The phylum Chaetognatha ("spiny jaws") consists of a few species (there are only about 120 of them) of obscure marine worms that are mostly planktonic predators, drifting with the ocean currents. Most people will have never seen one, and would be unimpressed if they had—most are small, only a few centimeters long and slender, and even the largest is only 15cm long.

They are also called arrowworms, and you can tell why from the drawing at the right. They are long and thin, with a set of lateral fins along the posterior part of their body. Their lifestyle involves floating passively most of the time, until something tasty like a fish larva or copepod drifts into their neighborhood, and then they dart forward rapidly and snag it with that head full of wicked spines (some species also secrete neurotoxins and poison their prey.)

Ah, a faceful of poisonous spines—this creature is getting more interesting already. It always seems that the closer you look at something in biology, the cooler it gets. Here are some closeups of the heads of chaetognaths:

These guys have a lot of potential to be great horror movie stars.

There's another reason that the chaetognaths are fascinating, though: their phylogenetic relationships are very, very murky. All the evidence suggests that their lineage split off from that of more familiar animals long ago, well before the Cambrian, at about the time of the deuterostome/protostome split. They aren't just monsters, they are primeval monsters, weird relics of an ancient world. Way cool!

The deuterostome and protostome distinction is thought to be a fundamental one, reflecting an early decision about how to organize the body plan. It refers to the fate of regions that are set up during gastrulation. In gastrulation, a two-layered body plan becomes a three-layered one by folding in a sheet of cells through an opening. The end result of gastrulation is that you've got an embryo with ectoderm (which will become skin and nervous system) on the outside, endoderm (future gut) deep inside, mesoderm (muscle and connective tissue) sandwiched in between, and a hole to the outside, called the blastopore. Where deuterostomes and protostomes differ is in what they do with the blastopore.

Protostomes, as they develop, turn the blastopore into their mouth opening. The term "protostome" means "primary mouth", and refers to the fact that they use the first embryonic opening as the mouth opening. Deuterostomes instead use that first opening as their anus, and have to subsequently create a second embryonic opening to form their mouth (hence the name, which means "secondary mouth"). The early body plans of deuterostomes and protostomes are both backwards and upside down relative to one another, which is why this is considered a primitive and fundamental distinction—wholesale reversals of the body plan are not something modern organisms can do lightly, and it had to have occurred way, way back when morphology was simpler and more ambiguous among our ancestral worms.

In case you are wondering, we chordates, along with our cousins the echinoderms (sea urchins and starfish) are deuterostomes: we turn our blastopore into our anus. More distant cousins, the annelid worms, molluscs, and arthropods, are all protostomes.

What about the chaetognaths? The chaetognath blastopore becomes the chaetognath anus, so by that property they are definitely deuterostomes. They have another feature, the coelom, or internal fluid-filled space that acts as a hydrostatic skeleton, that arises from the archenteron (another embryological cavity), which is also a deuterostome property…but it arises in a different way than it does in all other deuterostomes. Furthermore, other features like the organization of the musculature in the adult, the absence of a deuterostome-like ciliated larval stage, and preliminary molecular evidence, have all said these creatures have tighter affinities to the protostomes than to the deuterostomes. They are so confusing! Check your taxonomy textbooks, and you'll consistently find the chaetognatha tucked away in some head-scratching back chapter, lumped with the other ambiguous phyla, and mentioned with references to ongoing debates about their relationships.

Fortunately, molecular biology is giving us more characters to analyze and is helping to sort out the family tree. Helfenbein et al. have just published a paper in which they analyze the mitochondrial genome of a chaetognath and compare it to that of many other species, and what they've found is that these marine worms, despite forming their anus from their blastopore, are otherwise closer to the protostomes than they are to us deuterostomes. I won't go into the details—I'll just say that the paper also discusses many interesting peculiarities of the chaetognath mitochondrial genome, such as an extreme reduction—but will just jump directly to the conclusion of their phylogenetic analysis:

The species of chaetognath they studied, Paraspadella, is off on a branch by itself, but it's a branch off the lineage that led to protostomes in good standing, like molluscs and annelids and arthropods. How could this be?

One possibility is that chaetognaths are protostomes that, at some distant time in their history, flipped their body plan around. This would mean that deuterostomy evolved at least twice, once in our history, and once in arrowworm history. This seems a bit improbable.

The other, more likely possibility is that deuterostomy is the ancestral condition, and that the protostome lineage made the reversal once, after the chaetognath lineage had split off. That means that the ancestor at the yellow dot I've put in this diagram is the one that first evolved protostomy. It also means that we chordates can proudly say that we've conserved that initial pattern of development, and it was our crazy radical cousins the protostomes that at some point started eating with the hole we use to excrete with. I suppose we could get all snooty about it, except that it is that branch of the family tree that has gone on to become more diverse, more numerous, and generally more successful than our own.

Helfenbein KG, Fourcade HM, Vanjani RG, Boore JL (2004) The mitochondrial genome of Paraspadella gotoi is highly reduced and reveals that chaetognaths are a sister group to protostomes. PNAS 101(29)10639-10643.

Telford MJ (2004) Affinity for arrow worms. Nature 431:254-256.