BOTANICAL SUBJECTS. 333 



ia Hermione* the crown is scarcely recognizable, and is like that 

 of the cowslip, and "Forget-me-not " (corona obsoleta lutea). 



In the throat of the Hippeastrum again there is a fringe 

 which pro1>ably serves to shut out small insects from the nectar, 

 which is secreted further down. This is evidently the homologue 

 of the throat teeth of N. deHciens, which in N. Poeticus have 

 fused and become a shallow fringed cup. In the Jonquil this cup 

 has a level margin, but in other Narcissi and in the Daffodil it has 

 become a prominent part of the flower, with its margin divided 

 into six lobes. Then in many of the double forms of Daffodil 

 this same crown or cup is sj)lit up again into petals. This crown 

 appears to be nothing but a disk, situated outside the stamens ; 

 while in the Ruta and Orange, the disk is situated inside the 

 stamens. In Rhus Cotinus (Sumach) the stamens are inserted on 

 the disk. They are similarly inserted in Sanginsorba officinalis. 



The petaloid scales which emerge from the disk of Brexia 

 spinosa are of the same nature as the throat-fringes of Hippeastrum , 

 and so are the so-called nectaries of Parnassia palustris, which are 

 glandular fringes emerging from a disk at the base of the petals. 



In the Citrus and lluta the disk has a wavy margin, as shown 

 in Figs. 139 and 150, each crenation possibly representing a 

 suppressed stamen or group of stamens. If in these two the ovarv 

 were engulfed so as to become inferior, the disk would remain as a 

 cup, similar to the crown of the Narcissus, but i?iside the stamens. 



In the PcBonia Moutan this same cup is within the stamens 

 proper, and envelops the whole ovary, excepting the stio-matic 

 surfaces of the carpels. It often has teeth on its margin, and on 

 one occasion I found an anther on one of the teeth, which would 

 clearly indicate its origin. It is a fasciation of f^tamen Jilameuts, 

 antl its beautiful purple colour matches that of the lllaments proper 

 exactly. 



If we take the development of the stamens proper of the Citrus 

 according to J. B. Payer, we find that at first only Jive nipples 

 appear, alternating with the petals ; then on each side of these 

 original stamens (not as an inner whorl) another nipple appears 

 as if it were the tooth or lobe of a leaf, making fifteen stamens 

 with the original five now more developed. 



On the sides of these again other nipples apjjear, so as to make 

 twenty-five, each group forming a five-toothed leaf, each tooth 



*' Bot. Eeg.," V. 33, p. 2-?,. 



