464 
Saunders.—A Reversionary Character 
antero-posterior septum. (2) Each pair forms such a septum. (3) All four 
converge to the centre, forming four complete or incomplete septa which 
lie in the two diagonal planes. The third form, as it happened, was not 
seen in the Stock, but a Cheiranthus fruit figured by de Candolle 1 was of 
this type. 
This reversion to a multicarpellary condition, which has been observed 
in at least nineteen Cruciferous genera, 2 * in a few fruits, in an individual 
here and there, may sometimes prevail through a large part of the individual. 
In plants of Lepidium sativum , L., raised by de Candolle from seed of 
multicarpellary Abyssinian parents, he found trilocular fruits almost as 
abundant as those of normal structure. 2 In the two forms originally named 
Tetrapoma barbaraefolia, Turcz., and Holargidium Kusnetzowii, Turcz., but 
now recognized merely as many-carpelled variants of Nasturtium palustre , 
DC., and Drabaip hirta , L.), respectively, the exceptional fruits are exhibited 
throughout the whole or the greater part of the inflorescence. Offspring 
from seed of Tetrapoma individuals have been found to exhibit the rever¬ 
sionary character. According to Hooker, 4 however, this is not invariably 
the case, and as a criterion of species, therefore, the character is useless. 
The reappearance of the character in the second generation of de Candolle’s 
Lepidium plants grown in the spring, and its non-appearance in the third 
generation raised in the autumn, are of particular interest, as showing the 
dependence of the character upon conditions, as is indicated also in the 
Stock (see above, p. 457)- The herbarium material of Tetrapoma which 
I was able to examine had mostly four valves, alternating with four solid 
carpels forming complete septa, but as many as five or even six of each 
occurred in individual fruits. In the specimen figured by Baillon, however, 
the septa are incomplete. 5 6 The short style, as in comparable cases in 
Cheiranthus , showed a corresponding increase in thickness. From the 
strictly radial type of symmetry exhibited by these fruits, it is clear that 
this multicarpellary condition does not arise through the morphological 
abnormality known as ‘twinningRather it affords us a real glimpse of 
1 Loc. cit., PI. XVIII, Fig. 9. 
2 Alyssum incanum , L., A. libycum, Coss, Arabis alpina, I.., Brassica Napus, L., Brassica 
oleracea , L., Capsella Bursa-pastoris , L., Cheiranthus Cheiri , L., Draba nemorosa , L., Diplo- 
taxis muralis , DC., 1 ). tenuifolia, DC., Erophila vulgaris , DC., Erysimum pannonicum , Crantz, 
E . repandum , L., Fortunia Garcini , Shuttl., Iberis sempcrvirens , L , Lunaria annua , L., L. 
rediviva , L, Matthiola incana , R.Br., Megacarpaea , Pelt aria alliacea, Jacq., Raphanus sativus 
(caudatus) , L , Ricotia Lunaria , DC., Sinapsis arvensis , L., Thlaspi arvense , L. (For Erysimum 
repandum see Reichenbach, leones, ii, Tab. LXXIV, Fig. 4406 ; for the rest see references in Penzig’s 
Pflanzen-Teratologie.) 
5 Monstruosites vegetales, p 13. 4 Gen. PL, i, 1S62, pp. 83, ,967. 
5 Nat. Hist., iii, p. 184, Fig. 213. 
6 The phenomenon of ‘ twinning ’ is not uncommon among the Cruciferae. I have elsewhere 
suggested (Journ. of Genetics, vol. xi, 1, 1921) that the occasional references in the early literature to the 
occurrence of semi-double flowers in the Stock are probably to be thus explained, as also I have no 
doubt is Darwin’s statement that he had himself seen single and double flowers on the same Stock 
