12 



ment of the flower of the Canneae afforded direct proof that the 

 stamina only metamorphosed into petals in a more or less complete 

 manner from their first appearance, which impart to the corolla its 

 irregular aspect. The two outer ternary verticils are always develop- 

 ed one after the other, precisely as the calyx and corolla of dicotyle- 

 dons. This law, which I have verified in more than ten families, ap- 

 pears to be very general among monocotyledonous plants. In the 

 dicotyledons, the adult corolla of the Acanthaceae, Globulariae, Gesne- 

 riacese, Bignoniaceae and Goodeniaceae, which is frequently far from 

 regular, presents itself on its first appearance in the form of a small 

 cupule with five very equal and rounded teeth at the border, but this 

 state is more or less ephemeral according to the genera and species. 

 Very soon the unequal elongation of the divisions of the corolla, their 

 different degrees of adhesion or their partial atrophy, determine a 

 very marked irregularity. The same applies with respect to the flower 

 of Centranthus in the Valerianeae, to that of the Lobeliacese and of the 

 Scrophulariaceae. In this last family the corolla of the Calceolarias, 

 one of the most anomalous of the vegetable kingdom, is reduced at its 

 origin to a scooped-out cupola, which is very regular and furnished 

 with four equal minute teeth ; the nascent calyx likewise presents but 

 four divisions. 



The highly remarkable floral envelope of Begoniaceae likewise ap- 

 pears, at the period of its formation, as regards both male and female 

 flowers, in the form of a continuous ring, and exhibits at its circum- 

 ference five very equal small segments ; but there are some of them, 

 especially in the male flowers, which disappear entirely or which be- 

 come in part atrophied, so as to give to the coloured envelope that 

 peculiar structure which forms its principal character. 



From the facts detailed in my two memoirs and derived from the 

 study of genera with irregular flowers from twenty-five natural families, 

 I feel justified in deducing the following consequences: — 



1. The simple theory announced by DeCandolle as early as 1813, 

 according to which the irregular flowers should be referred to regular 

 types from which they appear to have degenerated, must be admitted 

 as true, although conceived a priori, and solely from the attentive ex- 

 amination of some cases of Peloria, or of flowers which have become 

 regular at the adult age. But if in the actual state of science, organo- 

 geny affords us a direct demonstration of this important principle of 

 botanical philosophy, I must add, that the symmetry of an irregular 

 flower even at its very origin does not always strictly exist; it is fre- 

 quently merely indicated by empty places where the absent organs 



