410 MISS M. BENSON—CONTRIBUTIONS ТО THE 
Examination of the fresh material was, as I have said, made as far as possible when it 
was gathered, and, although the exceeding delicacy of manipulation required of an em- 
bryologist, if trustworthy results are to be obtained, could scarcely be called into play 
when material was accumulating so fast, I soon found myself in contact with new and 
interesting forms which I had never met with before. It seemed advisable that I should 
ascertain the whole course of development of the embryo-sac and its contents in at least 
one particular group of genera and thus be sure of my ground іп oneconnexion. І chose 
the Cupuliferze, which, although more normal in many respects than some other groups, 
will prove, I think, on further description, to be an interesting point of reference when 
we pass to others. 
CUPULIFERZ. 
This group is referred to in the ‘Genera Plantarum’ of Bentham and Hooker as the 
Quercineze, under which name are placed the four genera Quercus, Castanopsis, Castanea, 
and Fagus. For embryological purposes it appears to me best to give some account of 
Fagus first. | 
Faeus. (Pl. LXVII. fies. 1-11; Pl. LXVIII. figs. 12, 18.) 
The species selected for observation was Fagus sylvatica, Linn., and to this I refer by its 
generic name simply. Early in April we Сап just distinguish the buds which contain 
the young flower-heads from the less bulky leaf-buds. If the young female flowers are 
now examined we find the ovules are already laid down, and rudiments of the inner 
integument are seen to be forming, partly at the expense of the epidermis. Simultaneously 
one or more contiguous epidermal cells at the apex of the young nucellus divide by 
periclinal walls, While this is takin g place there is differentiated immediately below them 
a layer of subepidermal cells, generally from three to six in number, as seen in a longi- 
tudinal section, which forthwith begin to divide by parallel straight walls in the same 
manner as the apical epidermal cells, 2. e. periclinally (figs. 1, 2). 
Although nearly the whole of the nucellus is constructed from these two layers, only 
the central strands maintain the regular mode of division and become strongly marked 
out from the peripheral tissue. Ina transverse section this core of the nucellus generally 
separates from the surrounding tissue and leaves a ring of nucellar tissue within the 
double ring of the transversely-cut integuments, As development proceeds the epidermal 
cap spreads over the whole upper surface of the nucellus and grows to be about thirteen 
cells deep, thus forming nearly a third of its bulk (fig. 9, е.р.). 
The central strands of the nucellus, as follows from what has been said, extend down 
to the level of the insertion of the inner integument, and are composed of large prismatic 
cells—often hexagonal in transverse section, and, especially in the central regions of the 
nucellus, elongated in the direction of its axis. It not unfrequently happens that in one 
strand certain partition-walls are delayed in development, and thus we get one long сей | 
corresponding to a tier of three or more cells. These long cells are often crowded out 
(fig. 10, 4.с.). It seems, however, that, at least occasionally, the embryo-sac may arise in 
this manner (fig. 8, /.с.). This central series of coaxial strands appears to be “ sporogenous 
tissue,” although it is very rare in Fagus to find more than one cell continuing its 
