easily be recognized by diluting a truly violet solution with water until the intensity of the colour corresponds with that of the sample, when one can see at once that, though the solution has been very much weakened, the colour still appears violet. A reddish, or, rather, yellowish-brown tint means that either the work has not been properly carried out, or else that the blood contained acids or alkalies in excess. The experiment must be repeated, otherwise it may happen that the existing conditions conceal a positive reaction. We shall deal with this point in fuller detail, when we return to the question of the sources of faulty observations.
Under certain conditions a special control test may be needed. Such would be the case in dealing with micro-organisms cultivated on a medium which could not be readily separated by centrifuging. In this case the germ-free medium must be boiled by itself, until the filtered boiled water gives no traces of coloration with ninhydrin. Then the cultures are prepared in exactly the same way, and the following tests are performed: (i) Serum alone; (2) serum + medium; and (3) serum + culture. Should the experiment (2) produce decomposition, then a positive reaction in experiment (3) would certainly not prove that the micro-organisms had been decomposed.
A very important control test, for proving the suitability of the organ or the substrate used, is the
