8 9 
Studies on Cyanophyceae. 
two vegetative cells on each side of an intercalary heterocyst 
become transformed into terminal heterocysts, whilst the later 
loses its connection with the rest of the thread. In this way the 
original filament has become split up into two smaller ones, 
bounded at the newly-formed extremities by the terminal hetero¬ 
cysts ; such filaments are frequently found, lying close together 
with a loose intercalary heterocyst between them. But new 
heterocysts may arise in the course of a thread at points, where 
there are no previous heterocysts. Their formation then takes 
place in the following way: An intercellular substance of a dark 
green colour is excreted between two adjoining cells, which 
gradually develop into terminal heterocysts. This intercellular 
substance has a flat bi-concave shape (fig. 8a), narrow in the middle 
and broadening out laterally; when the heterccysts are fully 
differentiated this substance has become quite colourless (fig. 8b). 
Occasionally (apparently when the excretion is formed near one end 
of the fiament) a heterocyst is developed only on the distal side, 
the short proximal portion of the filament being left without a 
terminal heterocyst. Brand (loc. cit. pp. 39, 50-51) was able to 
study the formation of such intercellular substance in connection 
with the formation of pseudo-branches and heterocysts in Tolypothrix 
penicillata, Thur. var. tenuis, Hansg. 1 and also mentions its occur¬ 
rence in connection with the development of hormogonia. In 
Nostoc commune (p. 52) finally Brand has observed such excretions 
appearing in connection with the disjointing of the heterocysts in 
old cultures. 
In older filaments of the Anabaena under discussion heterocyst- 
formation is much more abundant. 2 I have frequently observed as 
many as four terminal heterocysts side by side at one extremity of 
the filament and in close apposition to one another (figs. 4 and 9). In 
such cases the distal heterocyst is often most fully differentiated, is 
largest and has the thickest wall (cf. fig. 4); the following ones are 
generally smaller and show the typical characters to a less extent. 
It appears as though the functions of the first-formed heterocyst 
1 I can full}’ confirm Dr. Brand’s observations on this point. 
2 As far as I am aware, it is not a common tiling to find many 
lieterocysts side by side in the Nostocbineae. Bornet and 
Tliuret (’8o, p. 104 and Plate xxviii., Pig. 13) have described 
such a phenomenon in Nostoc ciniflomim, Tourn.; they also 
in their figure show the way in which two terminal 
lieterocj'sts appear on either side of an intercalary hetero¬ 
cyst and thus give rise to two new filaments. Palla 
(’93, p. 546) states that the heterocysts are “auffallend 
zahlreicli ” in Anabaena Azolla. 
