Keynote address - Stephen Jay Gould - In the Company of Animals - Transcript

Social Research, Fall, 1995

Now, that simple theme explains all the really weird homeotic mutations at the back end of Drosophila. The most famous of all is hithorax, the four-winged Drosophila. It looks as though it recovered an evolutionary past and has four wings again. But this is not true--for bithorax is just a loss-of-function mutation in which, because there is less gene product, the third thoracic segment--which ought to develop vestigial halteres--thinks it ought to be another second thoracic segment. And so it grows its third thoracic as though it were another second, so you have two seconds. And since seconds grow wings, you have a four-winged fly. It's not really recreating its ancestry.

And then we have an even weirder one, called hithoraxoid, the eight-legged fly. Insects have six legs. Here's one that seems to violate the definition of its class. But it is the same thing--a loss-of-function mutation. There is not enough gene product in the first abdominal segment. Therefore, it thinks--you see, we use backreading of intent-language all the time!--that is, the segment thinks it ought to differentiate as a supernumerary thoracic segment, and so it does. Instead of being a first abdominal, it differentiates as another third thoracic. The third thoracic has legs, so now we have eight legs instead of six.

So far this is just an insect story. But here is the great discovery of the last 10 years. The same genes exist in vertebrates. In fact, they exist in fourfold repetition. The whole sequence is repeated four times in vertebrates. This is probably why we do not have weird homeotic mutations in vertebrates, because there are four copies of all these genes. So if one of them mutates, the other three are still presumably expressing the normal state and can overwhelm it--whereas in insects there is only one copy, so if it mutates it expresses. But is that not stunning? I mean, Geoffroy was right after all. There is a fourfold repetition of gene clearly homologous to those of insects--they are 90 to 95 percent similar after 550 million years of separation from the insects. You might say, "Yeah, but so what? If it is differentiating segments in insects, what is it doing in vertebrates? If it is doing something totally different, then so what." Well, it is doing something similar in vertebrates--and that is the fascinating thing.

It turns out that vertebrate backbone segments are not the same thing as insect body segments--that is where Geoffroy was wrong. But what modern scientists had forgotten is something that all the great nineteenth-century embryologists knew--that the brain, the mid and hind brain, as it differentiates in embryology, develops as a set of segments, called rhombomeres. And you might say, "But it is all erased in the adult brain." But it is not, because the tongue structures and the cranial-nerve divisions are largely reflective of this old segmentation. In this slide, you can see four of the HOX genes--that is, the mammalian homologues of the invertebrate genes--and their anterior expression boundaries are not in the spinal column but in the rhombomeres. So, clearly, they are determining the rhombomeres, which are the homologues of the insect segments. It is just fascinating. And this slide is a mouse embryo showing that most of these genes--you can see them along the top--are expressing in the rhombomeres.