106 Oliver . — On the Obliteration of 
in the various species of Macrocystis, the sieve-plates being for 
the most part horizontally placed, dividing up the sieve-tube 
into members. As a rule sieve-plates will occur every '2-25 
mm. in any tube. These plates are perforated by large cir- 
cular or, more usually, polygonal pores. The plates indeed 
forcibly recall those in Cucnrbita. Sieve-plates are also found 
in great numbers on the longitudinal walls, occurring wherever 
two tubes run adjacently. These plates may be either on the 
radial or on the tangential walls. Fig. 14 is a longitudinal 
radial section through two sieve-tubes of M. pyrifera. Where 
they impinge on one another they are separated by a vertical 
sieve-plate. In the figure are seen sieve-plates (on which are 
developed thick callus-plates, coloured red) in all three planes. 
The protoplasmic content, p, lies contracted in the centre of 
the tube, spreading out however over the sieve-plates and 
forming the familiar Schlauchkopfe of German authors. 
It is to be noted that the horizontal sieve-plates are always 
larger than those on the vertical walls, and also that the per- 
forations of the former are wider than those of the latter. 
This is not surprising, for, assuming these sieve-tubes to 
function here as they do in higher plants, it is obvious that 
the requirements for transport of plastic material in such a 
plant as Macrocystis must be very much greater in a longi- 
tudinal than in other directions ; hence the larger horizontal 
sieve-plates with large perforations. These perforations have 
an average width of *003 mm. On the older sieve-plates, both 
horizontal and vertical, a well-developed callus-formation is 
found. This ultimately becomes very thick and completely 
obliterates the perforations. The thickness of the whole callus- 
development of both sides of any sieve-plate may reach as 
much as # o 75 mm., though in most specimens such a great 
thickness is not attained. In these sieve-tubes, as opposed to 
the trumpet-hyphae, there is no callus-formation in connection 
with the wall other than at a sieve-plate. This callus, as I 
shall show later on in this paper, gives reactions identical with 
the ordinary phanerogamic callus, and must be regarded as the 
same substance. The fully-developed callus-plates normally 
