14 Tannin in thc Fruit of tJie Persimmon [Sept. 



average condition of complexity; many are far too intricate for 

 graphic representation. 



Attention may be drawn to a few, more important details. The 

 canals are iisually tubulär ; this is seen to be the case in their optical 

 section. They are most frequently curved, sometimes sharply 

 so, and are usually expanded into cavities which are spherical, 

 oblate or prolate with reference to the canal as axis. Sometimes 

 they are hemispherical or hemispheroidal as the case may be (fig. 

 4a). In number, these cavities may be few or so many along a 

 Single canal as to appear like a string of beads. A considerable 

 number of canals may be, and normally are, confluent at some 

 central point, to form a cavity with flattened arms extending in 

 various directions. At its outer extremity a canal may end by 

 opening into another, or out upon the surface of the tannin-mass 

 (fig. 3). Here the mouth may be a minute pore no greater in 

 transverse measurements than the canal itself, or it may be funnel- 

 shaped, and either circular or a slit (fig. 2). It may open into a 

 depression or immediately at the surface. The slit-shaped mouth, if 

 it happens to be of that form, may extend so far around the tannin- 

 mass as to appear as a sulcus (fig. 2), and this may be so deep as to 

 cut half-way through the mass. Still again, just before reaching 

 the surface, a canal may divaricate several times, and thus form a 

 group of pores (fig. 3). Finally, a canal may become intricately 

 branched in the interior of the mass so as to resemble the galleries 

 of wood boring beetles, or so brauch and intertwine as to produce 

 an intricate web. These last named appearances have been seen 

 only in definitively matured tannin-cells. 



It requires only a slight realization of the condition which I 

 have tried to describe briefly, to render doubtful the explanation 

 which first occurred to me, namely, that these canals are caused by 

 the bridging Strands of protoplasm. The best explanation, I be- 

 lieve, is this, that the tannin-mass is composed of several, or a con- 

 siderable number of, originally separate gelatinous bodies (prob- 

 ably contained in separate vacuoles), which have extensively but not 

 entirely coalesced. The existence of separate tannin-masses, as 

 occasionally observed, in separate vacuoles (distinguishable on con- 

 traction by glycerol) may be adduced in support of this view (fig. 5). 



