Encyclopædia Britannica, Ninth Edition/Lake

LAKE. When a stream in its course meets with a depression in the land it flows into it and tends to fill it up to the lip of its lowest exit. Whether it succeeds in doing this or not depends on the climate. In the British Islands, and in most temperate and equatorial regions, the stream would fill the depression and run over, and the surplus water would flow on towards the sea. Such a depression, with its contents of practically stagnant water, constitutes a lake, and its water would be fresh. In warm dry regions, however, such as are frequently met with in tropical latitudes, it might easily happen that the evaporation from the surface of the depression, supposed filled with water, might be greater than the supply from the feeding stream and from rain falling on its surface. The level of the waters in the depression would then stand at such a height that the evaporation from its surface would exactly balance the supply from streams and rain. We should have as the result a lake whose waters would be salt. Lakes of the first kind may be considered as enlargements of rivers, those of the second kind as isolated portions of the ocean; indeed, salt lakes are very frequently called seas, as the Caspian Sea and the Dead Sea. The occurrence of freshwater lakes and salt lakes in the same drainage system is not uncommon. In this case the salt lake forms the termination. Well known examples of this are Lake Titicaca and the Desguadero in South America, and Lake Tiberias and the Dead Sea on the Jordan.

Distribution of Lakes.—Although there are few countries where lakes are entirely absent, still it requires little study to see that they are much more thickly grouped in some places than in others. Of the larger lakes, for instance, we have the remarkable group in North America, which together form the greatest extent of fresh water in the world. A similar group of immense lakes is found in Central Africa:—Lakes Victoria Nyanza and Albert Nyanza, whose overflow waters go to form the Nile; Lake Tanganyika, at the source of the Congo; and Lake Nyassa, on a tributary to the Zambesi. In Asia the largest fresh water lake is Lake Baikal, on the upper waters of the Lena. All these freshwater lakes of great size are at the sources of large and important rivers; the salt lakes in which Asia also abounds are at the mouths of large rivers, as the Caspian at the mouth of the Volga, and Aral Sea at the mouth of the Oxus.

Passing from the consideration of these larger lakes, which from their size may be considered inland oceans, and which therefore necessarily occur in small number, we find large numbers of lakes of comparatively small dimensions, and when we consider them attentively we find that they are reducible to a small number of species, and, as in the case of plants and animals, the distribution of these species is regulated chiefly by climate, but also by geological conditions. Perhaps the most important and remarkable species of lakes is that to which the Scottish lakes belong. They are generally characterized by occupying long narrow depressions in the valleys of a mountainous country in the neighbourhood of the sea, and in a temperate climate. On the sea-coast, lakes of this character are found in Norway, Scotland, Newfoundland, Canada, the southern extremity of South America, and the south end of the middle island of New Zealand; somewhat removed from the sea we have the Alpine lakes of Switzerland and Tyrol, and the great Italian lakes, all of which display the same features as those of Scotland or of Norway. In many flat countries lakes are extraordinarily abundant, as for instance in the north part of Russia and Finland, in the southern part of Sweden, in the northern parts of Canada, and on a small scale in the Hebrides.

Lagoons, found on all low sandy coasts, owe their origin to the shifting of the sand under the influence of the wind and tide. They are found at the mouths of large rivers, as on the Baltic and at the mouth of the Garonne.

In volcanic regions lakes are not uncommon, generally of a more or less circular form, and either occupying the site of extinct craters or due to subsidences consequent on volcanic eruptions; such are the Maare of the Eifel in Germany, and many lakes in Italy and in the Azores.

Lakes are not only widely distributed in latitude and longitude, they also occur at all elevations. Indeed, as a certain elevation above the sea produces an effect as regards climate equivalent to a certain increase of latitude, we find lakes existing in the centre of continents, and on high plateaus and mountain ranges, in latitudes where they would be speedily dried up if at the level of the sea. Many of the lakes in Scotland (as Lochs Lomond, Morar, Coruisk), of Norway, of British Columbia, and of southern Chili are raised only by a few feet above the level of the sea, and are separated from it often by only a few hundred yards of land, while in the Cordilleras of South America we have Lake Titicaca 12,500 feet, and in Asia Lake Kokonor 10,500 feet above the sea. Many lakes whose surface is raised high above the level of the sea are so deep that their bottom reaches considerably below that level.

Dimensions of Lakes.—The principal measurements connected with a number of lakes in different parts of the world, presented in the following table, will give a more precise idea of the size of the lakes than could be given by description alone:—

Name of Lake.	Mean Lati- tude.			Length.			Breadth (Max.).			Depth (Max.).	Height in Feet above the Sea of		Tempera- ture of Water at Bottom.
											Surface.	Bottom.
				Miles.			Miles.			Feet.			° F.
Superior	47°	45′	N.	350			100			978	627	−351	38·8
Michigan	44°		N.	320			80			840	594	−246	...
St Clair	42°	30′	N.	18			22			20	570	+550	...
Erie	42°		N.	220			48			204	564	+360	...
Titicaca	16°	30′	S.	90			30			924	12,500	−11,576	54·6
Kokonor	37°		N.	91			42			...	10,500	...	...
Baikal	53°		N.	330			40			4,080	1,360	−2,720	...
Balkash	46°		N.	280			25			238	72	+166	...
Caspian	42°		N.	600			50			3,600	−85	−3,685	44·6
Dead Sea	31°	30′	N.	45			10			1,308	−1,272	−2,580	...
Tanganyika	6°		S.	330			40			1,000	2,700	...	...
Como	46°		N.	48			2	·	5	1,356	670	−686	...
Geneva	46°	25′	N.	45			8	·	7	1,092	1,218	+126	41	·	7	to	43	·	5
Constance	47°	40′	N.	35			8			394	1,300	+906	39·6
Lomond	$\scriptstyle {\left\{{\begin{matrix}\ \\\\\ \\\ \\\ \\\ \\\ \\\ \\\ \\\ \\\ \\\ \ \end{matrix}}\right.}$	56° to 57° 30′ N.	$\scriptstyle {\left.{\begin{matrix}\ \\\\\ \\\ \\\ \\\ \\\ \\\ \\\ \\\ \\\ \\\ \ \end{matrix}}\right\}\,}$	20			4			630	25	−605	41	·	4	to	42
Morar				11			1	·	5	1,020	30	−990	40	·	8	to	41	·	4
Ness				23			1	·	3	774	50	−724	41	·	2	to	42	·	4
Lochy				10			1			480	93	−387	42			to	44
Katrine				7			0	·	8	480	364	−116	41·4
Tay				14	·	5	1			450	390	−60	43·9
Rannoch				14	·	5	1			378	668	+290	43·9
Ericht				14	·	5	0	·	8	330	1,153	+823	44·7
Tummel				2	·	5	0	·	5	120	450	+330	45·5
Garry				2	·	5	0	·	3	102	1,330	+1,228	53·9

From this table it will be seen that by far the largest continuous sheet of fresh water is the group of North American lakes, and of these Lake Superior is more than double the size of any of the others; this is principally due to its great breadth, as it is very little longer than Lake Michigan. Lake Superior communicates with Lakes Michigan and Huron, which are really branches of one and the same lake, by the St Mary’s river, the fall being 49 feet from Superior to Huron. Huron empties itself into Erie by the St Clair river, Lake St Clair, and finally the Detroit river. Lake Erie overflows by the Niagara river and falls into Lake Ontario, whence the water finally is conveyed to the sea by the St Lawrence. The area of the lakes together is in round numbers 100,000 square miles, and, if that of the St Lawrence and its estuary be added, the water area will be about 150,000 square miles, while the whole drainage area is only 537,000 square miles. Hence of the water conveyed by the St Lawrence to the sea, rather more than one-fourth falls on the surface of the water itself. Looking to their great extent, we should have suspected them to be much deeper than is found to be the case. The deepest, Lake Superior, is no deeper than Loch Morar in Inverness-shire. Comparatively shallow, however, as they are, the bottoms of them all, with the exception of Erie, are several hundred feet below the level of the sea. It has been supposed that in former times this chain of lakes formed an arm of the sea similar to the Baltic in Europe, and in support of this view we have the fact of the discovery of marine forms in Lake Michigan.

In Asia Lake Baikal is in every way comparable to the great Canadian lakes as regards size. Its area of over 9000 square miles makes it about equal to Erie in superficial extent, while its enormous depth of over 4000 feet makes the volume of its waters almost equal to that of Lake Superior. Although its surface is 1360 feet above the sea-level, its bottom is 2720 feet below it. A former connexion with the ocean has been claimed for this lake, owing to the fact that seals inhabit its waters. Other large lakes in Asia are mostly salt, and some lie wholly below the level of the sea. Thus the Caspian lies 85 feet below the Black Sea, and the bottom at its greatest depth is 3000 feet deeper. The Dead Sea is over 1300 feet deep, and its surface is 1272 feet below the Mediterranean, so that its bottom is 2580 feet below the level of the sea. In the Caspian seals are found. A former connexion with the Red Sea has been claimed for the Dead Sea, but this is disallowed by Peschel and others. The Jordan valley, with the Sea of Tiberias and the Dead Sea, lie on the line of an extensive fault, and it is claimed that this depression in the surface occurred with the production of the fault. Further evidence in support of the statement that the Dead Sea was never connected with the sea is of a negative character, and consists chiefly in the fact that marine forms have not been found in the waters of the Jordan or of Lake Tiberias, and that silver is absent from the waters of the Dead Sea.

A former connexion with the ocean is claimed for a number of the Swiss and Italian lakes by Dr Forel and Professor Pavesi, and the Norwegian lakes by Loven and Sars, on the ground of the occurrence of marine forms of the crustaceans and other classes. For a summarized account of these researches see Pavesi, Arch. de Genève, 1880, iii. 1.

Temperature of Lakes.—The earliest reliable temperature observations in lakes or seas are those of Saussure, and they are to be found in his charming Voyage dans les Alpes. He was the first to obtain thoroughly trustworthy observations in the deeper waters of the lakes. He used for this purpose an ordinary thermometer whose bulb was covered over with several thicknesses of cloth and wax, so as to render it very slowly conducting. He was in the habit of leaving it down fourteen hours, and then bringing it up as quickly as possible and immediately reading the temperature. He did not, however, trust to his thermometer not changing its reading while being brought up, but by an elaborate series of experiments he obtained corrections, to be applied when the thermometer had to be drawn through more or less water of higher temperature. His observations are collected in the following table along with those of Jardine in some of the Scottish lakes, at the beginning of the century:—

Name of Lake.		Date.	Temperature of		Depth.	Height above Sea.
			Surface.	Bottom.
			° F.	° F.	Feet.	Feet.
Geneva		February.	42·1	41·6	1,013	1,230
Neuchâtel		17th July.	73·7	41·4	346	1,304
Bourget		October 1784.	64·0	42·1	256	...
Annecy		14th May 1780.	57·9	42·1	174	1,426
Joux		...	55·6	51·3	85	350
Bienne		...	69·3	44·4	231	1,419
Constance		25th July 1784.	64·6	39·6	394	1,250
Lucerne		28th July.	68·4	40·8	640	1,380
Thun		7th July 1783.	66·2	41·0	373	1,896
Brienz		8th July 1783.	68·0	40·5	533	...
Maggiore		19th July 1783.	78·1	44·1	357	...
Lomond		8th Sept.	59·5	41·5	600	25
Katrine	$\scriptstyle {\left.{\begin{matrix}\ \\\\\ \ \end{matrix}}\right\}\,}$	7th Sept. 1812.	57·3	41·0	480	364
		3d Sept. 1812.	56·4	41·3	...	...

An exceedingly important and valuable series of observations was made by Fischer and Brunner^[1] in the Lake of Thun throughout the course of a whole year (March 1848 to February 1849). They used, after Saussure’s method, thermometers protected by non-conducting envelopes, which were pulled up as quickly as possible. The depth of the water where they observed was 540 feet, and they made a series of observations of the temperature at that depth, at the surface, and at eleven intermediate depths, and repeated the series of observations at eight different dates over the year. From these series, which afford the first information of the yearly march of temperature at different depths, we learn that the lake as a whole gains heat till the end of September, then loses it until the month of February, when it begins to warm again, though slowly. The maximum temperature occurs in October at depths from the surface to 70 feet, in November at depths from 70 to 120 feet, in December from 120 to 200 feet, and in February at 500 feet. As the whole yearly variation of the temperature at 200 feet is less than a degree, the epoch at which the greater depths attain their maximum and minimum temperatures cannot be certainly deduced from one year’s observations. The minimum temperature of depths from the surface to 80 feet is attained in the month of February, at greater depths in the month of March. During the course of the whole year the temperature at the bottom varied between 40°·7 and 40°·9 Fahr., and in the month of February the whole of the water from the surface to the bottom was between 40°·7 and 41° Fahr.

These and other observations showed that, from depths of 400 feet, the variation of temperature with increasing depth is quite insignificant, so that even though the lake might be 1000 feet deep the temperature at 400 feet is only one or two tenths of a degree different from that of the bottom; further, on many of the thermometers recently used, it is impossible to distinguish with certainty temperatures differing by less than half a degree, consequently it was not difficult to believe that in all deep lakes there is a considerable stratum of water which remains constantly at the same temperature, all the year and every year, and that in winter this stratum thickens so as often to fill the lake, and gets thinner again in summer. By the improvement of the instruments both of these suppositions have been shown to be erroneous. In summer and in temperate latitudes, however deep the lake may be, its temperature falls as the depth increases, first rapidly and then very slowly, and the bottom temperature observed in any summer depends on the nature of the winter which preceded it, and may vary from year to year by one to two degrees. It was also believed that the deep water of a lake preserved constantly the mean winter temperature or the mean temperature of the six coldest months of the year in the locality. This was deduced from some observations by Sir Robert Christison in Loch Lomond, who found the bottom temperature at Tarbet to be 41°·4 Fahr., agreeing with the mean of the six winter months as observed at Balloch Castle, which, however, is about 15 miles distant. Although the theorem may be accidentally true for Loch Lomond, it has been proved not to hold for other lakes. Thus Simony (Wien. Sitz. Ber., 1875, lxxi. p. 435) gives the following table, comparing the temperature of the bottom water in the Gmünder See with the winter (October to March) air temperature:^[2]—

	Winter Period. Mean Temperature.			Summer Period. Mean Temp.	Bottom Temp., Gmünder See.	Date of Observation of Bottom Temperature.
	Oct.–Mar.	Dec.–Feb.
	° F.	° F.		° F.	° F.
1867–68	37·5	32·9	1868	64·4	40·5	6th Oct. 1868.
1868–89	40·1	36·8	1869	63·1	40·5	1st Oct. 1869.
1869–70	35·0	29·3	1870	60·8	40·2	26th Sept. 1870.
1871–72	35·2	27·8	1872	62·2	40·0	3d Oct. 1872.
1872–73	41·0	35·0	1873	60·2	40·5	5th Oct. 1873.
1873–74	39·0	32·7	1874	61·9	40·4	25th Sept. 1874.
1874–75	33·8	28·2	...	...	39·1	10th April 1875.

It will be seen that, with the exception of the end of 1872, the mean winter temperature is below that of the bottom water, and generally very markedly so.

During 1877–81 observations have been made by the present writer on the distribution of temperature in lakes forming part of the Caledonian Canal. The monthly mean temperatures at Culloden and at Corran Ferry lighthouse, which cannot differ much in climate from Loch Ness and Loch Lochy respectively, have been supplied by Mr Buchan of the Scottish Meteorological Society. The bottom temperatures are those observed in the deepest part of the lakes, namely, 120 fathoms in Loch Ness, and 80 fathoms in Loch Lochy. The connexion between bottom temperature (as observed in the second week of August) and winter temperature can be judged of from the following table, where the mean temperatures of October to March, and also of November to April, are given:—

	Loch Ness.		Culloden.		Loch Lochy.		Corran.
	Surface.	Bottom.	Oct. to March.	Nov. to April.	Surface.	Bottom.	Oct. to March.	Nov. to April.
	°	°	°	°	°	°	°	°
1877	53·0	42·4	40·2	40·0	55·0	44·0	42·3	40·8
1878	59·0	42·3	41·6	40·9	61·0	43·7	42·7	42·5
1879	51·4	41·2	37·2	35·8	54·0	42·0	38·9	37·5
1880	57·0	42·4	41·0	40·8	57·6	43·8	42·0	41·9
1881	53·1	41·45	36·1	36·2	54·0	42·25	38·6	38·7

From this table it is apparent that the bottom temperature, even of lakes as deep as Loch Ness, is subject to considerable variation from year to year, that it depends on the temperature of the previous winter, and that it is usually higher than that temperature. The difference between the bottom temperature and the mean winter temperature is greater the lower the winter temperature is. It is further interesting to notice that the mean winter temperature of 1878–79 was about one degree higher than that of 1880–81, yet the bottom temperatures were 0°·25 lower in 1879 than in 1881, and this is no doubt due to the fact that the cold of 1878–79 was more continuous than that of 1880–81, when the actual temperatures observed were much lower. The temperature of the bottom water depends not only on the temperature of the previous winter, and on the depth of the lake; it also depends on the nature of the country where it lies, and especially on its exposure to winds. Winds drive the surface water before them, and if there were no return current it would be heaped up at the further end. The effect is to accumulate surface water at one end, and to draw on deeper water to make up the deficiency at the other end. Hence the prevailing direction of the wind impresses itself on the distribution of temperature in the water; and this is well shown in the distribution of temperature as determined from observations at five stations on the same day in Loch Ness in a summer after a warm winter, and in one after a cold winter. In Scotland, warm weather is associated with southerly and westerly winds, and cold weather with northerly and easterly winds. In the warm years we have accumulation of surface water at the north-eastern end, and of bottom water at the south-western end, producing in summer a higher mean temperature of water at the north-east, and a lower mean temperature of water at the south-west end. In cold years the reverse is observed. Thus in 1879, after a cold winter, the mean temperature of the first 300 feet of water at the south-west end of Loch Ness was 48°·8, and at the north-east end 44°·96, a difference of nearly four degrees. In 1880, after a comparatively mild winter, it was 48°·13 at the south-west end, and 47°·95 at the north-east end, or nearly identical temperatures. Even at stations a few hundred yards from each other, great differences are often observed in the temperatures observed at the same depth, and it is evident that the difference of density so produced must cause a certain amount of circulation. There can be but little doubt that, under the influence of the varying temperature of the seasons, and of the winds, the water of a lake is thoroughly mixed once a year. In lakes which do not consist of a single long trough like Loch Ness, but of several basins as Loch Lomond, the bottom temperature is different in the different basins, even when the depth is the same. Loch Lomond consists of three principal basins of very unequal depth:—the large expanse of water studded with islands at the lower end, the Balloch basin; the middle or Luss basin; and the upper and deepest or Tarbet basin. In the last we have 600 feet of water, in the Luss basin 200 feet, and in the Balloch basin a maximum of 72 feet of water. On 23d September 1876 the bottom temperature in the Tarbet basin was 41°·4, and in the Luss basin 46°·4. Loch Tummel, a much smaller lake, consists of three basins, each of them being from 100 to 120 feet deep, and in them we have bottom temperatures of 46°·3, 46°·9, and 45°·2, the lowest temperature being nearest the outlet.

It might have been expected that the bottom temperature in lakes similar as regards size and depth would be lower at greater elevations and higher nearer the sea-level. This does not, however, hold universally; thus Lochs Tummel and Garry are very similar in size and depth; they are only 12 miles from each other, but Loch Tummel is 450 feet and Loch Garry 1330 feet above the sea; yet at 102 feet in Loch Garry the temperature on the 18th August 1876 was 53°·9, and in Loch Tummel at the same depth on the 16th August 1876 it was 45°·4. The difference of elevation is nearly 900 feet, and, instead of the higher lake holding the colder water, its water is 8°·5 warmer than that of the lower one. Similarly in Loch Ericht, 1153 feet above the sea, the bottom temperature at 324 feet was 44°·7, and in Loch Rannoch, 668 feet above sea, at the same depth it was 44°·0. These examples will suffice to show that many circumstances concur in determining the temperatures of the waters of lakes. There is one factor which is often neglected, namely, the amount of change of water. This depends on the drainage area of its tributary streams, and necessarily varies greatly.

In comparing the bottom temperature in lakes with the mean temperatures of the coldest half of the year, we find that the two approach each other more nearly the higher these temperatures are. When the temperature of the air falls for a lengthened period below the temperature of maximum density of water (39°·2 Fahr.), then the mechanical effect produced is much the same as if the temperature had been raised. For, in virtue of the cooling above, the water will have no tendency to sink; it will rather tend to float as a cold layer on the surface of the warmer and denser water below. Were a lake comparable with a glass of water, that is, were its depth equal to or greater than its length or breadth, it would be possible to realize this ideal condition of things, which, until recently, was supposed to represent what really takes place when a lake is covered with ice, namely, that after the water has all been cooled to a uniform temperature of 39°·2 Fahr. further cooling affects only a small surface layer, which consequently rapidly freezes. If this were the case, we should expect to find the temperature of the water below the ice of a frozen lake increasing rapidly from 32° where it is in contact with the ice to 39°·2 at a short distance from it, and we should expect to find the remainder of the water down to the bottom at the same temperature. In fact, however, the depth of even the deepest lakes bears an insignificant proportion to their superficial dimensions, and temperature observations in summer show that the effective climate, that is, the climate in so far as it is effective for the purpose under consideration, varies much over the surface of even very small lakes. The variations in distribution of temperature produce variations in density which of themselves are sufficient to produce convection currents. Then, as a factor of climate, there are the winds, which are the main mixing agents, and also the movement in the waters caused by the inflow of water at different points and the removal of the excess at one point. The effect of these mechanical agents, winds and currents, is to propagate the air temperature at the surface to a greater depth than would otherwise be the case. At the same time it must be remembered that in seasons of great cold there is rarely much wind. If we reflect, however, on what must take place when there is a large expanse of open water in the middle of a country covered with snow, and exposed to the rigours of a winter night, we see that the air in contact with the surface of the water must get warmed and form an ascending current, its place being taken by fresh air drafted from the cold land surface, which not only cools the water but forces it out towards the middle, thus establishing a circulation consisting in broad lines of a surface movement from the sides to the middle of the lake, and a movement in the opposite direction below the surface. Even if the current of air were not sufficient of itself to produce a surface current in the water, it would do it indirectly. For, as it first strikes the water at the edges, the water there would get cooled most rapidly, and under suitable circumstances would form a fringe of ice; the water so cooled would be lighter than the warmer water farther out, and would have a tendency to flow off towards the middle, or with the current of air. Now, although, when compared with other seasons, there is in a hard frosty winter not much wind, still, even in the calmest weather there is almost always sufficient motion in the atmosphere to enable the meteorologist to state that the wind is from a particular quarter; this will assist the circulation which has just been described as taking place in a calm lake, though it will somewhat distort its effects. It will produce excessive cooling at the side nearest the wind, and, when the lake freezes, it will have a tendency to begin at the windward side.

The extent to which this circulation affects the deeper waters of a lake depends on local circumstances, and generally we may say that the more confined a lake is the more easily will it freeze, and the higher will be the mean temperature of its waters. In the very cold winter 1878–79 the writer was able to make observations on the temperature of the water under the ice in Linlithgow Loch and in Loch Lomond. In the following winter, which, though mild in Scotland, was excessively severe in Switzerland, Dr Forel made observations in the Lakes of Morat and Zürich, confirming the writer’s observations of the unexpectedly low temperature of the water. The freezing of so deep a lake as that of Zürich was a fortunate circumstance, because in it the bottom is actually at the temperature of maximum density. The majority of the lakes which freeze are so shallow as to admit of the whole of their water being cooled considerably below the temperature of maximum density.

The distribution of temperature in frozen lakes will be apparent from the table given below. Of the Lakes of Zürich and Morat and Loch Lomond the mean temperatures are in the order of their depth. Linlithgow is altogether peculiar. Its high temperature, which increased steadily all the time it was covered with ice, was due to chemical action amongst the filth which has been allowed to accumulate at its bottom. When the ice broke up the dead fish were taken away in carts.

Dr Forel gives the following particulars about the frozen Swiss lakes. “The Lake of Morat has a surface of 27·4 square kilometres and a maximum depth of 45 metres (147 feet); it is 1425 feet above the sea; and its mean latitude is 49° 56′ N. The ice overspread its whole surface suddenly in the night of the 17th to the 18th December, and it remained frozen till the 8th March. The Lake of Zürich has a superficies of 87·8 square kilometres, a maximum depth of 468 feet and altitude of 1338 feet, and a mean latitude of 47° 16′ N. Its congelation was gradual, and not sudden like that of the Lake of Morat. First the upper part of the lake was covered with ice between Männedorf and Wädensweil. At the end of December, the 28th, the ice covered it entirely, but only for a single day. On the 29th it thawed, and the lake remained partially free of ice until the middle of January. It froze over completely on the 22d January, and on the 25th the ice was 4 inches thick in the centre of the lake.” Of the larger Swiss lakes, Morat, Zürich, Zug, Neuchâtel, Constance, and Annecy were frozen in 1880; Thun is known to have been frozen four times, namely, in 1363, 1435, 1685, and 1695; Brienz has only once been frozen, in 1363; Lucerne freezes partially in very severe winters, and Geneva in its western and shallower part, whilst Wallenstadt and Bourget are not known to have ever been frozen.

Table of Temperatures in Frozen Lakes.

Depth (in feet).		Temperature in Degrees Fahr.
		Zürich, 25th Jan. 1880.	Morat, 23d Dec. 1879.	Lomond, 29th Jan. 1879.	Linlithgow.
					11th Jan. 1879.	25th Jan. 1879.
	3	...	...	33·00	35·90	36·00
	6	...	...	33·50	36·30	36·80
	18	...	35·06	33·95	36·90	37·80
(Bottom)	48	36·95	36·14	35·20	39·85	42·05
(Bottom)	65	37·25	36·30	36·30	...	...
	100	37·76	36·68	...	...	...
(Bottom)	150	38·39	37·04	...	...	...
	200	38·66	...	...	...	...
	300	38·84	...	...	...	...
(Bottom)	435	39·20	...	...	...	...
Mean		38·40	36·00	34·46	37·22	38·28

For further information on the temperature of frozen lakes, see Buchanan, Nature, March 6, 1879; Forel, Arch. de Genève, 1880, iv. 1; Nichols, Proc. Boston Soc. of Nat. Hist., 1881, xxi. p. 53.

Changes of Level.—As the water supply of lakes depends on the rainfall, and as this varies much with the season, and from year to year, we should expect, and indeed we find, fluctuation of level in all lakes. There are, however, other changes of level which are independent of the water supply, and which resemble tides in their rhythmic periods. They have long been known and observed in Switzerland, and especially on the Lake of Geneva, where they are known by the name of “seiches.” The level of the lake is observed to rise slowly during twenty or thirty minutes to a height which varies from a few centimetres to as many decimetres; it then falls again slowly to a corresponding depth, and rises again slowly, and so on. These movements were observed and much studied at the end of last century by Jallabert, Bertrand, and Saussure, and at the beginning of this century they formed the subject of an instructive memoir by Vaucher, who enunciated the following law connecting the seiches with the movements of the barometer. “The amplitude of seiches is small when the atmosphere is at rest; the seiches are greater the more variable is the atmosphere’s pressure; they are the greatest when the barometer is falling.” Vaucher recognized the existence of seiches in the Lakes of Geneva, Neuchâtel, Zürich, Constance, Annecy, and Lugano, and Dr Forel of Morges, from whose papers, published principally in the Bibliothèque Universelle et Revue Suisse during the last five years, the facts regarding the seiches have been taken, has observed them in every lake where he had looked for them. It is in every way likely that they are to be found in all lakes of notable extent and depth. They have been studied principally on the Lake of Geneva, where Dr Forel, at Morges, about the middle of the lake on the north shore, and M. Plantamour, at Sécheron, about a mile from Geneva on the north shore, have had self-registering tide gauges in operation for a number of years. In the writings of the Swiss observers the seiche is the complete movement of rise above and fall below the mean level, the amplitude is the extreme difference of level so produced, and the duration of the seiche is the time in seconds measured from the moment when the water is at the mean level until it is again at the mean level, after having risen to the crest and sunk to the trough of the wave. The amplitude of the seiches is very variable. At the same station and on the same day successive seiches are similar. When the seiches are small they are all small, when they are large they are all large. At the same station and on different days the amplitudes of the seiches may vary enormously. For instance, at Geneva, where the highest seiches have been observed, they are usually of such a size as to be imperceptible without special instruments; yet on the 3d August 1763 Saussure measured seiches of 1·48 metres, and on the 2d and 3d October 1811 the seiches observed by Vénié were as much as 2·15 metres. They are greater at the extremities than at the middle of lakes, at the head of long gulfs whose sides converge gently than at stations in the middle of a long straight coast, and in shallow as compared with deep lakes or parts of a lake. They also appear to increase with the size of the lake. The duration of the seiches is found to vary considerably, but the mean deduced from a sufficient number of observations is fairly constant at the same locality. Thus, for Morges, Dr Forel has found it to be for the half seiche 315±9 seconds. At different stations, however, on the same lake and on different lakes it varies considerably. Thus on the Lake of Geneva it is, for the complete seiche, 630 seconds at Merges, and 1783 seconds at Veytaux; on Lake Neuchâtel it is 2840 seconds at Yverdon, and 264 at Saint Aubin.

The curves traced by the gauge at Geneva have been subjected to a preliminary harmonic analysis by Professor Soret, and he has decomposed them into two undulations, the one with a period, from crest to crest, of seventy-two minutes, and the other with a period of thirty-five minutes, or a little less than half the larger period. As the amplitudes of the composing curves vary much, there is great variety in the resultant curves. Besides these two principal components, there are others which have not yet been investigated.

With regard to the cause of the phenomenon, Dr Forel attributes the ordinary seiches to local variations of atmospheric pressure, giving an impulse the effect of which would be apparent for a long time as a series of oscillations. The greater seiches, such as those of 1·5 metres, he attributed to earthquake shocks; but, as a very sensible earthquake passed over Switzerland quite recently without leaving the slightest trace on the gauge, he has abandoned this explanation, and is inclined to attribute them to pulsation set agoing by violent downward gusts of wind, especially at the upper end of the lake. M. Plantamour, who has devoted much attention to the same subject, assured the writer, in the summer of 1881, that he was completely at a loss for a satisfactory explanation of them.

Seiches have not been observed on the Scottish lakes, though there is little doubt that they would be found if sought for. There are, however, records of disturbances of some of the lakes, especially in Perthshire, of which the following may be cited as an instance.

A violent disturbance of the level of Loch Tay is reported in the Statistical Account of Scotland (1796), xvii. p. 458, to have occurred at Kenmore on 12th September 1784, continuing in a modified degree for four days, and again on 13th July 1794. Kenmore lies at the north-eastern end of the lake, where the river Tay issues from it. It lies at the end of a shallow bay. “At the extremity of this bay the water was observed to retire about 5 yards within its ordinary boundary, and in four or five minutes to flow out again. In this manner it ebbed and flowed successively three or four times during the space of a quarter of an hour, when all at once the water rushed from the east and west in opposite currents, . . . . rose in the form of a great wave, to the height of 5 feet above the ordinary level, leaving the bottom of the bay dry to the distance of between 90 and 100 yards from its natural boundary. When the opposite currents met they made a clashing noise and foamed; and, the stronger impulse being from the east, the wave after rising to its greatest height, rolled westward, but slowly diminishing as it went, for the space of five minutes, when it wholly disappeared. As the wave subsided it flowed back with some force, and exceeded its original boundary 4 or 5 yards; then it ebbed again about 10 yards, and again returned, and continued to ebb and flow in this manner for the space of two hours, the ebbings succeeding each other, at the distance of about seven minutes, and gradually lessening, till the water settled into its ordinary level. During the whole time that this phenomenon was observed the weather was calm. On the next and four succeeding days an ebbing and flowing was observed nearly about the same time and for the same length of time, but not at all in the same degree as on the first day.”

The above is the account given by the Rev. Thomas Fleming, at the time minister of Kenmore, who was an eye witness. It resembles in all essential particulars the descriptions of waves which accompany actual earthquakes, yet in his account he goes on to say—“I have not heard (although I have made particular inquiry) that any motion of the earth was felt in this neighbourhood, or that the agitation of the wave was observed anywhere but about the village of Kenmore.” It is well known that there were great seismic movements observed in Perthshire at the time of the Lisbon earthquake, and there is a tradition in the neighbourhood that Loch Lubnaig near Callander was largely increased in extent by the dislocations which took place.

In all lakes there are changes of level corresponding with periods of rain and of drought. They are the more considerable the greater the extent of country draining into them, and the more constrained the outflow. In the great American lakes, which occupy nearly one-third of their drainage area, the fluctuations of level are quite insignificant; in Lake Michigan the U.S. surveyors give as the maximum and minimum yearly range 1·64 and 0·65 feet. In the Lake of Geneva the mean annual oscillation is 5 feet, and the difference between the highest and the lowest waters of this century is 9·3 feet. The most rapid rise has been 3·23 inches (82 mm.) in twenty-four hours. A very remarkable exception to the rule that large freshwater lakes are subject to small variations of level is furnished by Lake Tanganyika in Central Africa. Since its discovery travellers have been much perplexed by the evidence and reports of considerable oscillations of level of uncertain period, and also by the apparent absence of visible outlet, while the freshness of its waters was of itself convincing evidence of the existence of an outlet. By the careful observations of successive explorers the nature of this phenomenon has been fully explained, and is very instructive. It has recently been visited by Captain Hore of the London Missionary Society, and it appears from his reports that the peculiar phenomena observed depend on the fact that the area of country draining into the lake is very limited, so that in the dry seasons the streams running into it dry up altogether, and its outlet gets choked by the rapid growth of vegetation in an equatorial climate. A dam or dyke is thus formed which is not broken down until the waters of the lake have risen to a considerable height. A catastrophe of this kind happened whilst Captain Hore was in the neighbourhood, and he noted the height of the water at different times near his station at Ujiji, and observed it fall 2 feet in two months It continued to fall until in seventeen months it had fallen over 10 feet. Taking the length of the lake at 330 miles, and the mean breadth at 30 miles, its surface is 9900 square nautical miles. If this surface be reduced 2 feet in sixty days, the water will have to escape at the rate of 137,500 cubic feet per second. The mean rate of discharge of the Danube is 207,000 cubic feet per second. Hence, without taking into account water which would be brought into the lake by tributaries during the two months, we require for cutlet a river at least two-thirds of the size of the Danube, and in the Lukuga such a river is found. When Stanley visited it the Lukuga was quite stopped up with dense growth, and no water was issuing; the lake was then rising; when Captain Hore visited it the lake was falling rapidly, and the Lukuga was a rapid river of great volume. One of the chief affluents to the lake was found to be discharging at the rate of 18,750 cubic feet of water per second; a few months later it was dry and the mouth closed with vegetation. During the dry season too the lake, with its 10,000 square miles of surface, is exposed to the evaporating action of the south-east trade wind, and when the supply is so insignificant this must be sufficient of itself to sensibly lower the level. Ordinarily then we might expect the lake to be subject to a yearly ebb and flow corresponding to the periods of drought and rains; and, from what we learn of the great fluctuations of rainfall one year with another, we should expect that during a series of dry years the obstructions to the outflow would gain such a head

...

---

1. Mém. Soc. Phys. Genève, xii. p. 255.
2. These air temperatures are those of the observatory at Vienna, corrected for difference of level.