4.44 OLA.CACE.<E. [Hvpogynous Exogens. 



of observations on the affinities of this Order, have lately been published by me. 

 From what has already been shown, we learn that other good distinctions are 

 derived from the peculiar modification of the disk ; and we are led to better notions 

 of the real affinities of the Order, by a more exact knowledge of the structure of 

 the ovarium and fruit. According to these views, the tribe Icacineae Benth. must 

 be altogether separated from the family, and for the reasons quoted in the next page, 

 established as a distinct order. The Olacacese, thus limited, are characterised in 

 the following manner : — Trees or shrubs with alternate, entire, coriaceous leaves, 

 without stipules ; flowers small, generally fragrant, hermaphrodite, or polygamous, 

 in close axillary panicles or corymbs. Calyx small, cup-shaped, entire, or slightly 

 toothed, persistent, sometimes becoming considerably enlarged with the fruit. 

 Petals generally 5, rarely 6, oblong, valvate in aestivation, with the summits apiculated 

 and inflected, margins sometimes adhering at base into a tubular form by the aggluti- 

 nation of the stamens, or often cohering in pairs by their margins nearly to the 

 summit, from the same cause, frequently furnished with hairs inside. Stamens 5 to 

 10, generally partly sterile, fertile 3 to 10, of which 5 or fewer are always opposite 

 the petals ; the sterile, generally alternate with them, are appendiciform ; filaments 

 shorter than petals, free in bud, but afterwards often partially agglutinated to them 

 by a nectariferous exudation : these are always inserted either upon the elevated 

 external margin, or outside of the conspicuous disk ; anthers innate, oblong, 2-lobed, 

 bursting longitudinally. Ovary seated within the disk, which in Opilia is divided 

 into 5 fleshy, linear lobes, alternate with the petals ; but in all other genera this disk 

 expands into a cup-shaped nectary, which is sometimes free from the ovary, and 

 partially adnate to the calyx, but more often wholly confluent with the ovary, and 

 free from the calyx, rarely free from both the ovary and the calyx : when this 

 assumes a cup-shape, the petals and stamens are borne on the external portion of its 

 limb. Ovary always quite superior to the calyx, but often imbedded in the disk, 

 frequently surmounted by a remarkable fleshy epigynous gland, that sometimes covers 

 its upper moiety, always unilocular at the summit, and incompletely 2- to 5-locular at 

 base, the placenta? arising in a free column from the axile line of junction of the 

 incomplete dissepiments, and erect in the summit, and sometimes extending into a 

 cavity of the style, but always free. Ovules generally one in each pseudo-cell, all 

 pendulous from the free central placenta. Style simple ; stigma more or less 

 clavate or imperfectly 2- to 5-lobed. Fruit somewhat drupraceous, frequently 

 enveloped by the enlarged ventricose calyx, and enclosing an indehiscent unilocular 

 1-seeded osseous putamen. Seed exutive, or wanting all integumental covering, 

 apparently suspended by a raphe-like thread proceeding from the base to the summit 

 of the cell, which, as in Santalacea?, is the withered remanet of the dissepiments and 

 placentary column. Albumen fleshy, having in the summit of its axis a small 

 embryo with short cotyledons and a superior terete radicle. 



" From the above characters, it is evident that the nearest affinity of the Olacacese is 

 to the Santalacea? on the one hand, and to the Styracea? and Humiriaceae on the other. 

 The Santalacea? differ only from the Olacacea? in the confluence of the calyx and 

 corolla into a more or less complete perigonium, and of this with the disk and ovary 

 in a more or less perfect degree. All those genera of the Santalaceae possessing 

 distinctly dichlamydeous flowers are therefore by me referred to Olacacese, with which 

 they accord in all essential respects. Cansjera is rejected from the order, its position 

 being near Thymelacea?. 



" I have also proposed to form a distinct alliance (the Cionosperrnea?) for those 

 families with polypetalous flowers, having a calyx generally free, sometimes confluent 

 with the corolla ; petals sometimes united at base by partial agglutination of the 

 filaments ; stamens, though often adherent to the petals, always originate from the 

 external surface of the disk ; an ovarium, with a simple style and stigma, is always 

 unilocular at summit, with ovules attached to a free central placenta, which he calls 

 a Cionosperm. Seeds either indutive, or exutive,* i.e. with or without the usual 

 integumental coverings. This alliance will comprise, first, those families having 

 indutive seeds, viz., Myrsinacete, Theophrastacea?, Styracese, Humiriaceae, iEgiceraceae, 

 and perhaps Aptandracea? ; second, those with exutive seeds, viz., Olacacese, 

 Santalacea?, and Viscacea?. The principle upon which this alliance is founded, 

 proceeds on the basis that we should look to the phenomena of the development of 

 the reproductive organs of plants as the ground on which all natural methods of 



* This expression is preferred to that of uok-id seeds, the application of which term, having been 

 made in various significations, might lead to confusion ; it has been used for the seeds of the 

 Coniferse ; it has been applied by Linnaeus to the gynobasic seeds of the Labiates ; it has been 

 adapted to instances like the present, and was employed by Bartling to denote the absence of the 

 vitelhis around a seed. — J. M. 



