440 



SCIENCE. 



[N. S. Vol. IV. No. 91. 



ture will readily suggest the Eanales among 

 the Thalamifloral Dicotyledons, and a closer 

 examination shows a remarkable similarity 

 of structure in not only the reproductive, 

 but also in the vegetative organs of the plants 

 of these two groups. After some years of 

 study given to a comparison of these groups 

 the author is more firmly convinced than 

 ever of their genetic relationship. They 

 show their relationship in their gross anat- 

 omy, the histology of their tissues and their 

 embryology. 



Allied to the Eanales are the Rosales, be- 

 ginnmg with the Ranunculus-like Potentillece, 

 and passing by easy steps to the simpler 

 LeguminoscB on the one hand ( Ccesalpiniacece 

 and Mimosacece) , and the Saxifragacece on the 

 other, and through the latter to Celastrales 

 and Myrtales. 



Here then, in the author's opinion, is the 

 point of divergence of the Monocotyledons 

 and Dicotyledons, represented by the Apo- 

 carpee of the former, and the Eanales and Eo- 

 sales of the latter. The similarities in struc- 

 ture between some Maerospermce and the Naia- 

 daceee in Bentham and Hooker's system, 

 noticed above,as between some of the families 

 (Naiadaceoi, Alismacece, Chloranthacew, etc.), 

 placed by Engler and Prantl at the begin- 

 ning of the two sub-classes, are hints as to 

 a natural arrangement, which it is strange 

 that these eminent systematists overlooked. 



The development of the cystocarp of Griffithsia 

 bornetiana : By Anna Aema Smith. 

 The development of the cystocarp of 

 Griffiithsia bornetiana was described, and it 

 was shown to agree in the main with that 

 in G. corallina as described by Janczewski, 

 except only one peripheral cell bears a car- 

 pogenic branch. The spores arise from a 

 cell cut off from the supporting cell of the 

 carpogenic branch after fertilization, the 

 influence of which is transmitted, appar- 

 ently, through the cells of the carpogenic 

 branch. Since the paper was published in the 



July number of the Botanical Gazette, and 

 because of the author's absence, it was read 

 by title. 



Morphology of the canna flower : By L. H. 



Bailey. 



The speaker called attention to the pre- 

 vailing asymmetry in the Scitaminese, and 

 remarked that groups of plants which show 

 marked irregularities in forms are nearly 

 always fertile subjects for plant-breeding. 

 The most nearly symmetrical flowers of this 

 order are found in the banana tribe, in 

 which 5 stamens are present and the 6th is 

 represented by a sterile filament. He ex- 

 hibited a banana fruit to show its 5-angled 

 form, and remarked that it is probable that 

 somewhere in its phylogeny this fruit had 

 lost its symmetry. He also called attention 

 to the three seedless loculi of the fruit, and 

 remarked that, although the plant is seed- 

 less, it still varies or it is the subject of 

 evolution, thus discrediting Weismann's 

 hypothesis ^that all progressive or perma- 

 nent variation arises through sexual union. 



In the ginger tribe the stamen is reduced 

 to one normal member. In the canna tribe 

 the stamen is presented by what is appar- 

 ently but a single loculus of the anther, the 

 other loculus being apparently developed 

 into a foliaceous organ. The remaining 

 stamens are represented by petal-like stami- 

 nodia, and these members make up the 

 showy part of the flower. The speaker ex- 

 hibited specimens of canna flowers, and 

 also charts, to show the very marked evolu- 

 tion in the form and size of the flower, and 

 more especially of the staminodia, and the 

 gradual increase in the size of the petal-like 

 appendage of the one fertile stamen. There 

 seems to be a considerable decrease in seed- 

 production in the modern cannas as com- 

 pared with the types of a generation and 

 more ago, and this decrease is probably as- 

 sociated with less pollen, or less eflicient 

 pollen, in the modern flowers. This ten- 



