MOEPHOLOGY OF BALSAMINACEiE. 161 



five-celled ovary has its carpels alternating with the stamens so 

 that the odd one is necessarily posterior instead of anterior, as in 

 the normal condition. 



Here therefore we have a case of " doubling" distinctly refer- 

 able to an absolute increase in the number of whorls of organs ; 

 for the regular alternation of the organs of successive whorls, 

 both in the normal and monstrous forms of the same flower, pre- 

 cludes the idea of any development of usually suppressed organs, 

 and of any dedoublement or chorisis, to which recourse is had, 

 perhaps too frequently, for the explanation of double flowers. 



Such multiplication of whorls doubtless occurs to a considerable 

 extent in cultivated plants, especially in genera with few stamens 

 and carpels. In the double Daff'odil there are found forty or fifty 

 petaloid organs, while the flower naturally contains only fifteen 

 organs ; and each piece exhibits a more or less perfect lobe at the 

 junction of the claw and limb, showing that there is no chorisis 

 causing the separate development of the coronal lobes. 



In most cases of doubling we find more or less of the organs 

 abnormally developed, rendering the conditions somewhat obscure ; 

 but in these Balsams the circles are formed of perfectly natural 

 structures ; and there is a point of physiological interest in the 

 throwing forward of the functions consequent upon the conversion 

 of the organs. The stamens are replaced by petals, the carpels by 

 stamens ; and an additional whorl of carpels is produced at the 

 summit of the axis. Generally speaking, the stamens are well- 

 developed; but now and then one or two are found sterile, or 

 surmounted by small petaloid lobes. 



These monsters are more favourable to Eoeper's than to Kunth's 

 view ; for, if what we regard as the anterior petal were a " double 

 sepal," we should expect to find a petal developed within it, oppo- 

 site its commissure, which was never the case in any of the very 

 numerous specimens examined. 



Somewhat related to the above metamorphoses are the con- 

 ditions described by Al. Braun in Delphinium (Pringsheim's Jahrb. 

 f. "Wiss. Bot. 1857, i. 206). In that genus the stamens and carpels 

 are members of a continuous spiral series, not of successive whorls. 

 In D. cardiopetalum the spiral is (approximately) f ; so that the 

 9th organ is opposite the 1st, &c. In cases where 16 stamens 

 were found, the first carpel, being the 17th organ of the series, 

 stood opposite stamens No. 9 and No. 1 ; the second carpel (18th 

 organ) was opposite stamen 10, &c. (fig. 4). In another case 18 

 stamens w^ere developed before the carpels appeared, in which case 



LINN. PEOC. — BOTANY. M 



