ANGIOSPERMS. 573 



the last remaining doubt on this points Cramer has shown in a number of 

 other instances that all stages of the metamorphosis of ovules occur when the 

 flower is developed in a monstrous condition, leading also to the conclusion that 

 the nucellus is a lateral development on the funiculus of the ovule. Malformations 

 of Delphinium elatuni, where the ovules spring from the margins of the carpels, show 

 that the carpel is transformed into a flat open pinnate leaf, the lobes of which are 

 the metamorphosed ovules. The nucellus here springs from the upper or inner 

 side of the lobe of the leaf which represents the transformed funiculus together 

 with the integument. In Meliloius, Primula chinensis, and Umbelliferae, Cramer 

 found the same to be the case ^. Relying on this and other facts, and on the 

 hypothesis that the ovule is never a terminal structure of the floral axis, Cramer^ 

 adopted the view that the ovule is either a metamorphosed leaf or part of a leaf 

 (a tooth or outgrowth of the upper surface). The ovule of Primulaceae and Com- 

 positse he considered to be a whole leaf, and he supposed that closer observation 

 would show the same to be the case in other flowers also, especially in those 

 where the flower is said to possess a solitary ' reputed terminal ovule,' as Urtica 

 (and Taxus), and perhaps also the Dipsacaceae and others. The nucellus would in 

 this case be a new formation on the surface of the ovular leaf, the funiculus would 

 correspond to the base of this leaf, and the integuments to its upper part, which is 

 folded once or twice in the form of a cup or hood round the nucellus. On the 

 other hand he would consider as only portions of the leaf (teeth or outgrowths of 

 the upper surface) all those ovules which spring singly or in numbers from the 

 margin or upper surface of carpellary leaves, as those of Cycadeae, Abietineae (?), 

 Liliaceae, Umbelliferae, Ranunculaceae, Resedaceae, Cruciferae, Leguminosae, &c. In 

 these cases the nucellus would be a new formation on the surface of the lobe, the 

 funiculus would correspond to its base, and the integuments to its upper part folded 

 once or twice round the nucellus in the form of a cup. Only in those few plants in 

 which the ovule has no integument would the naked nucellus or entire ovule corre- 

 spond to this lobe of the carpellary leaf. In the first edition of this book I expressed 

 my agreement with Cramer's view, but with a reservation with respect to Orchideae, 

 being especially influenced by the importance which I then attached to the morpho- 

 logical equivalency of the nucellus in all Phanerogams. Further reflection has, 

 however, deprived this reason of its importance; and I am the more induced to 

 ascribe diff'erent morphological significations to the ovules, according to their mode 

 of origin and their position, because (as has been shown by Magnus, Rohrbach, 

 Hanstein, and Schmitz *) in Piperaceae, Typhaceae, and Naiadeae the ovule is 

 actually the terminal structure of the floral axis, and in Naias this terminal ovule 

 is also anatropous. In these statements I not only find the confirmation of my 



' Schenk writes, ' that which appears plausible enough in the Compositoe is certainly not the 

 case in other families ; the ovule is not a lateral branch of the primitive rudiment, but this itself 

 developes into the ovule.' 



^ Compare also H. von Mohl, Vermischte Schriften, pi. I. figs. 27-29. 



^ Cramer, Bildungsabweichungen bei einigen wichtigeren Pflanzenfamilien und die morpho- 

 logiscbe Bedeutung des Pflanzeneies (Zurich 1869, p. 120), where the literature of this subject has 

 been carefully treated. 



* These researches have been already quoted; [see also Eichler, Helosideen, Bot. Zeitg. 1868, 

 and Stiasburger, Coniferen und Gnetaceen, 1872.] 



