Last updated: 21 May 2008
[Brian Davis, if you're reading this and wondering why you haven't heard from me: I lost your email address and all your emails when my computer died. Aargh.]
This page is intended to help role-players and authors of sci-fi and fantasy to, as the title suggests, create a world which resembles the Earth - that is, a world in which humans could live and develop societies similar to those with which we are familiar.
The key ingredient here is familiarity. Worlds which are substantially different from the Earth are certainly interesting, but they need more effort and imagination to create, and I'd personally prefer to channel my imagination into the story rather than the setting. People with alien tastes may nevertheless find some material of interest here anyway, particularly in the Astronomy and Geology sections.
The wildly different lengths of the sections reflect three factors:
There is plenty of relevant information out there on the World Wide Web, but of course you need to find it first, which takes time and effort; moreover, I'm not aware of any source which contains everything you need to know in sufficient detail. This page is ultimately intended as a source of enough information, or failing that, pointers to further information, to take the hassle and effort out of finding and using it. Most of the page is based on my experiences and discoveries while creating my own world. Some of you will no doubt find some sections irrelevant, while others may well find these same sections to be a source of hitherto unsuspected interesting ideas.
I can't claim to know everything about every subject I touch on here; the sign "[*]" indicates that, due to such a gap in my knowledge, I'm soliciting for more information from someone who knows more. Nor can I in all conscience claim infallibility; for which reasons I will gratefully accept corrections of errors. All contributions will be appropriately credited.
Note: to simplify the maths, all equations are given using normalised quantities, where the quantities are relative to the Earth, the Sun or the Moon as appropriate. This gets rid of several universal constants.
On the other hand, particularly if you're writing sci-fi, you might well want to consider what happens if you have a different type of Sun, or a planet with the average density of foam rubber. In this case, carry straight on; there are a lot of surprising restrictions which crop up where you might not expect them. Certain calculations have been simplified, although their conclusions are not particularly affected; there's little point in trying to be too exact.
I'll start this section big and steadily work inwards. The first assumption is that your world will be set in either this universe or one with the same laws of physics; if it isn't, many of the following equations will need to be changed.
Useful astronomy-related links include Curious about Astronomy, Phil Plait's Bad Astronomy site and the Planetary Society. The Voyager Project website (one of many, so I believe) has lots of stuff about our own Solar System. For further equations pelating to planetary mechanics, don't miss The World Builders' Cookbook.
My source for all these links also adds that "A good reference on resonance effects in planetary mechanics is at http://history.nasa.gov/SP-345/ch8.htm which is part of a very good online book on the Solar System formation and state at http://history.nasa.gov/SP-345/sp345.htm which is in turn from the NASA online histories page at http://www.hq.nasa.gov/office/pao/History/on-line.html -- this has so much good stuff that space history geeks (that'd be me) could be there for weeks." What you make of all these is up to you, but if you're paranoid about violating the laws of physics, they're well worth reading.
Have a look at StarGen, a program which creates "moderately believable planetary systems around stars other than our own".
magi = 2 * log10i + B - R * rand(100) / i
where magi is the magnitude of the i'th brightest star, R is a randomizing factor (the larger it is, the greater the deviation from a true logarithmic scale), and B is the magnitude of the brightest star in the sky. I used R = 0.01 (it should be small, *not* 1 as this page used to say!) and B = -1.0 to get satisfactory results.
This should give you a naturalistic distribution of the stars by brightness; now you need to place them on your celestial sphere. The right ascension (the celestial equivalent of longitude) can be totally random; the declination (analogous to latitude) should be the inverse cosine of a random number between -1 and +1.
Another addition to the celestial sphere is the Milky Way; as visible from the Earth, it forms a great circle in the celestial sphere because the Earth is in the galactic plane. If your planet is some way removed from the galactic plane, the Milky Way will form a smaller circle. In general, the density of stars will be greater closer to the Milky Way and less further away from it; the area looking towards the galactic centre (on Earth, this is in the direction of Sagittarius) will be particularly rich. This does not preclude placing bright constellations away from the Milky Way; you can place the brighter stars where you like, so the above equations should still apply.
Bear in mind that constellations are apparent groupings of stars which are really at widely differing distances; this is why it's meaningless to talk about "the Sagittarian Sector" (sci-fi writers please note), since in any given constellation there are stars which are closer to Earth than to the other stars in that constellation.
As far as other night-sky objects are concerned, external galaxies are more visible when you're looking away from the galactic plane; naked-eye galaxies are thus most commonly found far from the Milky Way. Globular clusters are generally found within or near to the galactic plane.
The colour of a star is of course a function of its temperature. The hottest stars are white or blue-white, the coolest are orange or red, and those in between are yellowish. The colours are actually only noticeable for the brightest stars; faint stars all look white. In general, the brightest stars tend towards the hotter end of the temperature range (classes B and A, with very occasional O); as the stars get fainter, types G and especially K become more common.
Because the Earth is rotating on its axis, the celestial sphere appears to rotate just over once per day. The "just over" - a result of the Earth moving along its orbit - causes the night sky to appear the same at any given time as it does slightly later the preceding night and slightly earlier the following night. The difference in time is calculated by dividing the length of the day by the length of the year; for the Earth it is 236.5 seconds per day.
Added to which, one correspondent mentions that "binary star systems will generally have too great a fluctuation in temperature to be habitable". For this reason I will assume one sun only here.
Our own Sun (spelt with a capital) is a main sequence star of spectral type G2 (yellow), which is pretty average in star terms. Its diameter is 1.39 million km, and the Earth orbits it at a mean distance of 149 million km (1 astronomical unit, or AU) in 365.25 days (1 Earth year) to complete one orbit. Note that 1 AU = 216 Sun radii.
Life on Earth has evolved because the Earth is at the right distance from the Sun to ensure that it receives the right amount of heat from the Sun. None of the other planets in the Solar System have developed "life as we know it" because they're either too close and thus too hot, or too far away and thus too cold. If you want your sun to be of a different spectral type from the Sun, there will be several knock-on effects to consider. There are several physical quantities which are relevant here, two of which are fundamental:
L = M3.5
D = M0.74
I = L / R2 (inverse-square law)
M × T2 = R3 (Kepler's third law)
Note also:
Surface temperature = M0.505
Lifetime = M-2.5
Now, for Earthlike planets I must be close to 1; according to Brian Davis, "recent work suggests very conservatively 1.1 > I > 0.53". Thus the feasible limits for R and T can be calculated:
Rmin = sqrt(L / 1.1)
Rmax = sqrt(L / 0.53)
Tmin = 0.53 × M2.125 Tmax = 1.1 × M2.125 From these can be calculated, for a star of any spectral type, reasonable year-lengths for a planet with Earthlike life orbiting around it. Using data from Norton's 2000.0 (18th edition), we get the following table. [I might redo this table sometime when Ihave the time to bring it in line with the new equations].
Type L M D Rmin Rmax Tmin (days) Tmax (days)
(main sequence)
O5 500000 40 14.72 674.20 971.29 1010981.52 1748158.60
B0 20000 18 6.01 134.84 194.26 134797.54 233087.81
B5 800 6.5 3.91 26.97 38.85 20063.49 34693.18
A0 80 3.2 3.02 8.53 12.29 5084.97 8792.77
A5 20 2.1 2.07 4.26 6.14 2219.26 3837.48
F0 6.3 1.7 1.53 2.39 3.45 1037.11 1793.35
F5 2.5 1.3 1.22 1.51 2.17 592.96 1025.34
G0 1.26 1.1 1.02 1.07 1.54 385.59 666.75
G5 0.79 0.93 0.96 0.85 1.22 295.48 510.93
K0 0.4 0.78 0.88 0.60 0.87 193.66 334.87
K5 0.16 0.69 0.77 0.38 0.55 103.56 179.08
M0 0.06 0.47 0.68 0.23 0.34 60.13 103.98
M5 0.01 0.21 0.42 0.10 0.14 23.47 40.58
(giants)
G0 32 2.5 6.01 5.39 7.77 2893.60 5003.52
G5 50 3.2 9.34 6.74 9.71 3574.36 6180.67
K0 80 4 14.70 8.53 12.29 4548.13 7864.49
K5 200 5 32.47 13.48 19.43 8087.85 13985.27
M0 400 6.3 66.12 19.07 27.47 12117.71 20953.57
(supergiants)
B0 250000 50 18.49 476.73 686.80 537669.89 929722.47
A0 20000 16 32.69 134.84 194.26 142974.38 247226.96
F0 80000 12.5 193.19 269.68 388.51 457518.01 791126.27
G0 6300 10 80.98 75.68 109.03 76041.58 131488.79
G5 6300 12.5 112.49 75.68 109.03 68013.66 117607.15
K0 8000 12.5 177.43 85.28 122.86 81359.49 140684.36
K5 16000 16 339.94 120.60 173.75 120941.60 209128.55
Sun 1 1 1.00 0.95 1.37 340.05 588.01
So, theoretically, your year may vary over a range of 23 days to a few thousand Earth-years; note that years of Earthlike length are only possible with Sunlike suns, and shorter years imply redder suns.
Brighter stars, giants and supergiants have shorter lifespans (3 billion years for F0, compared to 10 billion for the Sun). There's presumably a lower limit for the lifetime, below which the planet's atmosphere won't be able to become breathable before the star turns into a giant, but nobody seems to know what it is [*]. Stars dimmer than about K2 have tidal forces strong enough for the planet's rotation to be slowed down or stopped. This is what's happened with Mercury and Venus, but for different reasons; research at Weather on Tide-Locked Planets suggests that the day side might be able to support life.
A correspondent says:
"... if you want to create a group of stars with masses distributed the way you would see in a real-world group of stars, -ln(1-x)/ln(1.35), where x is a random number between 0 and 1, will do the trick. Most stars that come out of this are larger than the sun (2.3SM is about average), but the larger stars die so much more quickly than the smaller ones that there are already far more small stars in the galaxy than big ones."
I'd really like to know what effects the processes of planetary formation have on the distances of the planets from the sun [*]. In the meantime, the best I can offer is a method based on Bode's Law. This law relates the distances of the planets from the Sun to a simple formula, by which the distance of the i'th planet is given by:
Ri = 0.4 + 0.3 × 2i - 2
i.e. the distances in AU are ideally 0.55, 0.7, 1.0, 1.6, 2.8. 5.2, 10.0, 19.6, 39.2 and so on. Note however that Mercury's distance is 0.4, not 0.55; there is no planet at 2.8 AU from the Sun (we have the asteroids instead); and the Law puts Pluto where Neptune should be. Whether or not Bode's Law is a genuine physical law or the product of coincidence, you can still use it to generate a workable set of planetary distances by twiddling the numbers to your preferences. (I am informed that "Bode's Law works because planets tend to settle into orbits whose periods are in simple fractional relations: e.g. Neptune:Pluto::2:3 and Venus:Earth::8:13".) You can now work out the orbital periods of your planets with Kepler's third law:
M × T2 = R3
Having done this, you will also need to re-twiddle your distances to eliminate the possibility of two planets disturbing each other's orbits at the same point within the orbits. What this means is that the ratio of no pair of orbital periods must be close to the ratio of two small integers (e.g. 4/3, 3/2), unless the planets are far enough apart (how far? [*]). Once that's done, you can work out the synodic period (S) of each planet, which is the time taken by the planet to reach the same position relative to the sun and your own planet:
1/S = 1 - 1/T, or S = T / (T - 1)
This doesn't mean that every S years the planet returns to the same part of the sky (except as seen from from the Sun), because the home planet has also moved in that time; instead it means that the planet will be best visible every S years, and will have moved across the sky by an amount equal to the fractional part of S.
For example, consider Mars as viewed from Earth. For Mars, R is 1.52; T is thus 1.877, or 685 days, or one year and 10.5 months, giving a value of S of 2.14 years, or 781 days. This means that successive oppositions of Mars, when it is opposite the Sun as seen from the Earth, occur every 781 days, during which time it has moved 0.14 (the fractional part of S) of the distance across the sky from the previous opposition.
In general, it must be said that our own Solar System is believed to be typical of most solar systems. Thus it's highly probable that the outer planets of all solar systems are gas gaints, all with ring systems and large numbers of satellites. Note, too, that celestial mechanics dictate that neighbouring planets cannot approach each other closer than a certain limit without becoming perturbed and breaking up; this is the origin of the asteroids, and probably several of the moons of the planets beyond Earth.
Finally, you can work out how bright your planets will be in the sky; the equations here come from some pages from the National Solar Observatory Sacramento Peak. First of all, calculate M0 and M, the absolute and apparent magnitudes of your sun, from its luminosity (L) and distance from your home planet (R, which must be in kilometres):
M0 = 4.8 - 2.5 × log L
M = M0 - 5 × log (R / 308.6 × 1015), or M0 + 5 × log R - 72.447
For the Sun, these values are 4.8 and -26.8 respectively. Next calculate a useful constant C (for the Sun, 14.10):
C = M + 5 × log R, or 10 × log R - 2.5 × log L - 67.647
You can now calculate the magnitude m0 of a planet at 1 AU:
m0 = C - 2.5 × log (a × r2)
where r is the planet's radius in km and a is its albedo or reflectivity. The albedo depends on what the planet is made of; for rocky planets a is around 0.15, and for gas giants it's between 0.4 and 0.6. Venus, which is covered in highly reflective clouds, has an albedo of 0.65; the Earth's is about 0.4; and that of an icy planet would probably be 0.6 to 0.8.
At last! The apparent magnitude of your planet is given by:
mmax = m0 + 5 × log(d1 × d2) - 2.5 × log (0.5 + 0.5 × cos phase)
where:
For inferior planets, those closer to the sun than your planet, the phase angle is 180 degrees at inferior conjunction, i.e. when the planet is directly in front of the sun, and zero degrees at superior conjunction, when it's directly behind the sun. Obviously, you won't see inferior planets at either of these times; they're best seen around greatest elongation, when they're at their maximum distance from the sun in the sky. This distance, and the phase, are given by:
emax = sqrt(1 - d22)
phase = 180 - arccos d2
For superior planets, i.e. those outside the orbit of your home planet, the phase angle is rarely far from 180 degrees. Superior planets make complete circuits of the sky, including the interesting phenomenon of retrograde motion at opposition. This is particularly noticeable with Mars; as it reaches opposition, it slows down and stops, then moves backwards through its opposition, then stops and moves forwards again. Experimenting with a night-sky viewer, such as my Night Sky Applet, should help you to understand the process.
Note that planets with apparent magnitudes less than 6 will be invisible without optical aid, as is the case with Neptune and Pluto from the Earth. The value for Uranus is 5.5.
If you have a binary (or multiple!) sun, the same equations will work; the value of L in the initial equation should be the sum of the individual values of L for each individual sun.
Earth has, of course, only one Moon; there's no reason why your planets can't have many more. Of all the topics in this section, this one offers the greatest number of possibilities which are interestingly different from the Earth.
However many moons you have, you need to know the following about each of them:
The apparent diameter of a moon (i.e. its diameter as seen from the planet) is proportional to its actual diameter and inversely proportional to its distance from the planet; thus a moon half the size of the Moon and twice as far away will appear one-quarter the apparent diameter. This is why the Sun and the Moon appear about the same size: the Sun is roughly 400 times the diameter of the Moon, but also about 400 times further away.
Kepler's third law can be used to calculate the moon's distance from the planet given the length of the moon's orbital period, or vice versa. The formula here needs to be used in its full form:
G x (M + m) x T2 = 4 x pi2 x R3
in which all the quantities are used directly, rather than being normalised; thus M is the mass of the primary body (the planet) in kilograms, m is similarly the mass of the secondary body (the moon), R is the distance to the secondary body in metres, G is the universal gravitation constant (6.673 x 10-11 N m2 kg-2), and T is in seconds. (If you're dealing with a planet orbiting your Sun, the mass of the Sun is large enough that the mass of the planet can be disregarded.)
There is a lower limit to R - it's called Roche's Limit - below which the effect of the planet's gravity will pull the moon apart and form rings, as with Saturn. Roche's limit equals 2.45 × r x (P/p)1/3, where P and p are the densities of the planet and the moon respectively and r is the planet's radius. (The quantities don't need to be normalised, since the ratios of the densities are used, not their actual values, and Roche's limit is relative to the planet's radius anyway.)
The phases of your moons depend on the relative positions of the moons and the sun:
A moon will return to the same place in the sky relative to the stars once every orbital period, but with a different phase. The phases repeat after a period of time different from the orbital period, because during the moon's orbit the planet is also moving around the sun. If a year is y days and the moon orbits in m days, then the phases of the moon will repeat in t days, where 1/t = 1/m - 1/y. For the Moon, m = 27.1 and y = 365.25, so p = 29.27 days - about one month, which shouldn't be surprising.
For complicated reasons your moons will always be tide-locked to the planet, with their rotational periods the same as their orbital periods. This essentially means that you will always see the same side of the moons, as is the case with our own Moon; in other words, phases aside, the moons will individually always appear the same throughout their orbits.
One correspondent asked if it's possible for moons to have sub-moons orbiting around them. It turns out that the answer is "no", because this would require an impossible set of three separate tide-locks. Note that tide-locks only occur within a certain radius; thus the moon is tide-locked to the Earth, but the Earth isn't to the Sun.
Tides are caused by the gravitational pull of the sun and moons on the planet; more moons will produce more complicated tides. Here's a good page explaining how tides work. The magnitude of the tide a body causes at a point equals:
t = D3 × P
where:
t = magnitude of tide
D = apparent diameter of body as viewed from the point
P = density of body.
To finish with, here are some interesting phenomena of some moons in the Solar System, which you might want to emulate:
The length of the day on your planet is affected by one factor only, the speed of the planet's rotation about its axis; faster rotation results in shorter days, and slower rotation causes longer days. The speed of rotation has several other knock-on effects; for example, faster rotation will have the following effects, which I am unable to provide equations for all of as yet [*]:
For longer days, of course, all of these effects are reversed; Jordi Mas informs me that there is no upper limit to the length of the day. In particular, days which are too long will produce enough heat from the sun to kill off certain flora and fauna.
If you want to be precise, here's the maths. ob, the oblateness, is:
ob = (re - rp) / re
where re is the equatorial radius and rp is the polar radius. The upper limit for ob is:
obmax = (5 × pi2 × r3) / (G × M x T2)
where:
pi is, of course, 3.14159
r is the equatorial radius of the planet in metres
G is the universal gravitational constant, 6.67 x 10-11
M is the planet's mass in kilograms T is the length of the planet's day in seconds
The lower limit for ob is:
obmin = obmax × 0.315
You may want to experiment with retrograde rotation - i.e. what happens when the planet rotates "backwards" with respect to its orbit around the sun. Aside from making the sun and other objects move in the opposite direction, this would mean that the night sky would repeat its appearance slightly later each night, not earlier.
The axial inclination affects the heights above the horizon of all heavenly bodies; the greater the angle, the greater the variation in their positions. If the axial inclination is i degrees, at latitude L the height of the sun above the horizon will vary between i-L degrees and i+L degrees. Its maximum height in summer will also be i+L, while on the shortest day its maximum height will be L-i. The sun's changing height has a significant effect on climate, for which see later.
Gravity also affects the atmosphere, but here the upper atmosphere temperature is also important; Saturn's moon Titan, for example, has an atmosphere with a surface pressure 1.5 times that of the Earth. The surface pressure of a breathable atmosphere should probably be within 0.1 and 4 times that of the Earth.
Time for some more equations. The values here, all taken relative to the Earth, are:
M = P × R3
while surface gravity is related to mass and radius thus:
g = M / R2
Eliminating M, we get:
g = P × R
In other words, a planet with a radius twice the radius of the Earth will have to be half as dense to have the same gravity, and vice versa. An Earth-sized planet made of polystyrene will have a relative density of about 1, and thus a surface gravity one 5.5th that of the Earth's. If you were to jump on such a planet, you'd rise and fall very slowly. You'd also probably die trying to breathe the tenuous atmosphere, but that's another matter.
Here's a gravity-related mistake in a popular work of fiction, which only a pedant like me would notice. According to Karen Wynn Fonstadt's excellent Atlas of Pern, which accompanies the books written by Anne McCaffrey, ten degrees of latitude on the planet Pern equals about 80 miles, which indicates an equatorial circumference of 80 times 36 = 2880 miles, or a radius of 917 miles - 1/8.64 that of the Earth. Assuming that Pern has the same surface gravity as the Earth, this indicates that Pern has a density of about 43, twice the density of the densest known element, osmium, and thus physically impossible. Oops!
The gravity at the poles is always greater than that at the equator. For Earthlike planets, Jordi Mas provides the following information.
The variables, again normalised relative to the Earth, are:
ob = K × 0.00346 / (T2 × P).
If this is greater than 0.2, you have a very oblate planet for which the following formulae are not appropriate.
The polar radius and gravity are thus:
rp = re × (1 - ob)
gp = ge × (2.5 - K) × (1 - ob)
You can also work out the shape your planet will have, although it gets complicated! First of all, calculate its angular momentum using a formula somewhere within this paper. There are four cases to consider, based upon the momentum relative to two values X and Y:
The proportion of water to land on a planet's surface affects the carbonate-silicate cycle. Too much or too little water will cause this cycle to be unstable, which in turn will decrease the likelihood of a stable climate over geologic time, and thus the likelihood of Earthlike life.
Plate tectonics come into play here, too, although you don't need to worry about them too much; if you're interested, Wikipedia's article is very good. The areas where two plates meet are highly likely to feature mountain ranges (e.g. the Himalayas or Andes), volcanoes (the Mediterranean, Japan) and earthquakes (California). A correspondent points out that: "plate tectonics has one result worth remembering: you can only get high mountains on one side of a continent, since the newest mountains will be on the 'leading edge': compare the Rockies and the Appalachians".
If you haven't already done so, decide on a scale, so you know the size of the area the Map is supposed to represent. Start with your coastlines and the neighbouring islands, if you have them; offshore islands are usually formed by the same processes as the nearby coast, and so should have roughly similar-looking coastlines.
Next, draw your mountains and rivers. Unless you've got a good reason to do otherwise, mountains form irregular parallel chains, and are often continued offshore as islands. And don't forget that rivers always start high and flow downhill; and that most rivers are created by rainfall, which is highest on the windward sides of mountains.
For supplementary reading, author Holly Lisle has a workshop about mapmaking; it's oriented towards expediency rather than scientific accuracy and rigour, but might be useful if you're in a hurry. Take a look at her own Map, too. Mark Rosenfelder has created some lovely Maps for Virtual Verduria, with instructions for drawing Maps of similar quality; see also the Cartographers' Guild for information about how to draw nice Maps.
Once you've decided on the shapes of the land and sea, virtually everything else on your Map is dictated by the climate. Climates are affected by both large-scale and small-scale factors, for which reason it's probably a good idea to establish what the major land masses and seas are in the areas adjacent to your Map.
The climate of an area is defined as the weather conditions experienced by the area averaged over a long period of time; it is most conveniently described in terms of the yearly amount and patterns of two important and easily-observed factors, rainfall and temperature. These factors dictate the plants which grow, and in turn the animals which are found; these factors influence what kind of human cultures develop in the area. A desert society will be very different from one which inhabits a region with a cool temperate climate, for example.
By contrast, because there is so little rain in dry climates, rivers in areas with such climates will have formed elsewhere. In general, too, rivers are less frequent on the drier leeward sides of mountains.
Small axial tilt will reduce seasonal variation in any one place by reducing the variation in how direct the solar input is at each place. But, that will increase the variation in the total solar input at different latitudes, so you'll have a stronger temperature gradient: the difference in temperature between the equator and the poles will be greater. Temperature gradient is what drives weather, and you'll get stronger winds and other weather effects (all other things being equal).
Large tilt will have the opposite effect. The seasonal variation will be strong as a given latitude experiences a wider range of solar input, but this will also tend to spread heat more evenly across the planet, and reduce the temperature gradient, and make a calmer atmosphere (all other things being equal).
The secret to fine tuned control of your planetary weather is ocean currents. If you want small temporal and latitudinal temperature variation, you need strong North-South ocean currents working to distribute heat, and weak East-West ones that cannot effectively accumulate heat differentials.
Your planet should avoid long stretches of ocean where currents can flow East-West without running into land, especially near the equator and the poles. A major cause of instability in our current atmosphere is the West Wind Drift, the ocean current that circles Antarctica; it is almost completely flat, and precludes north-south transfer of heat. Also, you want lots of warm water moving around, avoid very large patches of unvaryingly warm water, which are hurricane breeding grounds.
Your planet should also avoid any very large land areas (like Asia). Land changes temperature faster than water; not actually because of the specific heat of water, but because in water, the heat gets distributed through a larger vertical section, whereas in land it all tends to accumulate in the top few centimeters. So large patches of land will create large areas of differential heating.
Islands feature species typical of their climatic regions, but in fewer numbers and often with idiosyncratic species. The absence of snakes in Ireland is due not to Saint Patrick, but to the simple facts that snakes don't cross water and didn't reach Ireland before it became an island. Moles and woodpeckers are other species which are absent from Ireland for the same reason; and distinctively Antipodean birds such as the cassowary, emu and kiwi evolved in isolation from those in the rest of the world.
Taking as a starting point the obvious fact that plants need water to grow, some useful generalisations follow. Most importantly, rainy climates will support many more species of plants and animals than dry climates; compare a rainforest to a desert and you'll get the idea.
Plants which grow in dry climates will develop to conserve precious water; this is why cacti are thick-skinned and why cork oak grows its thick spongy bark, for example. This fact also explains why conifers have smaller leaves than broadleaved trees, since small leaves lose less water through evaporation.
Coniferous trees are good examples of plants adapting to their climate for other reasons: their conical shapes allow heavy snowfalls to slide off onto the ground, and their strong branches are able to support the snow which remains. Their leaves, besides being small to conserve water, are also dark to absorb as much of the Sun's light as possible; sunlight is in much shorter supply in the cold climates in which conifers grow compared to the more temperate climates which support broadleaved forests.
Less obvious is the effect of landscape on the variety of plant species. North America has a much greater variety of tree species than Europe for two principal reasons: the orientation of the mountain ranges, and the effects of past Ice Ages. Essentially, as the ice encroached southwards during the Ice Ages, the trees in North America were able to retreat before the ice since the north-south mountains provided no real barrier; by contrast, in Europe the east-west mountains (the Alps and Carpathians) prevented all but the most hardy species from retreating southwards, with the Mediterranean Sea sealing the gaps.
Like plants, animals adapt to their environment. A particularly striking example of this may be found in snow climates (e.g. subarctic and humid continental), in which snow lies on the ground for periods of several months at a time; animals in these climates, such as the snow hare, ptarmigan and Arctic Fox, typically turn white in winter for camouflage. Animals in cold climates also evolve ways of retaining heat; seals and polar bears, for example, have layers of fat for this purpose.
Another good example is the fauna of the savannah climate. Here there are vast grasslands punctuated with occasional trees which have adapted to store water throughout the long dry season (when the grasslands turn to semi-desert), such as the bottle-shaped baobob tree. The grasslands support large numbers of herd animals such as gnu, impala, wildebeest and so on; in turn these herds support carnivores such as lions, cheetah and leopards. The wide open spaces allow the herd species and predators to evolve the ability to run fast to outrun each other. The huge herds of buffalo of the the American Great Plains lived there for similar reasons.
Cultures develop and evolve by interacting with other cultures and borrowing their ideas and inventions. This implies that cultures living in isolated regions, such as in mountainous areas or on islands, won't develop at the same speed as those on large flat plains. Plains are also easier to conquer and integrate into single cultural units; this explains not only why mountains make good natural borders, but also why there are only three countries in North America but over forty in Europe.
A very good read about the development of human cultures is Guns, Germs and Steel by Jared Diamond, which sets out to answer the question of why European cultures came to dominate the world, overtaking those of China and the Middle East. To simplify the book's main thrust somewhat drastically, the reason is ultimately down to the east-west orientation of Eurasia compared to the north-south orientations of Africa and the Americas, which provided Eurasia with much more land in temparate latitudes than any other landmass. This large amount of land greatly facilitated east-west diffusion of cultural developments, since little adjustment to different environments was necessary. By contrast, the diffusion of cultures through Africa and the Americas was hindered by the presence of deserts, dense rainforests, and the narrow mountainous land-bridges of Central America.
According to Diamond, China was eventually overtaken culturally by Europe because the more mountainous regions of Europe resisted homogenisation and preserved many competing cultures, which developed and, from a few centuries ago, exchanged ideas and inventions at a faster rate than in China. In particular, one reason why the Industrial Revolution began in Britain was that Britain was able to exchange cultural ideas with mainland Europe but was not hampered by wars on its soil. By contrast, it was easy for one culture to conquer the plains of China; this monolithic culture was not conductive to development at the same speed.
Another effect of the interaction of cultures in Europe was that resistances to diseases were spread quickly among the various peoples. The more isolated peoples in Central America did not share the same resistances; as a result, when the Spanish arrived in Central America, the native peoples suffered as much from European diseases as from their superior warcraft.
It's useful to know the populations of the places on the map. In general, the population density - the number of people in a given area - depends primarily on the quality of the soil and the level of farming technology; good soils in areas of reliable rainfall which can be ploughed with horse-drawn ploughs are likely to support much higher population densities than arid areas of steppe. Another factor is the security of the area - people don't generally tend to live in areas of land which are regularly ravaged by war. This page should help you calculate population densities; it's geared towards mediaeval societies and RPGs, but the basic principles should still be valid.
Place-names generally derive from local geographical features; for example Abertawe is "mouth of the river Tawe", and Sterrebeek means "stream of stars". Sometimes the names remain more or less unchanged down the centuries, as with these two. Other names change to varying extents, as with Dunfermline, which comes from the Gaelic dun fearum linn (I'm not sure of the spelling), which means "fort by the crooked stream", and York, which results from various types of phonetic change affecting the original Latin Eburacum.
Names of rivers tend to be particularly conservative; the Rhône in France, for example, has had the same name (subject to linguistic changes; the Romans knew it as the Rhodanum) since at least pre-Roman times.
If all you want from such a language is a way of naming places, you can get away with:
The mixture of names in any given part of your world reflects the cultures which have lived and fought there. The Great British names above, for example, come from areas settled by Anglo-Saxons, Vikings, P-Celts and Q-Celts.
Designing a Fantasy World, from everything2, is a very fine essay which covers many points I've skimped on above.
The Worlds in the Net site. This contains a list of useful links about world-building, of which Jesper Udsen's experience of designing a world and Rich Staats' essay are particularly good. Hunting around this site turns up plenty of other goodies, too.
Mark Rosenfelder's Virtual Verduria is an impressive constructed world, complete with attractive Maps and large amounts of absorbing detail.
Web Blackdragon is an online role-playing IRC channel. Don't miss the lovely Map.
Patricia Wrede's Worldbuilder Questions are useful pointers to things to think about.
Elizabeth Viau has an interesting online course in world-building, which contains plenty of scientific notes, although they aren't complete yet. The course is about creating planets in general, not just Earthlike ones.
The Alien Planet Designer webpage.
The Nocturne Research world-building website has archives with lots of interesting material, some of it relevant to the topics in this page.
The MythoPoet's Manual. Very good for culture and religion.
Epona, a planet in a constructed solar system. Very interesting and well thought out.
Here's a Quick'n'Dirty FAQ about science-related topics, some of which are relevant.
Occasionally you stumble across lecture notes for university courses which contain material of interest; here's a set about geography.
Not related to world-building, but rather to writing, are Holly Lisle's Forward Motion pages, and the very funny "What I Would Do If" lists on Chicken Soup for the Gamer's Soul.
Matthew White's website contains a lot of amusing and stimulating material, some of which (such as Climate in Mediaeval America) is of particular interest to world-builders. Other bits of it are barking mad, but good fun.
International recognition! Teresa Costa translated some of the astronomy section into Portuguese for her own world-building pages, which contain much else of value and interest.
A page of useful writer's resources is Creating Fantasy Worlds by Paul Nattress; it has many further links, and speaks very highly of this page too :-). One link particularly worth looking at is How to create a fantasy world, by the Australian author Sara Douglass.
Planetocopia takes the Earth as a basic model and extrapolates the consequences of various simple but drastic changes, to quote the ZBB member who mentioned the page.
Thanks too to everyone who sent in information but didn't want their name to appear here.
