We find difficult to examine the quantity of corrosion in tanks of a ship. For an example, when we examine reduction of thickness due to corrosion by drilling many test holes on the structure of the ship, we find that there are great differences in the result of each test. In the analysis of results of corrosion measurements on longitudinal bulkhead, transverse bulkhead, shell and side longitudinals of two oil tankers, the author recognized the significance of these fluctuations by dividing up and downward fluctuations and fluctuations in each tank from that of the whole, and moreover he investigated into significance in the fluctuations between structures. And he regarded the difference between whole fluctuations and these significant ones. As a result of the consecutive investigations, the following conclusions were obtained: 1. Quantity of corrosion in the same ship changes regularly according to up and downward position and location of each tank. 2. In the same ship, there is no essential difference among quantity of corrosion of each structure except shell. 3. As to the variance when the significant fluctuations are teken off, namely, the error, the rate of the fluctuations of corrosion which is the cause of the error is found to be shown by the normal distribution of the standard deviation σ_o=0.32.
