Although vertebrate immune systems have been commonly conceived as exquisitely developed to combat pervasiveness by pathogens, they are not infallible. The enigmatic expression of histocompatibility in vertebrates, the manifestation of natural chimerism, autoimmunity, malignancy, and other puzzling outcomes hint that immunity did not arise in evolution to fight infections and that this capacity is a late evolutionary appendage, owing its appearance to the redeployment of a system developed for other reasons. Allorecognition in the colonial tunicate Botryllus schlosseri serves here as a platform for a contending paradigm, advocating that immunity has developed as a surveillance machinery against and for purging of nascent selfish cells (stemmed from a kin organism or from transformed cells within the organism of origin). Defense against pathogens (always representing xenogeneic aliens) appeared later, revealing the multiplicity of newly developed phenomena. Allorecognition events characteristic of the Botryllus primitive immune system, such as fusion versus rejection, the morphological resorption with its expressed hierarchy, and the somatic/germ-cell parasitic outcomes, provide clues to the evolutionary basis of allorecognition. Recent work on Botryllus immunity that highlights the cost of littering individuality by somatic variants/allogeneic cells is discussed.