All these effects should be referable to the same group of causes. They may, in fact, be combined in one statement. All other fishes now extant, as well as all fishes existing prior to Cretaceous times, have a larger number of vertebræ than the marine shore fishes of the tropics of the present period. There is good reason to believe that in most groups of spiny-rayed fishes, those with the smaller number of segments are at once the most highly organized and the most primitive. This is true among the blennies, the sculpins, the flounders, the perches, and probably the labroid fishes as well. The present writer once held the contrary view, that the forms with the higher numbers were primitive, but the evidence both from comparative anatomy and from palæontology seems to indicate that among spiny-rayed fishes the forms most ancient, most generalized, and most synthetic are those with about 24 vertebræ. The soft-rayed fishes without exception show larger numbers, and these are still more primitive. This apparent contradiction is perhaps explained by Dr. Boulenger's suggestion that the prevalence of the same number, 24, in the vertebræ of various families of spiny-rayed fishes is due to common descent, probably from Cretaceous berycoids having this number. In this theory, perches, sparoids, carangoids, chætodonts, labroids, parrot-fishes, gobies, flounders, and sculpins must be regarded as having a common origin from which all have diverged since Jurassic times. This view is not at all unlikely and is not inconsistent with the facts of palæontology. If this be the case, the members of these and related families which have larger numbers of vertebræ must have diverged from the primitive stock. The change has been one of degeneration, the individual vertebræ being reduced in size and complexity, with a vegetative increase in their number. At the same time, the body having the greater number of segments is the more flexible though the segments themselves are less specialized. The primitive forms live chiefly along tropical shores, while forms with increased numbers of vertebræ are found in all other localities. This fact must be considered in any hypothesis as to the causes producing such changes. If the development of large numbers be a phase of degeneration the causes of such degeneration must be sought in the colder seas, in the rivers, and in the oceanic abysses. What have these waters in common that the coral reefs, the lava crags, and tide-pools of the tropics have not?