58 THE ROYAL SOCIETY OF CANADA 
in the bast.” He, therefore, takes the view that wherever there are no 
vertical canals present in the xylem, horizontal passages will also be 
lacking, which is, generally speaking, the case, but for reasons different 
from those given by Mayr. Exceptions to this rule are, however, 
recorded, as e. g., in Sequoia Burgessii (254, 45) where horizontal canals 
are present but vertical ones have not been observed. Moreover, the 
same is in reality the case wherever rays show a tendency to become 
multiseriate, without exhibiting a parallel tendency to the formation of 
vertical resin passages, for the multiserate ray is the direct precursor 
of the horizontal passage. 
The medullary ray canals cannot be said to take their origin from 
the vertical ones, for if anything of this nature occurs, it is the vertical 
ones which owe their origin not only to those rays containing a passage, 
but also to those which do not. Wherever there is a tendency in a 
plant towards the formation of multiseriate rays, the conditions calling 
this forth are also favorable for the initiation of either vertical or hori- 
zontal canals, and in the majority of cases of both. The ray does not owe 
its multiseriate character to the appearance of the canal in its midst, 
for the latter is only a secondary development induced by certain con- 
ditions which usually present themselves in the development of a ray 
of multiseriate character. Where such conditions are not at hand, no 
canal will be formed even though the ray be several cells wide. 
Fischer (9, 277) states that: ‘A view current since Goeppert is 
that a medullary ray which in its tangential direction consists of two 
or more rows of cells in the genera Pinus, Abies, Picea and Larix always 
encloses a large horizontal canal in its centre.” His own observations, 
however, showed that in Pinus abies, L., both in the stem and in the 
root, rays two and three cells in width, but not enclosing a resin canal, 
were present. In some cases, on the contrary, two and even three 
canals occurred in the same ray, some of the rays possessing canals 
being as many as five cells broad. “The resin passages do not always 
lie in the centre of the ray, and those which are multiseriate but contain 
no canal may show their greatest width either at one end or at both 
ends, and not necessarily in the centre.” 
Moreover, it frequently occurs that multiseriate rays are present 
in both fossil and existing forms, which do not exhibit any sign of a 
canal. This is well illustrated in Cupressinoxylon where some species 
possess one to three-seriate rays, as seen in tangential section, but are 
lacking in resin canals, and in species of Cupressus, Sequoia, Juniperus 
and many others where the rays are frequently two-seriate in part but 
do not exhibit a resin passage. It will be shown below that the presence 
or absence of canals in multiseriate medullary rays, and their position 
within the ray when present, depend solély on the tensions present in 
