Technical Comments

Comment on “A Jurassic ornithischian dinosaur from Siberia with both feathers and scales”

See allHide authors and affiliations

Science  24 Oct 2014:
Vol. 346, Issue 6208, pp. 434
DOI: 10.1126/science.1259983


Godefroit et al. (Reports, 25 July 2014, p. 451) reported scales and feathers, including “basal plates,” in an ornithischian dinosaur. Their arguments against the filaments being collagen fibers are not supported because of a fundamental misinterpretation of such structures and underestimation of their size. The parsimonious explanation is that the filaments are support fibers in association with badly degraded scales and that they do not represent early feather stages.

Godefroit et al. (1) attempt to dismiss the hypothesis of collagen fibers versus that of feathers in the ornithischian dinosaur Kulindadromeus with three main lines of arguments, which necessitate a reply.

1) “First, integumentary collagen fibers typically occur in layered arrays of parallel, densely packed fibers where fibers in successive layers are oblique to one another” [supplementary materials (SM) for (1)]. They are not typical but form a highly specialized design architecture associated with special biomechanical functions—for example, enabling stiffness and mobility either of the whole body (26) or of organ systems such as arteries (7), rectal sheaths, and linea alba (8). In humans, as in many other vertebrates, the dermal fibers are randomly (here, nonopposing orientations) organized. Even in animals with this specialist design, numerous layers of fibers may have the same orientation in one location of the animal and an alternating crossed-fiber orientation in another, dependent on function. Furthermore, fiber angles are rarely consistent but may change from low angles to high depending on the tension required in different parts of the body (4, 8). In fact, rather than consistent angles as alleged (1), in different zones in Kulindadromeus [figure 2, F and I, in (1)] varied filament angles are evident. As emphasized (6), the unique exposure of a crossed-fiber architecture in a 130-million-year-old Psittacosaurus fossil could only emanate “from freak combinations of circumstances.”

2) “[I]ntegumentary collagen fibers are typically on the order of several microns in diameter; the structures we describe are at least two orders of magnitude larger” [SM for (1)]. The integumental structures identifiable at light microscopy level in dinosaurs and ichthyosaurs and many extant vertebrates are not just several microns thick and are collagen fiber bundles, not fibers. Although most experts in this field refer to the components of the cross-fiber architecture as fibers, it is clear to workers involved that they are fiber bundles [see Pabst’s (3) table 2, “diameter of collagen fibres” in dolphin hypodermis range “208-469 μm” and white shark (Carcharodon carcharias) and ichthyosaur (Stenopterygius) dermis (4, 5), up to 1000 μm]. However, anticipating that workers unfamiliar with histological research might misinterpret the vernacular term of fibers, I consequently clearly defined its use (9) and reiterated it in a study (6) the authors attest knowledge of: “Note that as with most studies in which dermal fibre architectures have been investigated [e.g. sharks and dolphins in Motta (1977) and Pabst (1996), respectively], reference to the term ‘collagen fibre’ is one of convenience because, strictly speaking, observations at the light microscopic level show fibre bundles (approx. 50–250 μm across) rather than fibres (approx. 4–20 μm across).” Hence, it is of concern that the authors misinterpret the size of the integumental structures that they imply familiarity with by “at least two orders of magnitude” [SM for (1)].

3) “[M]orphology of the compound and ribbon-like integumentary filaments (Fig. 3D-H) is unknown for integumentary collagen” [SM for (1)]. Generally, figures allegedly showing feather-like structures are low-resolution images of badly degraded sections, confirmed by the need for interpretive drawings on the alleged vital structures (figure 3, C, F, and I, and figure S9C), which are highly subjective. Beyond alleging what the ribbon-like filaments are not—i.e., collagen—the authors do not say what they are (besides unsupported speculation). Notwithstanding, they have not explained why they cannot be among a multitude of alternatives—e.g., ribbon-like structures occur as structural collagen in blood vessels (7), linea alba, and rectal sheaths (8) and in connective tissue, the most abundant collagenous material in vertebrates (Fig. 1F). They could be plant or inorganic material in the proximity of the dead animal or even the degraded remains of very thin (highly degradable) ribbon-like scales—e.g., found in the ventral regions (aiding distension) of, for example, Naja mossambica (cobra) and Sinosauropteryx (10) (Fig. 1E and inset).

Fig. 1 Integument and decomposition.

Integument: (A and B) Scales of Jesus lizard along skin flap (A), back-lit, showing underlying collagen fibers. (C) Severely eroded scales (ghosts) in Psittacosaurus [also figure 2 in (6)] showing underlying fibers. (D) Pebble-like scales in Kulindadromeus, adjacent to “basal plates.” Decomposition: (E) Ribbon-like scales in ventral regions in Naja and Sinosauropteryx (inset) (10). (F) Bifurcating connective tissue bands around the gut of a juvenile ostrich, Struthio camelus.

Poor preservation renders the interpretation of filaments arising from a “basal plate” (1), given the enormous evolutionary implications, unsound. A parsimonious explanation is evident in figure 3A in (1). On the left are filaments associated with “basal plates” (both poorly preserved). On the right are somewhat less poorly preserved scales without filaments. The “basal plates” on the left are in fact severely degraded scales, but otherwise they resemble in size the scales on the right and the size class given by the authors for small scales, <3.5 mm, given variability of scale size in reptiles, even in close proximity (Fig. 1A). The explanation being self-evident, that more severe degradation and erosion have exposed the underlying filaments and resulted in loss of scales—and hence apparent greater spacing—in contrast to the reasons for proposing basal plates (1). In some places, all that may remain are the pigment impressions from the overlying scales (11) around the underlying filaments (1). We know that such filaments underlie scales in Basiliscus, the Jesus lizard (Fig. 1, A and B), and in Psittacosaurus R 497 [Fig. 1C; also figure 1B in (6)].

Given the severe level of degradation and distortion of soft-tissue structures in Kulindadromeus, the authors’ proposals of primary preservations of unprecedented structures allegedly depicting finite filament lengths, filament groups threaded through single and sometimes multiple “basal plates,” and filaments and plates preserved in three dimensions (1) are unjustified. The proposals are further weakened by a disregard for taphonomic tribulations (12) of a more than 150-million-year-old fossil and the complexities of tissue histology.


Stay Connected to Science

Navigate This Article