446 Bower . — Studies in the Phytogeny of the Fi lie ales. 
The series of drawings, shown in Fig. 3, i-vi illustrate the general 
structure of the rhizome of Metaxya at the departure of a leaf-trace, and of 
its attendant abaxial bud. The series reads from the apex downwards, and 
the orientation is as in the creeping rhizome. Fig. 3, 1, shows the leaf-trace 
in the petiolar base, approaching the foliar gap, while several roots are there 
seen traversing the cortex. In II a junction has been made near to, 
but not exactly at the lower margin of the leaf-gap, while it is seen that 
the incurved margin of the leaf-trace does not itself form the junction. In 
III, however, the flanges which projected inwards at the point of first 
junction have smoothed down, and slightly lower the opposite margin of 
the meristele joins the upper margin of the foliar gap. The section shown 
in IV traversed the bud which, as has been mentioned above, is frequently 
present on the abaxial side of the leaf-base. Its vascular supply, which is 
from the first solenostelic, arises, as in Lophosoria , as a diverticulum of 
the foliar trace, and the consequence is that the pith of the main axis is 
continuous with that of the runner. This is shown also in V, which is 
added as supplying a fact which is unusual in the more primitive soleno- 
stelic Ferns, viz. the interruption of the solenostele by a perforation which 
is not a foliar gap. It is, in fact, similar in nature to those * perforations * 
which are found in the more advanced types. This will be returned to later, 
as a fact of theoretical interest. Finally, as seen in vi, the irregularities 
caused by the insertion of the leaf and the runner being past, the regular 
solenostelic structure is resumed, and is interrupted only by the origin of 
the vascular supply to the roots. This comes off usually on the lower side 
of the creeping rhizome. 
Passing to details, the meristele in the petiole is defined by a well- 
marked endodermis, followed by a pericycle of one or two layers. The 
4 divergents ’ correspond to the convexities of the crinkled trace, and number 
from twenty to thirty in a full-sized trace. The protoxylem is not always 
clearly defined, while the tracheides often form only a single row. In the 
solenostele of the axis, which is delimited as before, the phloem is not profuse 
in quantity. The xylem-ring may consist of only three layers of tracheides, 
but in large rhizomes it usually comprises eight or more. There is no 
clearly defined protoxylem in the axis, and parenchyma cells are scattered 
through its xylem. 
The sori of Metaxya have never been examined with the attention they 
deserve, considering how divergent are the vegetative characters of this 
Fern from those of the genus with which it has usually been placed. In 
point of fact the analysis of Sir William Hooker, published in 1838, still 
remains the best hitherto given. Examining the mature sorus from above, 
it will be found that the number of sporangia is far in excess of that in 
Lophosoria , or of any species of Gleichenia , while its form is that of a flattened 
oval mass, of rather large area (Fig. 4). Countings of the number of 
