The earliest traces of animal life are found deeper still, beneath something like eighteen to twenty-five miles of rock, while below this level are the strata in which dwelt the earliest living things, organisms so small and simple that no trace of their existence has been left, and we infer that they were there because any given group starts in a modest way with small and simple individuals.

At the bottom, then, of twenty miles of rocks the seeker for the progenitor of the great family of backboned animals finds the scant remains of fish-like animals that the cautious naturalist, who is much given to "hedging," terms, not vertebrates, but prevertebrates or the forerunners of backboned animals. The earliest of these consist of small bony plates, and traces of a cartilaginous backbone from the Lower Silurian of Colorado, believed to represent relatives of Chimæra and species related to those better-known forms Holoptychius and Osteolepis, which occur in higher strata. There are certainly indications of vertebrate life, but the remains are so imperfect that little more can be said regarding them, and this is also true of the small conical teeth which occur in the Lower Silurian of St. Petersburg, and are thought to be the teeth of some animal like the lamprey.

A little higher up in the rocks, though not in the scale of life, in the Lower Old Red Sandstone of England, are found more numerous and better preserved specimens of another little fish-like creature, rarely if ever exceeding two inches in length, and also related (probably) to the hag-fishes and lampreys that live to-day.

These early vertebrates are not only small, but they were cartilaginous, so that it was essential for their preservation that they should be buried in soft mud as soon as possible after death. Even if this took place they were later on submitted to the pressure of some miles of overlying rock until, in some cases, their remains have been pressed out thinner than a sheet of paper, and so thoroughly incorporated into the surrounding stone that it is no easy matter to trace their shadowy outlines. With such drawbacks as these to contend with, it can scarcely be wondered at that, while some naturalists believe these little creatures to be related to the lamprey, others consider that they belong to a perfectly distinct group of animals, and others still think it possible that they may be the larval or early stages of larger and better-developed forms.

Still higher up we come upon the abundant remains of numerous small fish-like animals, more or less completely clad in bony armor, indicating that they lived in troublous times when there was literally a fight for existence and only such as were well armed or well protected could hope to survive. A parallel case exists to-day in some of the rivers of South America, where the little cat-fishes would possibly be eaten out of existence but for the fact that they are covered—some of them very completely—with plate-armor that enables them to defy their enemies, or renders them such poor eating as not to be worth the taking. The arrangement of the plates or scales in the living Loricaria is very suggestive of the series of bony rings covering the body of the ancient Cephalaspis, only the latter, so far as we know, had no side-fins; but the creatures are in no wise related, and the similarity is in appearance only.

Fig. 4.—Cephalaspis and Loricaria, an Ancient and a Modern Armored Fish.

Pterichthys, the wing fish, was another small, quaint, armor-clad creature, whose fossilized remains were taken for those of a crab, and once described as belonging to a beetle. Certainly the buckler of this fish, which is the part most often preserved, with its jointed, bony arms, looks to the untrained eye far more like some strange crustacean than a fish, and even naturalists have pictured the animal as crawling over the bare sands by means of those same arms. These fishes and their allies were once the dominant type of life, and must have abounded in favored localities, for in places are great deposits of their protective shields jumbled together in a confused mass, and, save that they have hardened into stone, lying just as they were washed up on the ancient beach ages ago. How abundant they were may be gathered from the fact that it is believed their bodies helped consolidate portions of the strata of the English Old Red Sandstone. Says Mr. Hutchinson, speaking of the Caithness Flagstones, "They owe their peculiar tenacity and durability to the dead fishes that rotted in their midst while yet they were only soft mud. For just as a plaster cast boiled in oil becomes thereby denser and more durable, so the oily and other matter coming from decomposing fish operated on the surrounding sand or mud so as to make it more compact."

It may not be easy to explain how it came to pass that fishes dwelling in salt water, as these undoubtedly did, were thus deposited in great numbers, but we may now and then see how deposits of fresh-water fishes may have been formed. When rivers flowing through a stretch of level country are swollen during the spring floods, they overflow their banks, often carrying along large numbers of fishes. As the water subsides these may be caught in shallow pools that soon dry up, leaving the fishes to perish, and every year the Illinois game association rescues from the "back waters" quantities of bass that would otherwise be lost. Mr. F. S. Webster has recorded an instance that came under his observation in Texas, where thousands of gar pikes, trapped in a lake formed by an overflow of the Rio Grande, had been, by the drying up of this lake, penned into a pool about seventy-five feet long by twenty-five feet wide. The fish were literally packed together like sardines, layer upon layer, and a shot fired into the pool would set the entire mass in motion, the larger gars as they dashed about casting the smaller fry into the air, a score at a time. Mr. Webster estimates that there must have been not less than 700 or 800 fish in the pool, from a foot and a half up to seven feet in length, every one of which perished a little later. In addition to the fish in the pond, hundreds of those that had died previously lay about in every direction, and one can readily imagine what a fish-bed this would have made had the occurrence taken place in the past.

From the better-preserved specimens that do now and then turn up, we are able to obtain a very exact idea of the construction of the bony cuirass by which Pterichthys and its American cousin were protected, and to make a pretty accurate reconstruction of the entire animal. These primitive fishes had mouths, for eating is a necessity; but these mouths were not associated with true jaws, for the two do not, as might be supposed, necessarily go together. Neither did these animals possess hard backbones, and, while Pterichthys and its relatives had arms or fins, the hard parts of these were not on the inside but on the outside, so that the limb was more like the leg of a crab than the fin of a fish; and this is among the reasons why some naturalists have been led to conclude that vertebrates may have developed from crustaceans. Pteraspis, another of these little armored prevertebrates, had a less complicated covering, and looked very much like a small fish with its fore parts caught in an elongate clam-shell.