998 Groom.—The Evolution of the Annual Ring" and 
was evergreen. But the existence of deciduous species of Castanea with 
solely narrow rays prevents us from fully relying on such a judgement. 
The change from or to an evergreen condition appears easy in the family, 
since it has occurred independently in three genera— Castanea , Nothofagns , 
and Quercus.) 
The view that the primitive condition in Quercus (and Alnus ) was that 
in which only narrow rays existed has been strongly supported by Bailey 
(TO (1), TO (2)) and Eames (TO) in a number of papers. Their view 
is that the broad high rays have become established by a compounding 
process. The evidence upon which they rely is geological, ontogenetic, and 
pathological. The geological evidence based upon the structure of the 
wood of a few isolated specimens showing ‘ false rays ’ (i. e. many fine rays 
closely aggregated), though suggestive, is not convincing, as the species 
examined were too few. Very interesting on the other hand is Eames’s 
discovery that in seedling oaks all the rays are narrow, and that later 
in life, in the later-formed rings, there is a more or less gradual c com¬ 
pounding, that is, grouping and fusion * of the narrow rays, until eventually 
the group of rays is represented by a single broad high parenchymatous 
ray, more or less free from interrupting fibres. Eames regards this evidence 
as important from the standpoint that ‘ the seedling is known to be a seat 
of ancestral characters \ And Bailey has found the same kind of phenomenon 
in certain species of Alnus , in which the inner annual rings of the seedling 
have merely evenly distributed uniseriate rays, while annual rings farther 
out possess aggregated rays. From the point of view of the assumption 
that the seedling is a seat of phylogenetically early structure features, these 
facts lose their significance if it be shown that a similar linking up of small 
rays to form large ones takes place in older parts. I find that such a linking 
up does take place, at least in connexion with some of the rays of the inner 
rings of the twigs of Quercus Robur . Again, in Quercus Robur , apart from 
the broad rays whose initiation commences at the centre of the stem, 
there are secondary 1 broad rays that take their origin in annual rings distant 
from the centre. How do these actually arise ? I have made no serial 
sections of such broad secondary rays, but the fact that in tangential sections 
of various oak woods there are to be seen in the same annual ring, side 
1 Both Bailey and Eames seem to employ the terms * primary ’ and 1 secondary ’ as synonyms 
of 1 broad ’ and ‘ narrow ’ in reference to the rays. A ‘ primary ray ’ is in reality one that establishes 
radial parenchymatous continuity between pith and the wood farther out, since it is initiated in 
connexion with the original ‘ primary rays 1 separating the vascular bundles. A ‘ secondary ’ ray 
is formed later and is not in parenchymatous continuity with the pith, but it may be as thick as 
a primary ray. The term ‘compound’ as applied to the broad rays by the same authors seems 
unfortunate, first because the term has already been employed in two different senses, both of which, 
however, imply that the ray is actually composite in structure ; to use the same term for an undivided 
ray of Quercus is not only to assume that it represents a fusion of separate rays, but also to render 
the term ‘ compound ’ as a descriptive term useless in case of other rays whose real structure is 
divided or aggregated. 
