There seems still a good deal of vagueness about the matter, and Corvo might well be given a thorough overhauling for vestiges of ancient times. This naturally should be extended to the submerged area close to the shore, for the outlying reefs and ridges may mark the site of lower lands where human work once went on and where its traces and relics may remain. In expanse the island probably was not always what we find it now, six miles in length by at most three in breadth (seven square miles in all, as most accounts compute it) with fringes of rock running off from the shore, “lifting themselves high above the water in one place, blackening the surface in another, and again sinking to such a depth that the waves only eddy and bubble over them.” Mr. Henriques says elsewhere: “In many of the islands, but especially in Flores, there are vestiges clearly indicating that formerly as well as lately parts of the island have sunk or rather disappeared in the sea.” He cites for instance a notable loss of land in the summer of 1847. There is reason to believe that Corvo has dwindled in this way much more, proportionately, than Flores. One striking indication is found in the comparison of the present map with those of the fourteenth and fifteenth centuries. For convenience sketches of these are appended (Fig. 23). The relative position of the islands is about the same in all. The form of Corvo varies from the pear shape of the Laurenziano map (1351),[283] and another shape[284] not much later slightly resembling an indented segment of a circle, to the three-lobed or clover-leaf form which was accepted as the final convention or standard and first clearly appears in the great Catalan atlas[285] of 1375, repeated by Beccario 1435[286], Benincasa 1482[287], and others; but all agree in making Corvo the main island and Li Conigi (Flores) a minor pendant. Corvo seems in every way to have commanded chief attention, and in size the difference was conspicuous and decisive. The difference certainly is great enough now, but conditions and proportions are reversed. Corvo has but one-eighth the area of Flores and less than one-tenth the population. In all ways it lacks advantages and conveniences, taking rather the place of a poor dependent. [Illustration: FIG. 23--Representation of Corvo on fourteenth- and fifteenth-century maps as compared with its present outline. (The sources may be identified from the text.)]