290 
Kusano, Further studies on Aeginetio. indiccL 
log'ous cases may perhaps be found in tbe root-bairs that develop 
previous to the formation of haustoria on th.e typical root of some 
hemi- and holo-parasites, sucli as Melanipyrum (Ledere du Sablon, 
1887), Lathraea (Heinricher, 1896, p. 381), Santalum (Barber, 
1906). In all these cases the root-hairs appeax to serve sin^ly 
for the fixation of the root of the parasite to the host. The 
cushion-cells in Cuscuta (Peirce, 1893) may be considered to 
perform the similar function. In Aeginetia it is quite obviop, 
already stated, that the hairs serve first of all as a „tentacle’, and 
after contact with the host, as a „prehensile organ ‘, besides drawing 
the seedling closer to the host. In function, therefore, they possess 
all the charaeters of a typical tendril (i. e., Omurbitmem), and 
hence I venture to propose for them the name of „hair-tendrils . 
In the root-system a similar function has already been known 
to appertain to the so-called root-tendrils (see Pfeffer, 1 , 
p 416). Th6y ar6 not, bowGVor, idontical morpbologically witn 
the haii-tendrils; for in typical root-tendrils the entire root plays 
a part of a tendril, while in hair-tendrils an appendage ot t e 
radicle comes into play. In origin, again, the hair-tendrils may 
be homologous to the papilla-like cells at the tip of the radicle 
in the seedling of Orohanäie (Koch, 1883, p. 189). However, m 
strnetnre and function tbe latter organ seems to be different irom 
tbe former sbowing a ratber dose resemblance to tbe cusbion-cells 
of Cuscuta. . . , 
Tbe kind of Stimuli required in causing tbe curvature ot tbe 
tendrils remains still unknown. But on tbe basis of my culture- 
experiments it seems bigbly probable tbat, unlike tbe true root- 
bairs (see Pfeffer, 1904, p. 459), mere contact witb sand- or 
soil-particles remains quite ineffectual, but tbat some cbemica 
Stimulus must be concerned, to wbicb tbe tip of tbe tendrils coming 
in contact witb tbe bost-root must respond. Tbat normal tendrils 
may respond to cbemical Stimuli bas already been ascertained by 
Correns (1896, p. 16). ^ ^ ^ i 
In almost all cases tbe globular cells do not appear to develop 
all into tbe bair-tendrils: some of tbem remain unebanged, wbue 
some are arrested from furtber development after reacbing tbe 
conical or papillae stage. As for tbe most probable ground of 
such variable development of tbe globular cells, my observations. 
of a number of seedlings bave led me to tbe conclusion tbat tbe 
number of tendrils tbat are formed in a seedling must depend moie 
or less upon tbe cbances of meeting witb an appropriate bost. 
In fact I bave found tbat wben a seedling came on contact witn 
a bost by a premature development of some tendrils tbe remainmg 
ones were more or less arcested from ^Üie^yetopment 
globular cells from forming further tendrils (Figs. 9,10,12), while 
when a seedling remained away from the host long enough ma^ 
tendrils were observed to develop at once and in fuU length, oi 
many globular ceUs to give rise to tendrils (Fig. 7). This fact 
makes it most probable that the seedling develops as many tendiils 
as possible in several directions until it finds out a host thus 
