Homework 3. Due Wed Feb 6, by noon.

 

1. Go back to Kevin Padian's testimony regarding homology, tasmanian wolves, kangaroos, etc.

Based on the traits presented, draw a phylogeny that shows the relationship between N. American wolves, dogs, tasmanian wolves, kangaroos, and oppossums. On the phylogeny, mark and label the appropriate branches with synapomorphic characters that unite subgroups.

4pts. Your phylogeny should show the groupings, and below each group you should show the synapomorphies. Important ones that clearly group the marsupials separately from the placentals are:

a) placenta; b) # of molars; c) presence/absence of palatal holes; d) reflected lamina; # of pre-molars; e) presence/absence of ear bulla; f) presence/absence of para-occipital process; g) carnassial teeth; position of lacrimal bone;
In addition, you could have added the "overall dog like morphology" that the creationists seized on.

 

 

The following questions refer to the squirrel article by Mercer et al.

In class we'll cover: What data were used to estimate the squirrel phylogeny? ; In figure 1, which branch do we have the least confidence in? Which one(s) do we have the most confidence in?

  1. 2pts. In figure 2, how did the authors determine the timing (in years (Ma) before present) of branching events? They used a molecular clock that was calibrated with the oldest known squirrel fossils, dated 36 Ma (mega-annum). [In the online supplement, one learns that they used only the IRBP gene because it exhibited more clock-like behavior than the mt rRNA genes.]

  2. 3pts. Aplodontidae (Mountain Beaver, found only on the PNW coast) is not a squirrel. Why is it included in the phylogeny? Aplodontia is an outgroup for the squirrel phylogeny. Based on other evidence (for example, fossil evidence), this group is thought to have branched off earlier than the squirrel clade, which makes it suitable as an outgroup.  Including an outgroup allows determination of the order of evolutionary change at each site in the DNA sequence. In turn, this allows determination of the temporal order of branching events (i.e. which squirrel clade branched off first, second, etc.).

  3. 3pts. Do the results suport the traditional claim that flying squirrels form a group relatively distant from all other squirrel lineages? No. Flying squirrels form a sister group to North American Tree squirrels, and are nested well within the squirrel family. This clearly indicates they are closely related to the N. American tree squirrels (and that they arose later in the evolution of squirrels rather than earlier). [If they had arisen earlier, they would be on a branch that comes off earlier, like that of Ratufa].

  4. Bishop thinks that the congruence of major diversification events and major changes in sea-level or formation of land bridges is remarkable. One reason is that this congruence required agreement of several independent types of data: 1) Identifying the nature and timing of geological/sea-level changes based on the geologic & biological (non-squirrel) record; 2) The branching pattern of the phylogeny, based on three different genes; and 3) the calibration of DNA substitution rates using fossil squirrel ancestors (note, this last item really requires several independently collected types of data – fossil position in geologic record, radiometric dating of fossil age, and DNA substitution rates on each branch).

    Questions : 4a) 4pts. What do these congruences suggest about factors affecting squirrel diversification? Your answer should refer to figure 3 as an example.
                The fact that these diversification events co-occur with major geographic changes suggests a causal relationship between geographic changes and squirrel diversification. [That’s right, squirrels, among the most amazing of nature’s engineers, caused major changes in sea-level and built land-bridges….;>)]. Each of these geographic changes allowed colonization of a new region by squirrels. These regions likely offered ecological opportunities with low competition for the colonists, and this likely resulted in the formation of new species (and genera). The fact that sea-level change later isolated island colonists prevented ongoing migration and also promoted divergence. 
                The correspondence between sea-level change, colonization, and diversification is beautifully illustrated by figure 3. The timing of the lowest sea level in the entire Pre-pleistocene Cenozoic [65mya-1.5mya] corresponds precisely to the timing (estimated by molecular clock) of the base of explosive diversification of squirrel genera on the Sunda shelf islands. Also at this time, the only lineage to cross Wallace's line branches off. [Wallaces Line dramatically separates Asian and New Guinea/Australian faunas].
                To sum up: In the words of the authors (last sentence in abstract), “These precise correspondences point to an important role for global change in the diversification of a major group of mammals.”


    4b) What does this study suggest about scientist's ability to reconstruct pre-historical events? You may wish to reference the 7 criteria for supporting scientific theories or hypotheses. 
    [Note: The best way to answer was to try to apply the criteria. It may be the case on questions like this that not all criteria apply. The study wasn't intended to be a test of the hypothesis that we can re-construct pre-historical events, because this fact has been demonstrated ad nauseum by scientists (and also by the actors on C.S.I.). However, but I wanted to get you to think about it re-constructing history and to apply our evaluation criteria.]
    4pts.
  1. If it better matches the data from a fair test.
    The phylogeny itself, including the scaling of branch lengths to represent time via a molecular clock, is a hypothesis about history. The test of that hypothesis is whether the diversification events correspond closely to major changes in earth history that allowed access to new geographic regions. It passes this test to an extraordinary degree.

  

  1. If it is confirmed by multiple independent fair tests
    Its not clear to me if there are multiple independent tests, or if we should just regard it all as one really strong test, but there are a number of facts that could be regarded as additional tests. For example, if lowered sea level really promoted colonization and diversification of squirrels on the Sunda Shelf islands, we might expect that other animals experienced the same thing. The authors point out the fact that flightless grasshoppers divergence has been pegged to the same time.  Another example is that the earliest known tree squirrel fossils in Africa correspond closely to the base of the diversification of these squirrels in Africa.

  2. If initially conflicting data can be shown to agree I'm not sure we have any cases of that here.

  3. If there are no conflicting lines of evidence
    One conflicting line of evidence is that Rheithosciurus, a giant squirrel from Borneo, nests very solidly within N. American tree squirrels. The authors choose to treat this as an anomaly in need of explanation, rather than an indication that the entire analysis is suspect [one certainly wonders if they mixed up a sample!]

  4. If the fair test that supports it is particularly strong
    In my opinion, the test for correspondence of branching events and land-bridge formation is an extremely strong one, because it required so many independent sources of data to correspond exactly. The timing of sea-level change and land-bridge formation comes from a completely separate scientific discipline with many independent authors, methods, and data sets. For the diversification times to correspond to geologic events, branch scaling via molecular clock required a) the right # of substitutions per branch; b) that the earliest squirrel fossil be found at a certain layer of the geologic column [also done by different scientific discipline]; and c) that the radiometric dating methods obtain accurate dates.

 

  1. If the alternatives are seriously defective conceptually
                I.e. no data suggest, or it is untestable
    They are, but this is not discussed in the paper.
  2. If the overall weight of evidence is greatly in its favor
    Again, it is, but its not discussed in the paper.