Do sex differences in autism symptomology explain the sex differences in diagnosis rates?

A recent study finds that, despite the fact that boys are four times more likely to be diagnosed with autism than girls are, there are no major differences in the constellation of symptoms shown by boys vs. girls, with the possible exception of restrictive repetitive behaviors (RRBs).

subsequent meta-analysis revealed only small and statistically non-significant overall effects, suggesting no robust sex differences in any domain examined here, with the possible exception of RRB where features were greater in males. Hence, on the basis of the published empirical literature to date, we conclude that there is currently no substantive evidence of sex differences in reciprocal social interaction and communication, social functioning, cognitive abilities or co-occurring conditions among autistic toddlers and young children, but that females may demonstrate fewer RRBs than males in early childhood.

However, given persistent concerns that there's a gender bias in autism identification that leads to undiagnosed autistic females missing on early interventions, the authors don't leave it at that. Instead, they raise two possibilities.

One is that autistic females are better at masking their social symptoms than autistic males are. Social masking, however, strikes me as (1) hard to measure; (2) something that all of us do, at least to some extent; and (3) something that may reflect social skills rather than social deficits.

Second, the authors propose that there may be a distinctive male "autism phenotype" that doesn't fit the standard criteria. This strikes me as problematic for two reasons:

First, the autism phenotype dates back to the late 1930s to early 1940s and to the clinical observations of Leo Kanner, one of the first child psychologists to identify the autistic condition as a constellation of symptoms. Kanner based this identification on a subset of the children who had been referred to his clinic. This subset, consisting of 8 boys and 3 girls, all shared a similar constellation of symptoms. Based on this shared symptomology, which amounted to a combination of social aloofness and restrictive, repetitive behaviors, Kanner coined the term "autism." 

It's unclear where gender bias would fit in here.

Second, ever since Kanner, autism has been diagnosed based only on symptoms. This means that autism is, essentially, a phenotype. This, in turn, implies that a phenotype that differs from what has long been thought of as the autism phenotype is something other than autism. 

In other words, a distinct autism phenotype for girls seems like a contradiction in terms.

This holds as well, incidentally, for the motor/apraxia disorder that facilitated communication proponents need autism to be. If there is such a condition, it's something other than autism.

One possibility the authors don't consider is that there are gender differences in restrictive interests and that stereotypically male restrictive interests are more likely to be recognized as such than stereotypical female ones are. Think electrical outlets and science fiction vs. horses and fantasy fiction. But while that might explain the gender difference in rates of identified RRBs, it doesn't explain away the gender difference in rates of identified social symptoms.