The genesis of manganese nodules in the ocean is a major subject attracting great interest from the scientific society. In more than one century, the understanding of their genesis has gone through the process from the knowledge of "colloid chemical action" in the early stage to the later understanding that microorganisms have important influence on the nodule formation. From 1991 to 1995, the authors had proposed the idea that ultra-microorganisms be the constructor of polymetallic nodules, which is different in essence from the traditional ideas. On that basis the authors have conducted an intensive study of the regular features shown by stromatolites (Minima and Admirabilis) and their constructors (Miniactinomyces chinesis and Spirisosphaerospora pacifica) , laminae of stromatolites and the growth of ultramicroorganisms, the relationships between the types of stromatolite and the main manganese minerals and elemental Fe and Mn, and direct mineralization of ultramicroorganisms. The results indicate the following: the stromatolite columns in the manganese nodule usually have three growth rhythms, i. e. laminae, laminaset and lamina belt; hyphae are dense in the bright lamina while they are sparse in the dark lamina; good corresponding relation exists between two types of stromatolite and ultramicroorganism fossil species; the variation in content of main manganese minerals is closely related to the types of stromatolite; Fe and Mn are distributed more regularly inside the stromatolite columns than in the interstices between them; the mineralization degree of nutrient hyphae is higher than that of reproductive hyphae ; the enrichment of Fe and Mn in the nodule directly resulted from the biochemical action of ultrmicroor-ganisms and the adherent action of sediments; and the periodic change of the global macroclimate and the mi-croenvironment of the ocean floor has influenced the flourishing and declining of microorganisms-all these resulted in the alternate appearance of bright and dark laminae. The authors believe that the rhythmic features shown by the laminae of stromatolites are hard to result from any colloid chemical action or sedimentary action, and that only the reproduction of microorganisms for generations can produce the almost unchanged pattern of the stromatolite columns. The above results further support the view of biogenesis of manganese nodules.