The development of the ovule, megaspore and megagametophyte in Saxifraga fortunei var. partita (Makino) Nakai was observed. The ovule is anatropous, bitegmic, and crassinucellate. Both integuments originate from the epidermis. The archesporium is considered to be multicellular. The primary sporogenous cell functions as the megaspore mother cell which forms a T-shaped tetrad. The chalazal member of the megaspore tetrad is functional and develops into a Polygonum-type embryo sac. In the pyriform synergids the filiform apparatus is observed, but any hook or indentations could not be recognized. The antipodal cells are detectable until the Helobial endosperm undergoes several nuclear divisions. Secondary multiplication of the nuclei or the cells of the antipodals could not be observed. THE FAMILY Saxifragaceae s.1.2 has been thought by many authors to consist of some heterogenous groups which might be divided into families. For example, Cronquist (1968) divided the traditional family Saxifragaceae into three families, the Hydrangeaceae, the Grossulariaceae, and the Saxifragaceae (sensu herbaceo), although he considered that his Saxifragaceae was still heterogenous and might be further divided. Hutchinson (1973) split it into 1 1 families and put Penthorum into Crassulaceae. Airy Shaw (1973), in his revision of Willis' "A Dictionary of the Flowering Plants and Ferns," gave 19 families which correspond to the Engler's Saxifragaceae. The development of the ovule and/or the embryo sac in the genus Saxifraga has been reported from time to time, e.g., by Vesque (1879) in S. ornata and S. ligulata, by Juel (1907) in S. granulata, by Pace (1912) in S. ligulata, S. sponhemica, S. cordifolia and S. crassifolia, by Schiirhoff (1925) in S. decipiens x granulata, by Chapman (1933) in S. virginiensis, by Harmsen (1939) in S. stellaris var. comosa, by Wiggins (1959) in S. hieracifolia and by Saxena (1964a) inS. diversifolia. In addition, there are some papers in which development is described incompletely. To 1 Received for publication 3 December 1980; revision accepted 2 April 1981. I am indebted to Dr. C. Kimura of the T6hoku University, Sendai, Japan, for his valuable criticism. 2 In this series of papers, as a matter of convenience, the delineation of the family Saxifragaceae s.l. is based on the Engler's system (Melchior, 1964). Major difference between Engler's Saxifragaceae and Bentham and Hooker's lies in the inclusion of the Cinoniaceae in the latter summarize these reports, the ovule in the genus Saxifraga is usually anatropous, unitegmic or bitegmic and always crassinucellate. The primary sporogenous cell is cut off from the single archesporial cell. It undergoes meiotic divisions to form a linear or less frequently a T-shaped tetrad. The chalazal member of the four megaspores is functional and gives rise to a Polygonum-type embryo sac. The synergids are hooked and exhibit the filiform apparatus in some species but neither the hook nor the filiform apparatus occur in others. The antipodals are mostly ephemeral and degenerate without further divisions but show secondary multiplication in S. diversifolia. This study, the first in a series of investigations, has been undertaken to furnish embryological data of the family Saxifragaceae s.l. and to contribute to a better understanding of the true relationships between the taxa within the family. The present paper deals with the development of ovule, megaspore and megagametophyte in S. fortunei var. partita (Makino) Nakai. MATERIALS AND METHODs-Many flower buds and flowers of Saxifraga fortunei var. partita were collected in Oct. 1975, from plants growing on wet rocks along one of the upper streams of the River Abu, called Ch6monkyo, Yamaguchi Prefecture, western Honshu, Japan. The materials, killed and fixed in FPA, were dehydrated and embedded in paraffin by the tertiary butyl alcohol method. Serial sections 7-12 ,um thick were cut and stained in Heidenhain's iron haematoxylin and fast