From his many years studying liverworts, mosses, and ferns, Bower concluded that they had evolved from algal ancestors. Bower’s hypothesis states, in essence, that the sporophyte generation (the conspicuous vegetative stage in familiar vascular plants) developed de novo from a haploid alga that lacked a diploid sporophyte generation but instead had merely a diploid zygote (a cell formed by the fusion of two gametes, such as sperm and egg). Before the evolution of embryos, this zygote would have immediately undergone meiosis (to relieve the diploid condition) and produced spores, the propagules of the next haploid generation. Growth of such a spore into a gametophyte is analogous to growth of an isolated human sperm or egg cell into a hypothetical haploid generation. Thus, the sporophyte generation first appeared as an added generation that came into existence as a result of delayed zygotic meiosis – sort of a delayed plant puberty. In other words, what might otherwise have become the new haploid cells of the next generation by chromosome reduction instead retained its diploid character and thus added, aà la Bower, a new generation to the life cycle. The final step of spore production still eventually occurred, but not until after the diploid cells had grown and developed into a new sporophyte generation, in essence an overgrown zygote.
Under Brower’s hypothesis, we suppose that, from the point of view of the gametophyte, the sporophyte generation is like a giant multicellular spore factory. For example, in Coleochaete pulvinata, a modern freshwater green alga, the surface of the mature zygote is covered by a layer of haploid cells, which form ingrowths that penetrate the zygote to provide nutrition. The protected diploid zygote in Coleochaete gives the aquatic alga advantages because many more spores can be produced from a single fertilization event than would be the case if the zygote hurried straight to meiosis and the formation of one of those four spore tetrads so common in the fossil record. Bower’s hypothesis remains to be tested, but if it is correct, the sporophyte generation (diploid cells) came to develop inside (and be protected by) the gametophyte generation (haploid cells) precisely because the arrangement ultimately benefited both generations.
An older, competing hypothesis dating back to 1874 held that the algal ancestor of embryophytes already had had alternation of two generations for a long time and was thus diplobiontic, as opposed to haplobiontic. Haplobiontic organisms, such as humans, have the gametes as the only haploid cells; diplobiontic organisms develop those haploid cells into a multicellular life stage. The diplobiontic hypothesis of 1874 is less favored now because it fails to explain how the sporophytes and gametophytes, which in modern diplobiontic green algae have no long-term physical connection, could have evolved the intimate physical connection, in both nutritional and developmental respects, shared by the haploid and diploid components of all embryophytes.
Bower’s other publications included Ferns (three volumes, published 1923-28) and Primitive Land Plants (1935). Bower was elected Fellow of the Royal Society in 1891 and was awarded the Linnean Medal in 1909, the Royal Medal in 1910, and the Darwin Medal in 1938, the latter “In recognition of his work of acknowledged distinction in the field in which Darwin himself laboured.”