2oG Contribution to a Statistical Study of the Cruciferce 
associated not only with aa increase in the male reproductive organs but also in 
an increase in the protective organs or perianth. 
(5) The gynwciwm consists of two sets of carpels within a single perianth. 
This is rather an anomalous group but is extremely interesting inasmuch as it 
contains a series of annectant forms linking group 2 to group 6. What we actually 
have here is a complete reduplication of the reproductive organs encased within 
a single series of protective organs. In some of the flowers examined with this 
structure it was rather difficult to determine the orientation owing to a torsion 
of the thalamus, but in the types figured on Plates IX and X (Fig. XCIV and 
Figs. XCV et seq.) the mode of origin of these is quite evident. Several 
important observations on these forms may be stated. 
(a) There are really two complete sets of reproductive organs and in one case 
(see Fig. XCVII) each of these is of the typical cruciferous structure. 
(b) Increase in the number of the reproductive organs is accompanied by an 
increase in the number of members in the protective organs. 
(c) The increase in the number of members of the reproductive organs is for 
the most part in the plane of division of the carpels, in other words, in the outer 
whorl of the calyx and its associated petals. 
(d) This is also the plane along which the separation of the reproductive 
organs has taken place. 
(e) This plane is the one which we have already shown in the statistical part 
to be the plane of greatest variability. 
(6) The gi/ncecium consists of two sets of two carpels tuithin separate perianths 
hut on one pedicel. 
In this group we reach the limit of variability in the material examined. In 
place of a single flower consisting of calyx, corolla, androecium and gynaecium we 
actually find two complete sets of all these organs, on one pedicel (see Figs. CIII — 
CVIII), while in one case (Fig. CIII) each of the two flowers has the typical 
cruciferous structure, so that were each of these separately examined it would 
undoubtedly be regarded as a normal flower. Yet we must bear in mind that, 
botanically considered, one flower and one flower only arises from a pedicel. Were 
this, therefore, an isolated example, and if no annectant forms existed, the departure 
might well be regarded as a " mutation," but a consideration of the numerous 
variations which we have already considered, taken in conjunction with group 5, 
only serves to emphasise the fact that " the vertical plane which passes through 
the partition wall of the two carpels and consequently separates the individuals 
of the pairs of stamens in the inner whorl and passes through the centres of the 
sepals of the outer whorl of the calyx is a plane along which this flower is in a 
state of flux and is the plane in which it is probable that the flower has changed, 
and is still changing, from some quite different ancesti'al form." 
