Studies on Cyanophyceae. 91 
in Eosin had practically no effect at all. A longer immersion of 
about a day, however, resulted in the granules inside the vegetative 
cells and spores taking on a deep red colour, whilst the general 
substance of the cell was far less deeply stained. 1 Even two or 
three days immersion produced no very visible effect on the older 
heterocysts, but in the young ones the terminal granule was stained 
just as deeply as those occurring in the ordinary cells. The 
collapsed cells were only very slightly affected by the stain. In a 
good many cases a few of the'granules inside some of the hetero¬ 
cysts had also apparently taken on a very faint colour; and in some 
cases a deep red granule was discernible within the cell-contents. 
The heterocysts of the blue-green Algae were at one time 
regarded* as subserving vegetative reproduction, and, as has been 
shown above, their formation is at times undoubtedly connected 
with the breaking up of a filament. Other authors have interpreted 
them as receptacles for the storage of reserve-substances, 3 whilst 
recently (1901, 1903), Brand has described the liberation of their 
contents as gonidia and their germination to form a new filament 
iii Nostoc commune and N. microscopicum ; l this latter author also 
ascribes a storage-function to them and considers that their 
contents may serve to replenish the exhausted adjoining cells. 
When the very marked differentiation, attained by the heterocyst, 
is taken into consideration, one can scarcely agree with Borzi 5 that 
“the heterocysts serve solely to interrupt and limit (the threads) at 
varying distances.” In the Anabaena under discussion the threads 
are almost invariably limited by a heterocyst, so that these cells 
would appear to be a necessary termination of the filament. When 
the filaments break up into two parts, heterocysts must be formed 
to limit the free ends of the new filaments. Similarly when 
branching takes place in a Tolypothrix, a heterocyst is formed just 
above the point, at which the continuity of the filament is inter¬ 
rupted by the formation of the branch (cf. also Brand, loc. cit. 
1 This was especially evident in cells which had almost homo¬ 
geneous contents with few or no granules. 
2 cf. Borzi ’7S. p. 239, and Hausgirg ’87, p. 124; and recently 
Kirchner ’98, p. 71. 
3 cf. Hieronymus, ’92, p. 483; Ilegler, 1901, p. 305. 
4 Brand, 1901, pp. 154, 155. Although lieterocysts were so very 
abundant in my matexhal I have not been able to confirm 
Brand’s observations with any certainty. Several cases 
were however found, in which it looked as though germin¬ 
ation of the contents of a heterocyst had taken place, but 
the stages were always too far advanced to recognise the 
typical structure of a heterocj'st. 
‘ Borzi ’78, p. 239. 
