304 PA THOGENIC BA CTERIA. The strongest serum ever obtained by the author con- tained 1400 units per cubic centimeter. As the quantity to be injected at each dose diminishes according to the number of units per cubic centimeter the serum contains, it is of the highest importance that the serums be as strong as possible. Various methods of concentration have been suggested, such as the partial evaporation of the serum in vacuo, but none have proved satisfactory. The latest suggestion comes from Bujwid,1 who finds that when an antitoxic serum is frozen and then thawed, it separates into two layers, an upper watery stratum and a lower yellowish one; the antitoxic value of the yellowish layer is about three times that of the original serum. Ehrlich asserts that 500 units are valueless: 2000 units are probably an average dose, and, as the remedy seems harmless, it is better to err on the side of too much than on that of too little. Fourteen thousand units have been administered in one case with beneficial results. The largest collection of statistics upon the results of antitoxic treatment in diphtheria in the hospitals of the world are probably those collected by Prof. Welch, who, excluding every possible error in the calculations, u shows an apparent reduction of case-mortality of 55.8 per cent." One of the most important things in the treatment is to begin it early enough. Welch's statistics show that 1115 cases of diphtheria treated in the first three days of the disease yielded a fatality of 8.5 per cent., whereas 546 cases in which the antitoxin was first injected after the third day of the disease yielded a fatality of 27.8 per cent. After the toxin has set up destructive organic lesions' in various organs and tissues of the body, no amount of neutralization will restore the integrity of the parts, so that the antitoxin must fail in these cases. The urticaria which sometimes follows the injection 1 Centralbl. /. Bakt. ^^,. Parasitenk., Sept., 1897, Bd. xxii., Nos. 10 and II,. p. 287.