PALMAE 449 



palmophyllum, fossil leaves from the Lower Jurassic, accompanying a 

 typical Jurassic flora, have been generally accepted as those of palms, 

 and the fossil leaves of Sanmigiielia from the Triassic of Colorado 

 closely resemble palm leaves. The leaves of Sanmigiielia are simple, 

 those of Propalmophylhim are costapalmate, a type intermediate be- 

 tween pinnate and palmate. If diese leaves are correctly interpreted, the 

 palms were already specialized in leaf structure by late Jurassic. 



The inflorescences show clear evidence of reduction from large pan- 

 icles to reduced, compacted, and fused smaller inflorescences, with 

 greatly modified phyllotaxy and the sinking of flowers in branches of 

 the axis. The range in position of inflorescences, from axillary among 

 the leaves to terminal on the tree and to axillary on the trunk — both 

 positions associated with monocarpy — is unusual. A remarkable spe- 

 cialization is the basipetal sequence in maturation of the inflorescences 

 on the trunks of the fishtail palms and in the bractlike scales of tlie 

 lorica. 



The flowers show a full series in reduction from bisexuality to uni- 

 sexuality. In typical unisexual genera, there are no vestiges of the lost 

 sporophylls; the staminate flowers of Phijtclephas and the Mauritieae 

 have no vestiges of a gynoecium. Supporters of the theory that the 

 unisexual flower is of two types — primitive and derived from the 

 bisexual flower — accept the interpretation of reduction where vestigial 

 organs are present, but deny it where they are absent. They should 

 consider the flowers of the palm family as a whole. Even within a genus, 

 vestigial carpels may be present or absent. Unisexuality has doubtless 

 arisen many times in different lines, but the existence of two types of 

 widely different origin within a family is unacceptable morphologically. 



The perianth is remarkable for absence of specialization; it usually 

 remains simple, with minor elaborations. 



The story of the androecium is one of reduction in stamen number, 

 with the establishment of whorled arrangement from indefinite or spiral 

 position. 



The number of carpels, three, is constant throughout the family, with 

 stages in connation and in reduction of fertility in two of them. Rarely, 

 even in apocarpous taxa, do all three carpels develop seeds. In the 

 sterile carpels, ovules may develop apparently normally until flowering 

 time; in other families, ovules that abort rarely develop beyond early 

 stages. In syncarpous genera, as in Cocos, the sterile carpels collapse 

 after fertilization and, in flattened condition, form part of the wall of 

 the fruit (Fig. 93). 



Syncarpy varies in method of carpel union; typically, the carpels are 

 closed and the placentation is axillary basal, but in Asiatic members 

 of the Calamineae, the partly open carpels are united by their margins 



