American Scientist

If you grew up visiting natural history museums, you probably think you have a good sense of what dinosaurs looked like, and the posed displays probably gave you an image of how they behaved. But recent studies of dinosaur eggs may shift your mental picture of these diverse creatures. In this issue’s Perspective column, “A New Picture of Dinosaur Nesting Ecology,” paleontologist Daniel T. Ksepka discusses findings of certain dinosaur eggs with soft shells, a feature more like modern turtle eggs than birds or crocodilians. Startlingly, it’s likely that the very first dinosaur eggs were also soft-shelled. It had long been assumed that all dinosaur eggs were hard-shelled, like their modern evolutionary descendants. It’s likely that hard-shelled eggs arose in three different lineages independently, rather than from a single ancestor. And some hard-shelled dinosaur eggs have been revealed to have had colors, a feature that arises when eggs are in partially exposed nests rather than buried underground, which is more of a birdlike trait. Perhaps it should not come as a surprise that dinosaurs would have some traits that are more like birds, other traits that are more reminiscent of reptilians, and yet others that are uniquely their own, not found in either. As Ksepka notes, “Evolution doesn’t always align with easy assumptions about early and modern traits.”

Shifting perceptions in research can alter the way that we look at many species. In “Is Garlic Mustard an Invader or an Opportunist?”, ecologist Michael Anderson takes a different view of what has commonly been thought to be an invasive plant, deserving of eradication, because early studies found diminished numbers and diversity of native plants in its presence. Much research has now been devoted to determining whether garlic mustard outcompetes local plants, or if it just took advantage of the sweeping changes in the ecosystem that humans have made over the past few decades. If it’s more of an opportunist, ecologists and land managers may need to rethink their perception of this plant as a weed that needs immediate and thorough removal. Anderson observes that “the spread of garlic mustard might be an effect rather than a cause of changes that are occurring in native ecosystems. The plant may largely be riding on the coattails of favorable environmental changes driven by humans.”

New data can also update the accuracy of mathematical models, which are often used to make predictions and to provide recommendations to decision-makers, so these simulations need to produce verifiable results as well as they possibly can. In “In Models We Trust—But First, Validate,” data scientist Daniel Solow discusses the different methods that can be used to evaluate the output of mathematical models, including common-sense reality checks, comparisons to experimental data, updating of parameters, showing confidence intervals, or performing sensitivity analysis. Solow reminds us that “Regardless of the amount of validation, one must accept some degree of uncertainty with model predictions, so some amount of skepticism is always warranted. The more you know about how a model is validated, the better you can decide how much to trust its predictions.”

How do you try to keep an open mind in your research, as trends shift? Write to us and let us know. —Fenella Saunders (@FenellaSaunders)