AbstractThe existence of fisheyes or gels in PVC final products has been a problem which the PVC industry has coped with for many years. The classical gel is a non‐porous PVC particle which cannot absorb plasticizer. Other finished product defects which may be classified as gels by a customer are unfused particles of higher molecular weight resin or resin gels and agglomerates caused by improper mixing procedures called compound gels. Although this paper will define and cite examples of resin gels and compound gels, it will deal primarily with classical gel. Over the years, many different gel tests have been used by PVC suppliers and consumers in order to define the gel level of a particular sample of resin. The most common of these tests is the test using a two roll mill to produce a colored mill sheet which is examined and the number of gels reported. This paper will use statistical methods to the factors which can introduce variability into the test results, including mill roll gap, mill speed, milling time, and temperature. In addition, the nature of the distribution of gels in a sample of PVC resin will be discussed. Finally, the paper will delve into new laboratory techniques which may lead to eventual automation of counting and reporting of gel levels in a resin sampl