4 i8 



FOURTH GROUP. SEED-PLANTS. 



I have shown that there is no duplication, but that the number of the rudimentary organs 

 in any given place depends merely on the space afforded, and on the size of the rudi- 

 ments. In other species of Rubus on the contrary it is the region of the torus in front 

 of the petals which grows most vigorously, and consequently more young organs are 

 found here than in front of the sepals. For the rather large amount of details which 

 have been collected on this subject the reader is referred to the author's treatise cited in 

 the note on the previous page, where also it is shown that the number and position of the 

 parts of the flower vary so much in the family of the Rosaceae that it is impossible to 



FIG. 348. Floral diagrams of various Kosaceae (carpels not included). /. Sibfraldia ciineata and some species also 

 of Agnmonia. II. Agrimoniaodorata; the fiist whorl of five stamens is followed by one of ten. ///. Potent-ilia; the 

 pentamerous corolla is succeeded by a whorl of ten stamens alternating with the ten stamens of a second whorl. //'". Rubus 

 Idaeits (spec al case), the outermost staminal whorl only shown ; the pentamerous corolla is followed by a whorl of ten 

 stamens, and from one to four stamens according to the growth of the zone of the floral axis are interpolated in the in- 

 tervals between each pair of the first stamens (not ote only as m ///), three at a, one at b, three at c, two at d, two at e, 

 two at/; four at g, two at h, three at ', two at k. 



frame a typical diagram. Similar facts are observed in other flowers also, as in Citrus and 

 Tetragonia, and in Alisma and Butomus among Monocotyledons, where in the same 

 way a whorl of three petals is succeeded by a whorl of six stamens. In the Fumariaceae 

 mentioned above it seems to me more natural to assume, not that there has been a 

 branching of the two stamens, but that after they were formed there was a sudden 

 diminution in the size of the rudiments of the stamens, and that therefore four were 

 formed instead of two and became connected in pairs with the two first formed, as 

 occurs in Agrtmonia (Fig. 348 //). This view appears to me to be most in accord with 

 the facts observed in allied plants, such as the Papaveraceae, and to be supported also 

 by the history of development. 



In considering the relative positions of the parts of flowers the obdiplostemonous 

 flower requires special notice. In many Dicotyledons the androecium is formed of two 

 whorls, one of which alternates with the sepals and is opposite to the petals (corolline 

 stamens), the other is opposite to the sepals (calycine stamens). Either the calycine or 

 the corolline stamens may form the outer whorl ; the former is the normal, that is, the 

 usual case of alternation, and the arrangement is said to be diplostemonous ; in the 

 latter the corolline stamens are outside, the calycine further in, so that the normal 

 alternation between the two whorls of stamens seems to be interrupted, and this is the 

 obdiplostemonous arrangement. The case of obdiplostemony has received various 

 explanations, all endeavouring to bring it into harmony with the normal arrangement of 

 alternation of parts. Sachs is of opinion that new members of the same kind are 

 formed on the same zone of the torus in the bud when still quite young between the 

 members already formed, in other words, that new members are interposed. This he 



