1. The Material Wood Is Made Of Would Sink in Water
Wood is built from millions of tiny tube-shaped cells. The walls of those cells are the "wood" in wood, and the hollow middle of each cell is called a lumen. Scientists have measured the specific gravity of the cell-wall material, which is its weight compared with an equal volume of water. It comes out at about 1.5, and the U.S. Forest Service's handbook uses 1.54. The value is nearly the same for every species of tree [1][2].
That means a cube of pure cell-wall material would weigh about one and a half times as much as a cube of water the same size. If a piece of wood were solid cell wall all the way through, it would sink.
Wood floats because it is not solid all the way through. So the real question becomes: how much of a piece of wood is empty space?
2. Most of a Piece of Dry Wood Is Empty Space, Not Wood
You can work this out from a wood's density, which is easy to measure. A wooden block that has dried in a room has a density you can write as a number relative to water, where water is 1. A table of woods published by Hokkaido, in northern Japan, lists the Japanese cedar called sugi at 0.38, the paulownia (kiri) at 0.30, and balsa at 0.16 [3]. The U.S. handbook agrees that balsa is at the very light end, with an oven-dry density of about 160 kilograms per cubic metre [1].
Now use the cell-wall figure of 1.5. For sugi, the cell wall takes up only 0.38 ÷ 1.5, which is about one quarter of the block. The other three quarters, roughly 75 percent, is gaps. Paulownia works out to about 80 percent gaps, and balsa to about 89 percent.
Treat these as estimates. They use one round value for the cell wall, and the densities are for wood that still holds some moisture, whose weight is counted as if it were cell wall. So the gap shares are rough guides, and the real figures shift a little depending on which numbers you pick. But the picture is clear. Balsa is not light because its material is lighter. Its material is as heavy as any other wood's. It is light because almost nine-tenths of it is gaps. Those gaps are mostly air in dry wood, and air weighs almost nothing next to water.
3. A Piece of Wood Floats When More Than a Third of It Is Gaps
Buoyancy gives a simple test. Water pushes up on anything in it with a force equal to the weight of the water the object pushes aside. A floating object has to have an average density lower than water's, counting everything inside it, gaps included. Put another way, its density has to be below 1.
If the cell wall is 1.5 times as heavy as water and the gaps weigh nearly nothing, the dividing line falls where gaps make up about one third of the volume, because 1 − 1 ÷ 1.5 is about 0.33. Wood with more gaps than that floats. Wood with fewer sinks.
Most familiar woods are far above that line: roughly three-quarters gaps for sugi, and nearly nine-tenths for balsa. That is why they float so easily. The line is also an estimate, since it depends on the 1.5 figure. A slightly different cell-wall value would move it a few points.
4. A Few Woods Sink Even With Gaps: Lignum Vitae and Other Dense Woods
Some woods sit below that line, and they sink even when dry. In the same Hokkaido table, lignum vitae has a density of 1.24, African blackwood 1.21, and king wood 1.20. A small Japanese oak called ubame-gashi measures 1.07 [3]. All are above 1, so they sink.
The arithmetic holds for them too. Lignum vitae is about 83 percent cell wall and about 17 percent gaps, which is well under a third. Even ubame-gashi, with roughly 29 percent gaps, is just on the heavy side of the line. On the other hand, the table shows zelkova (a hardwood common in Japan) at 0.69 and isu-no-ki at 0.90, so both still float.
So "wood floats" is a rule of thumb, not a law. What matters is how crowded the cell walls are.
5. What Happens When Water Slowly Fills the Gaps in Floating Wood
A floating stick is not a permanent float. The gaps are not sealed bubbles. They are the hollow insides of cells, and water can creep into them by capillary action, the same pulling force that lets a paper towel soak up a spill. The handbook explains that water clings to the cell walls and is drawn up the hollows, replacing the air as it goes [1].
Suppose all the gaps filled with water. The block would then be made of cell wall (about 1.5) and water (1.0), and the whole thing would be heavier than water. The Forest Service handbook includes a formula for the water content at which any wood sinks, and for the maximum water wood can hold: from about 267 percent of its dry weight for a very light wood down to about 44 percent for a very dense one [1]. The Japanese forest research institute in Hokkaido gives a similar formula [2].
Using that approach with the densities above, I estimate that a completely filled block of sugi would be about 1.13 times as heavy as the same volume of water, and balsa about 1.05. (The sum is the dry density plus the gap fraction, 1 − density ÷ 1.5, filled with water at 1: for sugi 0.38 + 0.75 ≈ 1.13.) Both are over 1, so both would sink. This is a ceiling, not a prediction: real wood usually does not fill every gap, and how long it takes depends on the species, the size of the piece, and the grain direction.
6. Why Driftwood Can Sink to the Bottom After Soaking for a Long Time
Researchers at Kyoto University's Disaster Prevention Research Institute describe the real-life version of this. Driftwood that has taken in water for a long time gains weight and is expected to sink to the bed of rivers and lakes. In heavy-rain floods, the wood can be swept up again and block dam gates and culverts, so the group has measured experimentally how much water force it takes to move such sunken wood [4].
Their abstract does not say how long soaking takes. I found no source that gives a number of days or years, so I will not guess.
A related fact explains why sunken wood can last. Wood that is completely underwater rots slowly because the water holds only about one-twentieth as much oxygen as air does, at 0 degrees Celsius, and the fungi that cause rot need oxygen [2]. Logs have been stored in ponds for exactly this reason.
7. A Safe Home Experiment: Compare Different Woods in a Bowl of Water
You can test this with wood you already have. Pick two or three pieces of different kinds, such as a wooden chopstick or craft stick, a toy building block, and a pencil. Do not use anything painted or varnished if you want to compare the wood itself. Float them in a bowl of water and look at how much of each piece sits above the surface. The pieces that ride higher are likely to have more gaps.
For a slower test, leave one wooden craft stick in water overnight. In the morning, look at the colour, the feel, and how high it floats. If you have a kitchen scale, weigh it before and again after wiping off the surface water. Do not expect it to sink in one night. Soaking takes much longer than that for most woods.
Water spills easily, so do this on a tray and ask an adult to help. You can also try to do the sugi arithmetic yourself: divide a density by 1.5, subtract the answer from 1, and compare it with the numbers in this article.