170 



extraordinary event among the Endogenae and deserves little 

 trust*. And it ought also to be noted, that in the flowers of 

 Musa above described, when 3 6-parted flowers passed into 

 2- or 4-parted, the leaf next the bract retained its relative po- 

 sition. This circumstance also appears to answer the question 

 whether the arrangement of the parts of the flower is determined 

 by the bract or by the axis, in so far that the former keeps fixed 

 the organ standing next above itf. On the other hand, to con- 



• Through a lucky accident I have been able to settle this matter. In our 

 lower savannahs is found a very elegant species of Epislephium, which I was 

 enabled to obtain in all stages of development, at the moment when I intended 

 to send off the above little essay. In this species the calyculus possesses three 

 largish teeth standing behind the outer segments of the perianth, a small one 

 under the labellnm and a pair of indistinct little teeth, one on each side of the 

 larger tooth next the bract. The rest of the arrangement agrees exactly with 

 that in other Orchidese : of other characters I have seen that the outer seg- 

 ments of the perianth never wholly cover the two lateral inner ones, but a 

 thick midrib remains exposed in each of the latter. The seeds, moreover, have 

 an appearance which seems to me unusual among the Orchidese ; the testa, 

 namely, is expanded into a wing running round the nucleus. The develop- 

 ment of the segments of the perianth is in agreement with the mode in which 

 they subsequently overlie one another. At each side of the little nodule which 

 is the first representative of the flower of this plant, we observe a little point, 

 the first trace of the sepals, and a little later the middle sepal ; at the same time 

 with the latter, the two lateral inner segments of the perianth. Then the 

 labellum appears, and almost simultaneously with that the anther. In this 

 flower also the anther is at first erect, although it subsequently lies upon the 

 summit of the column. Up to this time no trace of a calyculus is to be seen ; 

 it first presents itself clearly when the flower rises above the axil of the bract 

 and the boundary between the ovary and the segments of the perianth becomes 

 visible. The calyculus is persistent upon the fruit, while the other parts 

 separate from it at a very early period. 1 believe I am justified in concluding 

 from the foregoing, that the calyculus, when it presents itself in the Orchideae, 

 does not represent an external circle of organs, because (1) its segments do not 

 alternate with those of that standing above it ; (2) they originate later than 

 those ; and (3) because they persist upon the capsule, while the other parts 

 become detached. I should lay much stress upon the last reason, yet I think 

 that this calyculus must be regarded as analogous to that which may be 

 observed on the fruits of certain Com()Ositae, Dipsaceae, &c. 



f In the Dicotyledons also it may frequently be observed, that the sepal 

 which lies outside keeps its place in malformations or changes of numeral 

 relation. Thus in the Myrtaceae, Melastomaceae and Onagraceae (Jussieua), 

 Several species of the last vary uncommonly in their numerical relations ; one 

 segment of the calyx always lies next the bract. The same occurs in Osbeckia 

 glomerata when it presents ternary circles of organs, as sometimes though 

 rarely occurs. How far the assumption of the position of the floral organ lying 



