INTRODUCTION +4146The past 20 years have seen unprecedented progress toward solution of Darwin's "abominable mystery" of the origin of angiosperms, the plant group that has domi nated most terrestrial ecosystems since the middle of the Late Cretaceous, some 80 million years ago. Although studies of living plants have yielded important new insights (68,109, 124,144,145,147), the most striking advances have come from paleobotany, particularly from palynology (the study of spores and pollen) and new methods of analysis of the Cretaceous leaf record.Before the I 960s, most authors considered the fossil record of angiosperms unusu ally biased and unrevealing as a source of evidence on their origin and early evolu tion. Studies of Early Cretaceous leaf fl oras done around the turn of the century (18, 50) showed that angiosperms rose to dominance more gradually than had been previously believed (cf 8), but these studies produced identifications of early angios perm leaves with diverse modern genera and families that seemed to indicate a long period of prior diversifi cation. Such identifications, plus occasional reports of pre Cretaceous angiosperms and various analogical arguments (see 41), are the primary basis of Axelrod's theory (8-12) that the angiosperms originated and began to diversify extensively in tropical upland areas as early as the Permian or Triassic.In the apparent absence of critical fo ssil evidence, ideas on the origin of angios perms have been based largely on comparative morphology of modern plants. For example, the presence of such fe atures as vesselless wood, monosulcate pollen, free fl oral parts, and incompletely sealed carpels has been considered support for the concept that the dicot subclass Magnoliidae (128) or "ranaHan complex" includes the most primitive surviving angiosperms, and for the related euanthial theory of the fl ower-that the flower is a simple strobilus, an axis bearing micro-and megas porophylls (13-15, 33, 128, 136). Comparative evidence has also been used to argue that the angiosperms are a monophyletic group: They share unique advances such as double fertilization followed by endosperm formation; character correlations and morphological series allow departures from a euanthial, magnoliid prototype [such as the simple, unisexual fl owers of the wind-pollinated Amentiferae (Juglandaceae, .
Further
ANNUAL REVIEWS
366
DOYLEBetulaceae, etc)] to be interpreted as specializations within the angiosperms (33,128,136). However, these interpretations are not accepted by all authors. For example, proposes that the angiosperms evolved polyphyletically from gym nosperms with compound fe rtile branch systems called anthocorms, with groups of microsporangia or ovule-containing cupules borne on the secondary axes. Under this interpretation, the Amentiferae are as primitive as the Magnoliidae, the "fl ow-. ers" in some groups are homologous with whole anthocorms, in others with the secondary axes only, and "carpels" are sometimes derived by modification of cupules, sometimes by fusion of t...