I photographed a rather cooperative, large dragonfly today and after getting back to my desk and looking closer at the images I realized that the reflection of the Sun in its eyes produced large hexagonal bright spots that are:

  1. much larger/wider than if the eye surface were smooth specular reflector
  2. relatively uniform in brightness all the way to the edge
  3. have very sharp edges, transition from bright to dull red is very abrupt.

I'm aware that the components of compound eyes are usually at least roughly hexagonally packed, but what is happening here is something more than that.

What is it about a dragonfly's compound eye that can account for all three of these characteristics of reflected sunlight?

Presumably this is the same red dragonfly species as is seen in Identify these two large, colorful dragonflies in Taiwan? It's the same area and time of year.

red dragonfly with hexagonal spots in its eyes

Two more taken at different angles relative to the Sun. (click for larger)

red dragonfly with hexagonal spots in its eyes red dragonfly with hexagonal spots in its eyes


1 Answer 1


Here is the image in comments. The direction of the sun can be seen from the shadow of the left antenna. I just sketched this image to note some observations, it's the best answer that I can find at this time.

enter image description here

enter image description here source

  • $\begingroup$ Yes I think this provides a lot of insight, thanks! $\endgroup$
    – uhoh
    Sep 24, 2020 at 7:13
  • 1
    $\begingroup$ It's a bit of a steep optics challenge sorry it would be better to demonstrate using a 3D graph. Using blender there is probably an algorithmic option to place many components, so you'd just have to put in one ommatidea and then algoritmically copy it to a geode/spherical group. i don't know if the light comes from the lens or the cone, although the straight edges which are shorter are giving less reflection. $\endgroup$ Sep 24, 2020 at 7:49
  • $\begingroup$ I think that in Biology SE we can only go as far as identifying the anatomy/morphology. Consider adding the diagram in Figure 5 of the paper linked in this comment so readers can see the anatomy from an optical point of view? The interface between the bottom of the lens material (n=1.43 to 1.45) and the crystalline cone material (n=1.34) below it will also produce a small back-reflection, a bit like retro-reflective glass beads $\endgroup$
    – uhoh
    Sep 24, 2020 at 7:59
  • $\begingroup$ ok, i'll send you the mesh if you do the light render: i'd send an occluding mesh to wall off all the lenses, and one mesh made of thousands of lenses to be rendered in transparent. I can use 250 vertices per lens and render 10k lenses, and another mesh with cones is fine too. i can render the mesh although i don't do raytrace apps only wavefront. $\endgroup$ Sep 24, 2020 at 9:07
  • $\begingroup$ Thanks, and I appreciate your enthusiasm, but it's outside the scope of what I can do right now, and outside the scope of how we should use Stack Exchange. However if you't like to pursue it then you can take the information here and in the question and try a "How to model a dragonfly's eye's optics and reproduce this photo" type question in Blender SE and there's a chance someone there will find it interesting to try! They are quite an active, creative bunch! $\endgroup$
    – uhoh
    Sep 24, 2020 at 9:56

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .