The lovely stalk-eyed fly

Sphyrocephala beccarii

Here is a spectacularly pretty and weird animal: stalk-eyed flies of the family Diopsidae. There are about 160 species in this group that exhibit this extreme morphology, with the eyes and the antennae displaced laterally on stalks. They often (but not always) are sexually dimorphic, with males having more exaggerated stalks—the longer stalks also make them clumsy in flight, so this is a pattern with considerable cost, and is thought to be the product of sexual selection. The Sphyrocephala to the right is not even an extreme example. Read on to see some genuinely bizarre flies and a little bit about the development of this structure.

So here’s an exceptional example of eyes gone wild:

Female (centre) flanked by two male Teleopsis dalmanni. This species
has extremely exaggerated eyestalks and is sexually dimorphic for eyespan. T. dalmanni grow well under laboratory conditions and have
been much used in studies of sexual selection.

Like it says, these are a species used in studying the effects of sexual selection in populations. I’m more interested in how a fly builds such an amazingly warped head, and the paper I’m summarizing below has some tantalizing beginnings to the answer, but it’s early yet, and the specialties of this family haven’t emerged yet. The lesson so far is that the same genes involved in defining the structure of the Drosophila head are present here, used in novel ways. Similarities aren’t surprising any more, but it’s good to see anyway, because it does mean this is a tractable problem that can benefit from a comparative analysis.

We know in general how the adult fly’s head forms. It’s a remarkable process that involves scattered pieces of tissue that are sequestered in the larva and then expand and assemble into adult cuticular structures in the pupa, during metamorphosis. These pieces of tissue are called imaginal discs. The name has nothing to do with being imaginary, but instead refers to the adult form, which is called an imago by all those latin-loving entomologists.

One of these imaginal discs is called the eye-antenna disc, and when examined, you can find distinct regions in it associated with each of those organs. In the pictures below, a typical narrow-headed Drosophila and its eye-antennal disc is shown on the left; it looks a bit like two discs stuck together, and the one one top (with the blue stain for the distal-less gene product) is going to form the antenna, while the larger region below it will form the eye. I’ve written about the eye disc before — waves of coordinated activity are going to move across it to generate the array of ommatidia of the compound eye.

On the right is the head and eye-antennal disc of S. beccarii, a stalk-eyed fly. It’s about the same size, it has the same two-part structure, and it has a distal-less expression domain in the antennal disk. One difference is that they eye and antennal parts are separated by a discrete stalk of tissue rather than running together as they do in Drosophila. The relationship of the disc stalk to the adult head stalk is unclear, though—in diopsids, the adult eye and antenna both end up together at the end of the stalk, so the separation in the larva isn’t obviously consequential.

Comparison of head and eye-antennal imaginal
disc morphology in Drosophila and the stalk-eyed fly S. beccarii. A: Scanning
electron micrograph of Drosophila head. B: Scanning electron
micrograph of S. beccarii head. Note the antenna placed just
in front of each eye. C: Drosophila eye-antennal imaginal disc
stained for the protein product of the distal-less(dll) gene. DLL
expression is restricted to the primordial of the distal most
antennal segments in the anterior par t of the disc. The
posterior par t of the disc will give rise to the eye and most of
the head capsule. D: Eye-antennal disc of S. beccarii stained
for dll protein. As in Drosophila DLL expression is confined to
the distal antennal segments. The disc has anterior and
posterior par ts similar to those seen in Drosophila but, in stalk-
eyed flies, the two par ts are separated by a piece of
intervening of tissue (bracket). The size of this does not vary
noticeably between species, the discs of T. dalmanni are
morphologically indistinguishable from those of S. beccarii
despite the enormous differences in adult morphology.

To figure out what parts of the disc contribute to which parts of the adult head, the fate of regions of the disc was worked out by a series of transplantation experiments in which whole or partial discs were allowed to develop in a host, and then analyzed for where the donor bits ended up. This kind of analysis only tells us about the origin of the end points in development, and looking at the fate map, I’d be very interested in seeing more about the dynamic rearrangements of the tissue during the process. The posterior or eye part of the disc (in yellow) ends up in the compound eye, sensibly enough, but notice that it also contributes to the ocelli near the head midline, and to a prominent bristle on the stalk—it’s just smeared out along the whole extent of the stalk. Similarly, the anterior or antennal part of the disc (in blue) forms the antenna out on the end of the stalk, but also contributes to the palpus, in the animal’s mouthparts. That intervening stalk (I, in red) just seems to vanish.

(click for larger image)

Fate mapping the eye-antennal disc in T. dalmanni. Whole discs (AIP) or selected fragments are cut from a donor, injected into a
host third instar larva and allowed to go through metamorphosis before being extracted and analysed. The photograph shows some of the identifiable cuticular structures produced by a typical AIP implant. By comparing the range of structures
produced by each type of fragment, a fate map can be drawn up identifying which adult markers arise from the different regions of the
imaginal disc. A, anterior portion of the disc; I, intervening region between the anterior and posterior portions of the discs; P, posterior portion
of the disc; IB, inner verticle bristle.

Molecular markers can tell us more. In the diagrams below, the expression patterns of two important molecules that are involved in patterning the tissue, engrailed (en, in green) and wingless (wg, in red) are shown for Drosophila and T. dalmanni. They’re the same! Despite the overall differences in shape, the organization of the expression domains relative to each other are essentially identical in both species. Of particular interest is the region of the disc that will give rise to the stalk, which is indicated by the small bracket. The tiny collection of cells between the rightmost bands of wingless and engrailed expression is what produces the thin stretch of cuticle between the eye and ocelli in Drosophila, and it’s this same region that undergoes a dramatic expansion during metamorphosis to form the stalk in T. dalmanni.

(click for larger image)

Diagram showing the expression of molecular markers engrailed (en) and wingless (wg)
for the medial and lateral regions of the head capsule in Drosophila and T. dalmanni. A, C: Third instar eye-antennal disc and adult head of
Drosophila. En (green) and wg (red) are expressed in adjacent sectors of the antenna and in the future dorsal head region, indicated by the
bracket. En-expressing cells give rise to the medial structures including the ocelli in the adult fly, while wg-expressing cells give rise to the
latter most cuticle around the eye. The bracket marks the extent of the dorsal head territory between these two. B, D: Third instar eye-
antennal disc and adult head of T. dalmanni. As in Drosophila, en and wg are expressed in adjacent sectors of the antenna and in the future
dorsal head region, indicated by the bracket. We infer that, in the adult stalk-eyed fly, the par t of the disc bounded by these two expression
domains in B gives rise to the region indicated by the bracket in D.

At this point, we’re left hanging—there are these extreme morphological differences between stalk-eyed flies and our generic lab fruit flies, but the level of molecular analysis so far shows mostly great similarities in the genes involved. The authors propose to take advantage of common tools in Drosophila research, though, and extend them to puzzle out what the subtle differences between these two types of flies might be, and how they’re translated into a huge difference in form.

Warren I, Smith H (2007) Stalk-eyed flies (Diopsidae): modeling the evolution and development of an exaggerated sexual trait. BioEssays 29:300-307.


  1. #1 Torbjörn Larsson
    March 16, 2007

    I’m always having to spread the eyepieces on the binoc.

    Not only am I a bonehead (~ 1:1000 have thicker bones), but I’m a fathead too. In a semiconductor fab with ~ 100 persons, I had to constantly spread the eyepieces everywhere.

    The upside with having thick bones is that I haven’t broken them yet in spite of frequent opportunities. But the downside with wide eyes is that I apparently look reliable. Not a good thing for someone aspiring to be an ardent atheist and wanton womanizer.

    Now I must go and practice my pirate scowl again. Arrrhh!

  2. #2 Sachin.P.james
    January 9, 2009

    Dear Sir,

    I shall be thankful if you could provide me the distribution of stalked eye insects or else can you hint any reports from India regarding stalked eye insects. i have some specimens which was collected from Anamallais region Tamil Nadu India and i want to identify the specimen. Please give me a lead on this regard