[2] I have not studied these from an anthropological perspective and can therefore only echo Meek's report; it is worth noting that such differences are to some extent reflected in the lexicon (see Table Six).
[3] There is apparently some dispute as to whether Ndeba and Njerep are separate groups, but at least on the linguistic evidence this seems to be the case. The language spoken in most of Somié is referred to as Ju Ndeba or Ju Ba; Ju Njerep is now spoken by only five or so people in a small hamlet on the edge of the main village.
[4] The behaviour of velar C2 is not unusual; it is not yet clear, however, why a high vowel labial C2 combination should prove most conservative, most resistant to change.