Glacières; or, Freezing Caverns

ICE FORMATIONS AND WINDHOLES AT WATERTOWN.

At Watertown, New York, on the south side of the Black River, in the town itself, are some natural cracks or crevices in the limestone rocks. They are only a short distance from the New York Central Railroad station. The cracks enter the northern side of the railroad embankment, pass under the railroad tracks, and extend some distance back. In front of them are four cellars, used for storing beer kegs. The lessor, Mr. Ehrlicher, obligingly had the cellars opened for me, on the 12th of September, 1898. There was neither ice nor draughts in the cellars, and the temperature was normal. Mr. Ehrlicher said that in the spring there was ice in the cracks, but that it had all melted away as the result of the hot summer.

About four kilometers west of Watertown, on the south bank of the Black River, is the picnic ground of Glen Park, which is reached by trolley. The manager of the restaurant walked around the park with me. In one spot is a hollow or glen at the base of a small, much cracked limestone cliff, which has a northern exposure. The manager said that snow and ice usually lies in this place until June, not only among the broken rocks, but even in the open. Sometimes ice remains among the boulders all summer, but only near the front of the boulders, and by pushing in, one soon gets beyond it: we found none, a fact showing once more the effect of the unusually warm summer. On hot days, draughts issue from between the boulders, but as the day was cool, we did not notice any. The spot is well sheltered against the wind by a number of trees; and the shape of the hollow reminded me of the glen in front of the Eishöhle bei Roth.

Not one hundred meters from this hollow, is a little limestone cave, closed by a wooden door, which excludes any cold air in winter. The cave is lighted by electric lights, and is a narrow, crooked, descending fissure, a ganghöhle, where the marks of water action are plainly visible. At the bottom a little stream, evidently the active agent in forming the cave, ran through the fissured limestone. In the stream a large toad or frog was swimming about. There was nothing icy about the cave or the water, and the temperature was normal. Ice was never known to form in the cave. These two places, so close together, are an interesting confirmation that it is only where the outside cold can get in, that we find subterranean ice.

THE FREEZING CAVE AND FREEZING WELLS OF DECORAH.

Near Decorah, Iowa, is a freezing cavern, which is more frequently referred to in cave literature than is generally the case. I visited it on Friday, September the 30th, 1898, with an old English resident of Decorah, Mr. W. D. Selby-Hill. The cave is situated about one kilometer to the northward of Decorah, on the north bank of the Upper Iowa River, at the base of a bluff. It is some thirty to forty meters above the stream, and faces southward. It looks like a fault or fissure in the rocks, with the sides meeting a few meters overhead. It is a true cave, but probably in an early stage of formation, for there are no apparent traces of water action, nor any stalagmites nor stalactites. The absence of the latter may, however, be due to the fact that it is a periodic glacière. The rock is a white limestone, rich in fossils. The cave is some two to three meters in width and is rather winding, with a short arm or pocket branching out on the west side. The main cave runs back some thirty meters from the entrance. In one place it is necessary to stoop, to get past some overhanging rock slabs. By candle light, we went to the rear of the cave, and found it warm, dry, and free from ice. There were no draughts, possibly because the day was cool.

I looked in vain for tubular fissures, or indeed any fissures, through which water might freeze by pressure in its descent, as the believers in the capillary theory say it does. Nothing of the kind existed, and I wrote in my note-book: “Writing on the very spot about which this theory was started, I feel justified in asserting that the theory amounts to absolutely nothing and is entirely incorrect.”

Mr. Hill told me that there were two wells in the southern portion of Decorah Township, where ice was found in summer. I visited them both, but found no ice, and the temperatures normal. Mr. Hill said that one of the wells was dug about thirty years ago, and that the workman told him that the ground which he went through was frozen; and that at one place he struck an opening, from which came so strong a current of icy air, that it was hard to keep at work.

I talked to several persons afterwards. Inter alia, they told me that the bluff was a great place for rattlesnakes, sometimes big ones. They admitted also generally that they were puzzled about the formation of ice in the cave. Some claimed that the ice formed in summer—the old story once more. I met, however, Mr. Alois F. Kovarik of the Decorah Institute, who had made a series of regular observations for over a year and found that the ice begins to form about the end of March and beginning of April, and is at its maximum towards the beginning of June. Mr. Kovarik also told me, that he had found ice in one of the wells in the beginning of August.

This was an especially satisfactory trip to me, for it did away, once for all, with any possible belief that there was any basis of fact for the capillary theory. It also seems to me important to find that the ice of these freezing wells melts in summer. For it shows that their ice is due to the same causes as those which form the ice in the cave, and is another proof against the validity of the glacial period theory.

FREEZING ROCK TALUS ON SPRUCE CREEK.

On Spruce Creek, Huntingdon County, Pennsylvania, about four kilometers north of the Pennsylvania Railroad depot, is an ice bearing talus, known locally as the Ice Holes or Ice Caves. I visited this spot, on October the 5th, 1898, with Mr. Benner, of Spruce Creek. We walked up the pretty valley along the old Pittsburgh turnpike, at one place finding some papaw trees, whose fruit had a horrible sickening taste; then we crossed Spruce Creek by a footbridge and followed the other bank back for some five hundred meters, until we were nearly opposite the old Colerain Forge, which is located in a piece of land called by the curious name of Africa. About half way from the bridge we smelt a strange odor, which my companion thought came from a copperhead or rattlesnake: we did not investigate.

The freezing talus is situated at the foot of Tussey Mountain: it is big, and is composed of small sandstone (?) rock débris. The talus is at least thirty meters high and one hundred and twenty meters long. As I stood at the bottom, I was reminded strongly of the talus at the Dornburg. At the base were a number of small pits, evidently dug by man. From the interstices between the rocks, icy cold draughts issued in some places, and there was no doubt that there was plenty of ice beneath the stones. In one place we thought we could see ice, and I poked at the white substance with my stick, but I am not positive that it was ice. All over the talus, the temperature was strikingly colder than a few meters away, and in the pits we could see our breaths distinctly. Although I am not much of a botanist, yet it seemed to me that the flora immediately near the talus was somewhat different in character from that of the surrounding country.

Mr. Benner told me that he saw, three or four weeks before, plenty of ice in the pits; that they were made by farmers who formerly came to this spot to get ice; and that parties occasionally picnic here in the summer and make ice cream. He stated also that he saw, some years ago, a small cave or hole containing ice near Mapleton, Pennsylvania, but that it was destroyed by quarrying the rock away.

FREEZING GORGE NEAR ELLENVILLE.

On Sunday, October the 9th, 1898, with a young man from Ellenville, I visited the well known Ellenville Gorge, in the Shawangunk Range, Ulster County, New York. We left the hotel at eight-forty A. M. and reached the gorge, known locally as the Ice Cave, at ten-five A. M. It is about four kilometers northeast from Ellenville. The path rises steadily uphill and is of the roughest description; it is covered with loose stones, and looks as if it might become the bed of a mountain brook in wet weather.

I call this place a gorge, instead of a cave, because it is uncovered at the top, but probably originally it was covered. It is shaped like a pit cave minus a roof, and it reminded me of the Friedrichsteiner Eishöhle, and the Glacières de Saint-Georges and du Pré de Saint-Livres. It is entered by a long slope from the western end, the gorge turning northward further back. I estimated its width, at the bottom at some five to seven meters, at the top at some three to four meters; its length at some thirty meters and the deepest point we reached, at some twenty meters below the surface. These are guesses, however. In one place, a great rock slab overhangs the gorge. At nearly the lowest point of the rock floor, there is a hole which extends perpendicularly downwards some five or ten meters more; this opening is partly blocked up with fallen masses of rock which would make a further descent perilous. The north end of the gorge is also filled up with a mass of great broken rocks; in fact, the whole place is out of repair, as the rocks are cracked and creviced on both sides to a great extent. The rock is friable and seems to be all breaking up, or rather down, and I think there is some danger from falling stones, although I did not see any fall. There is a good deal of moss on the sides of the gorge, and on some ledges small evergreens are growing. The gorge is sheltered thoroughly from winds by its formation and position, and somewhat by the scrub forest surrounding it. There are several long, deep crevices a few meters further up the mountain side, and I think one of them is an extension of the main gorge.

We found no ice. It generally lasts till about the beginning of September; and Professor Angelo Heilprin, and Miss Julia L. Lewis, of Philadelphia, have found plenty of it in July and August. But the ice had evidently now been gone for some time, for the temperature at the bottom of the gorge was about 11° C. at ten-thirty A. M. This was but little colder than the temperature v outside, which at ten-fifteen A. M. was 14° C.

On returning to Ellenville, I learnt that there was another somewhat similar smaller gorge, some eight kilometers away, at a place called Sam’s Point. This, however, is said to retain only snow, while in the Ellenville gorge much ice is sometimes formed, and icicles a couple of meters long are said to hang on the sides of the cliffs. The proprietor of the hotel told me he had heard of a cave which contained ice not far from Albany, at a place called Carlisle, on the Delaware and Hudson Railroad.

FREEZING CAVE AND WINDHOLES NEAR FARRANDSVILLE.

I arrived at Farrandsville, Clinton County, Pennsylvania, early on Tuesday morning, October the 11th, 1898, and found a boy, who worked in a brick mill, as guide to the caves.[11] After emptying a small, flat bottomed boat of the water of which it was half full, we rowed across the Susquehanna River; then we walked up the road, along the river bank, for a couple of hundred meters, and struck up the so-called path to the caves. Although the whole of the mountain side was at the disposal of the road maker, no better plan seems to have suggested itself than to make the track go straight up. This saved making zigzags, yet the result is that the path is steep, and as it is rocky and slippery, it is hard travelling without bootnails or alpenstock.

As we went up, I noticed, in one or two places, cold draughts issuing from crevices in the rocks. We soon came to a hollow under a rock, where there were a number of cracks and crevices: the boy spoke of it as the lower cave. It is some sixty meters above the Susquehanna River and cold draughts flowed from the cracks, although we saw no ice. The cave was about twenty meters higher up. One could crawl into it for a couple of meters, and all round it the rocks are somewhat creviced; in fact, I think there are a good many cracks in the entire hill. There was no ice in sight in this hole, but a strong, cold draught poured from it. After an exposure of fifteen minutes the thermometer registered 6° C.; while outside, in the shade, it stood at 15° C. This decidedly sub-normal temperature proved unmistakably, in my opinion, the presence of ice a little further than we could see in. Both holes face about north and are sheltered, by their position and by the sparse forest which covers the ridge, against all winds except those from the north.

I talked to the postmaster and the railroad agent at Farrandsville on my return, and they stated that there was no ice in the hole in winter, but that it formed about April and remained over until towards September, showing that the cave is a normal glacière on a small scale.

GLACIÈRES NEAR SUMMIT.

In the search for coal, the mountains of the Appalachian Chain between the little town of Summit, and the neighboring village of Coaldale, Carbon County, Pennsylvania, were mined and tunneled in every direction. Owing to the caving in of some of these mines, depressions formed in certain places along the ridge in the upper surface of the ground, and in two of these hollows natural refrigerators occur. These were brought to my notice by Mr. C. J. Nicholson of Philadelphia, and I visited them on May the 5th, 1899, in company with two coal miners of Summit.

Starting from Summit, we passed across some rough ground under which there was a mine on fire; and the miners showed me the tops of two pipes sticking out of the ground, from which issued a smoke or steam, too hot to hold the hand in more than a few seconds. Going beyond through brushwood, for a couple of hundred meters, we came to the first glacière, which was also the nearest to Summit. It faced almost due north and looked as if it was formerly the entrance to a mine. It was fairly big, and my companions assured me that, until within about a year, ice was always found in it. Recently, however, part, of the rock roof fell in, blocking up the entrance with a mass of débris and making it unsafe to venture in. Formerly parties of tourists constantly visited this place, after coming over the Switchback, but this is no longer done and there has been some talk of cleaning away the broken rocks and making the glacière accessible. The men also said that occasionally people living in the neighborhood had dug out the ice for their own use.

The other glacière was a short distance further, in the direction of Coaldale. It is in a pit, which may have been the mouth of a disused shaft or only a depression resulting from a cave-in. A scrubby forest, which surrounds the hollow, acts as a windbrake. A rather steep slope leads down into the pit, and at the end passes under the wall of rock of the opposite side for a short distance, forming a small cave, which faces almost due south and whose floor is choked up with broken rock fragments. At the bottom of the slope we found some snow, and among the boulders a good deal of snow-ice as well as several long icicles hanging from the rocks. All the ice and snow lay on the north side of the rocks, or underneath them, so that it was in shady places where the sun could not reach it. The temperature was not at all uncomfortable, although somewhat cool and damp.

There was nothing in either glacière, to show that the ice was formed from any other cause than the drifting in, and melting and refreezing of the winter’s snow; and my impression is that the ice in the second glacière could not last through the summer.

THE SNOW HOLE NEAR WILLIAMSTOWN.

The Snow Hole near Williamstown (Massachusetts) is situated near the northern end of the Petersburgh Mountain of the Taghconic Range; it is slightly below the watershed on the Williamstown side, at an altitude of about seven hundred meters. The Snow Hole is in the State of New York, near the boundary between New York and Massachusetts. It is a long two hours’ drive from Williamstown, the last four kilometers or so, over an exceedingly steep and rough road, which is, in fact, nothing but an old logging road, and the worst I ever drove over except the road to Démenyfálva.

I visited the Snow Hole with my brother on Friday, September the 29th, 1899. It is surrounded by a dense forest, mainly of recent growth, which thoroughly shelters it from all winds. In shape and appearance it resembles the Gorge at Ellenville, except that it is smaller: its location on the ridge is not unlike that of the Friedrichsteiner Eishöhle. It is a narrow crack—or cave minus a roof—about fifteen meters long, six to seven meters deep and from two to five meters wide. It faces nearly north, and the bottom is in perpetual shadow. From the northern end, a gentle slope leads to the rear. The slope was a good deal blocked up by a big tree with large branches, which had fallen directly into the fissure. There was some moss or greenish mould on the rocks in places, and at the rear end of the slope there were some fissures in the rocks, into which one might perhaps have crawled a little farther, which formed a tiny cave. There was also a similar incipient cave at the northern end. I could not detect any draughts issuing from these rock fissures, and the air throughout was still, although the wind was blowing hard on the ridge. The rocks were moist in places and the air damp, but there was neither snow nor ice and the temperatures were normal. The driver told me that he had found plenty of snow in the base of the gorge some years ago in July; and he said that he had always heard that snow was found in the Snow Hole all the year round. All the conditions of the place, the shape of the fissure, and its sheltered northern exposition, are favorable to the retention of ice and snow, and it is not surprising that they remain over every spring.

ICY GULF NEAR GREAT BARRINGTON.

The Icy Gulf or Icy Glen is some eight kilometers from Great Barrington, Massachusetts. I have not been in it, but was told in October, 1899, by the farmers living near by, that after snowy winters, ice remains over through July. It must be similar to the Icy Glen at Stockbridge.

THE ICE BED OF WALLINGFORD.

The Ice Bed of Wallingford is situated about three kilometers to the east of Wallingford, Vermont. A drive of half an hour, over the Mount Holly and Hearburrow roads, takes the visitor to the entrance of a rough wood path, which, at a distance of three or four hundred meters, leads to the Ice Bed. This is a huge talus, at the base of the White Rock Mountain, whose cliffs rise steeply overhead for some three or four hundred meters. The talus, which was doubtless formed by a great slide at some distant date, consists of granite boulders, some of which are big ones. The ice-bearing portion may be some thirty or forty meters high vertically. It lies in a sort of gully or rock basin, and at the top is about thirty meters broad, tapering to a point at the bottom. The talus faces southward, and during a good part of the day the sun shines full upon it. A thin forest fringes the sides and grows round the bottom, but this can afford but little protection from the winds, especially to those from the south.

I visited this place on the 5th of October, 1899. There was a distinct drop in temperature as we neared the base of the talus, and a cool air drew gently down over the rocks. I think slight draughts issued from some of the crevices; but of this I am not sure. The temperature was sub-normal, about 8°, but hardly low enough to prove the presence of ice, although we could see our breaths distinctly. We looked carefully under a number of the boulders, but neither ice nor snow was visible. I was assured that ice was abundant there in the past July and August, and I should think it had melted away only shortly before my visit. My impression is, that this is a periodic glacière.

CAVES NEAR WILLIAMSTOWN.

On the eastern slope of the Petersburgh Mountain of the Taghconic Range in Massachusetts, at a good deal lower altitude than the Williamstown Snow Hole and about southeast of it are some caverns, which are but little known. A five or six kilometer drive from Williamstown takes the visitor to the base of the mountain, whence a rather steep ascent of about a kilometer and a half brings him to the caves, which are in the midst of a dense, scrub forest.

The caves were first entered, and possibly discovered, by Mr. W. F. Williams, of Williamstown, when a boy. Since then, he has visited them many times and explored them a good deal. They do not appear to have any name as yet, and it would seem only fitting to christen them after their explorer: the Williams Caves.

There are several unimportant holes in the immediate neighborhood of the two main caves. The latter lie side by side. The rock formation is the same as that of the Snow Hole, a dark gray slate with a few veins of quartz, and they are due also evidently to the same geological causes. It would seem as though the mountain had tended to open or crack at these spots and fallen apart. This seems probable, because wherever there is a projection on one side of the cracks, there is a corresponding hollow in the opposite side. After this, water action has come, and erosion and corrosion have worn out and carried away earthy matter, and slowly deepened and widened the fissures. The remarkable point in connection with the main caves, however, is that one is a normal cave and the other a periodic glacière.

I went with Mr. Williams to these caves on the 6th of October, 1899, and partially explored the glacière. On the way up, just as we left the carriage road, a fine, three-year-old buck, in his winter coat, came bounding out of the forest; on seeing us he stopped, and after taking a good look, quietly trotted off into the bushes.

The glacière is rather peculiar in shape and may be described as two storied. A long slope, set at an angle of some forty degrees, and covered with mud and dead leaves, leads down into the crack, which is from one to three meters in width. The first half of the slope is open to the sky; the last half is covered by the rock roof, and is a real cave. In this the floor is horizontal, the place forming a little chamber in which the daylight has almost vanished. At the exact summit of the slope a big tree grew most conveniently; and we tied to this one end of a twenty-meter Austrian Alpine Club rope, and by holding fast to it, and kneeling or sitting down in the mud in two or three places, the descent was easy enough. It was rather difficult to scramble up the slope again, however.

In the floor of the little chamber there are two holes, and, stepping over these, we stood at the rear end, about eighteen meters distant from the beginning of the slope. My companion now set some birchbark on fire and dropped it into the innermost hole, and we laid down in turn, flat on the rock floor, and craned our necks through the hole. Mr. Williams thought he could see ice below us. I looked down after him and found that I was looking into a lower chamber whose sides were invisible. The floor was some three meters below vertically, and on this the birchbark was burning brightly. I think I saw some ice, but I could not be sure, as there was too much smoke to see distinctly. My companion offered to go down through the hole and get some ice; a proposition I promptly vetoed, as had anything gone wrong, I could not possibly have given him any assistance, as there was no extra rope. Mr. Williams told me that he went down several times before in July or August, and always found ice on the slanting floor. He said he did not know how far this lower chamber extended, nor the length of the ice floor. One thing which makes me hesitate to think that we saw ice was, that the temperature of the chamber where we were was not at all icy; but probably—I had forgotten my thermometer—nearly normal.

When we stood once more by the tree at the top of the slope, the mouth of another cave was visible about two meters below us. Mr. Williams said it had never looked more than a little crack before, and that the opening was much bigger than at his last visit. It was directly under the slope by which we descended and it vanished into darkness. Its direction led straight towards the lower chamber, and it almost surely leads to it. It seems thus that there are two hollows, one directly above the other; and that the lower one is a glacière, while the upper one is not. The cold air of winter would naturally sink into the lower chamber, and the spring thaws would furnish plenty of drip, so that this place seems to answer every requirement of a cave glacière.

But the most interesting fact about these caves is that, while the shallower one is a glacière, the bigger and deeper one is not. This is situated about ten meters north of the glacière and the direction of the entrance is about the same. Mr. Williams has found snow and ice in May in the entrance pit as far as the daylight goes, but none beyond. I am inclined to think that the explanation of this is the fact that the cave is a ganghöhle or tunnel cave. Mr. Williams described it as a narrow passage with chambers, and at least a hundred meters long, and fifty meters in depth below the surface. The cold air sinks in a certain distance, but as the passage is narrow and long, and too winding for any strong draughts, the cold air which enters is soon neutralized by the supply of warmer air within and by contact with the rocks. I cannot help thinking that it is by some such explanation that we must hope to solve the problem of why certain caves are glacières and others in the immediate neighborhood normal caves; and the caves near Williamstown are exceptional in presenting the problem so patently.

PART II.

THE CAUSES OF SUBTERRANEAN ICE.

THE CAUSES OF SUBTERRANEAN ICE.

I.

Terminology.—Ice enduring the entire year is found, in temperate latitudes, in a variety of forms and in several different kinds of places. In some cases it is entirely above the surface of the earth; in others it is entirely beneath the surface of the earth. These are the extremes, and between them there are certain intermediate forms. The perennial ice above ground of temperate regions has gradually become known in English by the French word glacier, but strange to say, there is no term in use in English which accurately describes the perennial ice formations which are partially or completely underground. Thus the term “ice cave” is applied to a rock cavern containing ice, and the term “ice gorge” to a rock gorge containing ice. Both terms are misleading, because the character of the contents is mentioned before the nature of the geological formation. We say correctly enough “limestone cave” or “lava cave” and, in my opinion, we should apply the term “ice cave” in a similar manner to the hollows in the ice at the lower end of glaciers, whence the glacier waters make their exit. These are really “ice caves,” that is caves with sides and roof made of ice. Another trouble of the term “ice cave,” as applied to rock formations containing ice, is that it is not generic: not only is it incorrect, but also it is not comprehensive. It does not apply to mines, tunnels, wells, gullies, boulder taluses, or underground ice sheets. If “ice cave” is used, except in its true sense of glacier ice cave, it seems at least as though it should be so only for real caves which retain ice, as opposed to taluses and wells. Curiously enough, the Germans are just as inaccurate as ourselves, for their terms eishöhle and eisloch are absolute translations of our “ice cave” and “ice hole.” Indeed, there is no doubt that some of the incorrect notions about subterranean ice formations, are due to the inaccuracy of the terminology.

The only language, so far as I know, which has a correct and really generic term for subterranean ice formations, is the French in its word glacière. The French and Swiss say glacières naturelles of ice deposits formed naturally underground; and glacières artificielles of ice houses. Glacière naturelle is comprehensive and accurate. It covers all the rock formations and suggests also the mode of formation of the ice. It likewise implies the strong resemblance between natural ice deposits and artificial ice houses. It might be well, therefore, if the French term glacière were adopted as a generic term for all underground ice formations. As, however, there is little likelihood of this happening, the question arises as to the best English equivalent or equivalents. These seem to be “freezing cavern, freezing talus,” etc., “natural refrigerator” or “subterranean ice formation.” “Natural refrigerator” and “subterranean ice formation” are more generic than “freezing cavern, freezing well,” etc.; but the latter have the advantage of suggesting immediately that reference is made to the hollows of the earth which at times contain ice; and, therefore, they are the best terms, perhaps, which can be chosen in English.

Another point in the terminology of this subject has reference to subterranean hollows where draughts issue or enter. Such hollows are found in all parts of the world and are known usually in English as “blowing caves” or “cold current caves.” The Germans speak of them as windröhren or windlöcher. In my first paper about caves,[12] I used the word “windhole” which I translated from the German. The term “windhole” seems to me preferable to “blowing cave” or “cold current cave” in that it is more generic. It applies to taluses or boulder heaps, or in fact, to any hollows where draughts issue or enter, whether these hollows are genuine caverns or not.

It is necessary also to explain here that “glacière” and “windhole” are not synonymous terms. It must be understood that a glacière or natural refrigerator is a place where ice forms and endures in a subterranean or semi-subterranean situation; and that the presence of ice is the criterion of whether a place is or is not a glacière. Likewise it must be understood that a windhole or blowing cave is an underground hollow with at least two openings, and in which distinct draughts occur; and that the presence of draughts is necessary to constitute a place a windhole or blowing cave. A freezing cavern may or may not be a windhole, and a windhole may or may not be a freezing cavern.

Temperatures.—The phenomena of glacières are so closely connected with temperatures that it seems necessary at this point to mention some general facts in connection with subterranean temperatures, even if these still form a subject of some uncertainty, and one about which further observation is desirable. Subterranean temperatures may be grouped under three heads: 1, Ordinary or normal temperatures; 2, Temperatures above the normal or super-normal temperatures; 3, Temperatures below the normal or sub-normal temperatures.

1. In the great majority of caves, cellars and subterranean places of all descriptions, the temperature of the air is about the same, all the year round, as that of the ground. The frost of winter and the heat of summer penetrate the earth for some trivial distance, a few meters perhaps, and lower or raise the temperature of the ground temporarily. Below this there is a stratum where the temperature is found to vary but little the entire year and which, in a majority of cases, approximates the mean annual temperature of the district. Below this invariable stratum, the temperature generally rises slowly, not at exactly the same rate everywhere, but in a regular increase. This increase of temperature averages 1° C. for every 32 meters. As most caves and cellars are of small depth and as they take their temperatures from that of the ground, it follows that as a rule their temperatures are moderate and pleasant. And as the air of the majority of caves and subterranean hollows is about the same as the temperature of the surrounding rock, it is correct to call subterranean air temperatures closely approximating the ordinary temperature of the ground, ordinary or normal temperatures.

As already stated, with an increase of depth, there is, in almost all cases, a regular increase of temperature. For this reason, mines, which are much the deepest hollows reached by man in the surface of the earth, are, as a rule, warmer in the lower levels: if deep, they are also hot. And this is so generally the case that warmer temperatures at the bottom of mines may be considered as normal.

2. In a few hollows close to the surface, there are temperatures much above the normal temperature of the ground. Such places are rare and abnormal.[13] The heat is generally due to the presence of hot springs or to some volcanic action in the immediate neighborhood. In the case of one cave close to the surface, the heat is due to some limekilns which are situated immediately overhead.[14] Where these warm hollows are genuine caves it seems proper to call them “hot caves.”

3. In a number of places, there are abnormally low temperatures underground either for the whole or only for part of the year. Although commoner than hot caves, yet the underground places with low temperatures are also rare and abnormal. They may be divided into two groups: 1, Those where the temperatures are lower than the normal, without becoming low enough for ice to form; and 2, Those where the temperature sinks so low, that ice forms.

It is difficult to make definite divisions among the various forms of natural refrigerators, but it is correct, probably, to classify them under five heads, in accordance with the different kinds of formations of the hollows in the rocks:

1. Gullies, gorges, and troughs where ice and snow remain.

2. Soil or rocks overlaying ice sheets.

3. Taluses and boulder heaps retaining ice.

4. Wells, mines and tunnels in which ice sometimes forms.

5. Caves with abnormally low temperatures, and often containing ice.

1. Gorges and Troughs.—Gullies, gorges and basins which retain snow and ice are fairly numerous in mountain districts. They are generally ravines, or rock fissures, or hollows, in positions below the snow line where snow and ice are sufficiently protected, from sun and wind, to remain long after snow in the surrounding open country, at the same altitude, has melted away. Some of these gorges are small, some big. As a rule, they are deep and narrow.

In north-eastern Siberia, a form of permanent surface ice is found, which the Tungusses speak of as tarinnen, which means “ice troughs” or “ice valleys.”[15] These tarinnen are broad valleys, with either a horizontal floor or one sloping gently in the form of a trough, over which the ice is spread in the form of a sheet. The Tungusses assert that the ice in some of these troughs never wholly melts away, although it lessens in quantity from the beginning of May till the end of August, after which it once more increases.

Subterranean Ice Sheets.—In several places in different parts of the world there are underground ice sheets which extend over large spaces; they are common under the tundras of Alaska; and there are fine examples on Kotzebue Sound,[16] on the Kowak River,[17] and along the Yukon River.[18] The “Ice Spring” in Oregon seems to be a formation of the same kind. Several examples of these subterranean ice sheets are reported also from different parts of the Russian Empire.

A somewhat different kind of ice sheet was observed on Mount Etna. Sir Charles Lyell[19] speaks of it as a “glacier preserved by a covering of lava.” He says Signor Mario Gemmellaro satisfied himself that nothing but a flowing of lava over snow could account for the position of the glacier. Ice sheets somewhat similar to these are reported from Tierra del Fuego,[20] and probably also such sheets occur in Iceland; and enormous heaps of ice covered with sand are found on Mount Chimborazo.[21] On the northwestern coast of Greenland, glaciers, whose flow has stopped, were observed buried under an accumulation of moss and grass.[22]