Of the genera regarded as common, only two[47] or three are represented in the two regions by identical species, and these have a very wide distribution in the warm seas. Of the others, nearly all range to India, to the Cape of Good Hope, to Australia, or to Brazil. They may have ranged farther in the past; they may even range farther at present. Not one is confined to the two districts in question. As equally great resemblances exist between Japan and Australia or Japan and the West Indies, the case is not self-evident without fuller comparison. I shall therefore undertake a somewhat fuller analysis of the evidence bearing on this and similar problems with a view to the conclusions which may be legitimately drawn from the facts of fish distribution. =Differences between Japanese and Mediterranean Fish Faunas.=--We may first, after admitting the alleged resemblances and others, note that differences are equally marked. In each region are a certain number of genera which we may consider as autochthonous. These genera are represented by many species or by many individuals in the region of their supposed origin, but are more scantily developed elsewhere. Such genera in Mediterranean waters are _Crenilabrus_, _Labrus_, _Spicara_, _Pagellus_, _Mullus_, _Boops_, _Spondyliosoma_, _Oblata_. None of these occurs in Japan, nor have they any near relatives there. Japanese autochthonous types, as _Pseudoblennius_, _Vellitor_, _Duymæria_, _Anoplus_, _Histiopterus_, _Monocentrus_, _Oplegnathus_, _Plecoglossus_, range southward to the Indies or to Australia, but all of them are totally unknown to the Mediterranean. The multifarious genera of Gobies of Japan show very little resemblance to the Mediterranean fishes of this family, while blennies, labroids, scaroids, and scorpænoids are equally diverse in their forms and alliances. To the same extent that likeness in faunas is produced by continuity of means of dispersion is it true that unlikeness is due to breaks in continuity. Such a break in continuity of coast-line, in the present case, is the Isthmus of Suez, and the unlikeness in the faunas is about what we might conceive that such a barrier should produce. =Sources of Faunal Resemblances.=--There are two main sources of faunal resemblances: first, the absence of any barriers permitting the actual mingling of the species; second, the likeness of temperature and shore configuration on either side of an imperfect barrier. Absolute barriers do not exist and apparently never have existed in the sea. If the fish faunas of different regions have mingled in recent times, the fact would be shown by the presence of the same species in each region. If the union were of a remote date, the species would be changed, but the genera might remain identical.