[On the problems of the antituberculous drug resistance]

In this part, we will discuss exclusively about the technique of measurement. It is desirable especially for the proportion method that two inoculations with the same doses give the same number of colonies and when diluted, the colonies decrease exactly parallel to it degree of dilution. But this ideal is not necessarily realized and especially the difference between the dilution of 10(-3) and 10(-4) mg/ml on the drug free medium and that of 10(-2) and 10(-3) on the drug containing medium is fairly remarkable: the former is not so large but the latter sometimes varies tremendously with the drugs and their concentrations. So it is necessary to use always two inoculations for all specimens and the resistance is calculated on the medium inoculated with the same doses, for that, colonies on the control must be countable and suitable number that is 50-300. Further too concentrated solution like 10(-2) mg/ml must not be used for the proportion method. The selection of the most suitable two dilutions is specially important for this method. Original one is 10(-3) and 10(-5). Unfortunately we did not follow this formula and at first we adopted the combination of 10(-3) and 10(-4), then 10(-2) and 10(-4) finally 10(-2) and 10(-3) which was the worst. Now we searched for the dilution which gave the suitable number of colonies from our several thousand cases and found that 10(-4) was the most excellent and gave this number in 63%, next 10(-5) in 14%, and then 10(-3) in 7.2%, 10(-2) in only 1.0%. With the combination of 2 solutions 10(-3) and 10(-5)--original one--was most excellent (94.8%) then 88.4% with 10(-4) and 10(-5). The combination of 10(-2) and 10(-3) was worst, it gave only 8.9%. As the proportion method demands the most strict procedures for all cases, it is necessary as simplify as possible, thus in every day practice, only the drugs most frequently used (3 or 4 drugs) and one concentration for each drug could be examined. Further we discussed on the difference between DH-SM and SM, direct and indirect method and problem of plug.