Reconstructions of Ediacaran putative biomineralizers via a novel serial grinding and imaging technique

Abstract

550 million years ago [Ma], shallow water settings on Earth witnessed the rise of an entirely new kind of organism: the putatively shell-building metazoan. By the early Cambrian (beginning 539.6 Ma), when most modern phyla rst emerged, animals with hard parts were both engaging in an arms race and actively engineering their surroundings. Before we can understand why biomineralization evolved, or its immediate ecological, environmental, and/or evolutionary ramifications, we must understand the Ediacaran skeletal fossil record.

This record comprises four genera: Namacalathus, Sinotubulites, Cloudina, and Namapoikia. To date, the growth habits, environmental impacts, biological anities, and the modes and degrees of biomineralization of these organisms remain debated. Morphological analyses (including descriptions of the size, shape, and orientation of fossil elements), combined with comparisons to more recent organisms, can provide crucial insight into the paleoecology of Ediacaran putative biomineralizers.

Two-dimensional observations are subject to measurement error and misinterpretation, such that studies of morphology must be done using three-dimensional data. Unfortunately, Ediacaran putative biomineralizers largely are preserved as carbonate fossils within a carbonate matrix, precluding the use of traditional, density-based imaging techniques, such as X-ray computed tomography. A solution to this problem is to use serial grinding and optical imaging, a process in which a sample is incrementally polished and photographed to produce a stack of images for reconstruction.

Here, I present a new, semi-automated method for serially grinding, imaging, and reconstructing samples with low density contrast. This method comprises both a hardware and software component: data are first collected using a computer numerical control surface grinder that has been retrofitted with an imaging apparatus and then is processed through a neural-network based imaging pipeline. In addition to providing case studies that leverage this newly developed method, I apply it to the examination of two Ediacaran putative biomineralizers. Combining three-dimensional reconstructions of Cloudina and Namapoikia with field-based observations, I demonstrate that Cloudina likely was a weakly-to-non-biomineralized annelid and that Namapoikia was not an animal at all, but a microbial construction. On the basis of my findings, I suggest that rigid skeletons did not emerge until the midst of the Cambrian Period.