Units can be converted, so adding multiple different but convertible units does not add any value to the ontological model. Maybe it is more user-friendly to just have one (default) unit for phenotypes. The provided instance data can then be converted if necessary.
If a phenotype is subsequently used in mathematical formulas, we can assume that instances of this phenotype class are provided in its default unit.
ChristophB
commented
2 years ago 
Units can be converted, so adding multiple different but convertible units does not add any value to the ontological model. Maybe it is more user-friendly to just have one (default) unit for phenotypes. The provided instance data can then be converted if necessary.
If a phenotype is subsequently used in mathematical formulas, we can assume that instances of this phenotype class are provided in its default unit.