196 THE STRUCTURE OF FLOWERS. 



the other hand, in Macleya, Bicentra, Epimedium, Bartonia, 

 Impatiens, Ly thrum, Bracophyllum, MalacMum, Cerastium, 

 Primula, and Samolus the order is hasipetal, or from above 

 downwards. 



When the row of ovules is very nnmerons, it is the rule 

 that the point where they first begin to emerge is midway, 

 and the development takes place both upwards and down- 

 wards simultaneously. It is thus with Hellebortis and allied 

 genera with follicles, Capparis, Epilobium, Trifolium, Cajo- 

 phora, Lathyrus, Citriis, Passiflora, and the Monocotyledonous 

 orders, Iridacece and Amaryllidacem. Lythrum and Opuntia, 

 however, both of which have considerable rows of ovules, 

 develop them, as stated above, in a basipetal and basifugal 

 manner respectively. 



On examining Payer's numerous figures, I find that when 

 the order of development is from below upwards, the ovules 

 have their micropyles upward ; when they develop from 

 above downwards, the micropyles grow downwards. In 

 either case, occasionally the middle ones may be somewhat 

 horizontal, if they are somewhat numerous, as in Bartonia, 

 Spiraea, and Staphylea. When they are very numerous and 

 develop both ways from a point midway, then the ovules 

 may either tarn upwards or downwards ; the majority being 

 downwards in the proportion of nine to five. 



As a theoretical interpretation to account for the general 

 fact of the central ovules developing first when there are 

 long rows of them, it may be due to the carpel being com- 

 parable to a lanceolate leaf, where the longest and therefore 

 the most vigorous nerve-branch of the pinnate nerves is in the 

 middle. If the rows of ovules emerge from below upwards, 

 the carpel may be comparable to a more primitive type, as of 

 monocotyledons with a palmate foliage. Thus the only 

 exceptions I can find in Payer's figures of Monocotyledons are 



