Yes, the RLC 1031 is a form of notch filter. This filter topology is called an elliptic or Cauer filter. The added capacitor (C1031) in addition to the standard 4th order topology creates a notch with increases the filter rolloff substantially. It is possible to get up to 60dB/oct rolloff with as little as 5 components. This filter is optimised for 48dB/oct rolloff slopes. For more on Cauer filters, see http://ldsg.snippets.org/FILTERS/Cuadra/elliptic.php3 or varoius forum postings (do a search) on www.htguide.com.
These two drivers are not really well behaved in the crossover range. The 2214 has a pronounced peak at 1.9kHz, the 2416 has its resonance frequency near (with two ugly impedance peaks, one at 1350 Hz).
With 4th order filters I could not get rid of the 1.9kHz 2214 peak with a conventional LRC series notch filter in parallell with the driver without messing up filter slopes - probably because the notch is so close to the crossover frequency. Only an 8th order filter makes it (though you can still see it down at -27 db).
The broad 2 dB peak at 2 kHz in the summed frequency response is because of the tweeter. (See simulations below with individual curves. The simulation looks a little bit better that the measurement). The impedance peak at 1350 Hz at a point where the response is at its strongest makes a good design using passive components difficult. (An electronic filter might be easier, and is a future design option). Anyway, I fully agree that the 8th order maybe needs some more work.