email any corrections or suggestions (or alternative versions to stephn dot smith at yale dot edu) A good book chapter on parametric bootstrapping can be found at: Huelsenbeck, J. P., D. M. Hillis, and R. Jones. 1996. Parametric bootstrapping in molecular phylogenetics: Applications and performance. Pages 19-45 in Molecular Zoology: Advances, Strategies and Protocols (J. D. Ferraris and S. R. Palumbi, eds.). John Wiley & Sons, Inc., New York. Here begins the HOW-TO: Basically what one wants to do is use the orginal data set and create a tree as one would normally. 1. Choose the best model (modelgenerator, modeltest) 2. Run ML search. With a constrained and unconstrained tree (the constraints may be topological as they are in Ruedi et al. 1998 where one tree has a group constrained to be monophyletic and the other does not), in PAUP with the model you chose from the previous step. 3. Save the resulting trees into separate tree files. Be sure to save the distances (either with the menu under options, or with the command). 4. Record scores (parsimony) of each tree, and find the difference between the constrained and unconstrained trees. 5. Run Seq-Gen with the constrained ML tree and your model from step 1. Simulate a hundred new data sets (you may do more if you like, 1000?). You may want to add a paup block to make the runs go as a batch. 6. Analyze the simulated replicates in PAUP (constrained and unconstrained). Plot the frequency vs. differences (in steps) between constrained and unconstrained trees and plot the difference between the original trees (Your ML search). 7. You can also use extractscores to read paup log files from parametric bootstraps. 8. Analyze whether the difference was observed randomly from the simulation data this could be for example two of 100 with greater difference than yours not from simulated data, .02 < p < .03). modified from here