer whether predators in the forest would be able to locate the nests and whether the presence of artificial eggs would significantly affect predation rates. Plasticine is very malleable and allows for the identification of predators through bite marks , and so makes an ideal artificial egg material.  For the purposes of the pilot, forest edge was defined as within 10m of the edge of the forest, and forest interior was >100m away from any edge.  Each nest contained 1 quail egg, and half contained an additional plasticine egg, with half of the nests placed in the forest edge and half in the forest interior. Nests were placed in, or at the base of trees, where each tree is more than 15m away from any other selected tree. Arboreal nests were mounted to trees using steel wire where necessary, and were all accessible via branches capable of supporting the weight of predators. No attempts were made to disguise, or hide the nests either at ground or arboreal levels. Eggs were considered to be predated if they were missing, broken or otherwise interfered with (e.g bite marks). The pilot study found no significant difference between ground and arboreal nests (χ2 = 0.20991, df = 1, ρ = 0.6468) indicating the full study can use ground and arboreal nests, and that the arboreal nests are being predated. The experiment however did find a significant difference in predation rates between nests with and without the plasticine eggs ( χ2 = 17.681, df = 1, ρ < 0.0001), possibly due to the olfactory cues the plasticine was emitting, indicating either a different material for artificial eggs should be used, or steps should be taken to reduce the smell. It should also be noted that every day of the trial, at least 1 plasticine egg was completely eviscerated preventing predator identification.

Power Analysis

Due to the time limited nature of this experiment, it was imperative to carry out a power analysis to determine significance. The pilot study was too limited in scope to generate an accurate estimate of the effect size of edge effects, so effect sizes were gathered from a variety of papers who had performed similar experiments \cite{KOTZE_2007,Cooper_1998,Vaaland_Burkey_1993,TEMPLE_1988}. An average was taken which indicated a low effect size (w=0.256), and using the default values of  ρ = 0.05 and power 0.8, indicated that  showed N=119.8, indicating 120 nests need to be used for the full experiment.  

Materials and Methods

120 artificial birds nests made from conifer leaves and metal wire were acquired from a bullfinch breeder. 3 eggs were placed in each nest, 1 quail egg and 2 plasticine eggs. The additional plasticine egg was added to mitigate the risk of evisceration preventing predator identification. Plasticine eggs were made by hand to have the dimensions of a reference quail egg (3.5cm long, 2.8cm diameter), and were aerated for 3 days after manufacture to mitigate the strong smell. To reduce this further, plasticine eggs were sprayed with 4 layers of PlastiDip, which has been shown to reduce overestimation of mammalian predation rates \cite{Purger_2012}. Predators were identified using (?) reference photos of woodland predators skulls. The experiment was run for 3 weeks in May 2019.
We analysed the effect of edge on predation using  χand ANOVA using R \cite{team2013}.
Mayfield method? - Used to calculcate daily predation rates     http://www.prbo.org/cms/docs/terre/MayfieldHandout2006.pdf  (Other papers use this, might be worth understanding and applying)

Results

Discussion

One of the issues with this experiment is the use of artificial nests over natural ones, which has been shown to overestimate predation rates and does not mimic the seasonal variation in predation rates observed in natural nest experiments \cite{Zanette_2002}.