Corda, 1842). A large number of other species are known in other lands, and these have been recently described and figured by Mr. G. Massee in the “Annals of Botany” (vol. ix., 1895), wherein will be found an epitome of all that is known about the matter

In New Zealand we have a second species attacking the larva of the singing locust (Cicada). This species, known as Cordyceps sinclairii, was originally named by Berkeley in his “Flora of New Zealand” (1855): it is figured by him in his “Introduction to Cryptogamic Botany” (p. 73, fig. 17), and also by Taylor in his “New Zealand and its Inhabitants” (p. 647). Both these figures are rather crude; but, except for cursory references to it by Mr. Maskell and Mr. A. Hamilton in their papers on the vegetable caterpillar in the Transactions, I can find no further account of this species.

Massee, in his monograph above referred to, merely quotes Berkeley's description, who gave as its habitat, “Northern Island (New Zealand); in the larva of some orthopterous insect; amongst loose gravelly soil.” Hooker records it from a coleopterous larva.

Last month (August) I received two well-preserved specimens of this vegetable Cicada from the lightkeeper at Farewell Spit, who had been good enough to comply with a request of mine to preserve, in a bottle of formol that I had forwarded to him, any animals of interest that he met with. My thanks are due to him for these specimens, as well as for some very fine large Scalpellum, which appear to be new to science.

The two specimens of C. sinclairii are at different stages of development, and are so different in appearance from C. hugelii (= robertsii) that I looked into the literature of the subject, and, as a result, proceeded to investigate the matter.

The fungus, in this stage, makes its way out of the insect near the anterior end, as in all normal cases of Cordyceps hitherto recorded. It issues between the head and pronotum. The main branch grows straight forward for some distance and gives off branches right and left in a very characteristic fashion. At first the fungus is cylindrical, white in colour, with a certain amount of pink in places.^* In one specimen the-branches are few and short (figs. 1, 2); in the other (fig. 3) they are more numerous, and the whole aspect is much more elaborate, and resembles a deer's antlers. The branches lose their cylindrical shape and become irregular; some others broaden out, become more or less flattened, and are very

irregularly arranged. The pink colour is now more extensively developed and shades into a brown, the white being limited to the axils of the branches. These changes are connected with the maturity of the fungus. In addition to the main post-cephalic bundle of hyphæ, other smaller ones, in the older specimens, issue from the joints of the legs and from the sides of some of the segments.

A still further stage in maturity was represented by a dried specimen given to me by Mr. Hamilton. Here several long main stems issue behind the head, and one, behind the pronotum, and run side by side directly forward. These are less branched than in my specimen.

A comparison of the mature condition of C. sinclairii with that of C. hugelii shows very remarkable and easily recognisable differences (quite apart from the absence of branching in the latter), illustrating two distinct methods of spore-formation.

If an aerial fructifying branch of C. hugelii be examined even with the naked eye, it is seen to terminate in a long, velvety, thickened region, sharply marked off from the more woody base or stalk. This upper region owes its velvety appearance to a vast number of closely set yellowish vesicles (or perithecia), each of which, on further examination, may be found to contain eight long filamentous spores packed side by side. But the closest and most careful superficial examination of C. sinclairii shows nothing of the kind; there are no perithecia, the surface of the brown-pink region is quite smooth. If, however, this fungus be submitted to microscopical examination, either by the simple method of teasing up with needles, or, better, by cutting sections across it with a razor, it will be found that below this smooth surface, formed of closely arranged fungal hyphæ, innumerable spores occur, but of quite a different size and shape and arrangement. These spores are short and oval, quite irregularly arranged amongst the hyphæ. Further, they are produced in a manner quite different from that exhibited by C. hugelii. In fact, these two species of New Zealand Cordyceps illustrate the two chief modes of reproduction by spore-formation recognised by botanists as occurring in the group of fungi, and which generally occur in two different stages in the life-history of one and the same species of fungus.

The two different modes of spore-formation yield different kinds of spores—(1) Ascospores, formed by subdivision of the protoplasm inside a single cell or “ascus”; (2) Conidiospores, formed by constriction of a hypha, so as to form a row of spores arranged more or less like a string of pearls or beads.

Without entering into further details, it must suffice to state that the fructification of most of the species of Cordyceps with which we are acquainted is known only in the

ascospore stage (as in the vegetable caterpillar of New Zealand, C. hugelii), while nothing is known of the other stage—the conidiospore stage—of the life-history; but in a few European species, such as C. militaris and C. entomorhiza, both stages in the life-history of the fungus are known, and the relation of one to the other has been worked out. From experimental researches on these species we know that the conidial stage is a saprophytic form, which may grow on leaves, bark, wood, &c., and not necessarily on insects, and has been hitherto regarded as a distinct genus of fungus, to which the name Isaria had been given, while the ascospore stage is, of course, that condition known as infesting the insect. Thus C. militaris is the ascospore stage of a fungus of which the conidial stage has been known as Isaria farinosa; and C. entomorhiza,^* infesting the larvæ of various insects, has as its conidial stage I. densa, which normally occurs on the larva of the cockchafer (Melolontha). These two stages do not occur at the same period of the year, but it appears that a resting-stage, known technically as “sclerotium,” intervenes.

From these and other facts botanists consider it likely that each of the species of Cordyceps has its own “isarial” stage, though this can only be ascertained by experiments similar to those that have enabled the facts above mentioned to be determined.

Now, the chief interest in my own observations on C. sinclairii lies in the fact that it is the conidial stage of a fungus—a fact already noted, I may state, by Berkeley. We do not know its ascospore stage, and one is inclined to suggest that possibly the two fungi known as C. sinclairii and C. hugelii may be merely links in the chain of events in the life-history of one species of Cordyceps. But this can only be determined by direct experiments with living organisms—by sowing the spores of C. sinclairii (from the Cicada) in the tissues of the caterpillar, and ascertaining whether they give rise to C. hugelii. It is manifestly a difficult experiment to perform, and a considerable number of living fungi and insects would be required; and since we do not know with certainty what species of caterpillar (and whether more than one) is attacked by C. hugelii, it will be necessary to make a further investigation into the whole subject.^†