184 NOTICES OF BOOKS AND MEMOIRS. 



but not distinctly, doubly dentate ; teeth rarely gland-tipped ; end 

 leaflets 1-1|- in. long, f-|- in. broad, broadly rounded at the base. 

 Flowers 1-3, usually solitary ; peduncle naked, f- 1 in. long ; 

 calyx-tube broad oblong, naked ; segments f-f in. long, naked on 

 the back, with a distinct leafy point, some simple, some sparingly 

 pinnate. Corolla milk-w4iite, 18-21 lines across w^hen expanded ; 

 petals an mch long. Styles densely pilose. Fruit broad ovoid, 

 |— I in. long, turning colour at the end of August, deep crimson- 

 red, the spreading sepals fully persistent, still remaining attached 

 in October. Connects hibeniira (jlahra with canina, vars. sub-cristata, 

 Schultzii, and Hailstoni. — J. G. Baker. 



(To be contiuued.) 



S2(r les Caracteres et les Affinites des OUniees. By M. J. Decaisne. 

 Paris, 1877. 



In this pamphlet of fifteen pages M. Decaisne discusses in 

 much detail the structure and relations of the African genus, 

 Olinia, Thunb. (1799), which genus is very peculiar, and has been 

 treated as the type of a separate Natural Order by Walker Aruott 

 (1838), and other botanists. Opportunity has been taken of its 

 having recently flowered in the garden of the Paris Museum to 

 examine anew the flowers in the growing state, and to trace their 

 development ; and a full and clearly-figured analysis on an enlarged 

 scale is given on a plate appended to the pamphlet. 



In the first volume of the ' Genera Plantarum' (1867), Bentham 

 and Hooker placed the genus, to wiiicli they assign but a single 

 species, among the anomalous genera of Lythrane(E, from which 

 family it conspicuously differs by its sessile anthers, by its inferior 

 ovary, and by its definite ovules. But more recently (1876), M. 

 Baillon has ingeniously hit upon a different position for the genus 

 by attaching it to RhamnecB. A similar position had, however, been 

 previously adopted by DeCandoUe in the second volume of the 

 * Prodromus ' (1825.) This idea does not recommend itself to 

 M. Decaisne, and he finds that M. Baillon 's diagnosis of the genus 

 is erroneous in several important particulars. For instance, M. 

 Decaisne shows that the incurved scales, which alternate with and 

 are interior to the petals, and which arise from the top of the 

 calyx-tube, are not valvate in aestivation, as stated by M. Baillon, 

 but are quincuncial or imbricated. These scales M. Baillon con- 

 sidered to be the true petals, the sepals being represented by the 

 organs which M. Decaisne calls the petals ; and since the stamens 

 are placed immediately within the scales and opposite to them, the 

 affinity with RJiamnetr would easily be suggested. 



M." Decaisne places side by side the characters of Ilhanmeie and 

 of Olinia, and thus exhibits so many and important divergences 

 that it is impossible, on the face of the contrast, to conclude that 

 the two groups are closely connected. 



On the other hand, M. Decaisne discovers a complete agree- 

 ment between the floral structure of Olinia and that of several 



