ANONACEJE. 259 



of liber within, though not quite so distinctly marked out. In a 

 longitudinal and tangential section of the bark we find each bundle 

 forming a broken line, whose segments are pretty nearly equal and 

 inclined to one another at very obtuse angles, also nearly equal. 

 On examining one of these bundles we find that it touches the two 

 bundles at its sides alternately. The vertex of one of its angles 

 meets the vertex of an ano-le on the bundle to its rierht ; that 

 of its next angle meets that of an angle on the left-hand one. 

 The vertex of the third ansfle touches one on the ricrht acain, and so 

 on. Thus is formed a network with vertically- elongated, lozenge- 

 shaped meshes, something like a trellis, whose rhomboidal openings 

 are bounded by liber bundles, and are filled up with those transversely 

 elongated cells described above. This arrangement is represented 

 on the surface by an unbroken network with little vertical clefts, the 

 peculiar arrangement of which is often useful as indicating the bark 

 of an Anonad at a glance. 



All this, however, only refers to the true Anonacece, to plants 

 belonging to our three first series. But in Eupomatia, which is on 

 other grounds an aberrant type, there are also great differences in 

 the histological structure of the axis. In the bark of a young branch 

 of E. Benneftii F. Muell., we have found a thick parenchyma,' whose 

 cells are full of chlorophyll granules, or here and there contain a 

 homogeneous pink liquid ; while there are numerous independent 

 liber bundles, crescent-shaped in transverse section. But we no 

 longer find the liber forming lozenge-shaped meshes, nor its bundles 

 projecting; the outer surface of the bark is smooth, except for the 

 two decurrent parenchymatous crests, continuing the angular edges of 

 each petiole down the stem. The pith consists of a single sort of cells, 

 thin-walled and riddled with pores. ^ The wood alone retains the cha- 

 racter observed in certain Pol/jcarjjica, especially Brimydcce. The fibres 

 are thick-walled, and bear longitudinal rows of areolate pores, which 

 are circular, or more frequently elongated and oblique. At the point 

 of contact of two adjoining fibres we find enormous biconvex, lens- 

 shaped cavities, each resulting from the apposition of two areola? ; 



' Adansonia, ix. 21. similarity in the organization of the flower, and 



- Tliese considerable differences in the stem confirm the view Hut iwjaowff/ZffS are indubitably 



structure between the true AnonacecB and the more closely allied to the liloniminctuv than 



Eupomatieee correspond, as we know, to great dis- to the Anonacece tliemselves. 



s 2 



