I would like to add here that to help us in the decision between several contradictory opinions, there is still one criterion beside the hints enunciated above (which refer to the character of each individual statement). I mean the criterion whether the final opinion drawn from the evidence is compatible or not with the other well-established features of Australian sociology. When deciding to adhere to some view, which is not established by a unanimous and categoric opinion of all our informers, it is always necessary to put this view to the test of other well-established facts. There are some views which are quite incompatible with the general conditions of life in the Australian aboriginal society and with the resulting mode of living. As a good example of such deductive demonstrations we may quote the passage in Curr, where he arithmetically proves that the statement of Dawson about the Australian chiefs and their court cannot be true.[38] Another example is afforded by the interesting passage of Howitt quoted below _in extenso_ (pp. 113 and 114), which relates how the author thinks that our ideas on group marriage should be modified by what we know about the aboriginal mode of living and about the natural character of men. As a rule it is well always to try to ascertain whether our conclusion does not stand in contradiction with some part of our well-founded knowledge. Thus in practice it is always necessary to start with a crude series of facts, and in any attempt at criticism to be guided by the contradictions found in them. If then criticism and corrections, made according to our rules, remove the contradictions, we have another guarantee that our corrections were good. For if a series of statements, which at first sight seemed to present irreconcilable contradictions, do agree after we have applied to such of them as were either in a minority or appeared vague or came from uncertain sources, corrections or interpretations (the latter based on principles laid down quite independently), it may be concluded that our reason for applying the correction and the way in which we have done it, were sufficiently correct and justified. [38] Curr, _A.R._, i. pp. 53 _sqq._