Collectives as described in the accompanying paper. In a more complete ontology would be a subclass of Collection. Collectives are not mathematical sets but collectives of grains, e.g. grains of sand on a beach, cells in a tissue, etc. A convnient top entithy separating off other purely implementational classes etc. The ontology below is a severely abbreviated version of a roughly DOLCE compatible upper ontology supplemented by collectivity. Holes, lines, etc. After DOLCE A lobe of liver Here chosen as a representative determinate part. See paper. Anything that is a mixture of two or more amounts of matter From DOLCE. This version of the ontology takes a "Cognitivist" or "Multicative" view. The distinction between "Substance" and "Tissue" is taken from the FMA. From FMA The cells that make up the liver parenchyma, that is the major tissue of the liver. Differentiated as a substance rather than tissue following the FMA Liver parenchyma - following the FMA, the portion of the liver that is made of up parenchymal tissue. This may be one more level than is necessary. There is some question as to whether parencyma and parynchymal tissue need to be differentiated. To be discussed. The separation is made here as a demonstration. Eliding it would be easy. All relations represented as classes. Here only proportions. Usual biological definition of cell. Roughly from FMA. Biological physical objects. In many biomedical ontologies, the bio/non-bio distinction is useful and worth making. Probes are classes included to test the ontology. They are all defined and should have the correct entities under them when classified. The debugging framework in Protege OWL can be set up to test probes systematically. Red blood cells All red blood cells are part of, or at least begin as as part of collectives that are a portion of blood. The collective of red blood cells By axiom is said always to be a portion of a Blood. This notion needs to be taken normatively or in the sense of persistent parthood, e.g. that collectives of red blood cells always arise as portions of blood. Physical as opposed to non-physical entities. This demo ontology omits non-physical entities. As for red blood cell, but not filled in here for performance reasons Phsical objects with mass etc. Relation between Amount of matter and physical object in the cognitivist / multiplicative formulation of an ontology. The parts including ingredients, constituent parts and determinatge parts, but not granular parts. NB problems with removing proportion from domain. Domain and ranges constrained by universal properties in classes The relation between a grain and a collective. The boundary property that indicates a shift in level of collectivity ("granularity"). The loss of ndividual grains does not diminish the whole. Collectives have emergent properties not directly predictable from the grainsl. See documentation for relational properties. Parent relational property - as opposed to refining properties (roughly modifiers) as opposed to 'attributive properties' the properties required to link entities to their attributes in translations of UML models - roughly but not always refining properties. The parent generic part-whole relation A transitive relation. Neither antisymmetry nor reflexivity can be represented. Countable discrete parts such that the whole is diminished by the loss of one of them. See paper