After deciding how they wish to circumscribe species, biologists and paleontologists normally group species which are sufficiently similar or closely related into a genus, genera which are sufficiently similar or closely related into families, and so on. The motivations for this seem to be partly an innate tendency to place things in order, partly a desire to reflect the degrees of relationship between biological entities, and partly the convenience of being able to generalize about the different levels of a hierarchical classification - for example, statements can be made about monocotyledonous plants in general, or about palms, or about the various genera of palms.
The International Codes of Zoological and of Botanical Nomenclature impose (putatively) necessary constraints on the procedure of naming the entities of a taxonomic system.
A major problem of conventional taxonomic systems is that taxa are defined and named when they are first discovered - i.e., when they are least known - and that their definitions expand as more and more occurrences are found, exhibiting different ranges of variability. It is completely arbitrary whether a taxonomist encountering a new variant incorporates this into an existing species, or establishes a new one.
A record of a particular, named species therefore carries with it a great deal of uncertainty, which can be avoided only by providing a literature citation to the concept being applied in each case - e.g., Genus-species Gs in the sense or author A in year Y, with a literature reference.