The article introduces a new method called the Up-and-Down Method for testing sensitivity in small samples. This method helps determine the LD50 value, which is important in toxicology studies. The method involves testing different levels and sample sizes sequentially. The estimates obtained using this method have low error rates and are not greatly affected by the starting level or spacing between test levels. The actual sample size usually stays close to the intended size. The method is shown to be effective in two different analyses of variance designs.