A leaf has a problem it cannot solve. To make sugar it needs carbon dioxide from the air. The only doorway for that gas is a pore, and the same pore is the only doorway for water vapor going the other way. Every time a plant opens its stomata for carbon dioxide, it is billed for water.
This post is about the bill, and about the three different ways plant life has found to pay less of it. Where the water goes, and how it climbs the plant in the first place, is covered separately.
The Problem Is Not Water. It Is RuBisCO
Carbon dioxide sits in the atmosphere at about 425 parts per million. The enzyme that captures it, RuBisCO, is the most abundant enzyme on Earth. It works by adding carbon dioxide to a five-carbon sugar called ribulose-1,5-bisphosphate.
The difficulty is that RuBisCO is not selective. It also adds oxygen to that sugar, and the product cannot be used. The cell has to recover it, and recovering it costs energy and releases much of the carbon again. The whole wasteful circuit is called photorespiration.
The scale of it is not marginal. Roughly 25 percent of RuBisCO reactions add oxygen rather than carbon dioxide, which can cut photosynthetic output by about 25 percent in a plant using the oldest pathway. Each oxygenation costs one ATP and one NAD(P)H directly, and about 25 percent of the carbon that goes in comes back out as carbon dioxide. The ammonia produced also has to be detoxified, at a further cost.
 |
| RuBisCO works inside the chloroplast, in the light reactions’ neighborhood — Image: and3k and caper437, CC BY-SA 3.0, via Wikimedia Commons |
So RuBisCO needs two things at once: a high concentration of carbon dioxide at its site, and a low concentration of oxygen. Raising carbon dioxide inside the leaf means opening the pores wide. Opening the pores means losing water. That is the trade, and it is not an oversight in the design of a leaf. It is the mechanism working exactly as designed.
Why the pore cannot simply be narrower
Stomata are small adjustable openings in the leaf epidermis. Each one is flanked by a pair of guard cells that swell to open it and go slack to shut it. Across species they range from 10 to 80 micrometres in length.
 |
| The stomata are small even against a leaf hair, and a hair is a good deal bigger — Image: Louisa Howard, public domain, via Wikimedia Commons |
The temptation is to solve the problem by making the pores smaller or rarer, and plants do some of that. Many dicots carry more stomata underneath the leaf than on top. Most tree species carry them only underneath. Species with floating leaves put them on top instead, and submerged leaves may have none at all.
The problem is that a small pore still passes a lot of water, and it passes carbon dioxide badly. The same leaf interior that lets water vapor out is already saturated with it. Diffusion runs both ways through the same opening, and the more open the pore, the faster the loss. Shrinking the doorway reduces the bill and reduces the income at the same time.
So the three solutions that evolved do not shrink the pore. They change what passes through it.
C3: Pay the Full Price
The ancestral pathway is usually called C3 because the first product of carbon fixation is a three-carbon molecule. It needs no special anatomy and no extra machinery, and it is what most plants do.
- Carbon dioxide is fixed straight into RuBisCO, in mesophyll cells exposed directly to the air spaces inside the leaf.
- The stomata open as wide as the plant can afford, and the water goes.
- Some of that carbon dioxide is inevitably lost to photorespiration, because RuBisCO meets oxygen at the same site.
The cost is high in water. A plant relying on C3 alone loses about 97 percent of the water its roots take in.
This works well in cool, moist, shady conditions. The extra energy of a more efficient pathway would be wasted, and photorespiration is not a problem because it is too cold to matter.
C4: Buy the Carbon Cheaply, Then Release It
C4 carbon fixation is also called the Hatch-Slack pathway, named for the two Australian scientists who described it in the 1960s. It is an addition on top of C3 rather than a replacement for it.
The trick is to use a different enzyme for the first step. PEP carboxylase binds carbon dioxide much more effectively than RuBisCO does and does not react with oxygen at all. So the plant captures carbon with a cheap, accurate tool, and hands RuBisCO a concentrated solution later, in a place where oxygen is kept out.
- Carbon dioxide is fixed by PEP carboxylase into a four-carbon compound, oxaloacetic acid, in the mesophyll cells.
- That four-carbon compound is converted to malate or aspartate and moves into the bundle sheath cells.
- Inside the bundle sheath it is broken down again, releasing carbon dioxide at high concentration directly around RuBisCO, which suppresses photorespiration.
- The pyruvate left over returns to the mesophyll cell, where the energy of about two extra ATP per gross assimilation rebuilds the PEP.
 |
| A drawing traced from micrographs of a real leaf. The wreath-like ring of bundle sheath cells around each vein is the pump — Image: Ninghui Shi, CC BY-SA 3.0, via Wikimedia Commons |
The anatomy that makes it work is called kranz anatomy, from the German word for wreath. Each vein is surrounded by a ring of bundle sheath cells, itself ringed by mesophyll cells. The bundle sheath chloroplasts are different from the mesophyll ones and contain starch. A layer of suberin is often deposited at the boundary to reduce carbon dioxide leaking back out.
Because the plant is capturing carbon with an accurate enzyme, it can run its stomata less open than a C3 plant would need. Lower stomatal conductance means less water lost and higher water-use efficiency. C4 plants are also cheaper on nitrogen, because PEP carboxylase is less costly to make than RuBisCO.
What it costs, and what it bought
The extra ATP is real, and it is why C4 has not simply replaced C3. Where photorespiration is limited, by cool temperatures or shade, the C3 pathway is more efficient because it skips the pump. A C4 plant in a shady forest is running machinery it does not need.
There is a subtler version of the same problem inside the leaf. Tightening the barrier between mesophyll and bundle sheath raises efficiency in bright light, but it also raises carbon dioxide leakage in dim light, which brings photorespiration straight back. This is described as an inherent and inevitable trade-off, not something a better plant would solve.
The payoff is large even so. About 8,100 species use C4, about 3 percent of all plant species and 5 percent of plant biomass by mass. Those 3 percent of species carry out about 23 percent of all terrestrial carbon fixation. Maize, sugarcane and sorghum are C4 grasses, and about 46 percent of all grasses are C4. No tree above 15 metres uses the pathway at all.
CAM: Stop Paying During the Expensive Hours
Crassulacean acid metabolism takes the third approach. Rather than making the doorway more efficient, it changes when the doorway is open.
- At night the stomata open. Cool air means less water is lost, and carbon dioxide is captured by PEP carboxylase exactly as in C4.
- The carbon is stored as four-carbon malic acid in the cell vacuoles.
- During the day the stomata stay shut. The malate is moved into chloroplasts and broken down, releasing carbon dioxide around RuBisCO at high concentration.
- Photosynthesis then runs on internally stored carbon, with the plant sealed against the hottest and driest part of the day.
The mechanism was first found in the stonecrop family, which is where the name comes from.
The benefit is the timing. A plant that transpires while the sun is down and the air is cool loses far less water than one that must keep its pores open at noon.
 |
| Succulent Crassulaceae use the pathway CAM was named after — Image: Eric Hunt, CC BY-SA 2.5, via Wikimedia Commons |
CAM is more widespread than most people assume. It occurs in roughly 16,000 species across more than 300 genera and about 40 families, and probably that figure is an underestimate because many species show weak or facultative CAM. Over 99 percent of the known 1,700 or so cacti use it. So do pineapples, aloes and the ornamental Kalanchoe sold for drought-tolerant window boxes.
 |
| A CAM succulent that also solves reproduction by growing its own offsets — Image: Ixitixel, CC BY-SA 3.0, via Wikimedia Commons |
It has also escaped the desert more than expected. Some aquatic plants use CAM, and in those cases the limiting factor is not water. Carbon dioxide diffuses through water about 10,000 times more slowly than through air, so a submerged plant faces the opposite shortage: plenty of water and not enough carbon. Night-time uptake is a good solution to that problem too.
 |
| CAM is most concentrated where the growing season is hot and the rainfall is unreliable — Image: Kenricj, map adapted from Anderson 2001, CC BY-SA 3.0, via Wikimedia Commons |
The strangest version is facultative CAM. Some species, including trees in the genus Clusia, run C3 while water is plenty and switch to CAM when the dry season arrives. The plant is not committed to either strategy.
C4 and CAM are not rivals
Some species run both. In Portulaca the two pathways are fully integrated inside the same cells, with CAM metabolites fed directly into the C4 cycle.
The two are also connected in deep time. C4 appeared at least 62 separate times across 19 plant families. That makes it one of the clearest examples of convergent evolution in biology: the same solution invented over and over, from scratch, in unrelated lineages.
 |
| A cactus has no leaves to lose water from, so the stem does the photosynthesis and stores the water — Image: Tangopaso, derivative by Peter coxhead, public domain, via Wikimedia Commons |
Measuring the Price, and Why It Matters Beyond the Leaf
Water-use efficiency is the ratio of carbon gained to water lost. It sounds simple, and it is measured at three different scales that do not always agree.
| Level | What is measured | Also called |
| Leaf | rate of carbon assimilation against transpiration or stomatal conductance | instantaneous, or intrinsic WUE |
| Plant | dry biomass produced against water transpired | transpiration efficiency |
| Field | carbon and water fluxes measured across a whole crop or forest | measured by eddy covariance |
Water-use efficiency is defined at three scales, and the three definitions do not always agree with one another
Intrinsic water-use efficiency usually rises during soil drought, because stomata close. That is often read as drought tolerance, and the reasoning behind it is not straightforward. High water-use efficiency can mean a plant that saves water by closing early.
Low water-use efficiency can instead mean a plant that tolerates drought by some other route. That might be anatomy that resists cavitation, or roots that explore a wide volume of soil.
There is also a catch that matters for agriculture. The two things a crop farmer wants, more yield and less water loss, pull in opposite directions. If a water deficit lowered transpiration without also lowering photosynthesis, everyone would want it. In practice it does not, which is why decades of work trying to raise water-use efficiency in crops has not produced the straightforward win it promised.
Why this matters beyond the leaf
The scale of the evaporation is not a plant-level detail. Forests and crops move huge volumes of water into the air every day, and that flux is large enough to shape weather.
 |
| A large fraction of this flux comes back out as rain, which is why the forest controls its own rainfall — Image: NASA, courtesy Jeff Schmaltz, public domain, via Wikimedia Commons |
Because carbon and water move through the same doorway, plant physiology and the global carbon budget are the same problem viewed from different ends. The pathway a leaf uses decides how much carbon the plant locks up for each unit of water it spends.
 |
| Plants sit at the center of this, and stomata set the exchange rate both ways — Image: U.S. Department of Energy, public domain, via Wikimedia Commons |
It is also moving. Stomatal density responds to atmospheric carbon dioxide, and plants have consistently responded to its rise by producing leaves with fewer and less open stomata. Enrichment experiments find increased photosynthesis, reduced transpiration and raised water-use efficiency together, with simulated yield gains of about 5 to 20 percent at 550 parts per million. Simulated leaf photosynthesis rose 30 to 50 percent in C3 plants but only 10 to 25 percent in C4 ones, which is the same trade-off in miniature.
And there is an try to move the whole machinery. Rice is a C3 crop that feeds a large share of the world, and several projects are trying to convert it to C4 or to an intermediate C2 pathway. It has been observed that no large tree uses C4, so the ceiling on what the pathway can do is set by chemistry rather than by ambition.
Key Terms in Plain English
These are the words in this article that would send you to a dictionary. Each one is given here the way it is actually used above.
| Term | What it means here |
| Transpiration | the drying of water out of a plant as vapor through its pores. Not breathing, and not the same thing as evaporation from a wet surface |
| Stoma (plural stomata) | a small adjustable hole in a leaf, with two guard cells either side that open and shut it. The only doorway for both carbon dioxide and water |
| Photosynthesis | the job a plant does with sunlight: turning carbon dioxide and water into sugar, and giving off oxygen as a by-product |
| RuBisCO | short for ribulose-1,5-bisphosphate carboxylase/oxygenase. The enzyme that captures carbon dioxide. It also grabs oxygen by mistake, which is the problem this whole article is about |
| Photorespiration | the wasteful clean-up after RuBisCO grabs oxygen instead of carbon dioxide. It burns energy and releases back some of the carbon it captured |
| C3 plant | a plant that uses the original method of capturing carbon, straight into RuBisCO. Cheap to run, and expensive in lost water |
| C4 plant | a plant that captures carbon first with a cheaper enzyme and then pumps it to RuBisCO, so it can keep its pores narrower and lose less water |
| PEP carboxylase | the enzyme a C4 or CAM plant uses for the first step instead of RuBisCO. It grabs carbon dioxide accurately and ignores oxygen entirely |
| CAM | short for Crassulacean acid metabolism. A plant opens its pores at night, stores the carbon as an acid, and uses it during the day with the pores shut |
| Kranz anatomy | from the German word for wreath. The ring of cells around each leaf vein that lets a C4 plant keep carbon dioxide and oxygen apart |
| Mesophyll cell | an ordinary leaf cell near the surface, exposed straight to the air spaces inside the leaf |
| Bundle sheath cell | a cell in the ring around a vein. In a C4 plant it is where carbon dioxide is released back out for RuBisCO to use |
| Water-use efficiency | how much carbon a plant gains for each unit of water it loses. Measured at leaf, plant and field level, which do not always agree |
| Evapotranspiration | water leaving a landscape as a whole, from wet ground as well as from leaves. A forest moves an huge amount of it into the air |
These are the terms that carry the argument above. Every one of them is used here in the sense given here, and nowhere else.
Key Takeaways
- Transpiration is evaporation. Plants do breathe, but that is respiration, a separate process using the same pores.
- Stomata are the only route for both carbon dioxide in and water vapor out, so the two flows are inseparable.
- RuBisCO binds oxygen as well as carbon dioxide, which is why it needs high carbon dioxide and why photorespiration wastes about a quarter of the carbon it touches.
- C3 is the ancestral pathway. It costs the most water, losing about 97 percent of what the roots take up.
- C4 fixes carbon with PEP carboxylase first and pumps it to RuBisCO. That allows narrower stomata and better water-use efficiency, at the cost of about 2 extra ATP per gross assimilation.
- C4 is convergent evolution: at least 62 independent origins across 19 families. About 8,100 species, 3 percent of species, 23 percent of terrestrial carbon fixation.
- CAM opens the pores at night and seals them by day, storing carbon as malic acid. It occurs in roughly 16,000 species and in over 99 percent of cacti.
- CAM is not only a desert strategy. Aquatic plants use it because carbon dioxide diffuses about 10,000 times more slowly through water.
- Rising atmospheric carbon dioxide produces leaves with fewer stomata, and C4 plants gain less from the enrichment than C3 plants do.
Frequently Asked Questions
Do plants breathe?
Yes, but that is a different process. Respiration consumes oxygen and releases carbon dioxide to release energy from sugars. Transpiration is the evaporation of water through the stomata. Both use the stomata, which is why the two are so often confused.
What exactly is photorespiration?
When RuBisCO adds oxygen instead of carbon dioxide, the product cannot be used by the Calvin cycle. The cell recovers it through a network of reactions across chloroplasts, peroxisomes and mitochondria, spending energy to release much of the carbon again. About 25 percent of RuBisCO reactions do this.
What is the difference between C3 and C4 plants?
C4 plants capture carbon with PEP carboxylase first, then concentrate it around RuBisCO inside bundle sheath cells. That suppresses photorespiration and lets the plant run its stomata less open. C3 plants fix carbon directly with RuBisCO, which is simpler and cheaper but loses more water.
How is CAM different from C4?
C4 separates the reactions in space, between mesophyll and bundle sheath cells. CAM separates them in time, capturing carbon at night and using it during the day. Both end up with carbon dioxide concentrated around RuBisCO.
Why doesn't every plant use C4?
Because the pump costs energy. C4 spends roughly 2 extra ATP per gross assimilation, and where photorespiration is already limited by cool temperatures or shade, that spend buys nothing. Plenty of shady and cool-climate plants are better off as C3.
Is high water-use efficiency always a good thing?
No. In crops, higher efficiency and higher yield pull in opposite directions, because both come from the same stomatal behavior.
It also does not mean drought tolerance. A plant can have low water-use efficiency and still handle drought well by resisting cavitation or by exploring a wide volume of soil.
Can a plant switch between pathways?
Yes. Facultative CAM species, including some trees in the genus Clusia, photosynthesize as C3 while water is plentiful and switch to CAM in the dry season. Some Portulaca species run C4 and CAM together in the same cells.
Are C4 plants more advanced than C3 plants?
No. C4 is an adaptation to hot, bright, open conditions, and it has arisen independently at least 62 times, which is what convergent evolution looks like. In a cool, shaded forest C3 wins, because C4 would be paying for a pump it does not need.
Why do cacti have no leaves?
A leaf is a broad thin sheet built to absorb light, and a broad thin sheet is also built to lose water. A cactus keeps its photosynthetic tissue in a thick stem, which has a much lower surface-area-to-volume ratio, and uses CAM so it only needs to open its pores at night.
Sources: Wikipedia articles on C4 carbon fixation, Kranz anatomy, Crassulacean acid metabolism, CAM evolution, stomata, water-use efficiency, RuBisCO, photorespiration and photosynthesis. Images: Wikimedia Commons, with authors and licenses noted in each caption.
