This conclusion must hold so far as species are concerned, but the resemblance of the genera on the sides has a significance of its own. In 1880[59] Dr Günther expressed his views in still stronger language, claiming a still larger proportion of the fishes of tropical America to be identical on the two sides of the continent. He concluded that "with scarcely any exceptions the genera are identical, and of the species found on the Pacific side, nearly one-half have proved to be the same as those of the Atlantic. The explanation of this fact has been found in the existence of communications between the two oceans by channels and straits which must have been open till within a recent period. The isthmus of Central America was then partially submerged, and appeared as a chain of islands similar to that of the Antilles; but as the reef-building corals flourished chiefly north and east of these islands and were absent south and west of them, reef fishes were excluded from the Pacific shores when the communications were destroyed by the upheaval of land." =Conclusions of Evermann and Jenkins.=--This remark led to a further discussion of the subject on the part of Dr. B. W. Evermann and Dr. O. P. Jenkins. From their paper on the fishes of Guaymas[60] I make the following quotations: "The explorations since 1885 have resulted (1) in an addition of about 100 species to one or other of the two faunas; (2) in showing that at least two species that were regarded as identical on the two shores[61] are probably distinct; and (3) in the addition of but two species to those common to both coasts.[62] "All this reduces still further the percentage of common species. "Of the 110 species obtained by us, 24, or less than 21 per cent., appear to be common to both coasts. Of these 24 species, at least 16, from their wide distribution, would need no hypothesis of a former waterway through the Isthmus to account for their presence on both sides. They are species fully able to arrive at the Pacific shores of the Americas from the warm seas west. It thus appears that not more than eight species, less than 8 per cent. of our collection, all of which are marine species, require any such hypothesis to account for their occurrence on both coasts of America. This gives us, then, 1,307 species that should properly be taken into account when considering this question, not more than 72 of which, or 5.5 per cent., seem to be identical on the two coasts. This is very different from the figures given by Dr. Günther in his 'Study of Fishes.' "Now, if from these 72 species, admitted to be common to both coasts, we subtract the 16 species of wide distribution--so wide as to keep them from being a factor in this problem--we have left but 56 species common to the two coasts that bear very closely upon the waterway hypothesis. _This is less than 4.3 per cent. of the whole number._