Experiment 5 - Reconstructing a Trilobite Tree

Experiment 5 - Reconstructing a Trilobite Tree

Team: ZAME Members: Zach, Aaron, Marisa, and Evan

Written by Marisa and Zach, Put Together by ZAME

     In this lab, we were given the task of organizing fifteen trilobite species into a phylogenetic tree. In order to do this, we all put on our scientist caps and approached the situation as systematically as possible. This involved staring at the critters and trying to decide what should be considered an ancestral trait, what should be derived, and what should be analagous (something that shows up, disappears and shows up again). After scientifically discussing (mostly arguing but there was definitely logic involved) for twenty minutes we finally agreed on a selection of noteworthy traits. We took these traits and formed a character trait matrix. While we aimed toward parsimony (a word that felt like some kind of zen-like state we desperately wanted our tree to achieve) the matrix took a long time to fill out and felt pretty detailed. Here's what it looked like when we were done:

Specimen	Eye Presence	Genal Spine Presence	Terminal Pygidium Spike number	Shell Texture	Glabella Shape	Axial rings	Axial Ring Spaces
1	Y	Y	0	Smooth	P	Y	N
3	N	N	0	Smooth	P	N	N
4	Y	N	0	Smooth	W	Y	Y
5	Y	N	0	Smooth	S	Y	Y
6	Y	Y	1	Smooth	P	Y	N
7	Y	Y	0	Smooth	B	Y	N
9	N	N? (there was disagreement)	0	Smooth	W	Y	Y
10	Y	Y	2	Rough	B	Y	N
11	Y	Y	2	Smooth	P	Y	N
13	Y	Y	0	Smooth	B	Y	N
14	Y	Y	1	Rough	B	Y	Y
16	Y	Y	0	Smooth	B	Y	N
17	Y	Y	2	Smooth	P	Y	N
18	Y	N	0	Smooth	S	Y	Y
19	Y	Y	2	Rough	W	Y	Y

Table 1. Character Matrix for 15 different species of trilobite. Y = yes, N = no, W = warty, P = Pointy, S = square, and B = bulbous,

      We then went to work on trying to reorganize the matrix in different ways. Each time we tried to organize the data by choosing one particular trait, none of the others really fell into some kind of pattern. Before we knew it, lab was almost over and we were still deliberating. Zach, seeing that we only had 10 minutes left and realizing that we hadn't made any real headway shouted "Let's just attack this problem like a five-year-old would! How would a kid organize these?" It turned out to be the thing we needed because after all of us worrying about details and complex connections we hadn't really gotten anywhere. But after Zach's rallying cry, we formed things into groups through morphological similarities and simplified the trilobite traits and actually got somewhere. We created a prototype tree that at least gave us three really good nodes that were consistent and something we could work with. So after snapping some pictures of the tree, we left and attacked the character matrix with our new insight.

     Aaron, using our proposed first three nodes, reorganized the data into a tree that made alot of sense:

Specimen	Axial Rings	Eye Presence	Genal Spine Presence	Terminal Pygidium Spike number	Shell Texture	Glabella Shape	Sealed Axial Rings
3	N	N	N	0	Smooth	P	N/A
Synapomorphy: axial rings
9	Y	N	N?	0	Smooth	W	N
Synapomorphy: eye presence
4	Y	Y	N	0	Smooth	W	N
5	Y	Y	N	0	Smooth	S	N
18	Y	Y	N	0	Smooth	S	N
Synapomorphy: genal spine presence
7	Y	Y	Y	0	Smooth	B	Y
13	Y	Y	Y	0	Smooth	B	Y
16	Y	Y	Y	0	Smooth	B	Y
1	Y	Y	Y	0	Smooth	P	Y
Synapomorphy: terminal pygidium spike presence
6	Y	Y	Y	1	Smooth	P	Y
11	Y	Y	Y	2	Smooth	P	Y
17	Y	Y	Y	2	Smooth	P	Y
Synapomorphy: texture
14	Y	Y	Y	1	Rough	B	N
19	Y	Y	Y	2	Rough	W	N
10	Y	Y	Y	2	Rough	B	Y

Table 2. Reorganized data from original character matrix in Table 1 to create a design for a usable phylogeny tree.

Me, not dealing with numbers as well as the epidemiology-major (in our character matrix the trilobites are represented by their card number alone) and needing some pictures to look at put together a visual tree based on his suggested organization. I printed out the little guys, put them on a white board, drew some branches and voilà! We had a tree we could work with! Evan then took my picture and made the much nicer-looking version here:


Figure 1. Our completed phylogeny tree.

     So what can we say with our tree? We determined that the trilobite species 3, Peronopsis intersiricta, is the best outgroup among the species we were presented with. This decision is based on the basic characteristics that separate it from all of the other species present, especially the lack of more than two segments within its thoracic region and its lack of eyes (shared only by species 9, Calymene celebra). Looking at traits, eyes seem to be a basal/ancestral characteristic. All of our trilobite samples have eyes except for species 9, which is separated from most of the other species early in this phylogeny due to its lack of the other species' ancestral traits. This suggests that it is more closely related to the common ancestor shared with species 3, which also does not have eyes. An example of a derived characteristic with our tree is roughly-textured exteriors. This is something only present in species 10, 14, and 19 but no other species. These three species form a monophyletic group on our phylogenetic tree, separated from other species bearing pygidium spikes by this texturing.

Also according to our tree, the rear spine of species 6 is homologous to that of species 14, if you choose to accept that all pygidium spikes are equal. In one possible scenario of arrangement (see our first table posted), we considered having zero, one, or two pygidium spikes to be individual traits and would have split species off in this way, but it became too much detail and made no obvious sense. So we chose to treat pygidium spikes more simply, looking at whether a trilobite had any at all (regardless of number) or none. In this way, as stated before, this makes them homologous.

As for our white-rabbit-in-a-hat trait, we decided axial rings were the only trait to have a disappearing and reappearing act in our tree. We quantified this trait as the presence of spacing between axial rings. An example of this is shown in figure 1. Notice in this comparison that between species 14 (left) and species 16 (right):

Figure 2. Left: Species 14, Dalmanites verrucosus. Right: Species 16, Ogygopsis klotzi. Note the spacing between each axial ring in D. verrucosus in comparison to the sealed axial rings of O. klotzi.

Although we chose not to use this as a defining characteristic originally (it wasn't in any of our character matrices), a trend did appear related to this trait: the monophyletic tree composed of species’ 9, 5, 4, and 18 all have spaces which then disappear and don’t reappear until the last branch of species’ 10, 14, and 19. We now claim this as an analagous trait.

While all of us had a part in growing our tree lovingly and with care, it is always important in science to be reflective and take note of other people's work towards the same goal. As it turns out, the phylogenetic trees drawn by our group and of many of the other groups share many striking similarities. The most notable of these was the choice of outgroups. Every group that we compared to had the same species, Peronopsis intersiricta, as the outgroup, noting the lack of eyes and the dramatically simplified body. Consequently, in almost every tree the first trait to develop among trilobites after this species was eyes. Early development of genial spines and pygidium spikes were also common trends among other groups trees. However, our analysis appears to be unique in including spaces between the axial rings of different species. Although this never did end up influencing the development our tree, the feature appears to be conserved in several groups. It is also clear that many other groups used features of the glabella and the cephalon to help sort species. Although there are many interesting observable differences between these features in different species of trilobites, either the features were too diverse for us to consider (as with the cephalon), or the groupings never appeared to synchronize with the categorizing of other features (as with the glabella). As a result, in comparison to other groups trees ours is rather simplified (but we like it it that way).

Although it is difficult to say which tree is better, we are compelled to think that we did an excellent job. The simplification of our tree has a lot to do with our group’s ability to criticize itself (we did ALOT of scientific arguing). Many different traits were considered and several were in fact analyzed, but eventually anything that could not be simplified to a binary presence/absence was considered to be too subjective to truly include in the formation of our phylogeny. Where many groups contain separate branches for the unusual traits of individual species, the conservative approach that team ZAME took in our analysis yielded a simple, compact, and easily defended tree. Also, it is aesthetically pleasing. You know, like a bonsai.

Thanks for reading!

-ZAME