AND ELATERS OF MARCHANTIA POLYMORPHA. 107 
nature of these bodies being readily determined byiodine. The starch-granules frequently 
lie within the young tube, in such a manner that they may easily be mistaken for the 
rudiments of the spiral fibres, but they are quite distinct from these, and disappear before 
the fibres begin to be deposited. I believe that the accounts given by some authors of 
the formation of spiral fibres in spiral vessels from rows of minute granules are incorrect, 
and have arisen from observation of starch-granules lying in rows, often running obliquely 
across the tubes. As the tubes grow they enlarge more in length than in diameter, and 
appear as very long, slender filaments; the starch-granules, and finally the protoplasm 
disappear, and faint streaks, denoting the nascent fibres, are at length to be perceived 
upon the walls. These gradually become more and more distinct, until in the mature 
elaters they present themselves as strong, flattened bands. In Marchantia there are two 
fibres, and the ends of these are confluent at the extremities of the tubes in which they 
_ are contained. More properly speaking, therefore, the fibre is one endless fibre twisted 
upon itself; the best possible condition of structure for the purpose. We may represent 
the condition of the fibres by a piece of string doubled, and with its ends tied together ; 
this, when twisted up, unrolls immediately one end is set at liberty ; or, if both ends are 
let loose at once, the whole piece springs away as it unrolls, just as the elaters of Marchantia 
spring out when the capsules burst. In unrolling the fibre it tears up the membrane of 
the wall of the tube, and when the elaters are examined after they have been discharged, 
the fibres are found somewhat unrolled, and the torn membrane is often no longer to be 
detected. 
While the elaters are passing through these stages, the larger elongated cells exhibit a 
very remarkable series of changes, which appear to differ from everything that has yet 
been observed in analogous structures. They are at first filled, like the elaters, with a 
delicate, colourless protoplasm, in which float exceedingly minute granules ; this substance 
is coagulated even by water, and still more strongly by alcohol, acids, and iodine. It is 
apparently the same substance that occurs in all young cells which increase by self-division. 
I have found it unmixed, as here, in young hairs, in the parent-cells of pollen before the 
formation of the septa, in the confervoid body which grows out from the embryonal vesicle 
of the Orchidacee, in the Yeast Fungus, &c. In most cells it very soon becomes mixed 
with starch and chlorophyll vesicles. | 
“The elongated cells soon exhibit transverse streaks of a lighter colour, from the proto- 
plasm separating into a number of portions, and cross membranes are produced at these 
places, dividing the tubular cells into a row of cells, all of square form, except the two end 
ones, which are attenuated toward the free point, and thus appear triangular in the side 
view. I could not make out whether the septa were formed by gradual growing in of the 
membrane; if so, the process must go on very quickly. Neither could I detect a double 
membrane; but this must exist, as the cells afterwards separate from each other at these 
= points. Vertical septa often occur, producing a double row of cells from the original tube. 
M. Mirbel appears to have made his earliest observations subsequently to the breaking up 
of these rows of cells, and thus to have missed them. They are a constant phenomenon, 
and I know of no analogous structure, unless we compare them with the single rows of 
cells which first appear in the tissue of the anthers, and by subdivision become the parent- 
