Connecting the Silica Dots: Caroline's Scientific Journey

Caroline A.E. Strömberg, Ph.D.
Estella B. Leopold Assistant Professor in Biology and Curator of Paleobotany
University of Washington

As Albert Einstein famously stated, "imagination is more important than knowledge" when it comes to science. The following example shows, I think, that science also benefits from the willingness to take risks. This is what I (and my dissertation committee) did when I decided to try these mysterious things called "phytoliths" as a means to reconstruct the evolution of grasslands in the Great Plains of North America.

I had become interested in a classical problem in paleontology, namely the idea that the evolution of large, long-legged horses with tall cheek teeth was an adaptive response to the spread of grassland vegetation in the Great Plains. Since the days of T.H. Huxley, Darwin's bulldog, researchers had looked at the fossil record of horses and upheld it as the perfect example of natural selection to a changing environment. Sure, the horses show a convincing evolutionary sequence, but what do we really know about the vegetation? The problem was that 150 years of research had revealed that there were virtually no pollen or macrofossils from the Great Plains that could confirm that a change in vegetation from forest to grassland coincided with the evolution of (presumably) grass-eating horses. What to do? I started reading about grasses and grass fossils in search of any leads. I quickly realized that stable isotopes, which was the en vogue tool for grassland research at the time was not going to be helpful. However, in my readings, I stumbled upon a mention (in a paper by Joseph Thomasson) of something strange that I'd never heard of before: phytoliths, which are microscopic bits of silica that form in the tissues of living plants. At the time, the existence of phytoliths was unknown to most paleobotanists. Those familiar with phytoliths considered them worthless for deep-time paleobotany because there was a widespread notion that they wouldn't preserve in sediments older than a few thousand years. Still other workers thought that only phytoliths of grasses might be preserved, which would make them useless for reconstructing changes from forest to grassland. In contrast, archaeobotanists have long considered phytoliths preserved in soils and lakes as great tools for inferring past vegetation and human land use, among other things.

This is where I think my naivete gave me an advantage. Rather than taking the message conveyed in the literature at face value, I thought phytoliths are made of silica, so they should have a high potential of being preserved just like diatoms. They will be perfect for testing the hypothesis that grassland evolution drove horse evolution. I also found some old, published drawings of what looked like possible phytoliths found in Eocene sediments from the Great Plains and I met Francesca Smith (currently at Northwestern University), who was extracting late Miocene grass phytoliths from Great Plains sediments for isotope studies. This gave me confidence that my hunch was right, and I convinced my dissertation committee that it would be worth it to spend a year pursuing phytoliths as a potential source of paleobotanical data for the Eocene-Miocene of the Great Plains. Remarkably, they said yes!

I spent a summer driving around the Great Plains, from Montana in the north to Texas in the south and everywhere in between collecting sediment samples to test my ideas. The remainder of the year I was going to learn how to extract phytoliths from these samples. This was more easily said than done. I spent fall semester experimenting with methods published in the archaeological literature, but nothing worked, because the sediments I had were very different from the ones that archaeologists normally deal with. Finally, I got a phytolith expert, Dr. John Jones (then at Texas A&M), to kindly agree to show me his method of extraction. To my and Dr. Jones great excitement, his method worked- my first set of samples yielded beautifully preserved 17 and 28 million-year-old phytoliths from grasses and many other plants alike! I spent the rest of the year adjusting the method slightly and extracting more samples. As it turned out, phytolith assemblages were preserved throughout the record, from 40 to 9 million years ago. The paleobotanical record of the spread of grasslands was no longer missing!