334 



CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 



leaves of the main axis. Further, this hypothesis excludes 



the idea of a perianth in the grasses, and with this a very 



important clue to the connections between the grasses and 



other low Monocotyledons is overlooked. The second 



interpretation recognizes a perianth which is represented by 



the lodicul^. The two anterior members are present in 



most o-rasses, and the trimerous perianth is completed by a 



posterior member in the bamboos and certain other grasses. 



Neither of the hypotheses set up to explain the significance 



of the lodiculae is entirely satisfactory — the first explanation 



for the reasons given, and the second, mainly because of 



the position of the lodiculse^n most grasses. The relation 



of the lodiculae to the upper palet in Aveiia (and in some 



other grasses as well) in which there is a common origin 



for both, makes it difficult to look upon the lodiculae as 



members of a perianth, for this, if present, ought to originate 



and therefore be placed on a higher plane than the upper 



palet. This intimate connection of the lodiculae and upper 



palet is met also in Oryza, Zea, and Solenache (Eichler), 



The lodiculae vary greatly in number in the different 



grasses and assume various positions, as in the Paniceae 



they are mainly outside the upper palet, and in the Ehr- 



/laria-species similar bracts are associated with the lower 



glume. From this it seems reasonable to conclude that not 



all of the lodiculae can represent rudimentary perianths. 



In addition to this variation in position of the lodiculae there 



. are present in the rye (Eichler) and certain other grasses 



four anterior lodiculae placed on two planes; the two lower 



are connected with the upper palet (" stipular lodiculae") 



and the two upper, originating evidently on a higher plane, 



are regarded as being two members of a perianth ("peri- 



anthal lodiculae "). A comparison of the lodiculae oiAveiia 



with those of the rye shows them to be homologous to the 



stipular lodiculae of the rye, but in Avcna the perianthal 



lodiculge are wanting. In the bamboo a trimerous perianth 



of three lodiculae is present, and in addition to this, 'two 



stipular lodiculae also. The upper lodiculae of the bamboo, 



by reason of their position and number, have been rightly 



