Anonacece.] flora indica. % 87 



is also universal, but is met with in many of the neighbouring families. The 

 sepals always form a single verticil; and the petals, which never exceed six in num- 

 ber (in two rows), are in a few instances reduced to a single row by the suppression 

 of the inner series. In Magnoliacece they are generally much more numerous. The 

 anthers are always more or less extrorse, but the number of stamens is far from 

 constant, being in many genera reduced to 18, 15, 12, 9, and even as low as 6. 

 The ovaries are occasionally subdefinite, or even solitary, and the carpels are some- 

 times dehiscent. The valvate aestivation of the petals, which, when present, is the 

 most conspicuous character of the Order, is wanting in the Section Uvariea* 



The state of this comparatively little known Order is still very unsatisfactory, not- 

 withstanding that it has received the attention of many of the principal botanists of the 

 day, nor is it to be expected that the tribes and genera can be established on a proper 

 basis, till the species have been much more carefully and completely examined than 

 their very imperfect condition in herbaria has hitherto permitted them to be. Their 

 study, indeed, even under the most favourable circumstances, presents great difficul- 

 ties to the student of dried plants, from the minute size of the stamens and ovaries, 

 and from the bad state of preservation in which the flowers occur in herbaria. Though 

 the flowers are often large, they are generally more or less fleshy, and in drying 

 become much flattened and distorted, so that the restoration of the natural state is 

 almost impossible. The determination of the number of ovules is, in particular, a 

 very difficult matter, as the minute ovaries are always much compressed ; and their 

 walls are so brittle, that the dissection necessary for the isolation of the ovules can 

 only be effected by much patience, and with an abundance of materials. 



The number of species of Anonacece known to the older botanists was too small 

 to permit of any great progress being made by them towards the proper circum- 

 scription of the genera. These were first accurately defined, and the species care- 

 fully described, by Dunal, in a monograph of the Order, published in 1817. At 

 that time only 103 species of the Order were known, most of them very imperfectly. 

 Of these scanty materials M. Dunal has certainly made much ; and his work, which 

 has formed the foundation of all that has since been done, has been well characterized 

 by M. Alph. De Candolle as being a monument of talent and sagacity, considering 

 the period when it appeared. The 'Systema' and ' Prodromus' of De Candolle contain 

 no additions to the labours of Dunal, who had at his command all De Candolle's ma- 

 terials ; and since that period the Order has not been treated generally, except by 

 M. Alph. De Candolle, in a memoir in the fifth volume of the Geneva Transactions, 

 in which the additions to the Order, up to the year 1832, are reviewed. The num- 

 ber of known species is there stated at 204. 



Much attention has, however, been directed to the definition and arrangement of the 

 genera of Anonacea, in all the works which have been published of late on tropical 

 botany ; and so many remarkable forms have been figured, that much greater faci- 

 lities are now afforded for the correct appreciation of affinities, than were available 

 to the older botanists. The works of St. Hilaire, Martius, and Richard, on Ameri- 

 can Botany, and the 'Flora Java?' of Blume, have all contributed much to our know- 

 ledge of the Order. The careful analyses and excellent descriptions of the Eastern 

 forms in the last-mentioned work, in particular, have been of the greatest service 

 to us. 



From the time when the number and position of the ovules was first indicated by 

 Brown as an important character in Anonacece, in his remarks when founding the 

 genus Artabotrys, in the ' Botanical Register/ this character has been generally em- 

 ployed, not only for the distinction of genera, but also for the formation of the pri- 

 mary divisions of the Order. But though the number and position of the ovules is 

 nearly constant in each species, and therefore constitute most important characters 

 for the distinction of genera, the higher groups thus characterized appear to us un- 

 natural, and we therefore think it desirable to employ other characters for their 

 circumscription. Five aberrant tribes appear to be at once distinguishable by well 

 marked and easily recognizable characters. 



The first of these, which may be called Uvariea?, from its principal genus, has its 



