Before Darwin, others had already said that species change. What he actually gave us was a mechanism that needs no designer, no foresight and no goal.
It is simple enough to sound like a slogan — which is exactly why it has been misread for a hundred and sixty years. Some read it as "the strongest survive", some as "life is getting better", and some conclude that "an eye is far too precise to have come about by accident". All three are wrong, and each is wrong in its own way.
What follows is about one thing only: why, under certain conditions, change is not something that might happen but something you cannot avoid.
Here is the conclusion up front. Take any group of things. If all three of the following hold at once, the make-up of that group must drift over time, and nothing can stop it:
That is all. Notice that not one word of it mentions biology — so it is indifferent to the material. Cells can run it. So can words in a language, and jokes going round the internet.
The quickest way to feel why none of them is optional is to remove them one at a time.
Remove the differences. A room full of identical things: pick any of them, it makes no odds. The loop spins on empty.
Remove inheritance. A well-fed deer grows strong and outruns the wolf — but its fawns are not born strong because of it. Whatever advantage this generation was painstakingly sifted for is wiped at the next. Wasted sifting.
Remove the effect on offspring number. Who lives and who dies is pure chance. That is luck, not selection — and the difference between the two comes up again later.
So what does "the differences get passed on" look like when you get right down to it? There is a British moth that demonstrates it almost countably.
The peppered moth used to be pale and speckled, near-invisible on lichen-covered bark. Once soot from nineteenth-century industry blackened that bark, entirely black individuals started appearing. Where did the black come from? Someone traced it all the way to something very concrete: a piece of DNA that hops around on its own — some stretches of DNA do not sit still, but now and then copy themselves and insert somewhere else — and this one landed in the middle of a gene that controls when cells divide. It did not break the gene. It just made it work a little harder while the caterpillar was growing its wings, and the wings came out black.
Better still, that hop can be dated. Landing leaves a mark on the DNA around it, and the mark gets worn away generation after generation — how much survives tells you how long it has been. The estimate comes out at around 1819, while the first black moth anyone caught was in Manchester in 1848. Those two-odd decades are just what it takes to go from the only one in the world to common enough for a person to stumble across.
And what is "a bit blacker" actually worth? Someone spent six years releasing 4,864 live moths in the field, watching one by one which ones the birds took. Once the soot had cleared and the lichen grown back, a black moth's daily chance of being eaten ran about a tenth higher than a pale one's. That tenth alone accounts for how fast black moths disappeared from Britain after 1970.
So all three conditions are met: moths come darker and paler (differences), the shade is set by that transplanted piece of DNA and passed on (inheritance), and colour decides whether a bird takes you today (offspring number). The rest requires nobody's intervention.
Spencer coined the phrase in 1864; Darwin only took it up in the fifth edition, in 1869. It sounds like "the strongest survive", and that reading is thoroughly wrong — each of the two words hides a trap.
Take "fittest" first. It does not mean strong. It means how many offspring you leave on average, in this particular environment. A peacock's train is large and conspicuous, it cannot run and it cannot hide, so as strength it makes no sense at all — yet peacocks dragging one do leave more offspring. Change the environment and the same trait flips from asset to liability: bright colours are a courtship advertisement in the rainforest and a set of coordinates for predators on snow.
Now "survival". Staying alive is only the means: live as long as you like, and if you do not reproduce the books read zero. Salmon die right after spawning — a complete failure by longevity, quite a success by offspring left.
And there is a further detail, more lethal than either, that most people never consider: environments fluctuate, and what fluctuation rewards is not high average output but never hitting zero.
Suppose a strategy averages ten offspring a year. Sounds fine. But if every few years brings a total failure, then over the long run it is worth zero — these numbers multiply year on year, and a single zero anywhere wipes out everything accumulated before it. Which is why you see so many arrangements that look inefficient: one batch of seeds germinating across several years, insect eggs hatching in instalments, never everything at once. All of them deliberately lower the average to buy "not being wiped out".
One more objection worth heading off: "isn't survival of the fittest a tautology — whoever survives is by definition the fittest?" It is not, because fitness can be measured before the outcome is in: the mechanics of a beak tell you which seed hardness it can crack, the reflectance of a wing tells you the chance a bird spots it. Predict first, then watch who survives — which is exactly what was done with the Galápagos finches the Grants measured for forty years. In the 1977 drought the small soft seeds went first, 85% of the birds starved, and the survivors had visibly thicker beaks. Yet not one bird's beak ever got thicker; what changed was the proportion of thick-beaked birds. That is the hardest part of the whole idea to turn round in your head: selection never modifies an individual. It only changes proportions.
"An eye is so precise — how could it possibly have appeared by chance?" is probably the commonest objection there is, and its intuition is completely correct: the odds of drawing a working eye in a single go really are close to zero.
The trouble is the word "single". Evolution never runs a one-shot lottery: it saves each tiny improvement and draws again from the save. The gap between those two searches is astronomical.
In 1994 two researchers did the arithmetic properly. Start from a flat patch of skin that can do nothing but sense light, allow the shape to change by one percent per generation — a deliberately conservative rate — and run it all the way to a lensed, image-forming fish eye: about 360,000 generations. For an animal that breeds once a year, under four hundred thousand years. They pushed every assumption toward overestimating the time; the paper is simply titled "a pessimistic estimate".
Be clear what kind of number that is: it was calculated, not excavated — it shows geological time is more than sufficient, not that this is how eyes actually happened. The intermediate forms need no imagining though: eyespots, cup eyes, pinhole eyes and lensed eyes are all still in service among molluscs alive today, each good enough for the life its owner leads. To "what use is half an eye", the answer is: telling light from dark beats being blind.
Accumulation can also be watched directly in a lab. In February 1988 Lenski put a batch of E. coli into each of twelve flasks and has transferred them daily ever since — past seventy-five thousand generations. By generation fifty thousand they were leaving about 70% more offspring than their ancestor, and they had not stopped; the curve shows no sign of flattening.
One of those flasks, at around generation thirty thousand, learned something E. coli normally cannot do: eat citrate. Thawing the frozen samples and re-running from different points showed that the trick only became available once a few specific earlier mutations had accumulated. In other words, "you have to walk the path step by step" is measurable — not a metaphor.
Saving has a price, though: while it preserves the gains it also locks in the baggage — options taken early close off large numbers of later ones. Evolution climbs the highest point reachable from here, not the highest point on the map.