The subject of bacterial multiplication has begun once more to attract investigation. According to Regnier, David & Kaplan the method of the "viable count" may be responsible for serious errors. These workers, using Pseudomonas pyocyanea on a liquid medium, carried out a direct microscopic ("total") count and a "viable" count by the usual method of diluting, plating, and counting colonies. By the former method no latent period was observed, but the multiplica­ tion started immediately into the logarithmic phase. The viable count, however, disclosed the usual latent period; that is to say, no multi­ plication occurred for two hours. This seems to show that the latent period is an experimental artifact due to the transfer of cells from the liquid to the solid medium, and does not occur in the parent culture. Like Wilson (1922), the present authors found that, even after the latent period, when both methods of enumeration were giving a logarithmic rate of multiplication, the total count exceeded the viable. Hence it seems that in a growing culture a certain num­ ber of cells are capable of division whilst they remain in the liquid medium, but fail to divide when transferred to agar plates, and that this difference is so marked in the stages closely following inocula­ tion as to give rise to a deceptive appearance of latency 'which does not actually occur. A second and highly interesting phenomenon disclosed by the same investigators, is the influence of crowding on multiplication rate (Regnier & Kaplan). Comparing the rates of multiplication with large and small sowings, it was found that the duration of the logarithmic period was the same in both cases, as was also the maxi­ mal number of cells produced; that is to say that bacteria sown in large numbers multiply at a constant rate, but more slowly than those sown in small numbers. This difference is apparent from the be­ ginning and hence cannot be attributed to exhaustion of food ma-