, this is not the cause of a significant edge effect in this site. The majority of identified arboreal predations came from birds, which do not follow these linear geographic features, and generally have fewer dispersal contraints.
Autocorrelation of arboreal nest locations
One of the more unusual results from this experiment is the spatial autocorrelation between arboreal nest proximity and location, which is not present on ground nests. Particularly given that the dispersal constraints of terrestrial predators are usually larger than those of arboreal predators. As there is a smaller diversity of arboreal predators in the U.K., the autocorrelation could be due to a smaller number of individuals predating the nests, limited by their dispersal constraints. In conjunction, the ground nests were predated on by more individuals, disguising individual dispersal patterns. This could be an avenue of future study, however it should be noted the spatial autocorrelation was quite weak (r = 0.089), so the result may not be ecologically significant.
Predation intensity
Predation intensity at ground levels were significantly larger than those in arboreal nests, which supports some previous studies \cite{article,Piper_2004}, although this result is not universal \cite{articlea}. Ground nest predation rates in this experiment were very high compared to other artificial nest experiments, regardless of location \cite{article}. There are many potential reasons for this effect, as previously mentioned it may be due to an annual seasonal effect of spring increasing predation rates nationwide \cite{Malzer_2015}. The sharp edge of the forest may also play a role. The bordering open grassland is capable of supporting high densities of opportunistic generalist predators which can follow linear foraging routes of the edge. However, while this may have guided mesopredators to edge nests, but does not adequately explain the high interior predation rates \cite{Lahti_2001}.