[Footnote 56: Chamberlin and Salisbury: Geology, 1906, Vol. III, pp.

405-412.]

[Footnote 57: It may have retreated soon after reaching its maximum. If so, the general lack of thick terminal moraines would be explained. See page 122.]

CHAPTER X

CAUSES OF MILD GEOLOGICAL CLIMATES

In discussions of climate, as of most subjects, a peculiar psychological phenomenon is observable. Everyone sees the necessity of explaining conditions different from those that now exist, but few realize that present conditions may be abnormal, and that they need explanation just as much as do others. Because of this tendency glaciation has been discussed with the greatest fullness, while there has been much neglect not only of the periods when the climate of the earth resembled that of the present, but also of the vastly longer periods when it was even milder than now.

How important the periods of mild climate have been in geological times may be judged from the relative length of glacial compared with inter-glacial epochs, and still more from the far greater relative length of the mild parts of periods and eras when compared with the severe parts. Recent estimates by R. T. Chamberlin[58] indicate that according to the consensus of opinion among geologists the average inter-glacial epoch during the Pleistocene was about five times as long as the average glacial epoch, while the whole of a given glacial epoch averaged five times as long as the period when the ice was at a maximum.

Climatic periods far milder, longer, and more monotonous than any inter-glacial epoch appear repeatedly during the course of geological history. Our task in this chapter is to explain them.

Knowlton[59] has done geology a great service by collecting the evidence as to the mild type of climate which has again and again prevailed in the past. He lays special stress on botanical evidence since that pertains to the variable atmosphere of the lands, and hence furnishes a better guide than does the evidence of animals that lived in the relatively unchanging water of the oceans. The nature of the evidence has already been indicated in various parts of this book. It includes palms, tree ferns, and a host of other plants which once grew in regions which are now much too cold to support them. With this must be placed the abundant reef-building corals and other warmth-loving marine creatures in lat.i.tudes now much too cold for them. Of a piece with this are the conditions of inter-glacial epochs in Europe, for example, when elephants and hippopotamuses, as well as many species of plants from low lat.i.tudes, were abundant. These conditions indicate not only that the climate was warmer than now, but that the contrast from season to season was much less. Indeed, Knowlton goes so far as to say that "relative uniformity, mildness, and comparative equability of climate, accompanied by high humidity, have prevailed over the greater part of the earth, extending to, or into, polar circles, during the greater part of geologic time--since, at least, the Middle Paleozoic. This is the regular, the ordinary, the normal condition." ... "By many it is thought that one of the strongest arguments against a gradually cooling globe and a humid, non-zonally disposed climate in the ages before the Pleistocene is the discovery of evidences of glacial action practically throughout the entire geologic column. Hardly less than a dozen of these are now known, ranging in age from Huronian to Eocene. It seems to be a very general a.s.sumption by those who hold this view that these evidences of glacial activities are to be cla.s.sed as ice ages, largely comparable in effect and extent to the Pleistocene refrigeration, but as a matter of fact only three are apparently of a magnitude to warrant such designation. These are the Huronian glaciation, that of the 'Permo-Carboniferous,' and that of the Pleistocene. The others, so far as available data go, appear to be explainable as more or less local manifestations that had no widespread effect on, for instance, ocean temperatures, distribution of life, et cetera. They might well have been of the type of ordinary mountain glaciers, due entirely to local elevation and precipitation." ... "If the sun had been the princ.i.p.al source of heat in pre-Pleistocene time, terrestrial temperatures would of necessity have been disposed in zones, whereas the whole trend of this paper has been the presentation of proof that these temperatures were distinctly non-zonal. Therefore it seems to follow that the sun--at least the present small-angle sun--could not have been the sole or even the princ.i.p.al source of heat that warmed the early oceans."

Knowlton is so strongly impressed by the widespread fossil floras that usually occur in the middle parts of the geological periods, that as Schuchert[3] puts it, he neglects the evidence of other kinds. In the middle of the periods and eras the expansion of the warm oceans over the continents was greatest, while the lands were small and hence had more or less insular climates of the oceanic type. At such times, the marine fauna agrees with the flora in indicating a mild climate. Large colony-forming foraminifera, stony corals, sh.e.l.led cephalopods, gastropods and thick-sh.e.l.led bivalves, generally the cemented forms, were common in the Far North and even in the Arctic. This occurred in the Silurian, Devonian, Pennsylvanian, and Jura.s.sic periods, yet at other times, such as the Cretaceous and Eocene, such forms were very greatly reduced in variety in the northern regions or else wholly absent. These things, as Schuchert[60] says, can only mean that Knowlton is right when he states that "climatic zoning such as we have had since the beginning of the Pleistocene did not obtain in the geologic ages prior to the Pleistocene." It does not mean, however, that there was a "non-zonal arrangement" and that the temperature of the oceans was everywhere the same and "without widespread effect on the distribution of life."

Students of paleontology hold that as far back as we can go in the study of plants, there are evidences of seasons and of relatively cool climates in high lat.i.tudes. The cycads, for instance, are one of the types most often used as evidence of a warm climate. Yet Wieland,[61]

who has made a lifelong study of these plants, says that many of them "might well grow in temperate to cool climates. Until far more is learned about them they should at least be held as valueless as indices of tropic climates." The inference is "that either they or their close relatives had the capacity to live in every clime. There is also a suspicion that study of the a.s.sociated ferns may compel revision of the long-accepted view of the universality of tropic climates throughout the Mesozoic." Nathorst is quoted by Wieland as saying, "I think ... that during the time when the Gingkophytes and Cycadophytes dominated, many of them must have adapted themselves for living in cold climates also.

Of this I have not the least doubt."

Another important line of evidence which Knowlton and others have cited as a proof of the non-zonal arrangement of climate in the past, is the vast red beds which are found in the Proterozoic, late Silurian, Devonian, Permian, and Tria.s.sic, and in some Tertiary formations. These are believed to resemble laterite, a red and highly oxidized soil which is found in great abundance in equatorial regions. Knowlton does not attempt to show that the red beds present equatorial characteristics in other respects, but bases his conclusion on the statement that "red beds are not being formed at the present time in any desert region." This is certainly an error. As has already been said, in both the Transcaspian and Takla Makan deserts, the color of the sand regularly changes from brown on the borders to pale red far out in the desert. Kuzzil k.u.m, or Red Sand, is the native name. The sands in the center of the desert apparently were originally washed down from the same mountains as those on the borders, and time has turned them red. Since the same condition is reported from the Arabian Desert, it seems that redness is characteristic of some of the world's greatest deserts. Moreover, beds of salt and gypsum are regularly found in red beds, and they can scarcely originate except in deserts, or in shallow almost landlocked bays on the coasts of deserts, as appears to have happened in the Silurian where marine fossils are found interbedded with gypsum.

Again, Knowlton says that red beds cannot indicate deserts because the plants found in them are not "pinched or depauperate, nor do they indicate xerophytic adaptations. Moreover, very considerable deposits of coal are found in red beds in many parts of the world, which implies the presence of swamps but little above sea-level."

Students of desert botany are likely to doubt the force of these considerations. As MacDougal[62] has shown, the variety of plants in deserts is greater than in moist regions. Not only do xerophytic desert species prevail, but halophytes are present in the salty areas, and hygrophytes in the wet swampy areas, while ordinary mesophytes prevail along the water courses and are washed down from the mountains. The ordinary plants, not the xerophytes, are the ones that are chiefly preserved since they occur in most abundance near streams where deposition is taking place. So far as swamps are concerned, few are of larger size than those of Seistan in Persia, Lop Nor in Chinese Turkestan, and certain others in the midst of the Asiatic deserts.

Streams flowing from the mountains into deserts are almost sure to form large swamps, such as those along the Tarim River in central Asia. Lake Chad in Africa is another example. In it, too, reeds are very numerous.

Putting together the evidence on both sides in this disputed question, it appears that throughout most of geological time there is some evidence of a zonal arrangement of climate. The evidence takes the form of traces of cool climates, of seasons, and of deserts. Nevertheless, there is also strong evidence that these conditions were in general less intense than at present and that times of relatively warm, moist climate without great seasonal extremes have prevailed very widely during periods much longer than those when a zonal arrangement as marked as that of today prevailed. As Schuchert[63] puts it: "Today the variation on land between the tropics and the poles is roughly between 110 and -60F., in the oceans between 85 and 31F. In the geologic past the temperature of the oceans for the greater parts of the periods probably was most often between 85 and 55F., while on land it may have varied between 90 and 0F. At rare intervals the extremes were undoubtedly as great as they are today. The conclusion is therefore that at all times the earth had temperature zones, varying between the present-day intensity and times which were almost without such belts, and at these latter times the greater part of the earth had an almost uniformly mild climate, without winters."

It is these mild climates which we must now attempt to explain. This leads us to inquire what would happen to the climate of the earth as a whole if the conditions which now prevail at times of few sunspots were to become intensified. That they could become greatly intensified seems highly probable, for there is good reason to think that aside from the sunspot cycle the sun's atmosphere is in a disturbed condition. The prominences which sometimes shoot out hundreds of thousands of miles seem to be good evidence of this. Suppose that the sun's atmosphere should become very quiet. This would apparently mean that cyclonic storms would be much less numerous and less severe than during the present times of sunspot minima. The storms would also apparently follow paths in middle lat.i.tudes somewhat as they do now when sunspots are fewest. The first effect of such a condition, if we can judge from what happens at present, would be a rise in the general temperature of the earth, because less heat would be carried aloft by storms. Today, as is shown in _Earth and Sun_, a difference of perhaps 10 per cent in the average storminess during periods of sunspot maxima and minima is correlated with a difference of 3C. in the temperature at the earth's surface. This includes not only an actual lowering of 0.6C. at times of sunspot maxima, but the overcoming of the effect of increased insolation at such times, an effect which Abbot calculates as about 2.5C. If the storminess were to be reduced to one-half or one-quarter its present amount at sunspot minima, not only would the loss of heat by upward convection in storms be diminished, but the area covered by clouds would diminish so that the sun would have more chance to warm the lower air.

Hence the average rise of temperature might amount to as much at 5 or 10C.

Another effect of the decrease in storminess would be to make the so-called westerly winds, which are chiefly southwesterly in the northern hemisphere and northwesterly in the southern hemisphere, more strong and steady than at present. They would not continually suffer interruption by cyclonic winds from other directions, as is now the case, and would have a regularity like that of the trades. This conclusion is strongly reenforced in a paper by Clayton[64] which came to hand after this chapter had been completed. From his studies of the solar constant and the temperature of the earth which are described in _Earth and Sun_, he reaches the following conclusion: "The results of these researches have led me to believe: 1. That if there were no variation in solar radiation the atmospheric motions would establish a stable system with exchanges of air between equator and pole and between ocean and land, in which the only variations would be daily and annual changes set in operation by the relative motions of the earth and sun.

2. The existing abnormal changes, which we call weather, have their origins chiefly, if not entirely, in the variations of solar radiation."

If cyclonic storms and "weather" were largely eliminated and if the planetary system of winds with its steady trades and southwesterlies became everywhere dominant, the regularity and volume of the poleward-flowing currents, such as the Gulf Stream and the Atlantic Drift in one ocean, and the j.a.panese Current in another, would be greatly increased. How important this is may be judged from the work of h.e.l.land-Hansen and Nansen.[65] These authors find that with the pa.s.sage of each cyclonic storm there is a change in the temperature of the surface water of the Atlantic Ocean. Winds at right angles to the course of the Drift drive the water first in one direction and then in the other but do not advance it in its course. Winds with an easterly component, on the other hand, not only check the Drift but reverse it, driving the warm water back toward the southwest and allowing cold water to well up in its stead. The driving force in the Atlantic Drift is merely the excess of the winds with a westerly component over those with an easterly component.

Suppose that the numbers in Fig. 8 represent the strength of the winds in a certain part of the North Atlantic or North Pacific, that is, the total number of miles moved by the air per year. In quadrant A of the left-hand part all the winds move from a more or less southwesterly direction and produce a total movement of the air amounting to thirty units per year. Those coming from points between north and west move twenty-five units; those between north and east, twenty units; and those between east and south, twenty-five units. Since the movement of the winds in quadrants B and D is the same, these winds have no effect in producing currents. They merely move the water back and forth, and thus give it time to lose whatever heat it has brought from more southerly lat.i.tudes. On the other hand, since the easterly winds in quadrant C do not wholly check the currents caused by the westerly winds of quadrant A, the effective force of the westerly winds amounts to ten, or the difference between a force of thirty in quadrant A and of twenty in quadrant C. Hence the water is moved forward toward the northeast, as shown by the thick part of arrow A.

[Ill.u.s.tration: _Fig. 8. Effect of diminution of storms on movement of water._]

Now suppose that cyclonic storms should be greatly reduced in number so that in the zone of prevailing westerlies they were scarcely more numerous than tropical hurricanes now are in the trade-wind belt. Then the more or less southwesterly winds in quadrant A' in the right-hand part of Fig. 8 would not only become more frequent but would be stronger than at present. The total movement from that quarter might rise to sixty units, as indicated in the figure. In quadrants B' and D' the movement would fall to fifteen and in quadrant C' to ten. B' and D'

would balance one another as before. The movement in A', however, would exceed that in C' by fifty instead of ten. In other words, the current-making force would become five times as great as now. The actual effect would be increased still more, for the winds from the southwest would be stronger as well as steadier if there were no storms. A strong wind which causes whitecaps has much more power to drive the water forward than a weaker wind which does not cause whitecaps. In a wave without a whitecap the water returns to practically the original point after completing a circle beneath the surface. In a wave with a whitecap, however, the cap moves forward. Any increase in velocity beyond the rate at which whitecaps are formed has a great influence upon the amount of water which is blown forward. Several times as much water is drifted forward by a persistent wind of twenty miles an hour as by a ten-mile wind.[66]

In this connection a suggestion which is elaborated in Chapter XIII may be mentioned. At present the salinity of the oceans checks the general deep-sea circulation and thereby increases the contrasts from zone to zone. In the past, however, the ocean must have been fresher than now.

Hence the circulation was presumably less impeded, and the transfer of heat from low lat.i.tudes to high was facilitated.

Consider now the magnitude of the probable effect of a diminution in storms. Today off the coast of Norway in lat.i.tude 65N. and longitude 10E., the mean temperature in January is 2C. and in July 12C. This represents a plus anomaly of about 22 in January and 2 in July; that is, the Norwegian coast is warmer than the normal for its lat.i.tude by these amounts. Suppose that in some past time the present distribution of lands and seas prevailed, but Norway was a lowland where extensive deposits could acc.u.mulate in great flood plains. Suppose, also, that the sun's atmosphere was so inactive that few cyclonic storms occurred, steady winds from the west-southwest prevailed, and strong, uninterrupted ocean currents brought from the Caribbean Sea and Gulf of Mexico much greater supplies of warm water than at present. The Norwegian winters would then be warmer than now not only because of the general increase in temperature which the earth regularly experiences at sunspot minima, but because the currents would accentuate this condition. In summer similar conditions would prevail except that the warming effect of the winds and currents would presumably be less than in winter, but this might be more than balanced by the increased heat of the sun during the long summer days, for storms and clouds would be rare.

If such conditions raised the winter temperature only 8C. and the summer temperature 4C., the climate would be as warm as that of the northern island of New Zealand (lat.i.tude 35-43S.). The flora of that part of New Zealand is subtropical and includes not only pines and beeches, but palms and tree ferns. A climate scarcely warmer than that of New Zealand would foster a flora like that which existed in far northern lat.i.tudes during some of the milder geological periods. If, however, the general temperature of the earth's surface were raised 5 because of the scarcity of storms, if the currents were strong enough so that they increased the present anomaly by 50 per cent, and if more persistent sunshine in summer raised the temperature at that season about 4C., the January temperature would be 18C. and the July temperature 22C. These figures perhaps make summer and winter more nearly alike than was ever really the case in such lat.i.tudes.

Nevertheless, they show that a diminution of storms and a consequent strengthening and steadying of the southwesterlies might easily raise the temperature of the Norwegian coast so high that corals could flourish within the Arctic Circle.

Another factor would cooperate in producing mild temperatures in high lat.i.tudes during the winter, namely, the fogs which would presumably acc.u.mulate. It is well known that when saturated air from a warm ocean is blown over the lands in winter, as happens so often in the British Islands and around the North Sea, fog is formed. The effect of such a fog is indeed to shut out the sun's radiation, but in high lat.i.tudes during the winter when the sun is low, this is of little importance.

Another effect is to retain the heat of the earth itself. When a constant supply of warm water is being brought from low lat.i.tudes this blanketing of the heat by the fog becomes of great importance. In the past, whenever cyclonic storms were weak and westerly winds were correspondingly strong, winter fogs in high lat.i.tudes must have been much more widespread and persistent than now.

The bearing of fogs on vegetation is another interesting point. If a region in high lat.i.tudes is constantly protected by fog in winter, it can support types of vegetation characteristic of fairly low lat.i.tudes, for plants are oftener killed by dry cold than by moist cold. Indeed, excessive evaporation from the plant induced by dry cold when the evaporated water cannot be rapidly replaced by the movement of sap is a chief reason why large plants are winterkilled. The growing of transplanted palms on the coast of southwestern Ireland, in spite of its location in lat.i.tude 50N., is possible only because of the great fogginess in winter due to the marine climate. The fogs prevent the escape of heat and ward off killing frosts. The tree ferns in lat.i.tude 46S. in New Zealand, already referred to, are often similarly protected in winter. Therefore, the relative frequency of fogs in high lat.i.tudes when storms were at a minimum would apparently tend not merely to produce mild winters but to promote tropical vegetation.

The strong steady trades and southwesterlies which would prevail at times of slight solar activity, according to our hypothesis, would have a p.r.o.nounced effect on the water of the deep seas as well as upon that of the surface. In the first place, the deep-sea circulation would be hastened. For convenience let us speak of the northern hemisphere. In the past, whenever the southwesterly winds were steadier than now, as was probably the case when cyclonic storms were relatively rare, more surface water than at present was presumably driven from low lat.i.tudes and carried to high lat.i.tudes. This, of course, means that a greater volume of water had to flow back toward the equator in the lower parts of the ocean, or else as a cool surface current. The steady southwesterly winds, however, would interfere with south-flowing surface currents, thus compelling the polar waters to find their way equatorward beneath the surface. In low lat.i.tudes the polar waters would rise and their tendency would be to lower the temperature. Hence steadier westerlies would make for lessened lat.i.tudinal contrasts in climate not only by driving more warm water poleward but by causing more polar water to reach low lat.i.tudes.

At this point a second important consideration must be faced. Not only would the deep-sea circulation be hastened, but the ocean depths might be warmed. The deep parts of the ocean are today cold because they receive their water from high lat.i.tudes where it sinks because of low temperature. Suppose, however, that a diminution in storminess combined with other conditions should permit corals to grow in lat.i.tude 70N. The ocean temperature would then have to average scarcely lower than 20C.

and even in the coldest month the water could scarcely fall below about 15C. Under such conditions, if the polar ocean were freely connected with the rest of the oceans, no part of it would probably have a temperature much below 10C., for there would be no such thing as ice caps and snowfields to reflect the scanty sunlight and radiate into s.p.a.ce what little heat there was. On the contrary, during the winter an almost constant state of dense fogginess would prevail. So great would be the blanketing effect of this that a minimum monthly temperature of 10C. for the coldest part of the ocean may perhaps be too low for a time when corals thrived in lat.i.tude 70.

The temperature of the ocean depths cannot permanently remain lower than that of the coldest parts of the surface. Temporarily this might indeed happen when a solar change first reduced the storminess and strengthened the westerlies and the surface currents. Gradually, however, the persistent deep-sea circulation would bring up the colder water in low lat.i.tudes and carry downward the water of medium temperature at the coldest part of the surface. Thus in time the whole body of the ocean would become warm. The heat which at present is carried away from the earth's surface in storms would slowly acc.u.mulate in the oceans. As the process went on, all parts of the ocean's surface would become warmer, for equatorial lat.i.tudes would be less and less cooled by cold water from below, while the water blown from low lat.i.tudes to high would be correspondingly warmer. The warming of the ocean would come to an end only with the attainment of a state of equilibrium in which the loss of heat by radiation and evaporation from the ocean's surface equaled the loss which under other circ.u.mstances would arise from the rise of warm air in cyclonic storms. When once the oceans were warmed, they would form an extremely strong conservative force tending to preserve an equable climate in all lat.i.tudes and at all seasons. According to the solar cyclonic hypothesis such conditions ought to have prevailed throughout most of geological time. Only after a strong and prolonged solar disturbance with its consequent storminess would conditions like those of today be expected.

In this connection another possibility may be mentioned. It is commonly a.s.sumed that the earth's axis is held steadily in one direction by the fact that the rotating earth is a great gyroscope. Having been tilted to a certain position, perhaps by some extraneous force, the axis is supposed to maintain that position until some other force intervenes.

Cordeiro,[67] however, maintains that this is true only of an absolutely rigid gyroscope. He believes that it is mathematically demonstrable that if an elastic gyroscope be gradually tilted by some extraneous force, and if that force then ceases to act, the gyroscope as a whole will oscillate back and forth. The earth appears to be slightly elastic.

Cordeiro therefore applies his formulae to it, on the following a.s.sumptions: (1) That the original position of the axis was nearly vertical to the plane of the ecliptic in which the earth revolves around the sun; (2) that at certain times the inclination has been even greater than now; and (3) that the position of the axis with reference to the earth has not changed to any great extent, that is, the earth's poles have remained essentially stationary with reference to the earth, although the whole earth has been gyroscopically tilted back and forth repeatedly.

With a vertical axis the daylight and darkness in all parts of the earth would be of equal duration, being always twelve hours. There would be no seasons, and the climate would approach the average condition now experienced at the two equinoxes. On the whole the climate of high lat.i.tudes would give the impression of being milder than now, for there would be less opportunity for the acc.u.mulation of snow and ice with their strong cooling effect. On the other hand, if the axis were tilted more than now, the winter nights would be longer and the winters more severe than at present, and there would be a tendency toward glaciation.

Thus Cordeiro accounts for alternating mild and glacial epochs. The entire swing from the vertical position to the maximum inclination and back to the vertical may last millions of years depending on the earth's degree of elasticity. The swing beyond the vertical position in the other direction would be equally prolonged. Since the axis is now supposed to be much nearer its maximum than its minimum degree of tilting, the duration of epochs having a climate more severe than that of the present would be relatively short, while the mild epochs would be long.

Cordeiro's hypothesis has been almost completely ignored. One reason is that his treatment of geological facts, and especially his method of riding rough-shod over widely accepted conclusions, has not commended his work to geologists. Therefore they have not deemed it worth while to urge mathematicians to test the a.s.sumptions and methods by which he reached his results. It is perhaps unfair to test Cordeiro by geology, for he lays no claim to being a geologist. In mathematics he labors under the disadvantage of having worked outside the usual professional channels, so that his work does not seem to have been subjected to sufficiently critical a.n.a.lysis.

Without expressing any opinion as to the value of Cordeiro's results we feel that the subject of the earth's gyroscopic motion and of a possible secular change in the direction of the axis deserves investigation for two chief reasons. In the first place, evidences of seasonal changes and of seasonal uniformity seem to occur more or less alternately in the geological record. Second, the remarkable discoveries of Garner and Allard[68] show that the duration of daylight has a p.r.o.nounced effect upon the reproduction of plants. We have referred repeatedly to the tree ferns, corals, and other forms of life which now live in relatively low lat.i.tudes and which cannot endure strong seasonal contrasts, but which once lived far to the north. On the other hand, Sayles,[69] for example, finds that microscopical examination of the banding of ancient shales and slates indicates distinct seasonal banding like that of recent Pleistocene clays or of the Squantum slate formed during or near the Permian glacial period. Such seasonal banding is found in rocks of various ages: (a) Huronian, in cobalt shales previously reported by Coleman; (b) late Proterozoic or early Cambrian in Hiwa.s.see slate; (c) lower Cambrian, in Georgian slates of Vermont; (d) lower Ordovician, in Georgia (Rockmart slate), Tennessee (Athens shale), Vermont (slates), and Quebec (Beekmantown formation); and (e) Permian in Ma.s.sachusetts (Squantum slate). How far the periods during which such evidence of seasons was recorded really alternated with mild periods, when tropical species lived in high lat.i.tudes and the contrast of seasons was almost or wholly lacking, we have as yet no means of knowing. If periods characterized by marked seasonal changes should be found to have alternated with those when the seasons were of little importance, the fact would be of great geological significance.

The discoveries of Garner and Allard as to the effect of light on reproduction began with a peculiar tobacco plant which appeared in some experiments at Washington. The plant grew to unusual size, and seemed to promise a valuable new variety. It formed no seeds, however, before the approach of cold weather. It was therefore removed to a greenhouse where it flowered and produced seed. In succeeding years the flowering was likewise delayed till early winter, but finally it was discovered that if small plants were started in the greenhouse in the early fall they flowered at the same time as the large ones. Experiments soon demonstrated that the time of flowering depends largely upon the length of the daily period when the plants are exposed to light. The same is true of many other plants, and there is great variety in the conditions which lead to flowering. Some plants, such as witch hazel, appear to be stimulated to bloom by very short days, while others, such as evening primrose, appear to require relatively long days. So sensitive are plants in this respect that Garner and Allard, by changing the length of the period of light, have caused a flowerbud in its early stages not only to stop developing but to return once more to a vegetative shoot.

Common iris, which flowers in May and June, will not blossom under ordinary conditions when grown in the greenhouse in winter, even under the same temperature conditions that prevail in early summer.

Again, one variety of soy beans will regularly begin to flower in June of each year, a second variety in July, and a third in August, when all are planted on the same date. There are no temperature differences during the summer months which could explain these differences in time of flowering; and, since "internal causes" alone cannot be accepted as furnishing a satisfactory explanation, some external factor other than temperature must be responsible.

The ordinary varieties of cosmos regularly flower in the fall in northern lat.i.tudes if they are planted in the spring or summer. If grown in a warm greenhouse during the winter months the plants also flower readily, so that the cooler weather of fall is not a necessary condition. If successive plantings of cosmos are made in the greenhouse during the late winter and early spring months, maintaining a uniform temperature throughout, the plantings made after a certain date will fail to blossom promptly, but, on the contrary, will continue to grow till the following fall, thus flowering at the usual season for this species. This curious reversal of behavior with advance of the season cannot be attributed to change in temperature. Some other factor is responsible for the failure of cosmos to blossom during the summer months. In this respect the behavior of cosmos is just the opposite of that observed in iris.

Certain varieties of soy beans change their behavior in a peculiar manner with advance of the summer season. The variety known as Biloxi, for example, when planted early in the spring in the lat.i.tude of Washington, D. C., continues to grow throughout the summer, flowering in September. The plants maintain growth without flowering for fifteen to eighteen weeks, attaining a height of five feet or more. As the dates of successive plantings are moved forward through the months of June and July, however there is a marked tendency for the plants to cut short the period of growth which precedes flowering. This means, of course, that there is a tendency to flower at approximately the same time of year regardless of the date of planting. As a necessary consequence, the size of the plants at the time of flowering is reduced in proportion to the delay in planting.

The bearing of this on geological problems lies in a query which it raises as to the ability of a genus or family of plants to adapt itself to days of very different length from those to which it is wonted. Could tree ferns, ginkgos, cycads, and other plants whose usual range of location never subjects them to daylight for more than perhaps fourteen hours or less than ten, thrive and reproduce themselves if subjected to periods of daylight ranging all the way from nothing up to about twenty-four hours? No answer to this is yet possible, but the question raises most interesting opportunities of investigation. If Cordeiro is right as to the earth's elastic gyroscopic motion, there may have been certain periods when a vertical or almost vertical axis permitted the days to be of almost equal length at all seasons in all lat.i.tudes. If such an absence of seasons occurred when the lands were low, when the oceans were extensive and widely open toward the poles, and when storms were relatively inactive, the result might be great mildness of climate such as appears sometimes to have prevailed in the middle of geological eras. Suppose on the other hand that the axis should be tilted more than now, and that the lands should be widely emergent and the storm belt highly active in low lat.i.tudes, perhaps because of the activity of the sun. The conditions might be favorable for glaciation at lat.i.tudes as low as those where the Permo-Carboniferous ice sheets appear to have centered. The possibilities thus suggested by Cordeiro's hypothesis are so interesting that the gyroscopic motion of the earth ought to be investigated more thoroughly. Even if no such gyroscopic motion takes place, however, the other causes of mild climate discussed in this chapter may be enough to explain all the observed phenomena.

Many important biological consequences might be drawn from this study of mild geological climates, but this book is not the place for them. In the first chapter we saw that one of the most remarkable features of the climate of the earth is its wonderful uniformity through hundreds of millions of years. As we come down through the vista of years the mild geological periods appear to represent a return as nearly as possible to this standard condition of uniformity. Certain changes of the earth itself, as we shall see in the next chapter, may in the long run tend slightly to change the exact conditions of this climatic standard, as we might perhaps call it. Yet they act so slowly that their effect during hundreds of millions of years is still open to question. At most they seem merely to have produced a slight increase in diversity from season to season and from zone to zone. The normal climate appears still to be of a milder type than that which happens to prevail at present. Some solar condition, whose possible nature will be discussed later, seems even now to cause the number of cyclonic storms to be greater than normal. Hence the earth's climate still shows something of the great diversity of seasons and of zones which is so marked a characteristic of glacial epochs.

FOOTNOTES:

[Footnote 58: Rollin T. Chamberlin: Personal Communication.]

[Footnote 59: F. H. Knowlton: Evolution of Geologic Climates; Bull.

Geol. Soc. Am., Vol. 30, 1919, pp. 499-566.]