What is the default method of splitting in decision trees?

The default splitting method in decision trees is often based on measures like Gini impurity (for classification) or mean squared error (for regression).