The Scandinavian kitchen-middens, or shell-mounds, carry us further back into this early neolithic period. The shell-mounds which are found in great numbers along the Baltic shore of Denmark are often of great size. They are formed of an accumulation of shells of oysters, mussels, and other shell-fish, bones of wild animals, birds, and fish, all of existing species, with numerous implements of flint or bone, and occasional fragments of coarse pottery. They are decidedly more archaic than the lake-dwellings, showing a much ruder civilization of savages living like the Fuegians of the present day, in scanty tribes on the sea-shore, supported mainly by shell-fish, supplemented by the chase of wild animals.

The dog was their only domestic animal, and their only arts the fabrication of rude pottery and implements of stone and bone, unless it can be inferred from the occasional presence of bones of cod and other deep-sea fish, that they possessed some form of boat or canoe, and had hooks and lines or nets. These mounds must have taken an enormous time to accumulate, for they are very numerous, and often of great bulk, some of them being 1000 feet long, 200 feet wide, and ten feet thick. How long such masses must have taken to accumulate must be apparent when we consider that the state of civilization implies a very scanty population. It has been calculated that if the neolithic population of Denmark required as many square miles for its support as the similar existing populations of Greenland and Patagonia, their total number could not have exceeded 1000, and each mound must have been the accumulation of perhaps two or three families. Ancient, however, as these mounds must be, they are clearly neolithic. They are sharply distinguished from the far older remains of the palæolithic period by the knowledge, however rude, of pottery and polished stone, and still more by the fauna, which is entirely recent, and from which the extinct animals of the quaternary period have disappeared; while the position of the mounds shows that only slight geological changes, such as are now going on, have occurred since they were accumulated. Similar mounds, on even a larger scale, occur on the sea-coasts of various districts in Europe and America, but they afford no indication of their date beyond that of great antiquity.

The peat-mosses of Denmark have been appealed to as affording something like a conjectural date for the early neolithic period in that country. These are formed in hollows of the glacial drift, which have been small lakes or ponds in the midst of forests, into which trees have fallen, and which have become gradually converted into peat by the growth of marsh plants. It is clearly established that there have been three successive ages of forest growth, the upper one of beech, below it one of oak, and lowest of all one of fir. The implements and relics found in the beech stratum are all modern, those in the oak stratum are of the later neolithic and bronze ages, and those in the lowest, or fir-horizon, are earlier and ruder neolithic, resembling those found in the older lake-villages and shell-mounds. Now beech has been the characteristic forest tree of Denmark certainly since the Roman period, or for 2000 years, and no one can say for how much longer. If the speculations as to the origin of the Aryan race in Northern Europe are correct, it must have been for very much longer, as the word for beech is common to so many of the dialects into which the primitive Aryan language became divided. The stages of oaks and firs must equally have been of long duration, and the different stages could only have been brought about by slow secular variations of climate during the post-glacial period. Still this affords no reliable information as to specific dates, and we can only take Steenstrup's calculation of from 4000 to 16,000 years for the formation of some of these peat-bogs as a very vague estimate, and this only carries us back to a time when Egypt and Chaldæa must have been already densely peopled, and far advanced in civilization.

On the whole, it seems that the neolithic arrow-heads found in Egypt, and the fragments of pottery brought up by borings through the deposits of the Nile, are the oldest certain human relics of the neolithic age which have yet been discovered, and these do not carry us back further than a possible date of 15,000 or 20,000 years b.c.

Nor is there any certainty that any of the neolithic remains found in the newer deposits of rivers and the upper strata of caves go further, or even so far back as these relics of an Egyptian stone period. All that the evidence really shows is, that while the neolithic period must have lasted for a long time as compared with historical standards, its duration is almost infinitesimally small as compared with that of the preceding palæolithic period. Thus in Kent's Cavern neolithic remains are only found in a small surface layer of black earth from three to twelve inches thick; while below this, palæolithic implements and a quaternary fauna occur in an upper stalagmite one to three feet thick, below it in red cave earth five to six feet thick; then in a lower stalagmite in places ten or twelve feet thick, and below it again in a breccia three or four feet thick. This is confirmed by the evidence of all the caves explored in all parts of the world, which uniformly show any neolithic remains confined to a superficial layer of a few inches with many feet of palæolithic strata below them. And river-drifts in the same manner show neolithic remains confined to the alluvia and peat-beds of existing streams, while palæolithic remains occur during the whole series of deposits while these rivers were excavating their present valleys. If we say feet for inches, or twelve for one, we shall be well within the mark in estimating the comparative duration of the palæolithic and neolithic periods, as measured by the thickness of their deposits in caves and river-drifts; and as we shall see hereafter, other geological evidence from elevations and depressions, denudations and depositions, point to even a higher figure.

In going back from the neolithic into the palæolithic period, we are confronted by the difficulty to which I have already referred, of there being no hard-and-fast lines by which geological eras are clearly separated from one another. Zoologically there seems to be a very decided break between the recent and the quaternary. The instances are rare and doubtful in which we can see any trace of the remains of palæolithic man, and of the fauna of extinct animals, passing gradually into those of neolithic and recent times. But geologically there is no such abrupt break. We cannot draw a line at the culmination of the last great glaciation and say, here the glacial period ends and the post-glacial begins. Nor can we say of any definite period or horizon, this is glacial and this recent.

A great number of palæolithic remains and of quaternary fossils are undoubtedly post-glacial in the sense of being found in deposits which have accumulated since the last great glaciers and ice-caps began to retreat. Existing valleys have been excavated to a great extent since the present rivers, swollen by the melting snows and torrential rains of this period of the latest glacial retreat, superseded old lines of drainage, and began to wear down the surface of the earth into its present aspect. This phase is more properly included in the term glacial, for both the coming on and the disappearance of the periods of intense cold are as much part of the phenomenon as their maximum culmination, and very probably occupied much longer intervals of time. In like manner, we cannot positively say when this post-glacial period ended and the recent began. Not, I should say, until the exceptional effects of the last great glacial period had finally disappeared, and the climate, geographical conditions, and fauna had assumed nearly or entirely the modern conditions in which we find them at the commencement of history. And this may have been different in different countries, for local conditions might make the glacial period commence sooner and continue later in some districts than in others. Thus in North America, where the glaciation was more intense, and the ice-cap extended some ten degrees further south than in Europe, it might well be that it was later in retreating and disappearing. The elevation of the Laurentian highlands into the region of perpetual snow was evidently one main factor of the American ice-cap, just as that of Scandinavia was of that of Europe, and it by no means follows that their depression was simultaneous. It would be unwise, for instance, to take the time occupied in cutting back the Niagara gorge by a river which only began to run at some stage of the post-glacial period, as an absolute test of the duration of that period all over the world. Indeed, the glacial period cannot be said to have ended and the post-glacial begun at the present day in Greenland, if the disappearance of the ice-cap over very extensive regions is to be taken as the test.

Any approximation to the duration of the post-glacial period in any given locality can only be obtained by defining its commencement with the first deposits which lie above the latest glacial drift, and measuring the amount of work done since.

This has been done very carefully by the officers of the Geological Survey and other eminent authorities in England and Scotland, and the result clearly shows that since the last glaciation left the country buried in a thick mantle of boulder-clay and drift, such an amount of denudation, and such movements of elevation and depression have taken place, as must have required a great lapse of time. The most complete attempt at an estimate of this time is that made by Mr. Mellard Read of the Geological Survey, from the changes proved to have occurred in the Mersey valley.

In this case it is shown that the valley, almost in its present dimensions, must have been first carved out of an uniform plain of glacial drift and upper boulder-clay by sub-aërial denudation; then that a depression let the sea into the valley and accumulated a series of estuarine clays and silts; then an elevation raised the whole into a plain on which grew an extensive forest of oak rooted in the clays; this again must have subsided and let in the sea for a second time, which must have remained long enough to leave a large estuarine deposit, and finally the whole must have been raised to the present level before historical times. The phenomenon of the submerged forest is a very general one, being traced along almost all the sea-coasts of Western Europe, where shelving shores and sheltered bays favour the preservation of patches of this primæval forest. It testifies to a considerable amount of elevation and subsequent depression, for its remains can be traced below low-water mark, and are occasionally dredged up far out to sea, and stately oaks could not have flourished unless more or less continental conditions had prevailed.