mm term behavior I can't explain

[drh20drh20]

In the vignette, one of the model terms in the first formula is:

    ~mm("role", levels = I(roleset),
        levels2=~.%in%list(list(row="Father",col="Mother"),
                           list(row="Father",col="Child"),
                           list(row="Mother",col="Child"))) +

...yet only the Father-Mother statistic is included in the model:

> summary(f.wd)
                                             N(1)~edges 
                                                   1300 
                                 N(I(n <= 3)TRUE)~edges 
                                                    162 
                                 N(I(n >= 5)TRUE)~edges 
                                                    520 
                       N(1)~mm[role=Father,role=Mother] 
                                                    199 
N(1)~F(nodematch("role",levels=I("Child")))~nodecov.age 
                                                   4752 
                                            N(1)~kstar2 
                                                   2926 
                                          N(1)~triangle 
                                                    934 
                              N(I(n >= 6)TRUE)~triangle 
                                                    195 

[krivit]

The rows and columns for "Child" were swapped: they need to be alphabetical.