448 Browne . — A Fourth Contribution to our Knowledge of 
above, the protoxylem seems to be generally in contact with part of 
the metaxylem. In E. arvense , in which the tracts of parenchyma separat- 
ing the internodal strands of the cone are relatively narrow, the metaxylem 
develops chiefly in a lateral direction, and is of slight radial extent, so that 
it is usually separated from the protoxylem by parenchymatous cells. But 
in this species I have observed that in certain bundles metaxylem tracheides 
are in contact with the carinal canal for a considerable part of the internode. 
In E . hyemale and E. sylvaticum the relations between the protoxylem and 
the metaxylem in the cone vary between continuity and separation. In 
E. sylvaticum (in my specimens at least) the protoxylem usually abuts on 
the metaxylem, but in E. hyemale either condition seems about equally 
common. Often in a single section of a bundle the metaxylem abuts 
on one of the sides of the carinal canal, and is separated from the other by 
a parenchymatous cell. In E. giganteum, in which the metaxylem is of 
relatively considerable depth, the protoxylem of the cone seems always 
to be continuous with the metaxylem. Continuity of metaxylem and 
protoxylem is characteristic also of the internodes of the cone of E. debile 
and E. variegatum , though occasional indications of a local separation 
of protoxylem and metaxylem by one or two parenchymatous cells may be 
observed. 
The above remarks may be summarized by saying that the separation 
or continuity of metaxylem and protoxylem depends chiefly on the relative 
radial extent of the former. Where the development of metaxylem is 
chiefly in a lateral direction (e. g. E. arvense and E. maximum ) separation 
of metaxylem and protoxylem in the internodes is the result. Where, as in 
E . giganteum, the metaxylem (whether disposed in wide or narrow bundles) 
is of considerable depth it remains in contact with the protoxylem. Though 
the internodal strands in the cones of E. debile and E. variegatum are 
on a small scale, so that the actual number of tracheides present is small, the 
radial extent of the metaxylem is considerable compared with its width. 
The continuity of protoxylem and metaxylem in the internodes of the 
cones of these species is, therefore, not surprising. In the species in which 
the protoxylem may be either continuous with or separate from the 
metaxylem continuity usually occurs when rather more metaxylem is 
present radially or, more rarely, if the band of metaxylem is slightly 
incurved. If the vascular system of the cone is a reduced siphonostele 
it would seem to follow that the separation of protoxylem and metaxylem 
is a derivative character, resulting from poor development of the metaxylem 
radially. It should be pointed out that although it is possible that in some 
cases parenchymatous cells may have been destroyed during the formation 
of the carinal canals, yet this possibility does not allow of our regarding the 
metaxylem abutting on the canal as originally separate from the proto- 
xylem ; for in many cases the metaxylem at the edge of the canal is 
