Figure 2. 3D scan energy at different spatial scales relative to clutch
size: x1/8, x1/4, x1/2, x1, x2, x4. The left-hand side shows the
original height map and associated ROIs: eggs, clutch, rim and
background. More elevated (positive) regions are shown as lighter, while
less elevated regions are shown as darker. The right-hand images show
the separated spatial scales with the position of the clutch marked by a
black arrow. The eggs are most visible at x1/4 scale, the scrape at x1
scale and the nest elevation at x4 scale.
Clutch Occlusion &
Visibility:
For each depth map, occlusion maps were created for 16 different
observer orientations around the azimuth of the nest, from
0o-337.5o, in
22.5o intervals. To calculate occlusion, for each
pixel in the clutch, the shallowest/minimum elevation angle (mA) that
allowed it to be un-occluded was calculated from each of the 16
bearings, given the 3D depth profile, see Figure 1. Elevation angles
above a pixel’s mA allow that pixel to be visible. For observer
elevation angles between 0.50 - 600we measured the visibility. The visibility at a given polar angle was
equal to the mean percentage of pixels un-occluded across the 16
bearings. The distance required to achieve viewing angles was calculated
at fox height [0.4m] and a matrix of corvid flight heights [1.6m,
3.2m, 6.4m, 12.8m, 25.6m].
Colour Metrics:
Luminance ΔS and colour ΔS (JND) from the local (nest) and distal
(background) surrounding each clutch was modelled for corvid vision and
fox vision as a metric of camouflage, using the mica toolbox (Jacobs et
al., 1993; Martin, 2017; Moher Alsady et al., 2016; Vorobyev and Osorio,
1998). Where higher ΔS values correspond with a poorer match. The
Siddiqi method was used for ΔS luminance (Weber fraction 0.2) and RNL
model for ΔS colour (Weber fraction of most abundant cone of 0.05) (Lind
et al., 2013; Moher Alsady et al., 2016; Pretterer et al., 2004; Siddiqi
et al., 2004; Vorobyev and Osorio, 1998). For each observer, we used the
most phylogenetically relevant systems known. These were the common
peafowl Pavo cristatus, for the corvid vision, and the red foxVulpes vulpes , for the fox vision (Jacobs et al., 1993;
Malkemper, 2014; Malkemper and Peichl, 2018; Ödeen and Håstad, 2013).
ΔS values were measured for images acuity corrected for the hypotenuse
distance required for a given series of polar viewing angles
[1.875o, 2.5o,
3.75o, 5o, 7.5o,
10o, 15o, 20o,
30o and 40o], when at the height
of the model observers (fox 0.4m, corvid 3.2m) (van den Berg et al.,
2020). The polar viewing angles for corvid vision were adjusted post-hoc
to the matrix of values used for clutch occlusion [1.6, 3.2, 6.4, 12.8
and 25.6 m], by calculating the polar angle and horizontal distance
required to produce the same observer distance. Acuity correction was
carried out using the known peak resolving power, magpie Pica
pica 33.33 cp/d and red fox Vulpes vulpes 8 cp/d (Malkemper,
2014; Martin, 2017).