[KrrscH] CERTAIN STRUCTURES IN THE PTERIDOPHYTES 399 
cases large nuclei were present in both cells, proving that thylose forma- 
tion is only a phase of ordinary cell division induced under certain 
conditions. 
The thyloses filled up nearly the whole space left vacant by the 
protoxylem in the process of canal formation, but in most cases a con- 
siderable portion of the original cavity remained, some of the protoxylem 
elements floating around in this. 
The canals described above are present in the Equisetaceæ in the 
form of the well-known carinal canals, which are formed in precisely 
the same manner as in Pteris, but are generally more regular in outline 
Strasburger in his description of the anatomy of Equisetum Telmateja, 
speaks of them as follows (19, 431): “ Each vascular bundle has on its 
inner edge a carinal canal, which has the same origin as the frequently 
occurring, schizogenous, water-conducting, intercellular passage in the 
vascular tissue of the Monocotyledons. Into these cavities project rings 
of the narrow, primary xylem vessels which become disorganised on 
separation.” 
Thyloses are not found in these normally, but they can be induced 
to appear under certain experimental conditions. Strasburger gives the 
following account of their occurrence in Equisetum Telmateja (19, 437). 
“In decapitated stems (Shoots) set up in water and protected from 
evaporation, drops of water squeezed out of the carinal canals after a 
short interval. After a period of twelve hours the bounding wood paren- 
chyma cells put forth prolongations of a globular nature into the carinal 
canals, and there consequently ensued in these a formation of thyloses 
very similar to that customary in the lumen of vessels. Regardless of 
the fact that the wood parenchyma cells bordering on the canals could 
have curved out on their whole surfaces, still they mostly put forth only 
localised, bladder-like prolongations, several of these in a row frequently 
jutting out into the cavity.” 
The mode of thylose formation agrees essentially with that observed 
in Pteris aquilina, except in the occurrence of localised projections, which 
phenomenon was not noted in that form. That the occurrence of thy- 
loses under such conditions is what would naturally be expected, will 
be demonstrated below. 
The canals are also encountered in the Lycopodiacee. According 
to De Bary (3, 338) one such intercellular canal is present between the 
xylem and phloem of most species of Isoetes, while in Isoetes Engelmanni 
there are usually three. He states that their origin is not clear. 
Russow (15, 140) has the following in this connection :—“ In the 
vascular bundle of Isoetes lacustris there occurs one lacuna (canal) 
