The LD50 test: some considerations of precision.

Substantial improvements in the statistical aspects of the LD50 test are possible. These include modifications to the design of the experiments (including the use of sequential methods) and the replacement of outdated techniques of statistical analysis with more appropriate methods. With such improvements, it would be possible to reduce appreciably the number of animals used (to about 10 or 12) while still achieving precision similar to that obtained with present methods.