A data set consists of 145 observations that range from 56 to 490 What size class interval?

The naive answer to the question is 30. That assumes that the observations are more or less uniformly distributed across the range and, if that is the case, you should get around 5 observations per class.

It also assumes that your interest in the observations is uniform: you are as interested in values near 60 as you are in values near 480. If you were only really interested in values above 450, you could class all of 56 to 449 in one big class and split the rest into smaller classes.

It is also important to see if the distribution is uniform. If it is skewed in either direction, it would make more sense to have smaller classes where the observations were more dense and wider classes where they were sparse.