tion.
Precisely the same arrangement is found in the four-class tribe. The
four-class are therefore simply a systematisation of the terms of
kinship in use under the two-phratry system.

Comparing now the eight-class with the four-class system, we do not see
at a glance the essential principle of the former. The clue is given by
the fact that classes I and IV, II and III in phratry A, I and II, III
and IV in phratry B, are what we have termed a couple, that is to say
stand in the relation of parent and child alternately. Marriage being
between classes of corresponding numbers, it follows that
Kumara-Bulthara and Appungerta-Umbitchana are the maternal and paternal
grandparents of the man EGO. The grandparents of his wife are in the
same classes but with reversal as regards the sex. Bulthara is the
cousin of Appungerta, Kumara of Umbitchana and so on. We see therefore
that, just as among the Dieri, a man may not marry his cousin, but must
marry his second cousin, to use ordinary terms, which in this case are
not misleading.

Looking now at the Ngerikudi system, we see that elder and younger
sisters are distinguished in the generations of EGO and his parents.
Possibly they are the eight-class tribe of Queensland to which Dr Howitt
alludes. If not, we have in them a tribe one stage earlier than the
southern Arunta, who have their four classes divided but as yet without
any corresponding names.

The Dieri rule is that of the eight-class tribes. The person designated
as the proper spouse for a male is his mother's mother's brother's
daughter's daughter, in other words, the grandchildren of brother and
sister intermarry. This, as we have already seen, is precisely the
effect of the eight-class rules. We are therefore confronted with three
possibilities. Either the Dieri regulations are aberrant or they have
introduced these rules under the influence of the neighbouring
eight-class system; or the eight-class organisation is a systematisation
of the Dieri rule, adopted perhaps to facilitate the determination of
marriageableness or otherwise in the case of persons residing at some
distance from each other and therefore less likely to be acquainted
with genealogical niceties than the members of a small community. Now if
the second of these hypotheses is correct, it is by no means clear why
the Dieri, having in view the attainment of the object of the
eight-class system, did not simply adopt it; for this we can find no
reason;