I previously drew attention to the recent study that argues fossilized embryos are really fossilized bacteria. PZ Myers does a nice job of outlining the several lines of evidence that made it look like these fossils were embryonic in origin:

The cells were dimpled and shaped by adjoining cells, suggesting a flexible membrane—not a cell wall. This rules out algae, fungi, and plants.

The number of cells within each specimen was usually a power of 2. This is something we typically see in cleaving embryos, the sequence from 1 to 2 to 4 to 8 to 16 cells.

They were big. Typical somatic cells in animals are 5-10 µm in diameter, but ova can be a millimeter or more in diameter, and individual blastomeres (the cells in the cleavage stage embryo) can be several hundred µm across. These cells and the whole assemblage were in that size range.

The individual cells were uniform in size, as seen in many cleavage stage embryos, and contained organelles arranged in a consistent pattern.

They were often found encapsulated in a thin membrane, similar to the protective membrane around embryos.

In his original blog, Dissecting embryos from half a billion years ago , Myers focuses primarily on the similarities.

There are suggestive elements, such as the fact that some of the organelles are consistently aligned relative to the cleavage planes, suggesting that they might be nuclei, which do the same thing in modern embryos. Numerous smaller blobs suggest lipid granules, another common feature of developing cells. In particular, compare the “A” figures to “B”; “A” is a two celled embryo from Doushantuo, while “B” is a two-cell sea urchin embryo. The similarities are striking. (emphasis added)

and

The feature that jumps out at us from these fossils are their similarities to modern organisms at early stages of development. (emphasis not added)

However, Myers does raises a caveat.

However, the authors also not a significant absence, a difference from modern embryos: even in the later stage fossils of a thousand cells or more, there is no sign of epithelialization. This is a major lack. One of the hallmarks of the development of modern metazoans, from sponges on up, is the early formation of sheets of cells—even the terms diploblast and triploblast refer to the number of tissue layers formed, and in derived organisms like the zebrafish what we watch in early development is the formation of ectodermal and mesodermal sheets of tissue, and their movements and interactions. These embryos don’t exhibit any sign of doing that!

Yet notice how he interprets it.

The absence of such a key feature suggests that these animals are very primitive, and that what we’re seeing in this collection are the preserved remains of the embryos of stem-group metazoans.

Thus, a crucial fact that could be signalling the misidentification of these fossils is interpreted to mean the animals were simply “primitive.” It’s not clear how one defines “primitive,” but the intuitive notion can always be invoked.

In his new blog, Myers raises “concerns” that were not raised in his original blog. Yet in each case, notice there is a handy reponse.

Concern: “There are some concerns about the interpretation, though. One troubling aspect of their distribution is that they are all only in the cleavage stage: we don’t see any gastrulas, the stage at which embryonic cells undergo shape changes and begin to move in a specific, directed manner.”

Response: “Studies of taphonomy (analyses of the processes that lead to fossilization) have shown that these later stages are particularly difficult to preserve, which potentially explains why we’re seeing a biased sample.”

Concern: “Another unusual bias in the sample is that all of the embryos exhibit that regularity of division that produces equal-sized blastomeres—yet many invertebrate embryos have early asymmetric cleavages that produce recognizable, stereotyped distributions of cells.”

Response: “That asymmetry could be a feature that evolved late, but at the same time, some of the fossils were described as resembling molluscan trefoil embryos. Why aren’t the examples of early asymmetry translated into a later asymmetry?”

So what lessons might we draw from this example?

First, the example should make it clear that evidence is a function of perspective. It is the concepts that we bring to the table that allow us to convert raw data into something we call evidence. In this case, the evidence amounts to perceived similarities and analogies, whether it is comparing the fossils to modern-day sea urchin embryos or modern-day bacteria. Essentially, something that “looked like” an animal embryo now also “looks like” bacteria.

Second, despite the several lines of evidence in support of the embryo identification, and the responses that could be raised to the objections/concerns, it is now clear that the hypothesis is probably false. As one commentator on Myers’ site wisely noted:

Wow! I’m a high school physics teacher, trying to get my students to act like scientists, and not blindly accepting everything their teachers say. I’m going to give them this example in the next class, showing them that no matter how much evidence there is supporting a hypothesis or a theory, there could always be more evidence somewhere that changes everybody’s mind.

In other words, prior to the Bailey et al. study, one could posture as if science had indeed determined these fossils to be embryonic in origin and browbeat those who didn’t agree (let me mention that Myers did NOT do this). Thus, it is always important to remember the provisional and tentative nature of science when we explore our evolutionary past.

[See also Lessons from a Humble Fungus]