FIGURE LEGENDS
Figure 1. (A) Defended and (B) undefended morphs of D. pulex .
Animals are third instar juveniles exposed to either 1µl
ml-1 or 0µl ml-1 (control) ofC. flavicans predator cue. The enlarged image shows a close-up of
the neckteeth defence.
Figure 2. Image digitisation and Procrustes superimposition. (A) First,
images were prepared for landmarking in Microsoft Paint. This process
identified a total of six landmarks for each photo (open circles)
including (1) the centre of the eye (head top), (2) the neckteeth
defence (or the corresponding area in the controls), (3) the base of the
tail, (4) the dorsal (back) midpoint, (5) the ventral (belly) midpoint
and (6) the rostrum (snout). (B) Second, the landmarks (closed circles)
were applied in R to create digitised images that captured key aspects
of shape (links). (C) Third, Procrustes superimposition was performed to
align digitised images in terms of size, position and orientation. The
digitised images in grey differ from the image in black in terms of
size, position and orientation. (D) The final set of digitised images.
Figure 3. The first set of three D. pulex phenotypic trajectories
along a gradient of predation risk. The principal components summarise
changes in neck shape (PC1) and head height (PC2), which are visualised
by the deformation grids along the axes. Individual data points for each
specimen are shown by the small points, whereas the large points
correspond to the mean phenotype for each treatment. The colour
indicates the level of risk (0, 0.1, 0.25, 0.5, 0.75 and 1 µl
ml-1) and the lines connect each treatment in order of
increasing predation risk. The shape of the points refers to the
specific clone used in the experiment (circles – Carlos, squares –
Cletus, triangles – Chardonnay).
Figure 4. The second set of three D. pulex phenotypic
trajectories along a gradient of predation risk. The principal
components summarise changes in belly bulge (PC3) and head height (PC2).
See the description of figure 3 for the legend.
Figure 5. The third set of three D. pulex phenotypic trajectories
along a gradient of predation risk. The principal components summarise
changes in neck shape (PC1) and belly bulge (PC3). See the description
of figure 3 for the legend.
Figure 6. Modularity and integration tests. (A) Histogram of CR
coefficients from simulations of random partitions of the head and body,
with the observed CR coefficient indicated by the arrow. (B) PLS scores
for the head and body, with the line of best fit and deformation grids
which visualise the shape identities and how they change along the axes.
(C) The same as (A), but for the front and back. (D) The same as (B),
but for the front and back. The specific landmarks involved in the tests
are represented by the diagrams above the plots. If there was
modularity, the observed CR coefficient would be significantly less than
one, and if there was no integration, the line of best fit would look
like either a horizontal or vertical line.