It may be that the original home of Gleichenia was in the far North at a time when climatic conditions were very different from those which now prevail. Gleichenia, like many other northern plants, retreated to more southern regions where, in the warmer countries of the world, many species still flourish widely separated in space and time from the place of their birth. The ferns so far mentioned have a more or less extended distribution at the present day. In the case of _Pteridium aquilinum_, the cosmopolitan Bracken Fern, wide range would seem to be correlated with comparatively recent origin; on the other hand, the facts of palaeobotany show that the wide distribution of Osmunda, a type of fern which differs in many important respects from members of the family (Polypodiaceae) to which the Bracken belongs, is not inconsistent with an exceptionally ancient family-history. There are, however, certain genera of ferns which afford remarkable examples of restricted geographical distribution associated with great antiquity. The island of Juan Fernandez, 420 miles off the coast of Chili, the home for four years of Alexander Selkirk (to whose adventures we owe Defoe's creation of Robinson Crusoe), is interesting also from a botanical point of view. The vegetation of this oceanic island, 20 square miles in area with basaltic cliffs rising to a height of 3000 ft. above the sea, includes more than 40 species of ferns, eight of which occur nowhere else. One of these endemic ferns is _Thyrsopteris elegans_, the only representative of the genus; it is readily distinguished by its large and graceful fertile fronds, examples of which may occasionally be seen on a plant of this species in the Royal Gardens at Kew: the sporangia are produced in circular cups which replace the ordinary leaflets on the lower branches of the frond and hang from the short axis like miniature clusters of grapes. It is noteworthy that among the fragmentary remains of the fern vegetation of the Jurassic flora in England and in other parts of Europe specimens occur with fertile segments practically identical with those of the Juan Fernandez species. Students of fossil plants are occasionally led away by the temptation to identify imperfect specimens with rare existing species to which they exhibit a superficial resemblance, and this is well illustrated by the frequent use of the generic name Thyrsopteris for Jurassic and Lower Cretaceous ferns which are too imperfect to be determined with any degree of certainty We have, however, satisfactory grounds for the assertion that the Juan Fernandez fern affords a striking confirmation of the truth of Darwin's dictum that 'Rarity, as geology tells us, is the precursor to extinction.' In this remote oceanic island, for reasons which we cannot explain, there lingers an isolated type which belongs to another age.