It is worth noting, too, that these tadpoles, like all other very low vertebrates, are mud-haunters; and the most primitive among adult vertebrates are still cartilaginous mud-fish. Not much is known geologically about the predecessors of frogs; the tailless amphibians are late arrivals upon earth, and it may seem curious, therefore, that they should recall in so many ways the earliest ancestral type. The reason doubtless is because they are so much given to larval development. Some ancestors of theirs--primæval newts or salamanders--must have gone on for countless centuries improving themselves in their adult shape from age to age, yet bringing all their young into the world from the egg, as mere mud-fish still, in much the same state as their unimproved forefathers had done millions of æons before. Similarly, caterpillars are still all but exact patterns of the primæval insect, while butterflies are totally different and far higher creatures. Thus, in spite of adult degeneracy in the ascidian and adult progress in the frog, both tadpoles preserve for us very nearly the original form of their earliest backboned ancestor. Each individual recapitulates in its own person the whole history of evolution in its race. This is a very lucky thing for biology; since without these recapitulatory phases we could never have traced the true lines of descent in many cases. It would be a real misfortune for science if every frog had been born a typical amphibian, as some tree-toads actually are, and if every insect had emerged a fully formed adult, as some aphides very nearly do. Larvæ and embryos show us the original types of each race; adults show us the total amount of change produced by progressive or retrogressive development. XI. _AMONG THE HEATHER._