The Yaos hold that it should not be administered to a free person without grave cause—_i.e._ unless suspicion is definitely directed to some one person. The Angoni and Makololo chiefs, whose government (at least as regards the subject tribes) was more despotic, used to order wholesale _mwavi_-drinkings, trying a whole village or district to discover the supposed culprit. In these cases, the poison was often taken by proxy and given to fowls or dogs—each animal being tied by a string to the person whom it represented, and whose guilt or innocence was decided by its death or recovery. There was a _cause célèbre_ of this kind at Chekusi’s kraal, about twenty years ago. Chekusi’s mother had been suffering from rheumatism, and had been treated for it by Dr. Henry, of the Livingstonia Mission. Shortly afterwards, unfortunately, she committed suicide in a fit of depression, and, naturally, things looked very black for the doctor. However, there were two other possible culprits, and a trial took place with three sets of fowls, one for the Europeans at Livlezi Mission, one, I think, for Mponda, with whom Chekusi was then more or less at war, and the third for the other suspected party. The proceedings were complicated by the number of victims demanded by the importance of the occasion: had one been allotted to each defendant, the decision would have been quite clear; as it was, some of each set died and some recovered, and the issue was such hopeless confusion that the case was dropped without arriving at a verdict. The poison used throughout the Shiré Highlands, on the Lake, and by the Angoni appears to be the same; it is the pounded bark of a tree known to science as _Erythrophleum guineense_. Its effect is fatal within an hour or two, unless it causes sickness; this symptom is therefore held to be a sign of innocence. Its different action on different people probably arises from the strength of the dose being varied by accident or design.