Mereology (from the Greek μερος, ‘part’) is the theory of parthood relations: of the relations of part to whole and the relations of part to part within a whole. Java is an object oriented language where every entity could be represented as an object and the presence of part to whole relations and part to part within a whole are present in different ways. The three most common mereological relationships in Java are the binary class relationships of Association, Aggregation and Composition. These binary class relationships are very easy to represent using the modelling languanges such as UML, but hard to detect in Java source code as there is a discontinuity between the source code of a program in Java and the correspoding representation in UML. This project aims to propose an algorithm that will detect the binary class relationships in Java programs and apply the algorithm in some big open source projects to gather statistics on the three mereological relations taken into account for this project. Statistics will be gathered from the application of the algorithm on the big open source projects. The development of the algorithm will be a step forward to bridging the gap between the representation of the binary class relationships in Java programs and their representation in UML. Besides the statistics gathered from the application of the algorithm on some big open source projects will provide useful information on some of the best practices used to implement binary class relationships in Java. Thus the results obtained from the project will be quite useful to writing better dependable software systems.