46 THE CELL-WALL. 
brane, destined to constitute a partition-wall, for open communications such as the 
above. For segments of the wall of various sizes are made from the beginning with 
sieve-like perforations, as is shown in fig. 104, which represents diagrammatically 
portions of tubular cells called “ sieve-tubes.” The pores are crowded close together 
on the perforated areas of the walls of the sieve-tubes, and their dimensions are 
relatively broad and short. ‘Thus, when two neighbouring protoplasts reach out to 
one another through these pores, that is to say, when there is continuity of the 
protoplasm of the two cell-cavities, the connecting filaments, which pass through the 
pores and which fill them completely, are short and thick and have the appearance 
of pegs or stoppers. 
But in many cases the pores through which adjoining cell-cavities communicate 
are drawn out to a great length, forming infinitesimally slender passages. They are 
situated close together in great numbers and penetrate transversely through the 
thick cell-walls (fig. 10°). Neighbouring protoplasts may be brought equally well 
into mutual connection by means of these canals, or perhaps it would be better to 
say that their connection may be equally well maintained. For it is very probably 
the case that in the first rudimentary partition-wall, which is produced between the 
products of division of a protoplast, minute spots remain open and are occupied by 
connecting threads common to both halves of the protoplasm as they draw apart. 
Then in proportion as the partition-wall between the two protoplasts, produced by 
the division, becomes thicker, the openings take the form of fine canals, and the con- 
necting filaments are modified into long and exceedingly fine threads which fill the 
canals. These protoplasmic threads pierce through the thickened cell-wall in the 
same way as a dozen telegraph-wires might be drawn through a partition from one 
room into another. Often a number of protoplasts living side by side and one 
above the other are linked together by filaments of this kind, which radiate in all 
directions. 
This species of connection, of which an intelligible idea is given by fig. 10°, 
escaped the notice of observers in former times owing to the extraordinary minute- 
ness of the canals, and delicacy of the protoplasmic filaments. Another method of 
communication between protoplasts in adjoining cells has, on the other hand, been 
long known and often described, its phenomena being very striking and visible when 
only slightly magnified. The connection referred to is that which is afforded by 
the formation of so-called “vessels.” By vessels the older botanists understood 
tubes or utricles, arising from the dissolution of the partition-walls between a series 
of cells. Either the partition-walls in a rectilineal row of cells vanish, in which 
case long straight tubes are produced; or portions of the walls of cells arranged at 
different angles to one another are dissolved, and then tubes are formed having an 
irregular course, and sometimes branching or even uniting, so as to make a net-work. 
In instances of the first kind the lateral walls of the series of cells which are to lose 
their transverse partitions are previously thickened and made stiff by the proto- 
plasts, which also provide them with various mouldings and panellings, and above all 
with bordered pits. This task accomplished, the protoplasts forsake the tubes, whose 
