SECONDARY CHANGES OUTSIDE THE ZONE OF THICKENING. 549 



again by those in centripetal order. The latter process may then be subsequently 

 repeated from time to time. This is the case in most Pomaceae, Virgilia, Solanum 

 Dulcamara, Hamamelis, Platanus, and Acer campestre. In the majority of the forms 

 investigated, however, the appearance of at least one layer of phelloderm immediately 

 after the first divisions is the rule, and division thus takes place in the orders termed by 

 Sanio centripetal-intermediate (Aronia rotundifolia, Fagus, and Rhamnus Frangula), 

 centrifugal-intermediate, and centrifugal-reciprocal. The two latter orders of succes- 

 sion occur, for example, in Staphylea pinnata, and all the species enumerated on p. 548, 

 after the mark ; — . 



Sanio describes in Viburnum Opulus an exquisite example of individual variations. In 

 summer the first layer of periderm is here formed in the centrifugal-reciprocal order, the 

 reversal of the centrifugal succession taking place after 3-5 divisions. In internodes, 

 which only develope their periderm later, in September, the reversal takes place more 

 rapidly, immediately after the second division (centrifugal-intermediate order) ; those 

 internodes, finally, which only begin to form cork late in autumn, show a purely centri- 

 petal succession. The relations which here manifestly exist between the change in the 

 process of development and the action of external agents (warmth, light, &c.) may be 

 recommended to more accurate physiological investigation. Various similar changes, 

 which cannot always be brought into relation with external influences, are described by 

 Sanio, /. c. 



The number of the phellodermal layers, which arise in the structures under considera- 

 tion, is in most species very small, in comparison with the cork-layers which appear in 

 the same space of time. In each radial row only one or two phellodermal cells cor- 

 ; respond to numerous cork-cells, even after several years ; in periderms which grow for a 

 very long time, e. g. Fagus, this relation constantly becomes more striking as time goes 

 on ; in a branch of this tree four years old Sanio figures two phellodermal layers to more 

 than seven cork-layers, and found in branches twelve years old only two or three layers of 

 phelloderm, while the number of the cork-layers, though not exactly stated, had certainly 

 increased considerably. 



Here also, however, numerous deviations from the usual rule occur, generally according 

 to species and genera. In most investigated species of Salixy each initial epidermal cell 

 produces in the first year one cork-cell externally, and one phellodermal cell internally ; 

 between the two there is a central meristematic cell, with its wall thickened on the out- 

 side, and immediately becoming cuticularised on the external thickened surface. In 

 this central meristematic cell the same division and differentiation as in the initial epi- 

 dermal cell is repeated in the second year, and the same process takes place in each 

 succeeding year, starting from the meristematic cell for the time being, until the forma- 

 tion of bark begins at a later period. An abundant formation of phelloderm — six layers 

 by the third year — was found by Sanio in Quercus Suber^ where, however, the formation 

 of cork-cells is also very abundant. Canella alba and Cinnamodendron corticosum may 

 also be mentioned here ; the thin old cortex of these plants, which is used commercially, 

 shows (when not externally abraded), in so far as the structure can be recognised from 

 the seriation of the elements in question, massive zones of phelloderm, which may be 

 over twenty layers thick ; these lie inside bulky layers of cork, and are separated from 

 them by a layer of meristem. The elements of the phellodermal rows are almost cubical, 

 and consist chiefly of stone-scierenchyma, but with the latter thin-walled, unlignified cells, 

 sometimes containing starch, sometimes clustered crystals, are intermixed in the most 

 various ways. My material did not allow of a more accurate determination of the origin of 

 these suberous and phellodermal structures ; according to observations on single pieces 

 of cortex their development in the deeper cortical layers is at any rate not improbable. 



As is evident from the preceding statements, and from those in Sect. 24, the 

 quantity of cork-layers originally formed, and of those produced anew by the meristem 

 to replace the layers which peel off during the progress of growth in thickness, is 



