how to produce different ensemble object by bag algorithm

Using fitensemble(X,y,'bag',NLearn, Learners) can produce an ensemble object. Changing Learners can obtain different objects. So I use

tmp=templateTree('MaxNumSplits',10);

ens=fitensemble(X,y,'bag',100,tmp,'type','regression');

I got error message: 'MaxNumSplits is not a valid parameter name'. Where do I make mistake?

What are the common ways that can produce different ensembles in bag method?