1) Tuberculosis contacts and nurses-in-training are tested with tuberculin every 4 months. If and when positive, they are then serologically tested every 4 months. 2) Danger signs of approaching tuberculous disease are seen when successive Inhibitive or Fixation reactions produce certain positive or significant serological graphs. A graph showing variation portends more imminent danger than if merely positive. It is explained that this is understandable with immunological reactions. Recent marked increase in tuberculin sensitivity adds to the precariousness of the situation. 3) When Danger Signs are seen in the graphs of nurses on full-day duty, we order sleep in their daily hours off, and the deletion of four of their six late leaves per month. This usually dissipates the danger, and tuberculous disease fails to develop. If however, the danger does not disappear after several months of this enforced rest, we change the prescription to one, two or three months in bed at home. With contacts on ordinary 7 or 8 houra-day work, advancing the average lights-out hour from 12 p. m. to 11 or from 11 to 10 is usually sufficient; alternate week-ends in bed, may be added. Failure in taking the rest results sooner or later in the development of tuberculosis in most of these individuals. 4) Three examples of a safe graph are shown: no danger signals, no extra rest taken, and no tuberculosis developed. Ten examples of danger warning graphs are shown, where extra rest was taken, the danger signals abated, and no disease developed. A dozen danger warning graphs are then shown where the prescribed rest was not taken, and tuberculosis did develop. 5) The two main factors in the success of the method are, firstly, precision in the interpretation of tests, and secondly, firmness in prescribing rest on a purely hypothetical basis and in the absence of symptoms or any manifest abnormality.