Many times, a forensic artist will come across the problem of an edentulous skull: how can we do a facial approximation when there are no teeth? How do you know how the mandible will articulate to the skull? Do it incorrectly, and you could significantly alter the shape of the unidentified person's face.
A guideline first offered in Grant's Atlas of Anatomy (and perpetuated in subsequent forensic art textbooks) was to take a pencil or dowel, pass it through the mandibular notch, under the pterygoid process (or "plate") of the sphenoid, and out through the opposite notch. This method is known to forensic artists as the "pencil trick" and, theoretically, would give an estimate for the proper placement of the mandible, like this:
It would be great if this worked all the time...unfortunately, it just doesn't.
I know this because I decided to test it out, all in the name of science, and pure curiosity. I'm the type of person that tends to question things, so when I'm in a room full of skulls, the wheels started turning. It didn't take long to discover that using a pencil as a guideline gives no better estimate for articulating the mandible than pure guesswork alone.
Here's the problem: this method is dependent on two things: the shape of the wings of the pterygoid:
...and the shape of the mandibular notch. The more shallow the mandibular notch, the less room there is for the pencil to fit:
...and a deeper mandibular notch like this gives more room for the pencil to fit:
Add that to the pterygoid, and there's no telling what you'll come up with. Here's what happens (below) when the notch is shallow, and the pterygoid is simply not cooperating. There's no way that pencil is fitting in there, without the mouth gaping open wildly. I purposely used a skull that still had it's teeth, so you could see for yourself that the articulation is off:
Here's another skull, same problem:
And the opposite can happen when the notch is very deep, and/or the pterygoid process isn't very pronounced, or any number of varaible in between...there's so much room, the mandible can almost fit over the top teeth. Think about it: if you were handed this skull below, and there were no top teeth, you'd mistakenly think that the pencil "worked" because it fit. Sure, it "fits" ....but it doesn't give the proper articulation:
I can tell you, it didn't take long for these problems to become glaringly apparent. After the fourth or fifth skull, I knew I was on to something. I went through about 40 skulls in one session, and probably as many more after that testing it out. A pencil worked sometimes, but more often, it didn't.
There really wasn't any reason for it to work in the first place. If the pencil were replacing a similar size and shaped structure in the skull that was absent, I could see a reason for it. That would make sense to me. But it's not, it's arbitrary.
So, what should you do instead? Take into consideration the age of the skull, and the condition of the teeth, and proportional canons of the face. If possible, consult with a forensic odontologist, in addition to an anthropologist. You should always be working with a forensic anthropologist when you do facial approximations anyway, whether they are 3D or 2D, so this shouldn't be a problem (hint, hint).
Anthropologists can get clues by looking at the wear and tear on the condyle and mandibular fossa where it articulates. Depending on the condition of the skull (resorption of the bone due to ante-mortem tooth loss, or post-mortem animal and rodent activity, etc) the anthropologist might only be able to give you an educated guess. But that's better than a Ticonderoga pencil!