4. Observe during life or until the completion of twenty-eight days, and in the event of death occurring during that period, make a complete post-mortem examination. 5. Repeat the experiment at least once. In the event of a positive result estimate the minimal lethal dose of "killed" culture for each of the species of animals experimented upon. (b) Extracellular or Soluble Toxins: 1. Filter the cultivation through a porcelain filter candle (Berkefeld) into a sterile filter flask, arranging the apparatus as in the accompanying figure (Fig. 160). 2. Inoculate mice, rats, guinea-pigs, and rabbits subcutaneously with that quantity of toxin corresponding to 1 per cent. of the body-weight of each respectively, and observe, if necessary, until the completion of one month. 3. Inoculate a "control" tube of bouillon with a similar quantity and incubate, to determine the freedom of the filtered toxin from living bacteria. 4. In the event of a fatal termination make complete and careful post-mortem examinations. 5. Repeat the experiments and, if the results are positive, ascertain the minimal lethal dose of toxin for each of the susceptible animals. The estimation of the _m. l. d._ of a toxin is carried out on lines similar to those laid down for living bacteria (_vide_ page 316) merely substituting 1 c.c. of toxin as the unit in place of the unit "loopful" of living culture. It frequently happens, during the course of casual investigations that a bouillon-tube culture is available for a toxin test whilst a flask cultivation is not. In such cases, Martin's small filter candle and tube (Fig. 161) specially designed for the filtration of small quantities of fluid, is invaluable. This consists of a narrow filter flask just large enough to accommodate an ordinary 18 × 2 cm. test-tube. The mouth of the tubular Chamberland candle 15 × 1.5 cm. is closed by a perforated rubber cork into which fits the end of the stem of a thistle headed funnel, whilst immediately below the butt of the funnel is situated a rubber cork to close the mouth of the filter flask. When the apparatus is fixed in position and connected to an exhaust pump, the cultivation is poured into the head of the funnel and owing to the relatively large filtering surface the germ free filtrate is rapidly drawn through into the test-tube receiver. ~Raising the Virulence of an Organism.~--If it is desired to raise or "exalt" the virulence of a feebly pathogenic organism, special methods of inoculation are necessary, carefully adjusted to the exigencies of each individual case. Among the most important are the following: 1. _Passage of Virus._--The inoculation of pure cultivations of the organism into highly susceptible animals, and passing it as rapidly as possible from animal to animal, always selecting that method of inoculation-e. g., intraperitoneal--which places the organism under the most favorable conditions for its growth and multiplication.