88 
F. E. Fritsch. 
actual connection with the neighbouring cell-wall.—There is, 
however, another type of intercalary heterocyst, which is not rare 
in older threads. This has a more oblong shape and each end is 
drawn out into more or less of a point (cf. fig. 5 of an extreme case) 
with a very distinct pore, whilst the wall is uniformily thickened. 
Some of the terminal heterocysts occasionally also exhibit such 
characters. I have not been able to follow up the development of 
this type of heterocyst; possibly it arises by a transformation of 
the larger, oblong vegetative cells. 
It will be found, that this account of the development of the 
heterocyst scarcely agrees with Brand’s (1903, pp. 39 and 40). 
His observations were apparently chiefly made on Nostoc commune 
and on species of Tolypothrix and it is of course possible, that the 
details of the development may vary in different forms. No 
mention is made of the terminal granule, which 1 have found so 
very characteristic of early stages, whilst his account says that 
“ Die innere Mundung des Porus umgibt sich bald mit einer 
urspriinglich nur leicht angedeuteten Membranverdickung.” I do 
not quite understand whether this refers to the above-mentioned 
lip-like thickenings or whether it is meant to indicate a thickening, 
extending right across the pore. Further on we find the state¬ 
ment : “ Die auffallendsten Prominenzen gehoren der Membran 
iiberhaupt nicht an, sondern vverden durch gewisse etwas starker 
lichtbrechende Korner gebildet.” This evidently refers to the 
granules, I have described (cf. Brand’s figs. 4, 6, 10 and 11), but 
which, as far as I understand, according to Brand only appear later 
on in the fully-developed heterocyst. 
The contents of the heterocysts in later stages generally 
occupy only the central portion of the lumen, being separated by a 
space, more or less wide, from the lateral walls (cf. figs. 3, 9); in 
most cases, however, they are in connection with one (fig. 3) 
or both (fig. 1) of the end-walls in the terminal and intercalary 
heterocysts respectively. Detached heterocysts alone have the 
contents contracted together in the middle, not being in contact 
with the wall at any point (fig. 4). 
In the young filaments (in which there is no indication of 
spore-formation) the heterocysts occur singly and at considerable 
intervals from one another (fig. 1). The filaments are almost 
invariably bounded at one or more usually at each end by terminal 
heterocysts and intercalary ones occur in the course of the longer 
filaments. As the filament gets older it often happens, that the 
