ABSTRACT
Extrapolating microevolutionary models does not always provide satisfactory explanations for phenotypic diversification on million-year time scales. For example, short-term evolutionary change is often modeled assuming a fixed adaptive landscape, but macroevolutionary changes are likely to involve changes in the adaptive landscape itself. A better understanding of how the adaptive landscape changes across different time intervals and how these changes cause populations to evolve has the potential to narrow the gap between micro- and macroevolution. Here, we analyze two fossil diatom time series of exceptional quality and resolution covering time intervals of a few hundred thousand years using models that account for different behaviors of the adaptive landscape. We find that one of the lineages evolves on a randomly and continuously changing landscape, whereas the other lineage evolves on a landscape that shows a rapid shift in the position of the adaptive peak of a magnitude that is typically associated with species-level differentiation. This suggests phenotypic evolution beyond generational timescales may be a consequence of both gradual and sudden repositioning of adaptive peaks. Both lineages are showing rapid and erratic evolutionary change and are constantly readapting towards the optimal trait state, observations that align with evolutionary dynamics commonly observed in contemporary populations. The inferred trait evolution over a span of a few hundred thousand years in these two lineages is therefore chimeric in the sense that it combines components of trait evolution typically observed on both short and long timescales.
ABSTRACT
Fossil evidence indicates that modern assemblages of temperate nonmarine planktonic diatoms began near the middle/late Miocene boundary when the genus Actinocyclus, an important constituent of lacustrine planktonic diatom assemblages during the early to middle Miocene, was replaced by genera of the family Stephanodiscaceae. This floral turnover has been confirmed in many regions of the world, except eastern Asia where taxonomic data about early and middle Miocene planktonic diatom assemblages have until recently been scarce. Our analysis of Lower and Middle Miocene lacustrine diatomaceous rocks in Japan confirms that species of nonmarine Actinocyclus were important constituents of lake phytoplankton there as well. The appearance of nonmarine Actinocyclus species near the beginning of the Miocene may have resulted from the introduction of euryhaline species into lacustrine environments during a highstand of sea level at that time. Similarly, it is possible that species of Stephanodiscaceae evolved from marine thalassiosiroid ancestors that invaded high latitude lacustrine environments during multiple Paleogene highstands, resulting in a polyphyletic origin of the family. The turnover from nonmarine Actinocyclus to Stephanodiscaceae genera near the middle/late Miocene boundary may be linked to a contemporaneous increase in silica concentrations in lakes caused by active volcanism, increased weathering of silicate rocks due to orogeny, and the expansion of C4 grasslands. This turnover may also have been influenced by enhanced seasonal environmental changes in the euphotic zone caused by the initiation of monsoon conditions and a worldwide increase in meridional temperature gradients during the late Miocene. Morphological characteristics of Stephanodiscaceae genera, such as strutted processes and small size, suggest their species were better adapted to seasonal environmental changes than nonmarine species of Actinocyclus because of their superiority in floating and drifting capabilities and possibly metabolism, intrinsic growth rate, and reproductivity. As climates deteriorated during the late Miocene, Stephanodiscaceae species may have spread from high latitudes to temperate lakes where they diversified, ultimately displacing Actinocyclus.
