We demonstrate that, given a limited amount of genotyping resources, the power of a TDT study can be increased substantially by genotyping within families sequentially. By sequential genotyping we mean that one parent in a family should be typed and then the decision on whether to continue genotyping the family is based on this parent's genotype. If it is decided to continue genotyping then a further decision is made about whether to genotype the offspring once the second parent has been genotyped. We show that, for a given power, reductions in sample size of over 80% are possible, even for the most robust selection strategies. We discuss the practical application of such sequential genotyping and illustrate its potential using a real data set.