They have use Z-test to solve this. But if variance of population is unknow(which is the case here), we use t-test. So why they hav used z-test? . Also if we use t-test, we don't have values of t99, what to do in such cases?
I think for large samples (when the population variance is unknown) we can use z-test in place of t-test. Since for large sample s^2 is a good estimator of sigma^2.. see page 9 of chapter 12.