To explore new mechanisms for planktonic thin layer formation, particle and continuum models of gyrotactically swimming phy-toplankton are embedded in simulations of a dynamically unstable stratified shear layer. Two trapping mechanisms are observed in the developing Kelvin-Helmholtz (K-H) billow train. Within the K-H billows, a particle can remain preferentially in downwelling regions, cancelling its upward swimming motion. In the braids that separate the billows, intense shear defeats the gyrotactic stabi-lization mechanism and causes cells to tumble. Particle and continuum models are compared statistically to reveal both consistencies and weaknesses in each. A scaling based on Reynolds number and swimming speed is used to predict the maximum concentration generated by an instability event. Although K-H billows are short lived in comparison with planktonic thin layers observed in the coastal oceans, the resulting trapping causes aggregation at rates which could account for the observed cell concentrations.