I’m going to introduce you to either a fascinating question or a throbbing headache in evolution, depending on how interested you are in peculiar details of arthropod anatomy (Mrs Tilton may have just perked up, but the rest of you may resume napping). The issue is tagmosis.

The evolutionary foundation for the organization of many animal body plans is segmental—we are made of rings of similar stuff, repeated over and over again along our body length. That’s sufficient to make a creature like a tapeworm or a leech (well, almost—leeches have sophisticated specializations), but there are further steps involved in making a fly or a spider or a human. There is an arrangement of positional information along the length of an animal, so one segment can recognize whether it is near the head or the tail, and the acquisition of new patterns of gene expression based on that positional information that cause the development of specialized structures in different segments. That process of specializing segments is called tagmosis. It’s how a fly forms mouthparts in head segments, legs and wings in thoracic segments, and no limbs at all in abdominal segments.

The relationships between segments and how they are specialized are key features in identifying patterns of descent in the arthropod clade. An analysis of those elements in an obscure group, the pycnogonids, has uncovered a surprising relationship—they seem to be related to well known Cambrian organism. You’ll have to read through to the end to discover what it is.

Tagmosis isn’t such a difficult concept so far. Add to the specializations the idea that segments can be reduced and lost (maybe) and more commonly fused with adjacent segments, though, and it starts to get messy when you look at an organism and try to puzzle out what segment is what. You may look at an adult fly’s head and see a smooth and relatively seamless dome, but that structure is assembled embryonically from six segments.

I think.

One other property of tagmosis is that analyzing it triggers the most passionate, heated, long-running arguments. I write “six segments”, and could say more about the identity of each of those segments, and somewhere there is an arthropod taxonomist and embryologist who will wax wroth and prepare a lightning bolt with which to smite me low. Let me just say outright that I am not an expert in arthropod anatomy, and everything I write here about the details is provisional and subject to correction by real experts. I’m just going to recite the rough pattern as many of us understand it, and I’m sure that anyone with knowledge much deeper than mine will happily enlighten and correct me. Please don’t kill me. I have to finish paying my childrens’ tuition.

Anyway, the general pattern of segment organization in an insect’s head is that the first segment forms something called the acron, the second has the antennae, the third the mouth, the fourth the mandibles, and the fifth the maxillae, and the sixth the labium. Look at the weird tangle of dangly bits in the bottom half of a grasshopper’s face for instance, and what you see are highly modified limbs all scrunched together.

That’s what insects are like. What about other arthropods? This is where it gets interesting. There are different specializations, different patterns of organization, in different arthropod groups. These differences would be useful cues to sort out their evolutionary histories, since modifications of the head and brain and feeding structures are rather fundamental to the organisms’ life styles, if only we were confident of the homologies. Here, for instance, is a table of arthropod segment and appendage identities, taken from Raff’s Shape of Life(amzn/b&n/abe/pwll).

Segment # Onychophora Trilobita Chelicerata Uniramia Crustacea
1 Antenna Acron Acron Acron Acron
2 Mouth, jaw Antenna Cheliceraea Antenna 1st antenna
3 Slime palp Mouth, 1st pair legs Mouth, pedipalps Mouth, premandibularb Mouth, 2nd antenna
4 Trunk limbs 2nd pair legs 2nd pair legs Mandibles Mandibles
5 3rd pair legs 3rd pair legs Maxillae 1st maxillae
6 Trunk limbs 4th pair legs Maxillaec 2nd maxillae
7 5th pair legs 1st thoracic app.d 1st thoracic app.
8 No appendages 2nd thoracic app.e 2nd thoracic app.
9 3rd thoracic app. 3rd thoracic app.
10 Abdominal app.f Abdominal app.

aChelicerae arise postorally in development, migrate to preoral position

bEmbryonic only

cLabium in insects, lost in millipedes

dFirst pair of legs in insects, maxillipedes in centipedes, and collum with no legs in millipedes

eSecond pair of legs in insects, first pair of wings

fAbdominal appendages lost or reduced in insects, trunk legs in millipedes and centipedes

That’s Raff’s consolidation of observations from several sources, so again, please don’t kill me if you favor a different set of homologies. It’s an example.

Resolving the homologies here would help us understand how these groups are related to one another. It’s a tricky business (especially with all those ferocious taxonomists taking pot-shots at you), but the way to get at the answers is straightforward: hard work. Examining a wide array of organisms. And using new tools to see deeper into the tissues and molecules of these highly modified head segments.

So here’s an example of exactly those strategies. It’s a study of a rather obscure group of organisms, the pycnogonids or sea spiders. They’re an ancient group of bizarre arthropods (here’s a photo of a 425 million year old sea spider; it’s very fun, a color stereopair. Cross your eyes and see it in 3 dimensions!) that have long been considered a sister group to the chelicerates, which includes spiders and horseshoe crabs and scorpions. They have a thin stalk of a body and long spindly limbs, and on their heads they have a prominent pair of appendages called chelifores. These appendages have long been thought to be homologous to the chelicerae, or “fangs”, of spiders, but at the same time there have been enough doubts that no one was quite confident enough to call them chelicerae, so chelifores they are. If they are homologous to chelicerae, though, that would help confirm that the pycnogonids are closely related to spiders. If they aren’t homologous, there’s going to be a bit of a scramble as the arthropod family tree gets reorganized.

Here’s a drawing of a larval (left) and adult (right) pycnogonid. As you might know, one way to probe the origins of a structure is to look embryonically and examine the relationships in a relatively simple, unelaborated form.


a, The three appendages of the protonymphon larva (shown) correspond to the cephalic appendages of the adult pycnogonid. b, The adult male cares for embryos until hatching (Nymphon rubrum).

A couple of charming facts about pycnogonid child-rearing: It’s the father who’s in charge of tending to the embryos, and my favorite detail of all, the larvae consist of just the primitive head of the animal—no thorax, no abdomen, no legs. Just a head. The appendages in the drawing on the left are all the head appendages, and only later as it matures does the larva sprout a body. I think that is so cute.

The question is whether those big chelifores are homologous to spider chelicerae. How would we figure that out? In the table up above, you can see that spider chelicerae form on the second head segment…so we want to see to which segment the chelifores belong.

One way to work that out is to identify the neuromeres that innervate it. The insect brain is made up of lobes or ganglia called neuromeres associated with each segment, that extend nerves into the appendages also associated with that segment. In adults, these ganglia tend to run together and fuse, making their relationships harder to sort out, but their positions in the embryo are much clearer. The neuromere in segment 1 is called the protocerebrum, that in segment 2 is the deuterocerebrum, and the one in segment 3 is the tritocerebrum. If the pycnogonid chelifore is homologous to a spider chelicera, it should be innervated in the same way, by nerves from the deuterocerebrum in segment 2.

The answer is illustrated in the complicated figure below. The photograph is of immunostained neurons in the larval head, which are grouped into ganglia. The diagram in the top half labels the clumps of cells; in orange is PR, the protocerebrum, and two pairs of small nodes below, labeled A2G and A3G respectively, are the deutero- and tritocerebrum. Can you tell which ganglion has nerves (in red) running into the chelifores (LCH and RCH)?

a, Diagram of the protonymphon seen from an oblique posteriolateral view based on reconstructions from confocal stacks (b?f). The CNS consists of four pairs of ganglia connected by commissures across the midline. The oesophagus runs through the proboscis between the left (LCH) and right (RCH) chelifore, and ends incompletely at the posterior ganglia (PG). The first neuromere is the protocerebrum, consisting of anteriolateral cheliforal ganglia (CG) connected by a prominent supraoesophageal protocerebral commissure (PR), and ocular nerves (ON). b, High magnification of the protocerebrum stained for tubulin (red) and serotonin (green) showing the cheliforal ganglion (arrows) and bifurcating ocular nerves (b, arrowheads). Circumoesophageal connectives run posteriorly from the PR, leading to the second (A2G) and third (A3G) ganglia that innervate the second and third appendages (c). c, High magnification (α-tubulin, grey scale) of A2G and A3G and the fibrous appendicular commissures conecting the ganglia across the innervated oesophagus (O). d, Depth-coded image (α-tubulin): colours range from warm (red) indicating dorsal, to cooler (blue) indicating ventral. e, Transverse optical section (α-tubulin, grey scale) showing the circumoral shape of the protonymphon ‘neuropil ring’ in relation to the oesophagus (O); the anterior bifurcating cheliforal nerves (arrowheads) target the cheliforal ganglia at the top of the ring. f, Same view of the CNS as in e, stained for serotonin (green). Immunoreactivity is visible in the cheliforal ganglia (arrows), along the circumoesophageal connectives, the suboesophageal appendicular commissure, and separately in the posterior ganglia (PG). Note background staining in the tripartite luminal surface of the oesophagus (O). Scale bars, 25 µm.


Yep. The chelifores are not innervated by the deuterocerebrum, so this piece of evidence suggests that they are not homologous to chelicerae. They are innervated by the protocerebrum. They are something different. They seem to be homologous to certain other organs found in ancient stem-group arthropods known from the Cambrian, Kerygmachela and Leanchoilia and the best known of them all, Anomalocaris, all of which had a large structure on the front of their heads called the great appendage, which as near as can be distinguished in the fossil material, is part of segment 1 and was probably innervated by the protocerebrum.

I think it’s really cool that a distant cousin and descendant of Anomalocaris is still doddering about in the ocean depths.

As I mentioned above, this observation does shake up the arthropod family tree a bit. Below are two different cladograms. The right side is the current view, with the Pycnogonida and Chelicerata as sister clades, more closely related to each other than to the Mandibulata, insects and crustaceans. This view assumes chelicerae and chelifores are homologous structures. On the left side is the cladogram if they are not homologous, and the chelifore is instead homologous to the great appendage of the stem arthropods. That would mean the Pycnogonida are linked to the base of the arthropod tree, and are the most primitive branch of that distinguished group.


The tree to the left reflects the Cormogonida hypothesis, with Pycnogonida as sister group to remaining extant arthropods. The tree to the right reflects the classical hypothesis of pycnogonids as sister group to chelicerates.

With pycnogonids as the surviving member of the most basal arthropods, we can also infer something about the organization of the ancestral arthropod’s head and brain.

…our results support previous models of head evolution that predict that the original arthropod bore an acronless four-segmented head, encapsulating a tripartite circumoral brain rotated in an axial position, reminiscent of that found in onychoporans, nematodes and other cycloneuralians.

That’s one of the beautiful things about evolutionary biology. The data fit together so well, and every new observation brings us a clearer picture of our planet’s evolutionary history. We can still see echoes of the ancient world in the molecules of life.

Maxmen A, Browne WE, Martindale MQ, Giribet G (2005) Neuroanatomy of sea spiders implies an appendicular origin of the protocerebral segment. Nature 437(7062):1144-8.


  1. #1 anne
    October 20, 2006

    Whoa. I think you just caused a few of my brain cells to explode. I remember learning all about pycnogonids in the older days of my marine biology career. I thought they were the coolest things (despite having the term “spider” attached to them). Truly fascinating. Thank you for the post!

  2. #2 Nescio
    January 26, 2007

    Reading Raff’s list of segment homologies above, I wondered if it wouldn’t make greater sense to shift the onychophorans’ one step downward, leaving it without a homologue of the acron. Would I be correct in reading the “acronless” comment you quote at the end as supporting a such interpretation?

  3. #3 Rob Clack
    November 7, 2008

    I too, love pycnogonids, and like you, think it’s so cool that they’re related to Anomalocaris et al. I also love the casual way you drop that last paragraph in, as if it were almost an afterthought.

New comments have been temporarily disabled. Please check back soon.