Laminaria saccharina . 
305 
the interior of the tube. Later, this also is often obliterated and the cavity 
becomes entirely blocked. 
b. Secondary sieve-tubes, (i.) Surface views. A study of the earliest 
stages in the development of the terminal sieve-plates, both in the young 
material and in the sieve-tubes found on the periphery of the medulla in 
the older stems, yields very similar results. In the youngest sieve-plates 
there is a uniform distribution of the threads. Fig. 5 > Ph XIX, repre- 
sents a slightly older and very much more frequent case in which the 
centre of the plate is still occupied by single threads, but towards the edge 
little groups of four or five threads are present ; between the centre and 
the periphery groups are seen composed of two or three threads. Other 
cases are found in which the centre of the plate is in this last stage, 
while groups of four to seven threads are present on the periphery. In 
Fig. 6, PI. XIX, is shown an older plate, just before callus formation, 
in which the threads are distributed throughout in groups, five being the 
usual number, though sometimes as few as four or as many as seven 
threads may be present in one group. These stages were seen in pre- 
parations stained with Safranin, Safranin and Aniline blue, Aniline blue only, 
and Benzyl blue, so that there seems little chance that any of them can be 
an artificial product. 
Numerous estimations were made of the numbers, both of the separate 
threads in the youngest sieve-plates and of the groups of threads in the 
later stages, and there is no doubt that the two quantities broadly corre- 
spond. It seems impossible, therefore, that the groups are formed by the 
aggregation of the original single threads ; rather, it seems likely that each 
group has arisen from the division of a single thread. This conclusion is 
also supported by the fact that groups of two or three threads are often 
found in an intermediate position between the earliest arrangement of separate 
threads and the final resulting groups of five or more. 
In a surface view of a sieve-plate callus is first seen as a small ring sur- 
rounding each of the threads which make up such a group, and a young 
sieve-plate in this stage shows numerous small circles each made up of 
four to six blue points, as shown in Figs. 8 and 9 b , PL XIX. By lateral 
fusion this group of separate spots soon gives rise to a single ring (Figs. 8 
and gc, PI. XIX). 
All this time the threads remain distinct, though some change takes 
place in their composition, for, simultaneously with the advent of the callus 
they stain more deeply than before with protoplasmic dyes. 1 Soon after 
the fusion of the separate callus-spots into a single ring, alterations take place 
very rapidly at the centre of each group of threads. The final result is that 
a single slime-string is formed, surrounded by its ring of callus, in place of the 
original group of independent threads (Fig. 7 and Figs. 8 and gc, PI. XIX). 
Before arriving at this stage two or three smaller slime-strings are often 
1 Cf. Hill, 1 . c., 1901, II, p. 589. 
