Symbolic machine learning methods induce explicitly represented symbolic models from data. The models can thus be inspected, modified, used and verified by human experts and have the potential to become part of the knowledge in the respective application domain. Applications of symbolic machine learning methods to ecological modelling problems are numerous and varied, ranging from modelling algal growth in lagoons and lakes (e.g. in the Venice lagoon) to predicting biodegradation rates for chemicals. This paper gives an overview of machine learning applications to ecological modelling, focussing on applications of symbolic machine learning and giving more detailed accounts of several such applications.