tropical Pacific. The species specially referred to in this paper are Dendrobium Cunninghamii Lindl., Earina mucronata Lindl., and Earina suaveolens Lindl.

If transverse sections of the roots of Earina or Dendrobium be examined with the microscope it will generally be found that some of the cortical cells contain each a yellowish-brown mass. Sometimes these masses are few in number and small in size, but often they are present in abundance and are conspicuous objects in the section. They are most plentiful in the outer layers of cortical cells, are seldom found in close proximity to the stele. Close examination of these masses shows that they vary greatly in size, in shape, and in the materials composing them. In colour and denseness, too, they show variation, so that no two are alike, even in the same cross-section. It will be of assistance in description to refer all these masses to one or another of three chief classes, into which they may be divided, the first class containing masses which consist almost wholly of a more or less dense coil of well-defined fungal hyphae; the second comprising those which are dense and yellowish in colour, and in which no traces of hyphae are observable; and a third class consisting of masses which are intermediate in structure between those of the first two classes. The bodies belonging to the last class are usually composed partly of numerous granular bodies and partly of what are evidently degenerating fungal hyphae.

In the bodies of the first type the hyphae are thin-walled, almost colourless, often swollen slightly in places, and are usually colied up in a more or less dense mass. They are septate, the division walls being sometimes very numerous, and they often branch freely. A vacuolated appearance is nearly always observable, and sometimes the hyphae are seen to have collapsed in places.

The masses of the second class, although varying greatly in shape, generally have their edges well defined. They are usually of a yellowish colour, and the substance of the mass appears homogeneous, no signs of fungal hyphae being observable. Each of these masses is usually connected with the walls of the cell in which it occurs by one or more narrow strands, which often appear to consist of protoplasm, but which are sometimes fungal hyphae in a state of degeneracy. Sometimes the strands belonging to the masses in adjacent cells appear to be continuous through the dividing cell-walls, so that the strands would seem to connect the yellowish bodies with one another.

The masses of the third class contain numerous degenerating hyphae, the outlines of which are often difficult to distinguish, and mingled with these are often large numbers of tiny granular bodies. Sometimes small starch-grains are present among these, but they are never abundant. Often the central portion of these masses is of a yellowish colour, and more dense than the outer portions, which consist largely of degenerate hyphae; in fact, if it were not for the presence of this outer hyphal layer the mass could be correctly referred to class 2. It should have been stated at the outset that it is not possible absolutely to delimit these classes from one another, as they are connected by masses of intermediate structure.

In the cortical cells containing the bodies just described the nucleus is frequently much swollen, and it is generally observed to be in close contact with the mass in the cell to which it belongs—sometimes, indeed, it appears to be partly included in the mass. Usually bodies belonging to all three of the above types will be seen in any one section, but one class of mass

invariably predominates. In most cases the bodies coming under the second head will be found to be most numerous.

An examination of the velamen tissue of these orchid-roots shows that it almost invariably contains many colourless or brownish hyphae, which branch freely and wander through its cells, forming a loose network of mycelial threads. The hyphae are septate, and it would sometimes appear from their great variation in thickness that more than one species of fungus was present. This is undoubtedly often the case. At maturity the velamen is a dead tissue, and no coils of hyphae or yellowish masses such are seen in the cortical cells are present in it. Where the hyphae are present in abundance in the velamen that tissue is sometimes observed to be in a state of decay; indeed, in the older portions of some roots it has entirely disappeared, and the exodermis is then the outermost layer.

The passage-cells of the exodermis are often occupied by a scanty coil of hyphae formed by threads which have entered from the velamen. Sometimes a mass of granular bodies occupies each of these cells, which, unlike those of the velamen, contain living contents. Where extensive thickening has taken place in the exodermal cells it is not always easy to demonstrate the connection existing between the hyphae of the transfusion-cells and those of the velamen, owing to the presence of a felty mass which often guards these cells on their outer sides. The transfusion-cells of the exodermis are the ones which admit the fungal hyphae to the cortical tissues. Having entered one of these pasage-cells, a hypha usually forms a loose coil, and then extends out into the cortex, where it may branch, and each branch is generally seen to end in one of the yellowish masses above described. Sometimes after forming a coil in one cell a hypha enters an adjacent cell, where it forms another coil, and it may do this in several cells in succession. The hyphae which are found in the cortex are seldom normal, and they generally present an unhealthy and impoverished appearance. This ill-nourished condition is indicated by the presence of vacuoles, and often by the collapse of the hyphal walls. The hyphae of the velamen never have this unhealthy appearance.

Where the roots of these orchids come into contact with humus, as they usually do in the crevices of the dark of the supporting tree, or in the black material formed by the decay of lichens, mosses, &c., it is noticed that numbers of hairs, much resembling the root-hairs of ordinary terrestrial plants, are frequently developed. Sometimes among the other normal hairs some will be seen which have become flattened, and then twisted in a regular spiral fashion. The cavity of each hair usually contains one or more fungal hyphae, which are continuous with those in the velamen.

The foregoing description applies in all essential respects to both species of Earina and Dendrobium Cunninghamii, in which the velamen is usually a well-marked tissue, and the roots of which are structurally very similar. In the New Zealand species of Bulbophyllum and in Sarcohilus adversus the velamen is poorly deveoped, being seldom more than two cells seep. In Sarcochilus the hyphae of the velamen are very scanty, but the yellowish masses in the cortical cells are often large and abundant.

It is by no means easy to discover the exact physiological significance of sturctures such as those briefly described, although such problems are of supreme interest. It is well known that the roots of many plants growing in soils rich in humus often live in intimate connections are believed to be beneficial to both organisms, and to such an alliance between the roots of one of the

higher plants and a fungal mycelium the name “mycorhiza” is given. Mycorhizae are of two kinds—those in which the fungal elements do not penetrate the root, but merely form a dense mat round it, and those in which the hyphae enter the root and form coils within its cells. There can be no doubt that the relationship existing between the roots of our epiphytic orchids and the fungal filaments comes under the head of mycorhzic associations. It is evident, moreover, that the mycorhizae in their case are endotrophic. Orchids seem to be particularly prone to the formation of such alliances, and it is therefore not surprising to find that some New Zealand orchids posess mycorhizae. Among mycorhizae hitherto described is that of a British saprophytic orchid, Corallorhiza innata, which was dealt with by Messrs. Hanna and Jennings in a paper published in the “Proceedings of the Royal Dublin Society” (1898).

There is little fixity of opinion as to the exact nature of the benefits accruing to the partners in these mycorhizic associations. The view of Hanna and Jennings with regard to Corallorhiza is that the hairs found on the rhizime of the plant are produced with the object of attracting the fungal hyphae into the rhizome, so that the orchild can use them to augment its supply of food-materials. According to these authorities, the coils of hyphae on the orcdhid's cells are gradually absorbed as food by the protoplasm, starch-grains appearing in abundance as the hyphae disappear. A view held by Groom and Janse with regard to endotrophic mycorhize in general is that the fungus is digested by the root, therebhy supplying combined nitrogen. Other investigators—e.g., Hiltner and magnus—hold similar views.

In the case of the New Zealand orchids it is not probable that the hairs on their roots are produced as fungus-trapping organs. In some cases it is certain from the manner in which the hyphae branch that they are leaving and not entering, the root. In many instances spores have been seen in the hairs, and in more that one case the cavity of a hair at its extremity was observed to be densely packed with spores. What seems more probable is that the hairs perform, for a time at least, the functions which root-hairs perform in ordinary terrestrial plants—viz., absorption of water and mineral food. As above mentioned, the velamen at maturity is a dead tissue: its cells after having thickend their walls in a curiously intricate fashion lose their living contents, so that the hairs springing from them must also die, and presumably become functionless. It seems probable that no special significance is to be attached to the presence of the hyphae in the hairs. That the fungi sometimes use them as a means of entering the root cannot be doubted, but it is very likely that the hyphae are capable of penetrating the velamen at almost any point, and that their chief mode of entrance is by direct penetration of the outer cells of that tissue.

It is well known that humus is invariably penetrated in all directions by the hyphae of many species of fungi, and these are believed to assist in breaking down the complex organic compounds of the humus, thereby bringing a large amount of plant-food into an available condition. The humus, from which these epiphytic orchids must obtain practically all their food-materials except carbon, is formed, as before stated, largely by the decay of lower plants, chiefly lichens and mosses. It is unlikely that the carbon of the humus is utilized by the orchids, as they are all well supplied with chlorophyll, and it would therefore seem that their supply of carbon from atmospheric sources would be adequate to their needs. It cannot be said that their supply of available nitrogen is by and means