WE have already discussed the first causes of nature, and allnatural motion, also the stars ordered in the motion of the heavens,and the physical element-enumerating and specifying them and showinghow they change into one another-and becoming and perishing ingeneral. There remains for consideration a part of this inquirywhich all our predecessors called meteorology. It is concerned withevents that are natural, though their order is less perfect thanthat of the first of the elements of bodies. They take place in theregion nearest to the motion of the stars. Such are the milky way, andcomets, and the movements of meteors. It studies also all theaffections we may call common to air and water, and the kinds andparts of the earth and the affections of its parts. These throwlight on the causes of winds and earthquakes and all theconsequences the motions of these kinds and parts involve. Of thesethings some puzzle us, while others admit of explanation in somedegree. Further, the inquiry is concerned with the falling ofthunderbolts and with whirlwinds and fire-winds, and further, therecurrent affections produced in these same bodies by concretion. Whenthe inquiry into these matters is concluded let us consider whataccount we can give, in accordance with the method we have followed,of animals and plants, both generally and in detail. When that hasbeen done we may say that the whole of our original undertaking willhave been carried out.

   After this introduction let us begin by discussing our immediatesubject.

   We have already laid down that there is one physical element whichmakes up the system of the bodies that move in a circle, and besidesthis four bodies owing their existence to the four principles, themotion of these latter bodies being of two kinds: either from thecentre or to the centre. These four bodies are fire, air, water,earth. Fire occupies the highest place among them all, earth thelowest, and two elements correspond to these in their relation toone another, air being nearest to fire, water to earth. The wholeworld surrounding the earth, then, the affections of which are oursubject, is made up of these bodies. This world necessarily has acertain continuity with the upper motions: consequently all itspower and order is derived from them. (For the originating principleof all motion is the first cause. Besides, that clement is eternal andits motion has no limit in space, but is always complete; whereasall these other bodies have separate regions which limit one another.)So we must treat fire and earth and the elements like them as thematerial causes of the events in this world (meaning by materialwhat is subject and is affected), but must assign causality in thesense of the originating principle of motion to the influence of theeternally moving bodies.

   Let us first recall our original principles and the

   Fire, air, water, earth, we assert, originate from one another,and each of them exists potentially in each, as all things do that canbe resolved into a common and ultimate substrate.

   The first difficulty is raised by what is called the air. What arewe to take its nature to be in the world surrounding the earth? Andwhat is its position relatively to the other physical elements. (Forthere is no question as to the relation of the bulk of the earth tothe size of the bodies which exist around it, since astronomicaldemonstrations have by this time proved to us that it is actuallyfar smaller than some individual stars. As for the water, it is notobserved to exist collectively and separately, nor can it do soapart from that volume of it which has its seat about the earth: thesea, that is, and rivers, which we can see, and any subterranean waterthat may be hidden from our observation.) The question is really aboutthat which lies between the earth and the nearest stars. Are we toconsider it to be one kind of body or more than one? And if morethan one, how many are there and what are the bounds of their regions?

   We have already described and characterized the first element, andexplained that the whole world of the upper motions is full of thatbody.

   This is an opinion we are not alone in holding: it appears to bean old assumption and one which men have held in the past, for theword ether has long been used to denote that element. Anaxagoras, itis true, seems to me to think that the word means the same as fire.For

   For the same opinions appear in cycles among men not once nor twice,but infinitely often.

   Now there are some who maintain that not only the bodies in motionbut that which contains them is pure fire, and the interval betweenthe earth and the stars air: but if they had considered what is nowsatisfactorily established by mathematics, they might have given upthis puerile opinion. For it is altogether childish to suppose thatthe moving bodies are all of them of a small size, because they soto us, looking at them from the earth.

   This a matter which we have already discussed in our treatment ofthe upper region, but we may return to the point now.

   If the intervals were full of fire and the bodies consisted offire every one of the other elements would long ago have vanished.

   However, they cannot simply be said to be full of air either; foreven if there were two elements to fill the space between the earthand the heavens, the air would far exceed the quantitu required tomaintain its proper proportion to the other elements. For the bulkof the earth (which includes the whole volume of water) isinfinitesimal in comparison with the whole world that surrounds it.Now we find that the excess in volume is not proportionately greatwhere water dissolves into air or air into fire. Whereas theproportion between any given small quantity of water and the airthat is generated from it ought to hold good between the totalamount of air and the total amount of water. Nor does it make anydifference if any one denies

   So it is clear that neither air nor fire alone fills theintermediate space.

   It remains to explain, after a preliminary discussion ofdifficulties, the relation of the two elements air and fire to theposition of the first element, and the reason why the stars in theupper region impart heat to the earth and its neighbourhood. Let usfirst treat of the air, as we proposed, and then go on to thesequestions.

   Since water is generated from air, and air from water, why areclouds not formed in the upper air? They ought to form there the more,the further from the earth and the colder that region is. For it isneither appreciably near to the heat of the stars, nor to the raysrelected from the earth. It is these that dissolve any formation bytheir heat and so prevent clouds from forming near the earth. Forclouds gather at the point where the reflected rays

   After this exposition of the difficulties involved, let us go onto lay down the truth, with a view at once to what follows and to whathas already been said. The upper region as far as the moon we affirmto consist of a body distinct both from fire and from air, but varyingdegree of purity and in kind, especially towards its limit on the sideof the air, and of the world surrounding the earth. Now the circularmotion of the first element and of the bodies it contains dissolves,and inflames by its motion, whatever part of the lower world isnearest to it, and so generates heat. From another point of view wemay look at the motion as follows. The body that lies below thecircular motion of the heavens is, in a sort, matter, and ispotentially hot, cold, dry, moist, and possessed of whatever otherqualities are derived from these. But it actually acquires orretains one of these in virtue of motion or rest, the cause andprinciple of which has already been explained. So at the centre andround it we get earth and water, the heaviest and coldest elements, bythemselves; round them and contiguous with them, air and what wecommonly call fire. It is not really fire, for fire is an excess ofheat and a sort of ebullition; but in reality, of what we call air,the part surrounding the earth is moist and warm, because itcontains both vapour and a dry exhalation from the

   However, it may well be that the formation of clouds in that upperregion is also prevented by the circular motion. For the air round theearth is necessarily all of it in motion, except that which is cut offinside the circumference which makes the earth a complete sphere. Inthe case of winds it is actually observable that they originate inmarshy districts of the earth; and they do not seem to blow abovethe level of the highest mountains. It is the revolution of the heavenwhich carries the air with it and causes its circular motion, firebeing continuous with the upper element and air with fire. Thus itsmotion is a second reason why that air is not condensed into water.

   But whenever a particle of air grows heavy, the warmth in it issqueezed out into the upper region and it sinks, and other particlesin turn are carried up together with the fiery exhalation. Thus theone region is always full of air and the other of fire, and each ofthem is perpetually in a state of change.

   So much to explain why clouds are not formed and why the air isnot condensed into water, and what account must be given of thespace between the stars and the earth, and what is the body that fillsit.

   As for the heat derived from the sun, the right place for aspecial and scientific account of it is in the treatise about sense,since heat is an affection of sense, but we may now explain how it canbe produced by the heavenly bodies which are not themselves hot.

   We see that motion is able to dissolve and inflame the air;indeed, moving bodies are often actually found to melt. Now thesun's motion alone is sufficient to account for the origin ofterrestrial warmth and heat. For a motion that is to have thiseffect must be rapid and near, and that of the stars is rapid butdistant, while that of the moon is near but slow, whereas the sun'smotion combines both conditions in a sufficient degree. That most heatshould be generated where the sun is present is easy to understandif we consider the analogy of terrestrial phenomena, for here, too, itis the air that is nearest to a thing in rapid motion which isheated most. This is just what we should expect, as it is thenearest air that is most dissolved by the motion of a solid body.

   This then is one reason why heat reaches our world. Another isthat the fire surrounding the air is often scattered by the motionof the heavens and driven downwards in spite of itself.

   Shooting-stars further suffix to prove that the celestial sphereis not hot or fiery: for they do not occur in that upper region butbelow: yet the more and the faster a thing moves, the more apt it isto take fire. Besides, the sun, which most of all the stars isconsidered to be hot, is really white and not fiery in colour.

   Having determined these principles let us explain the cause of theappearance in the sky of burning flames and of shooting-stars, andof 'torches', and 'goats', as some people call them. All thesephenomena are one and the same thing, and are due to the same cause,the difference between them being one of degree.

   The explanation of these and many other phenomena is this. Whenthe sun warms the earth the evaporation which takes place isnecessarily of two kinds, not of one only as some think. One kind israther of the nature of vapour,

   The cause of these shooting-stars is sometimes the motion whichignites the exhalation. At other times the air is condensed by coldand squeezes out and ejects the hot element; making their motionlook more like that of a thing thrown than like a running fire. Forthe question might be raised

   When the phenomenon is formed in the upper region it is due to thecombustion of the exhalation. When it takes place at a lower levelit is due to the ejection of the exhalation by the condensing andcooling of the moister evaporation: for this latter as it condensesand inclines downward contracts, and thrusts out the hot element andcauses it to be thrown downwards. The motion is upwards or downwardsor sideways according to the way in which the evaporation lies, andits disposition in respect of breadth and depth. In most cases thedirection is sideways because two motions are involved, a compulsorymotion downwards and a natural motion upwards, and under thesecircumstances an object always moves obliquely. Hence the motion of'shooting-stars' is generally oblique.

   So the material cause of all these phenomena is the exhalation,the efficient cause sometimes the upper motion, sometimes thecontraction and condensation of the air. Further, all these thingshappen below the moon. This is shown by their apparent speed, which isequal to that of things thrown by us; for it is because they are closeto us,

   Sometimes on a fine night we see a variety of appearances thatform in the sky: 'chasms' for instance and 'trenches' and blood-redcolours. These, too, have the same cause. For we have seen that theupper air condenses into an inflammable condition and that thecombustion sometimes takes on the appearance of a burning flame,sometimes that of moving torches and stars. So it is not surprisingthat this same air when condensing should assume a variety of colours.For a weak light shining through a dense air, and the air when it actsas a mirror, will cause all kinds of colours to appear, but especiallycrimson and purple. For these colours generally appear whenfire-colour and white are combined by superposition. Thus on a hotday, or through a smoky, medium, the stars when they rise and set lookcrimson. The light will also create colours by reflection when themirror is such as to reflect colour only and not shape.

   These appearances do not persist long, because the condensation ofthe air is transient.

   'Chasms' get their appearance of depth from light breaking out ofa dark blue or black mass of air. When the process of condensationgoes further in such a case we often find 'torches' ejected. Whenthe 'chasm' contracts it presents the appearance of a 'trench'.

   In general, white in contrast with black creates a variety ofcolours; like flame, for instance, through a medium of smoke. But byday the sun obscures them, and, with the exception of crimson, thecolours are not seen at night because they are dark.

   These then must be taken to be the causes of 'shooting-stars' andthe phenomena of combustion and also of the other transientappearances of this kind.

   Let us go on to explain the nature of comets and the 'milky way',after a preliminary discussion of the views of others.

   Anaxagoras and Democritus declare that comets are a conjunction ofthe planets approaching one another and so appearing to touch oneanother.

   Some of the Italians called Pythagoreans say that the comet is oneof the planets, but that it appears at great intervals of time andonly rises a little above the horizon. This is the case with Mercurytoo; because it only rises a little above the horizon it often failsto be seen and consequently appears at great intervals of time.

   A view like theirs was also expressed by Hippocrates of Chios andhis pupil Aeschylus. Only they say that the tail does not belong tothe comet iself, but is occasionally assumed by it on its course incertain situations, when our sight is reflected to the sun from themoisture attracted by the comet. It appears at greater intervalsthan the other stars because it is slowest to get clear of the sun andhas been left behind by the sun to the extent of the whole of itscircle before it reappears at the same point. It gets clear of the sunboth towards the north and towards the south. In the space between thetropics it does not draw water to itself because that region isdried up by the sun on its course. When it moves towards the southit has no lack of the necessary moisture, but because the segment ofits circle which is above the horizon is small, and that below it manytimes as large, it is impossible for the sun to be reflected to oursight, either when it approaches the southern tropic, or at the summersolstice. Hence in these regions it does not develop a tail at all.But when it is visible in the north it assumes a tail because thearc above the horizon is large and that below it

   These views involve impossibilities, some of which are common to allof them, while others are peculiar to some only.

   This is the case, first, with those who say that the comet is one ofthe planets. For all the planets appear in the circle of the zodiac,whereas many comets have been seen outside that circle. Again morecomets than one have often appeared simultaneously. Besides, iftheir tail is due to reflection, as Aeschylus and Hippocrates say,this planet ought sometimes to be visible without a tail since, asthey it does not possess a tail in every place in which it appears.But, as a matter of fact, no planet has been observed besides thefive. And all of them are often visible above the horizon togetherat the same time. Further, comets are often found to appear, as wellwhen all the planets are visible as when some are not, but areobscured by the neighbourhood of the sun. Moreover the statementthat a comet only appears in the north, with the sun at the summersolstice, is not true either. The great comet which appeared at thetime of the earthquake in Achaea and the tidal wave rose due west; andmany have been known to appear in the south. Again in the archonshipof Euclees, son of Molon, at Athens there appeared a comet in thenorth in the month Gamelion, the sun being about the wintersolstice. Yet they themselves admit that reflection over so great aspace is an impossibility.

   An objection that tells equally against those who hold this theoryand those who say that comets are a coalescence of the planets is,first, the fact that some of the fixed stars too get a tail. Forthis we must not only accept the authority of the Egyptians who assertit, but we have ourselves observed the fact. For a star in the thighof the Dog had a tail, though a faint one. If you fixed your sighton

   Enough has been said, without further argument, to show that thecauses brought forward to explain comets are false.

   We consider a satisfactory explanation of phenomena inaccessibleto observation to have been given when our account of them is freefrom impossibilities. The observations before us suggest the followingaccount of the

   When the matter begins to gather in the lower region independentlythe comet appears by itself. But when the exhalation is constituted byone of the fixed stars or the planets, owing to their motion, one ofthem becomes a comet. The fringe is not close to the stars themselves.Just as

   Now when this matter gathers in relation to a star the cometnecessarily appears to follow the same course as the star. But whenthe comet is formed independently it falls behind the motion of theuniverse, like the rest of the terrestrial world. It is this fact,that a comet often forms independently, indeed oftener than roundone of the regular stars, that makes it impossible to maintain thata comet is a sort of reflection, not indeed, as Hippocrates and hisschool say, to the sun, but to the very star it is alleged toaccompany-in fact, a kind of halo in the pure fuel of fire.

   As for the halo we shall explain its cause later.

   The fact that comets when frequent foreshadow wind and droughtmust be taken as an indication of their fiery constitution. Fortheir origin is plainly due to the plentiful supply of that secretion.Hence the air is necessarily drier and the moist evaporation is sodissolved and dissipated by the quantity of the hot exhalation asnot readily to condense into water.-But this phenomenon too shall beexplained more clearly later when the time comes to speak of thewinds.-So when there are many comets and they are dense, it is as wesay, and the years are clearly dry and windy. When they are fewerand fainter this effect does not appear in the same degree, thoughas a rule the is found to be excessive either in duration or strength.For instance when the stone at Aegospotami fell out of the air-ithad been carried up by

   That there are few comets and that they appear rarely and outsidethe tropic circles more than within them is due to the motion of thesun and the stars. For this motion does not only cause the hotprinciple to be secreted but also dissolves it when it is gathering.But the chief reason is that most of this stuff collects in the regionof the milky way.

   Let us now explain the origin, cause, and nature of the milky way.And here too let us begin by discussing the statements of others onthe subject.

   (1) Of the so-called Pythagoreans some say that this is the pathof one of the stars that fell from heaven at the time of Phaethon'sdownfall. Others say that the sun used once to move in this circle andthat this region was scorched or met with some other affection of thiskind, because of the sun and its motion.

   But it is absurd not to see that if this were the reason thecircle of the Zodiac ought to be affected in the same way, andindeed more so than that of the milky way, since not the sun onlybut all the planets move in it. We can see the whole of this circle(half of it being visible at any time of the night), but it shows nosigns of any such affection except where a part of it touches thecircle of the milky way.

   (2) Anaxagoras, Democritus, and their schools say that

   This, too, is obviously impossible. The milky way is alwaysunchanged and among the same constellations (for it is clearly agreatest circle), whereas, since the sun does not remain in the sameplace, what is hidden from it differs at different times. Consequentlywith the change of the sun's position the milky way ought to changeits position too: but we find that this does not happen. Besides, ifastronomical demonstrations are correct and the size of the sun isgreater than that of the earth and the distance of the stars fromthe earth many times greater than that of the sun (just as the sunis further from the earth than the moon), then the cone made by therays of the sun would terminate at no great distance from the earth,and the shadow of the earth (what we call night) would not reach thestars. On the contrary, the sun shines on all the stars and theearth screens none of them.

   (3) There is a third theory about the milky way. Some say that it isa reflection of our sight to the sun, just as they say that thecomet is.

   But this too is impossible. For if the eye and the mirror and thewhole of the object were severally at rest, then the same part ofthe image would appear at the same point in the mirror. But if themirror and the object move, keeping the same distance from the eyewhich is at rest, but at different rates of speed and so not always atthe same interval from one another, then it is impossible for the sameimage always to appear in the same part of the mirror. Now theconstellations included in the circle of the milky way move; and sodoes the sun, the object to which our sight is reflected; but we standstill. And the distance

   Again, we can see the milky way reflected at night in water andsimilar mirrors. But under these circumstances it is impossible forour sight to be reflected to the sun.

   These considerations show that the milky way is not the path ofone of the planets, nor the light of imperceptible stars, nor areflection. And those are the chief theories handed down by othershitherto.

   Let us recall our fundamental principle and then explain ourviews. We have already laid down that the outermost part of what iscalled the air is potentially fire and that therefore when the airis dissolved by motion, there is separated off a kind of matter-and ofthis matter we assert that comets consist. We must suppose that whathappens is the same as in the case of the comets when the matterdoes not form independently but is formed by one of the fixed stars orthe planets. Then these stars appear to be fringed, because matterof this kind follows their course. In the same way, a certain kindof matter follows the sun, and we explain the halo as a reflectionfrom it when the air is of the right constitution. Now we mustassume that what happens in the case of the stars severally happens inthe case of the whole of the heavens and all the upper motion. Forit is natural to suppose that, if the motion of a single starexcites a flame, that of all the stars should have a similar result,and especially in that region in which the stars are biggest andmost numerous and nearest to one another. Now the circle of the zodiacdissolves this kind of matter because of the motion of the sun and theplanets,

   We have now explained the phenomena that occur in that part of theterrestrial world which is continuous with the motions of the heavens,namely, shooting-stars and the burning flame, comets and the milkyway, these being the chief affections that appear in that region.

   Let us go on to treat of the region which follows next in orderafter this and which immediately surrounds the earth. It is the regioncommon to water and air, and the processes attending the formationof water above take place in it. We must consider the principles andcauses of all these phenomena too as before. The efficient and chiefand first cause is the circle in which the sun moves. For the sun asit approaches or recedes, obviously causes dissipation andcondensation and so gives rise to generation and destruction. Nowthe earth remains but the moisture surrounding it is made to evaporateby the sun's rays and the other heat from above, and rises. But whenthe heat which was raising it leaves it, in part dispersing to thehigher region, in part quenched through rising so far into the upperair, then the vapour cools

   The exhalation of water is vapour: air condensing into water iscloud. Mist is what is left over when a cloud condenses into water,and is therefore rather a sign of fine weather than of rain; formist might be called a barren cloud. So we get a circular process thatfollows the course of the sun. For according as the sun moves tothis side or that, the moisture in this process rises or falls. Wemust think of it as a river flowing up and down in a circle and madeup partly of air, partly of water. When the sun is near, the stream ofvapour flows upwards; when it recedes, the stream of water flows down:and the order of sequence, at all events, in this process alwaysremains the same. So if 'Oceanus' had some secret meaning in earlywriters, perhaps they may have meant this river that flows in a circleabout the earth.

   So the moisture is always raised by the heat and descends to theearth again when it gets cold. These processes and, in some cases,their varieties are distinguished by special names. When the waterfalls in small drops it is called a drizzle; when the drops are largerit is rain.

   Some of the vapour that is formed by day does not rise highbecause the ratio of the fire that is raising it to the water thatis being raised is small. When this cools and descends at night itis called dew and hoar-frost. When the vapour is frozen before ithas condensed to water again it is hoar-frost; and this appears inwinter and is commoner in cold places. It is dew when the vapour hascondensed into water and the heat is not so great as to dry up themoisture that has been raised nor the cold sufficient (owing to thewarmth of the climate or season) for the vapour itself to freeze.For dew is more commonly found when the season or the place is warm,whereas the opposite, as has

   Both dew and hoar-frost are found when the sky is clear and there isno wind. For the vapour could not be raised unless the sky were clear,and if a wind were blowing it could not condense.

   The fact that hoar-frost is not found on mountains contributes toprove that these phenomena occur because the vapour does not risehigh. One reason for this is that it rises from hollow and wateryplaces, so that the heat that is raising it, bearing as it were tooheavy a burden cannot lift it to a great height but soon lets itfall again. A second reason is that the motion of the air is morepronounced at a height, and this dissolves a gathering of this kind.

   Everywhere, except in Pontus, dew is found with south winds andnot with north winds. There the opposite is the case and it is foundwith north winds and not with south. The reason is the same as thatwhich explains why dew is found in warm weather and not in cold. Forthe south wind brings warm, and the north, wintry weather. For thenorth wind is cold and so quenches the heat of the evaporation. But inPontus the south wind does not bring warmth enough to causeevaporation, whereas the coldness of the north wind concentrates theheat by a sort of recoil, so that there is more evaporation and notless. This is a thing which we can often observe in other placestoo. Wells, for instance, give off more vapour in a north than in asouth wind. Only the north winds quench the heat before anyconsiderable quantity of vapour has gathered, while in a south windthe evaporation is allowed to accumulate.

   Water, once formed, does not freeze on the surface of the earth,in the way that it does in the region of the clouds.

   From the latter there fall three bodies condensed by cold, namelyrain, snow, hail. Two of these correspond to the phenomena on thelower level and are due to the same causes, differing from them onlyin degree and quantity.

   Snow and hoar-frost are one and the same thing, and so are rainand dew: only there is a great deal of the former and little of thelatter. For rain is due to the cooling of a great amount of vapour,for the region from which and the time during which the vapour iscollected are considerable. But of dew there is little: for the vapourcollects for it in a single day and from a small area, as its quickformation and scanty quantity show.

   The relation of hoar-frost and snow is the same: when cloudfreezes there is snow, when vapour freezes there is hoar-frost.Hence snow is a sign of a cold season or country. For a great dealof heat is still present and unless the cold were overpowering itthe cloud would not freeze. For there still survives in it a greatdeal of the heat which caused the moisture to rise as vapour fromthe earth.

   Hail on the other hand is found in the upper region, but thecorresponding phenomenon in the vaporous region near the earth islacking. For, as we said, to snow in the upper region correspondshoar-frost in the lower, and to rain in the upper region, dew in thelower. But there is nothing here to correspond to hail in the upperregion. Why this is so will be clear when we have explained the natureof hail.

   But we must go on to collect the facts bearing on the origin ofit, both those which raise no difficulties and those which seemparadoxical.

   Hail is ice, and water freezes in winter; yet hailstorms occurchiefly in spring and autumn and less often in the late summer, butrarely in winter and then only when the cold is less intense. And ingeneral hailstorms occur in warmer, and snow in colder places.Again, there is a difficulty about water freezing in the upper region.It

   Some think that the cause and origin of hail is this. The cloud isthrust up into the upper atmosphere, which is colder because thereflection of the sun's rays from the earth ceases there, and upon itsarrival there the water freezes. They think that this explains whyhailstorms are commoner in summer and in warm countries; the heat isgreater and it thrusts the clouds further up from the earth. But thefact is that hail does not occur at all at a great height: yet itought to do so, on their theory, just as we see that snow falls moston high mountains. Again clouds have often been observed moving with agreat noise close to the earth, terrifying those who heard and sawthem as portents of some catastrophe. Sometimes, too, when such cloudshave been seen, without any noise, there follows a violenthailstorm, and the stones are of incredible size, and angular inshape. This shows that they have not been falling for long and thatthey were frozen near to the earth, and not as that theory wouldhave it. Moreover, where the hailstones are large, the cause oftheir freezing must be present in the highest degree: for hail isice as every one can see. Now those hailstones are large which areangular in shape. And this shows that they froze close to the earth,for those that fall far are worn away by the length of their falland become round and smaller in size.

   It clearly follows that the congelation does not take

   Now we see that warm and cold react upon one another by recoil.Hence in warm weather the lower parts of the earth are cold and in afrost they are warm. The same thing, we must suppose, happens in theair, so that in the warmer seasons the cold is concentrated by thesurrounding heat and causes the cloud to go over into watersuddenly. (For this reason rain-drops are much larger on warm daysthan in winter, and showers more violent. A shower is said to bemore violent in proportion as the water comes down in a body, and thishappens when the condensation takes place quickly,-though this is justthe opposite of what Anaxagoras says. He says that this happens whenthe cloud has risen into the cold air; whereas we say that ithappens when the cloud has descended into the warm air, and that themore the further the cloud has descended). But when the cold hasbeen concentrated within still more by the outer heat, it freezesthe water it has formed and there is hail. We get hail when theprocess of freezing is quicker than the descent of the water. For ifthe water falls in a certain time and the cold is sufficient to freezeit in less, there is no difficulty about its having frozen in the air,provided that the freezing takes place in a shorter time than itsfall. The nearer to the earth, and the more suddenly, this processtakes place, the more violent is the rain that results and thelarger the raindrops and the hailstones because of the shortness oftheir fall. For the same reason large raindrops do not fall thickly.Hail is rarer in summer than in spring and autumn, though commonerthan in winter, because the air is drier in summer, whereas inspring it is still moist, and in autumn it is beginning to grow moist.It is for the same reason that hailstorms sometimes occur in thelate summer as we have said.

   The fact that the water has previously been warmed

   It is for the same reason that rain falls in summer and not inwinter in Arabia and Ethiopia too, and that in torrents and repeatedlyon the same day. For the concentration or recoil due to the extremeheat of the country cools the clouds quickly.

   So much for an account of the nature and causes of rain, dew,snow, hoar-frost, and hail.

   Let us explain the nature of winds, and all windy vapours, also ofrivers and of the sea. But here, too, we must first discuss thedifficulties involved: for, as in other matters, so in this notheory has been handed down to us that the most ordinary man could nothave thought of.

   Some say that what is called air, when it is in motion and flows, iswind, and that this same air when it condenses again becomes cloud andwater, implying that the nature of wind and water is the same. So theydefine wind as a motion of the air. Hence some, wishing to say aclever thing, assert that all the winds are one wind, because theair that moves is in fact all of it one and the same; they maintainthat the winds appear to differ owing to the region from which the airmay happen to flow on each occasion, but really do not differ atall. This is just like thinking that all rivers are one and the sameriver, and the ordinary unscientific view is better than ascientific theory like this. If all rivers flow from one source, andthe same is true in the case of the winds, there might be some truthin this theory; but if it is no more true in the one case than inthe other, this ingenious idea is plainly

   We find analogous views about the origin of rivers. It is thoughtthat the water is raised by the sun and descends in rain and gathersbelow the earth and so flows from a great reservoir, all the riversfrom one, or each from a different one. No water at all isgenerated, but the volume of the rivers consists of the water thatis gathered into such reservoirs in winter. Hence rivers are alwaysfuller in winter than in summer, and some are perennial, others not.Rivers are perennial where the reservoir is large and so enoughwater has collected in it to last out and not be used up before thewinter rain returns. Where the reservoirs are smaller there is lesswater in the rivers, and they are dried up and their vessel emptybefore the fresh rain comes on.

   But if any one will picture to himself a reservoir adequate to thewater that is continuously flowing day by day, and consider the amountof the water, it is obvious that a receptacle that is to contain allthe water that flows in the year would be larger than the earth, or,at any rate, not much smaller.

   Though it is evident that many reservoirs of this kind do exist inmany parts of the earth, yet it is unreasonable for any one torefuse to admit that air becomes water in the earth for the samereason as it does above it. If the cold causes the vaporous air tocondense into water above the earth we must suppose the cold in theearth to produce this same effect, and recognize that there not onlyexists in it and flows out of it actually formed water, but that wateris continually forming in it too.

   Again, even in the case of the water that is not being formed fromday to day but exists as such, we must not suppose as some do thatrivers have their source in definite subterranean lakes. On thecontrary, just as above the earth small drops form and these joinothers, till finally the water descends in a body as rain, so too wemust suppose that in the earth the water at first trickles togetherlittle by little, and that the sources of the rivers drip, as it were,out of the earth and then unite. This is proved by facts. When menconstruct an aqueduct they collect the water in pipes and trenches, asif the earth in the higher ground were sweating the water out.Hence, too, the head-waters of rivers are found to flow frommountains, and from the greatest mountains there flow the mostnumerous and greatest rivers. Again, most springs are in theneighbourhood of mountains and of high ground, whereas if we exceptrivers, water rarely appears in the plains. For mountains and highground, suspended over the country like a saturated sponge, make thewater ooze out and trickle together in minute quantities but in manyplaces. They receive a great deal of water falling as rain (for itmakes no difference whether a spongy receptacle is concave andturned up or convex and turned down: in either case it will containthe same volume of matter) and, they also cool the vapour that risesand condense it back into water.

   Hence, as we said, we find that the greatest rivers flow from thegreatest mountains. This can be seen by looking at itineraries: whatis recorded in them consists either of things which the writer hasseen himself or of such as he has compiled after inquiry from thosewho have seen them.

   In Asia we find that the most numerous and greatest rivers flow fromthe mountain called Parnassus, admittedly the greatest of allmountains towards the south-east. When you have crossed it you see theouter ocean, the further limit of which is unknown to the dwellersin our world.

   All other rivers would be found to flow in the same way, but we havementioned these as examples. Even where rivers flow from marshes,the marshes in almost every case are found to lie below mountains orgradually rising ground.

   It is clear then that we must not suppose rivers to originate fromdefinite reservoirs: for the whole earth, we might almost say, wouldnot be sufficient (any more than the region of the clouds would be) ifwe were to suppose that they were fed by actually existing wateronly and it were not the case that as some water passed out ofexistence some more came into existence, but rivers always drewtheir stream from an existing store. Secondly, the fact that riversrise at the foot of mountains proves that a place transmits thewater it contains by gradual percolation of many drops, little bylittle, and that this is how the sources of rivers originate. However,there is nothing impossible about the existence of such placescontaining a quantity of water like lakes: only they cannot be bigenough to produce the supposed effect. To think that they are isjust as absurd as if one were to suppose that rivers drew all theirwater from the sources we see (for most rivers do flow fromsprings). So it is no more reasonable to suppose those lakes tocontain the whole volume of water than these springs.

   That there exist such chasms and cavities in the earth we are taughtby the rivers that are swallowed up. They are found in many parts ofthe earth: in the Peloponnesus, for instance, there are many suchrivers in Arcadia. The reason is that Arcadia is mountainous and thereare no channels from its valleys to the sea. So these places getfull of water, and this, having no outlet, under the pressure of thewater that is added above, finds a way out for itself

   The same parts of the earth are not always moist or dry, but theychange according as rivers come into existence and dry up. And sothe relation of land to sea changes too and a place does not alwaysremain land or sea throughout all time, but where there was dry landthere comes to be sea, and where there is now sea, there one day comesto be dry land. But we must suppose these changes to follow some orderand cycle. The principle and cause of these changes is that theinterior of the earth grows and decays, like the bodies of plantsand animals. Only in the case of these latter the process does notgo on by parts, but each of them necessarily grows or decays as awhole, whereas it does go on by parts in the case of the earth. Herethe causes are cold and heat, which increase and diminish on accountof the sun and its course. It is owing to them that the parts of theearth come to have a different character, that some parts remain moistfor a certain time, and then dry up and grow old, while other parts intheir turn are filled with life

   If the sea was once pushed out by rivers and encroached upon theland anywhere, it necessarily leaves that place dry when it recedes;again, if the dry land has encroached on the sea at all by a processof silting set up by the rivers when at their full, the time must comewhen this place will be flooded again.

   But the whole vital process of the earth takes place so graduallyand in periods of time which are so immense compared with the lengthof our life, that these changes are not observed, and before theircourse can be recorded from beginning to end whole nations perishand are destroyed. Of such destructions the most utter and suddenare due to wars; but pestilence or famine cause them too. Famines,again, are either sudden and severe or else gradual. In the lattercase the disappearance of a nation is not noticed because some leavethe country while others remain; and this goes on until the land isunable to maintain any inhabitants at all. So a long period of time islikely to elapse from the first departure to the last, and no oneremembers and the lapse of time destroys all record even before thelast inhabitants have disappeared. In the same

   Men whose outlook is narrow suppose the cause of such events to bechange in the universe, in the sense of a coming to be of the world asa whole. Hence they say that the sea being dried up and is growingless, because this is observed to have happened in more places nowthan formerly. But this is only partially true. It is true that manyplaces are now dry, that formerly were covered with water. But theopposite is true too: for if they look they will find that there aremany places where the sea has invaded the land. But we must notsuppose that the cause of this is that the world is in process ofbecoming. For it is absurd to make the universe to be in processbecause of small and trifling changes, when the bulk and size of theearth are surely as nothing in comparison with the whole world. Ratherwe must take the cause of all these changes to be that, just as winteroccurs in the seasons of the year, so in determined periods therecomes a great winter of a great year and with it excess of rain. Butthis excess does not always occur in the same place. The deluge in thetime of Deucalion, for instance, took place chiefly in the Greek worldand in it especially about ancient Hellas, the country about Dodonaand the Achelous, a river which has often changed its course. Here theSelli dwelt and those who were formerly called Graeci and nowHellenes. When, therefore, such an excess of rain occurs we mustsuppose that it suffices for a long time. We have seen that some saythat the size of the subterranean cavities is what makes some riversperennial and others not, whereas we maintain that the size of themountains is the cause, and their density and coldness; for great,dense, and cold mountains catch and keep and create most water:whereas if the mountains that overhang the sources of rivers are smallor porous and stony and clayey, these rivers run dry earlier. Wemust recognize the same kind of thing in this case too. Where suchabundance of rain falls in the great winter

   Since there is necessarily some change in the whole world, but notin the way of coming into existence or perishing (for the universeis permanent), it must be, as we say, that the same places are not forever moist through the presence of sea and rivers, nor for ever dry.And the facts prove this. The whole land of the Egyptians, whom wetake to be the most ancient of men, has evidently gradually comeinto existence and been produced by the river. This is clear from anobservation of the country, and the facts about the Red Sea suffice toprove it too. One of their kings tried to make a canal to it (for itwould have been of no little advantage to them for the whole region tohave become navigable; Sesostris is said to have been the first of theancient kings to try), but he found that the sea was higher than theland. So he first, and Darius afterwards, stopped making the canal,lest the sea should mix with the river water and spoil it. So it isclear that all this part was once unbroken sea. For the same reasonLibya-the country of Ammon-is, strangely enough, lower and hollowerthan the land to the seaward of it. For it is clear that a barrierof silt was formed and after it lakes and dry land, but in course oftime the water that was left behind in the lakes dried up and is nowall gone. Again the silting up of the lake Maeotis by the rivers hasadvanced so much that the limit to the size of the ships which can nowsail into it to trade is much lower than it was sixty years ago. Henceit is easy to infer that

   Again, this process of silting up causes a continuous currentthrough the Bosporus; and in this case we can directly observe thenature of the process. Whenever the current from the Asiatic shorethrew up a sandbank, there first formed a small lake behind it.Later it dried up and a second sandbank formed in front of the firstand a second lake. This process went on uniformly and withoutinterruption. Now when this has been repeated often enough, in thecourse of time the strait must become like a river, and in the end theriver itself must dry up.

   So it is clear, since there will be no end to time and the worldis eternal, that neither the Tanais nor the Nile has always beenflowing, but that the region whence they flow was once dry: fortheir effect may be fulfilled, but time cannot. And this will beequally true of all other rivers. But if rivers come into existenceand perish and the same parts of the earth were not always moist,the sea must needs change correspondingly. And if the sea is alwaysadvancing in one place and receding in another it is clear that thesame parts of the whole earth are not always either sea or land, butthat all this changes in course of time.

   So we have explained that the same parts of the earth are not alwaysland or sea and why that is so: and also why some rivers are perennialand others not.