86 Mr. B. Clarke on Relative Position ; 



carpel (see Part III. Compositse). It may be interesting, however, 

 here to add, that in Aste?' and Centaurea the ovule arises from 

 the base towards the posterior side of the ovaiy always more or 

 less distinctly. 



In such Cichoracese as I have examined, the raphe is for the 

 most part or always lateral (that is, towards one side of the ovary), 

 in no instance posterior ; but as the carpels in this section of 

 Compositse are right and left the axis, the position of the raphe 

 might be expected to be different. In Calytrix virgata the I'aphe 

 is not so completely averse as in the other instances, being in- 

 termediate between lateral and averse ; but in Berberis vulgaris I 

 have since ascertained it is always next the dorsum of the carpel. 



Causes of the Variations. 



The cause of the first of these variations has already been de- 

 monstrated by Dr. Schleiden [loc. cit.), and each of the others 

 being also especially deserving of attention as influencing the 

 value of the characters derived from the position of the raphe, 

 I would suggest the following as the most frequent, if not unex- 

 ceptionable. 



1. That a single ovule pendulous with the raphe averse, being, 

 as first observed by Mr. Brown, an erect ovule pressed or grow- 

 ing downwards, may result from the cavity of the ovary elonga- 

 ting in that direction, while its upper part remains stationary. 

 As tending to show that it may be produced by pi'essure, I have 

 met in Sassafras officinale with an instance in which the ovule 

 had apparently forced its way through the upper part of the 

 ovary as it was growing from its external surface, — the ovary 

 having again closed and the cavity still remaining, but empty. 

 It may be further suggested, however, that it is only when an 

 erect ovule has the raphe next the placenta, that it has the raphe 

 averse when it thus becomes pendulous. 



2. That a single pendulous ovule with the raphe lateral is an 

 ovule originally extending horizontally from the placenta with 

 the raphe lateral (as in Ranunculacese, where the ovules are 

 numerous, and in Cucurbitacese), and subsequently becoming 

 pressed downwards in consequence of the ovary elongating in 

 that direction more than upwards. By the raphe lateral in ho- 

 rizontal ovules, it is to be understood that it is not on the upper 

 surface towards the stigma, nor beneath it towards the base of 

 the ovary. 



3. That a single pendulous ovule with the raphe next the pla- 

 centa may possibly be an ovule originally extending in a hori- 

 zontal direction from the placenta, having the raphe on the 

 under surface, and should perhaps be regarded as the only truly 



