la .and. cusku di'e
The book is quite clear that ro as a quantifier is importing (16.8, as pc has just pointed out on Jboske). Like you, my preference would have been for nonimporting ro, but I can't see any grounds for overriding the book -- it's not inconsistent or 'broken' onthis point.
It sure is inconsistent on this point. According to the book, 'ro pavyseljirna xirma cu blabi' is false, since 'ro pavyseljirna' has existential import, and thus 'naku ro pavyseljirna xirma cu blabi' is true, since it is the negation of a false statement. According to ch. 16 sec. 11, this is exactly equivalent to 'su'o pavyseljirna xirma naku blabi', which is false, since once again it claims existence of unicorns, and so either the book allows contradictions, and should be called 'the complete zenban language', or we can disregard that silliness about 'ro' having existential import, and use 'ro' as is standard in mathematics at least (whether or not that is the standard use in logic, as pc seems very certain that it is not).
mu'o mi'e .adam.