The evolution of phenotypic traits is usually studied on generational times or across species on million-year timescales. We bridge this conceptual gap by using high density sampling of a species lineage, Microporella agonistes (Bryozoa, Cheilostomatida), over 2 million years of its evolutionary history, to ask if trait-fitness associations are consistent with evolutionary trait models often applied to phenotypic time series. We use average fecundity and competitive outcome as two different fitness components, where competitive outcome is a proxy for partial survival. Examining three quantitative traits in multivariate analyses, we present evidence that some traits experienced substantial selective pressures, in part controlled by past environments. A complex interplay of resource competition with an altering set of competitors and past temperatures, has contributed to the changing patterns of phenotypes within the focal species. A comparison with congeneric species living in the same regional community suggests that size traits are more temporally variable and less constrained than shape traits. Our analyses also show that while controls on phenotypes are complex and varied in time, ecological and evolutionary processes that unfold on shorter time scales are not inconsistent with macroevolutionary patterns observed on longer timescales.

Fossil and contemporary bryozoan colonies were collected over several fossil field expeditions to the Whanganui Basin of North Island New Zealand and dredges in Cook Strait and Taranaki Bight. The bryozoans were encrusting on hard substrates, mainly bivalve shells. Images of bryozoans were taken with Scanning Electron Microscopy and data of colonies and images extracted using a combination of deep-learning (DEEPBRYO) and manual measurements/collection. All the scanning electron microscope images associated with the work are supplied here and image names are associated with detailed metadata.

Funding provided by: The Research Council of NorwayROR ID: https://ror.org/00epmv149Award Number: 314499 Funding provided by: European Research CouncilROR ID: https://ror.org/0472cxd90Award Number: 724324