To obtain precise information for computing the indices of tuberculosis situation in a community, with passage of time, reliance has been placed on tuberculosis infection rates obtained by carrying out tuberculin surveys. In most developing countries, covered extensively by BCG vaccination without prior tuberculin testing, the tuberculin test has problems of interpretation for demarcating the infected persons from the uninfected. To overcome the problem, therefore, the test results are analysed among persons who do not show a BCG scar and are, thus, considered as normal population. In this paper, an attempt is made to show that BCG vaccination not always lead to the formation of a scar, and also that the scar resulting from BCG vaccination may fade away with time and the person, thus, may be wrongly included in the unvaccinated group. It has also been found that there is greater fading of scars in the younger age groups: in children 0-2 years of age, upto 52% of the scars faded away within 21 months of vaccination. This proportion steadily decreased to about 8% in the 10-14 years age group.
The implication of the finding is that the demarcation line between uninfected and infected persons may require to be shifted from survey to survey, based on the distributions among the 'no scar' population. Moreover, in a totally vaccinated community, the differences of reactions may provide the answer to the problem of identifying the newly infected persons.

It has been postulated that BCG scar disappears in a good number of children and some of the vaccinated children will get included in the non- vaccinated group and cause difficulty in interpreting the results of tuberculin test. It was decided to analyse information on BCG scar status in the younger population of a rural community in 3 taluks of Bangalore district with an objective to find out whether disappearance of BCG scar is dependent on the age of the child, size of post-vaccination induration at initial survey and tuberculin sensitivity status of children in whom BCG scar has disappeared, in comparison with children in whom the BCG scar has not disappeared. In all, 1095 children aged 0 to 14 years were found with BCG scar in 119 randomly selected villages during an epidemiological survey done in 1961 at the time of intake. Following two groups of children were studied for disappearance of the scar. Of them, a) 796 children who had BCG scar at the first survey, and whose BCG scar status was available at 4th survey, b) 299 who showed no BCG scar at first survey but were found with BCG scar at 2nd survey and whose BCG scar status was available at 4th survey.

Of the BCG scars recorded at intake, 26.4% and 32.5% disappeared subsequently during three and a half and five year periods respectively. The waning of BCG scars was independent of age of the child and tuberculin sensitivity status at intake. Tuberculin sensitivity status in children in whom scar had disappeared was the same as that found in children in whom scar had persisted at intake and after five years. The misclassification of children, in whom scars have disappeared, as unvaccinated leads to a difficulty in interpreting the results of tuberculin test done for the purpose of computation of the Annual Risk of Infection. Further, the extent of misclassification increases in proportion with the increase in BCG coverage of the population. This finding justifies the practice of identifying the demarcation level on the basis of the distribution of tuberculin induration sizes for classifying the infected persons in a population in each survey.