When we analysed differential rent, we started from the premise that the worst land pays no ground-rent — or, put more generally, that land pays ground-rent only where the individual production price of its product lies below the production price that governs the market, so that a surplus profit arises in this way and turns into rent. The first thing to note is that the law of differential rent, as differential rent, is completely independent of whether that premise is right or wrong.

Call P the general production price, the one that governs the market. For the product of the worst type of land, A, this P is exactly A's own individual production price. In other words, the price A's product fetches pays back the constant and variable capital used up in producing it, plus the average profit (= profit of enterprise plus interest).

So on A the rent is zero. On the next better type of land, B, the individual production price is P’, and P > P’: the market price P pays more than the actual production price of B's product. Let P - P’ = d. This d, the amount by which P exceeds P’, is the surplus profit that the farmer on class B makes, and it turns into rent, which he has to pay to the landowner. For the third class of land, C, let the actual production price be P’’, with P - P’’ = 2d: this 2d turns into rent. In the same way the fourth class, D, has the individual production price P’’’, and P - P’’’ = 3d, which turns into ground-rent; and so on.

Now suppose that for class A the premise is false — the premise that its rent = 0, so that the price of its product = P + 0 — and that A, too, pays a rent, = r. Then two things follow.

Firstly: the price of class A's product would not be governed by its price of production. It would contain an excess over it; it would be P + r. Here is why. Take the capitalist mode of production in its normal state, so that the extra r the farmer hands to the landowner comes out neither of wages nor of capital's average profit. Then the farmer can pay it only by selling his product above its price of production. That excess would be a surplus profit for him, if he did not have to give it up to the landowner as rent.

The market price that governs the whole product on the market, from every type of land, would then not be the price of production that capital in general brings in, in every sphere of production — outlays plus the average profit. It would be the price of production plus the rent: P + r, not P. For the price of class A's product marks, in general, the limit of the governing general market price, the price at which the whole supply can be brought to market; and to that extent it governs the price of that whole product.

But secondly, in this case, even though the general price of the land's product would be substantially modified, the law of differential rent would in no way be abolished by this. For if the price of class A's product, and with it the general market price, = P + r, then the price of classes B, C, D etc. would also = P + r. But since for class B P - P’ = d, (P + r) - (P’ + r) would likewise = d; for C, (P + r) - (P’’ + r) would = 2d; and finally for D, (P + r) - (P’’’ + r) would = 3d, and so on. So differential rent would be the same as before and would be governed by the same law, even though the rent would contain an element independent of this law and would rise generally along with the price of the land's product.

It follows that, however things may stand with the rent of the least fertile types of land, the law of differential rent is not only independent of it; the only way to grasp differential rent itself in keeping with its character is to set the rent of land class A = 0. For whether it is = 0 or > 0 makes no difference as far as differential rent is concerned, and indeed does not enter the account.

So the law of differential rent is independent of the result of the investigation that follows.

If we now ask further what the premise rests on — the premise that the product of the poorest land, class A, pays no rent — then the answer has to be this. If the market price of the product of the land, say of corn, has risen high enough that an extra advance of capital placed in class A land pays the customary price of production, and so yields the capital the customary average profit, that condition is enough for the extra capital to be invested in class A land. That is, the condition is enough for the capitalist to put new capital in at the customary profit and to valorize it in the normal way.

Note here that even in this case the market price must stand higher than A's price of production. For as soon as the extra supply is there, the relation between demand and supply has plainly changed: before, the supply was not enough; now it is enough. So the price must fall. For it to be able to fall, it must have stood higher than A's price of production. But the newly cultivated class A land is less fertile, and that is why the price does not fall back as low as it was when class B's price of production still governed the market. A's price of production is the limit — not for a temporary rise of the market price, but for the relatively permanent one.

If on the other hand the land newly brought into cultivation is more fertile than class A, which governed the price until now, and yet is only enough to cover the extra demand, then the market price stays as it was. And in this case too, the question whether the lowest class of land pays a rent comes to the same thing as the one to be pursued here. For here too the premise that class A land pays no rent would be explained by the fact that the market price is enough for the capitalist tenant farmer to cover exactly the capital laid out plus average profit — in short, that the market price delivers him the price of production of his commodity.

In any case the capitalist tenant farmer can cultivate class A land under these conditions, so far as he has to decide as a capitalist. The condition for the normal valorization of capital on class A land is now present. But from the premise that the tenant farmer could now invest capital on class A land in line with the average conditions for valorizing capital, even if he had no rent to pay, it by no means follows that this class A land is now simply at his disposal. The fact that he could make the customary profit if he paid no rent is no reason at all for the landowner to lend him his land for nothing, and to be so philanthropic toward a business friend as to give him free credit. What such a premise contains is an abstraction from landed property, the abolition of landed property — whose very existence is a barrier to the investment of capital and to valorizing it on the land however one likes. And that barrier in no way falls just because the tenant farmer reflects that the level of corn prices would let him extract the customary profit from his capital by working class A land, if he paid no rent — that is, if he could in practice treat landed property as not existing. But the monopoly of landed property — landed property as a barrier to capital — is presupposed in differential rent; without it the surplus profit would not turn into ground-rent and would not fall to the landowner instead of the tenant farmer. And landed property remains a barrier even where rent as differential rent falls away, that is, on class A land.

If we look at the cases in a country of capitalist production where capital can be invested in the land without any rent being paid, we will find that they all involve a factual, if not legal, abolition of landed property — an abolition that can only take place under quite specific circumstances, accidental by their nature.

Firstly: if the landowner is himself the capitalist, or the capitalist is himself the landowner. In this case, once the market price has risen enough to get the price of production out of what is now class A land — that is, replacement of capital plus average profit — he can farm his own plot himself. But why? Because landed property sets him no barrier to investing his capital. For him the land can be simply an element of nature, so he can let himself be guided solely by what the valorization of his capital demands — by capitalist considerations. Such cases do occur in practice, but only as an exception. Just as capitalist cultivation of the land presupposes a separation between functioning capital and landed property, it rules out self-farming by the landowner as the rule. It is plain at once that this is purely accidental. If the increased demand for corn requires a greater extent of class A land to be cultivated than is in the hands of self-farming owners — if, that is, part of it must be leased if it is to be cultivated at all — then this hypothetical abolition of the barrier that landed property puts in the way of investing capital falls away at once. It is an absurd self-contradiction to begin from the split between capital and land, between tenant farmer and landowner, which belongs to the capitalist mode of production, and then to turn round and take it as the rule that landowners farm their own land — to exactly the extent, and wherever, capital would get no rent out of farming the land if no landed property stood over against it independently. (See the passage in Adam Smith on the rent of mines, quoted below.) This abolition of landed property is accidental. It may occur, or it may not.

Secondly. Within the whole complex of a single lease there may be individual pieces of land that pay no rent at the given level of market prices — that are in fact let for nothing — but the landowner does not look at them that way, because what he has in view is the total rent of the land he has leased out, not the particular rent of each separate piece. In this case, as far as the non-rent-bearing pieces of the lease are concerned, landed property falls away for the farmer as a barrier to investing capital — and it falls away by contract with the landowner himself. But he pays no rent on those pieces only because he pays rent on the land to which they are an accessory. What is assumed here is precisely a combination in which the worse land type A does not have to be resorted to as a separate, new field of production in order to supply the missing amount; instead it forms only an inseparable strip running inside the better land. But the case that has to be investigated is precisely the one where tracts of type-A land are farmed independently, and therefore have to be leased independently, under the general preconditions of the capitalist mode of production.

Thirdly. A farmer can invest extra capital on the same lease even though, at the existing market prices, the extra product he gets this way brings him only the price of production and yields him the ordinary profit, without enabling him to pay an additional rent. So with one part of the capital invested in the land he pays ground-rent, and with the other part he does not.

But you can see how little this supposition solves the problem. If the market price (and at the same time the fertility of the soil) enables him to get a surplus yield with the extra capital that, like the old capital, yields him a surplus profit on top of the price of production, then he pockets that surplus profit himself for the duration of the lease. But why? Because, as long as the lease lasts, the barrier of landed property to investing his capital in the land has fallen away.

Yet the mere fact that, to secure him this surplus profit, additional worse land must be independently taken up and independently leased proves irrefutably that investing additional capital on the old land does not suffice to produce the increased supply that is needed. The one assumption excludes the other. Now one could indeed say that the rent of the worst land type A is itself a differential rent, compared either with land the owner cultivates himself (though this occurs only as a chance exception) or with the additional capital investments on the old leases that yield no rent. But this would be, first, a differential rent that did not arise from differences in the fertility of the soil types, and that therefore did not presuppose that land type A pays no rent and sells its product at the price of production. And second, whether additional capital investments on the same lease yield rent or not is just as immaterial to whether newly cultivated class-A land pays rent or not as it is, for example, for setting up a new independent factory whether another manufacturer in the same line puts part of his capital into interest-bearing paper because he cannot fully valorize it in his own business, or whether he adds extensions here and there that bring him less than the full profit but more than the interest. To him that is a side matter. New establishments, by contrast, must yield the average profit, and are set up on that expectation.

Admittedly, the additional capital investments on the old leases and the additional cultivation of new class-A land set limits to one another. How far extra capital can go on being invested on the same lease, under worse conditions of production, is set by the competing new investments on class-A land; and, the other way round, how much rent class-A land can yield is capped by the competing extra investments on the old leases.

But all these false evasions still do not solve the problem, which, put simply, is this. Suppose the market price of corn (which in this investigation stands for every product of the soil) were high enough for parts of land class A to be taken into cultivation, and for the capital invested in those new fields to fetch the price of production of the product — that is, replacement of capital plus average profit. Suppose, then, that the conditions for the normal valorization of capital on class-A land were present. Is that enough? Can this capital then really be invested? Or must the market price rise so far that even the worst land, A, yields a rent? In other words, does the landowner's monopoly impose on the investing of capital a barrier that, from a purely capitalist standpoint, would not exist without the existence of this monopoly? From the terms of the question itself it follows that if, for instance, there are additional capital investments on the old leases that at the given market price yield no rent but only the average profit, this fact by no means settles the question whether capital can now really be invested on class-A land, which would likewise yield the average profit but no rent. For that is precisely the question. That the additional capital investments which yield no rent do not satisfy the demand is proved by the necessity of bringing in new class-A land.

If the additional cultivation of A land takes place only insofar as it yields rent, that is, more than the price of production, then only two cases are possible. Either the market price must stand high enough that even the last additional capital investments on the old leases yield surplus profit, whether the farmer pockets it or the landowner. This rise in price and this surplus profit of the last additional investments would then be a consequence of the fact that A land cannot be cultivated without yielding rent. For if the price of production — the mere yielding of average profit — were enough for cultivation, the price would not have risen so far, and competition from the new lands would have set in as soon as they merely yielded these prices of production. Capital investments on A land that likewise yield no rent would then compete with the additional capital investments on the old leases that yield no rent.

Or else the last capital investments on the old leases yield no rent, but the market price has nevertheless risen high enough for A land to be taken up and to yield rent. In this case, the additional capital investment that yields no rent was possible only because A land cannot be cultivated until the market price allows it to pay rent. Without this condition, its cultivation would have begun at a lower price level; and those later capital investments on the old leases, which need the high market price to yield the ordinary profit without rent, could not have taken place. At the high market price they of course yield only the average profit. At a lower price, which would have become the governing one along with the cultivation of A land as its price of production, they would not have yielded this profit, and so, on this supposition, would not have taken place at all. A's rent would thus indeed form a differential rent, as compared with those capital investments on the old leases that yield no rent. But that A's tracts form such a differential rent is only a consequence of their not being accessible to cultivation at all unless they yield a rent; that is, of the fact that the necessity of this rent — a rent not, in and for itself, conditioned by any difference in the types of land — holds, and forms the barrier to the possible investment of additional capitals on the old leases. In both cases, A's rent would not be the simple consequence of rising corn prices, but the opposite: the fact that the worst land must yield rent before its cultivation is allowed at all would be the cause of corn prices rising to the point where this condition can be fulfilled.

Differential rent has something odd about it: here landed property only intercepts a surplus profit — one the farmer would otherwise pocket, and under certain circumstances really does pocket for as long as his lease runs. Landed property is only the cause of a transfer of part of the commodity price. That part comes about without any help from landed property — it comes instead from competition fixing the production price that governs the market — and it resolves into surplus profit. Landed property moves this part of the price from one person to another, from the capitalist to the landowner. But here it is not the cause that creates this part of the price, or the price rise that part presupposes.

The other case is different. If the worst land, A, cannot be cultivated — even though cultivating it would yield the production price — until it yields a surplus above that production price, a rent, then landed property is the creative basis of this price rise. Landed property itself has produced rent.

Nothing is changed if, as in the second case we looked at, the rent now paid by land A forms a differential rent when compared with the last extra capital invested on old leaseholds, which pays only the production price. For it is only because land A cannot be farmed until the governing market price has climbed far enough for A to pay a rent that the market price rises to a point that pays the last capital investments on the old leaseholds only their production price, though a production price that at the same time yields a rent for land A. And that A has to pay rent at all is here the cause of the differential rent that forms between land A and the last capital investments on the old leaseholds.

When we say that land of class A pays no rent — on the assumption that the corn price is governed by the production price — we mean rent in the categorical sense of the word. If the farmer pays a lease-price that is a deduction, whether from the normal wages of his workers or from his own normal average profit, then he pays no rent: nothing that is an independent part of the price of his commodity, distinct from wages and profit. It has already been noted that this happens all the time in practice. To the extent that farm labourers' wages in a country are generally pushed below the normal average — so that a deduction from wages, a part of the wage, regularly ends up in rent — the farmer on the worst land is no exception. In the same production price that makes cultivation of the worst land possible, this low wage is already a constituent item, and so selling the product at the production price does not enable the farmer of that land to pay a rent. The landowner can also lease his land to a worker who is content to pay over to someone else, in the form of rent, all or the greater part of what the sale price gives him above his wages. In all these cases, however, no real rent is paid, even though a lease-price is. But where relations exist that correspond to the capitalist mode of production, rent and lease-price must coincide. And it is precisely this normal situation that is to be examined here.

If even the cases considered above — where capital investments on the land can really be made, inside the capitalist mode of production, without yielding rent — decide nothing for our problem, then pointing to colonial conditions decides still less. What makes a colony a colony — and here we are speaking only of agricultural colonies proper — is not just the mass of fertile land lying in its natural state. It is rather the fact that this land is not appropriated, not subsumed under landed property. This is what makes the enormous difference between the old countries and the colonies as far as the land goes: the legal or actual non-existence of landed property, as Wakefield correctly notes, and as Mirabeau père, the Physiocrat, and other older economists had discovered long before him. It is quite immaterial here whether the colonists simply appropriate the land, or whether they pay the state, under the name of a nominal land price, in fact only a fee for a valid legal title to the land. It is also immaterial that colonists already settled are legal owners of land. In fact, landed property forms no barrier here to the investment of capital, or of labour without capital; the seizure of one part of the land by already settled colonists does not shut new arrivals out from the possibility of making new land a field for their capital or their labour. So if the task is to examine how landed property affects the prices of land's products, and rent, where it restricts the land as a field of investment for capital, then it is utterly absurd to speak of free bourgeois colonies, where neither the capitalist mode of production in agriculture exists nor the form of landed property that corresponds to it — where the latter in fact does not exist at all. This, for example, is what Ricardo does in the chapter on ground-rent. At the opening he says he wants to examine the effect of the appropriation of land on the value of land's products, and right after that he takes the colonies as his illustration, assuming there that land is present in a relatively elementary state and that its exploitation is not restricted by the monopoly of landed property.

Owning land in law, and nothing more, does not bring the owner any ground-rent. What it does give him is the power to keep his land out of use until economic conditions allow it to be used in a way that yields him a surplus — whether for farming proper or for other productive purposes, such as building. He cannot add to or take away from the total amount of this field of employment; what he can change is how much of it is on the market. That is why, as Fourier already remarked, it is a characteristic fact that in every civilized country a relatively large share of the land always stays out of cultivation.

Suppose, then, that demand calls for new land to be broken up — say, land less fertile than any farmed so far. Will the landowner let it for nothing just because the market price of farm produce has climbed far enough for capital put into this land to pay the farmer its price of production, and so the ordinary profit? Not at all. The investment has to yield the landowner a rent; he will not let the land until a lease-price can be paid to him. So the market price must have climbed above the price of production, to P + r, far enough for a rent to be paid to the landowner. But since landed property, on our assumption, brings its owner nothing unless it is leased — is economically worthless to him — a small rise of the market price above the price of production is enough to bring new land of the worst kind onto the market.

Now the question is this. Ground-rent on the poorest land cannot be derived from any difference in fertility. Does it follow that the price of what that land produces is necessarily a monopoly price in the ordinary sense — or a price in which rent enters the way a tax does, except that here the landowner collects the tax instead of the state?

That this tax has its given economic limits goes without saying. It is bounded by extra capital invested on the old leaseholds, by competition from foreign farm products (assuming these can be imported freely), by competition among landowners, and finally by what consumers need and can afford to pay. But that is not what is at issue here. What is at issue is whether the rent paid by the poorest land goes into the price of its product — which, on the premise, is what governs the general market price — the same way a tax goes into the price of the commodity it is imposed on: as an element independent of that commodity's value.

It by no means necessarily follows, from the rent of the poorest land, that the price of its product is a monopoly price in the ordinary sense, or a price that takes in rent the way a price takes in a tax. That has been claimed only because, until now, the difference between the value of commodities and their price of production had not been grasped. We have seen that a commodity's price of production is by no means the same as its value. It is true that prices of production, taken all together, are governed only by the total value of commodities; and it is true that, other things being equal, the prices of production of the different kinds of commodity move solely as their values move. Even so, it has been shown that a commodity's price of production can lie above its value or below it, and coincides with its value only by exception.

So if agricultural products sell above their price of production, that by itself is no proof that they also sell above their value — any more than industrial products selling on average at their price of production proves that they sell at their value. It is possible for agricultural products to sell above their price of production and below their value, while on the other side many industrial products bring in only their price of production because they sell above their value.

How a commodity's price of production stands to its value depends on one thing only: the proportion between the variable part and the constant part of the capital that produces it — the organic composition of that capital. Suppose the capital in some sphere of production has a lower composition than the average social capital: its variable part, laid out on wages, is larger relative to its constant part, laid out on the material conditions of labour, than it is in the average capital. Then the value of its product must lie above its price of production. The reason: such a capital employs more living labour, so, with labour exploited to the same degree, it produces more surplus-value, and so more profit, than a slice of the average social capital of the same size. The value of its product therefore lies above its price of production, since that price is only the capital replaced plus the average profit, and the average profit is smaller than the profit actually produced in this commodity. The average social capital produces less surplus-value than a capital of this low composition.

With a capital of higher composition than the social average it is the other way round: the value of the commodities it produces lies below their price of production — as it generally does with the products of the most developed industries.

When the capital in some sphere of production has a lower composition than the average social capital, that is, to begin with, just another way of saying that the productive power of social labour in that sphere is below the average level. For the level that productive power has reached shows itself in the relative weight of the constant part of capital over the variable part, or in the steady decline of the share of a given capital that is laid out on wages. Conversely, a higher composition of capital in a particular sphere expresses a development of productive power above the average level.

Leaving aside genuinely artistic work, which by the nature of the case falls outside our subject, it goes without saying that different spheres of production, given their technical peculiarities, require different proportions of constant and variable capital, and that living labour must take up more room in some and less in others. In extractive industry, for example, which must be strictly distinguished from agriculture, raw material drops out altogether as an element of constant capital, and even auxiliary material plays a significant role only here and there. In mining, on the other hand, the other part of constant capital, fixed capital, plays a significant role. Even so, here too the progress of development can be measured by the relative growth of constant capital compared with variable.

If capital in agriculture proper has a lower composition than the average social capital, that would, on the face of it, express the fact that in countries of developed production agriculture has not advanced as far as manufacturing industry. Leaving aside all the other economic circumstances — and some of them are decisive — such a fact would already be explained by this: the mechanical sciences, and above all their use in practice, developed earlier and faster than chemistry, geology and physiology. These developed later — some of them only very recently — and their application to agriculture, above all, came later still. It is, besides, an undisputed fact, and one long known, that progress in agriculture itself steadily shows up as the constant part of capital growing relative to the variable part. Whether in a particular country of capitalist production — England, for instance — agricultural capital really has a lower composition than the average social capital is a question that only statistics can settle, and for our purpose there is no need to go into it in detail. In any case, this much is settled in theory: only on that premise can the value of agricultural products lie above their price of production — only, that is, if a capital of a given size produces more surplus-value in agriculture than an equally large capital of average social composition does, or, which comes to the same thing, sets in motion and commands more surplus labour (and so employs more living labour altogether).

So for the form of rent we are examining here — which can arise only under this assumption — it is enough to make the assumption. Where the hypothesis falls away, the form of rent that corresponds to it falls away too.

Take the bare fact that farm products carry a value higher than their price of production. By itself that fact would in no way be enough to explain a ground-rent that does not depend on differences in the fertility of the soil or on successive investments of capital on the same soil — in short, a rent conceptually distinct from differential rent, which we can therefore call absolute rent. A good many manufactured products have a value above their price of production and still throw off no surplus over the average profit, no surplus profit that could be turned into rent. The opposite is the case. The very existence and the concept of the price of production, and of the general rate of profit it contains, rest on commodities not being sold at their value. Prices of production come out of an equalization of commodity values: once the capital values consumed in the different spheres are replaced, the total surplus-value is distributed not in the proportion in which it was produced in each sphere and so sits in its products, but in proportion to the size of the capitals advanced. Only in this way does an average profit arise, and with it the price of production of commodities, whose characteristic element that average profit is.

It is the constant tendency of capitals, through competition, to bring about this equalization in the distribution of the surplus-value that total capital produces, and to overcome every obstacle to it. Their tendency is therefore to tolerate only such surplus profits as arise, whatever the circumstances, not from the difference between the values of commodities and their prices of production, but from the general price of production governing the market and the individual prices of production that differ from it — surplus profits that accordingly arise not between two different spheres of production but within each sphere, and so leave the general prices of production of the different spheres, that is the general rate of profit, untouched; they rather presuppose the transformation of values into prices of production and the general rate of profit. This presupposition, however, rests — as discussed earlier — on the total social capital being shared out among the different spheres in proportions that keep changing: on capital constantly flowing into some spheres and out of others, on its being able to move from one sphere to another — in short, on its moving freely between the spheres, each of them open as a field of investment to the independent parts into which the total social capital is divided. What is assumed here is that no barrier, or only an accidental and temporary one, prevents the competition of capitals from reducing value to price of production — say, in a sphere whose commodities have a value above their price of production, or whose surplus-value is above the average profit — and so from distributing that sphere's extra surplus-value proportionally among all the spheres capital exploits.

But if the opposite happens — if capital meets an alien power that it can overcome only partly or not at all, a power that restricts its investment in particular spheres and permits it only on conditions that wholly or partly shut out that general equalization of surplus-value into average profit — then in such spheres a surplus profit would plainly arise from the excess of commodity value over its price of production, and that surplus profit could be turned into rent and, as rent, become something standing on its own over against profit. And it is as just such an alien power and barrier that landed property faces capital in its investments in land, or the landowner faces the capitalist.

Here landed property is the barrier. It permits no new capital investment on land not yet cultivated or not yet leased unless it levies a toll — that is, unless a rent is demanded — even though the land newly brought under cultivation is of a kind that yields no differential rent, and even though without landed property it could have been cultivated at a smaller rise in the market price, so that the governing market price would have paid the man who farms this worst land only his price of production. But because landed property stands in the way, the market price has to climb until the land can pay an excess over its price of production — that is, a rent.

Since the value of the commodities produced by agricultural capital stands above their price of production, by assumption, this rent — apart from a case to be examined in a moment — forms the excess of value over the price of production, or a part of that excess. Whether the rent equals the whole difference between value and price of production, or only a greater or lesser part of it, would depend entirely on how supply stands in relation to demand and on how much land is drawn into new cultivation. As long as the rent is not equal to the excess of the value of agricultural products over their price of production, a part of that excess would always enter the general equalization and proportional distribution of all surplus-value among the individual capitals. As soon as the rent equals the excess of value over the price of production, that whole part of the surplus-value over and above the average profit would be withdrawn from this equalization.

But whether this absolute rent equals the whole excess of value over the price of production, or only a part of it, agricultural products would always be sold at a monopoly price — not because their price stood above their value, but because it stood at their value, or below their value and yet above their price of production. Their monopoly would lie in this: that they are not levelled down to the price of production, as other industrial products are, whose value stands above the general price of production. One part of value and of price of production is a factually given constant — the cost price, the capital consumed in production = k. So the difference between them lies in the other, variable part, the surplus-value, which in the price of production = p, profit, that is the total surplus-value reckoned on social capital and on each individual capital as a proportional part of it, but which in the value of the commodity is equal to the actual surplus-value that this particular capital has produced, and forms an integral part of the commodity values it has created. If the value of a commodity stands above its price of production, then the price of production = k + p and the value = k + p + d, so that p + d = the surplus-value contained in it. The difference between value and price of production is therefore = d, the excess of the surplus-value this capital has produced over what the general rate of profit assigns to it.

Several things follow from this. The price of agricultural products can stand above their price of production and still fall short of their value. Up to a point, their price can also rise lastingly before it has reached their value. It is only because of the monopoly of landed property that the amount by which their value exceeds their price of production can become a determining element of their general market price. And finally, in this case the rent is not the effect of the product growing dearer: the product grows dearer because of the rent. If the price of the product from a unit of area of the worst land = P + r, then every differential rent rises by the corresponding multiple of r, since by assumption P + r becomes the governing market price.

Suppose the non-agricultural social capital had an average composition of 85c + 15v, with a rate of surplus-value of 100 per cent: its price of production would then be 115. Suppose agricultural capital were made up of 75c + 25v: at the same rate of surplus-value the value of its product, and the governing market value, would be 125. If the agricultural and the non-agricultural product evened out at an average price (for brevity we take the total capital in the two branches to be the same size), the total surplus-value would be 40, which is 20 per cent on the 200 of capital, and each product would sell at 120. Under an equalization to prices of production, then, the non-agricultural product would, on average, fetch a market price above its value and the agricultural product one below its value. If agricultural products sold at their full value, they would be 5 higher, and industrial products 5 lower, than under the equalization. If the state of the market does not let agricultural products sell at their full value — at the whole excess over their price of production — the effect lies between the two extremes: industrial products would sell somewhat above their value, and agricultural products somewhat above their price of production.

Landed property can push the price of agricultural products above their price of production. But how far the market price climbs above the price of production, towards their value, is not up to landed property: it depends on the general state of the market. So it also depends on the market how far the surplus-value produced in agriculture beyond the given average profit turns into rent, or instead goes into the general equalization of surplus-value to the average profit. Either way, this absolute rent, which arises from the excess of value over price of production, is no more than a part of agricultural surplus-value: that surplus-value turned into rent and intercepted by the landowner — just as differential rent arises when surplus profit is turned into rent and intercepted by landed property, at the general governing price of production. Only these two forms of rent are normal. Outside them, rent can rest only on a monopoly price proper, which depends neither on the commodities' price of production nor on their value, but on what the buyers need and are able to pay; examining it belongs to the theory of competition, where the way market prices actually move is looked into.

Suppose all the land in a country that can be farmed were leased out — taking the capitalist mode of production and normal conditions generally for granted. Then there would be no land that did not yield rent. But there could be capital investments, individual portions of the capital laid out on the land, that yielded no rent. For once the land is leased, landed property stops acting as an absolute barrier to the capital investment that is needed. It carries on as a relative barrier even then, in this respect: at the end of the lease the capital worked into the soil falls back to the landowner, and that sets the farmer very definite limits. Only in this case would all rent turn into differential rent — and not differential rent determined by differences in the quality of the soil, but by the difference between the surplus profits that a particular piece of land yields after the last capital investments on it, and the rent that would be paid for leasing the worst class of land. Landed property is an absolute barrier only in so far as being let onto the land at all, as a field for capital, requires paying tribute to the landowner. Once that permission has been given, he can no longer set any absolute limits to how much capital is laid out on a given piece of land. With building a house, a barrier is always set by a third party's landed property in the ground the house is to be built on. But once that ground has been leased for building, it depends on the lessee whether he puts up a tall house or a low one.

If agricultural capital had, on average, the same composition as the average social capital, or a higher one, absolute rent would disappear — always in the sense worked out here, that is, the rent that differs both from differential rent and from rent resting on a monopoly price proper. The value of the agricultural product would then no longer lie above its price of production; agricultural capital would set no more labour in motion than non-agricultural capital, and so would realize no more surplus labour either. The same would happen if, as cultivation progressed, the composition of agricultural capital came level with that of the average social capital.

At first sight it seems a contradiction to assume two things together: on the one hand, that the composition of agricultural capital rises — its constant part growing against its variable part — and on the other, that the price of the land's product rose high enough for new land, worse than any farmed so far, to pay a rent: a rent which, in this case, could come only from an excess of the market price over the value and the price of production, in short, only from a monopoly price of the product.

Here a distinction has to be made.

First: in looking at how the rate of profit is formed, we saw that capitals which, technically regarded, are composed in the same way — that is, set the same amount of labour going in proportion to machinery and raw material — can still be composed differently, because the constant parts of those capitals have different values. The raw material or the machinery can cost more in the one case than in the other. To set the same mass of labour going (and by assumption that is what is needed to work up the same mass of raw material), a larger capital would have to be advanced in the one case than in the other. With a capital of 100, for instance, I cannot set the same amount of labour going if the raw material that also has to be paid for out of the 100 costs 40 in one case and 20 in the other. But that these capitals are still technically composed in the same way would show itself at once if the price of the dearer raw material fell to the price of the cheaper. The value ratios between variable and constant capital would then have become the same, although nothing had changed in the technical relation between the living labour applied and the mass and nature of the conditions of labour applied.

Conversely, a capital of lower organic composition could, simply through a rise in the values of its constant parts, appear to reach the same level as a capital of higher organic composition — looked at from the standpoint of value composition alone. Take one capital of 60c + 40v, because it uses a great deal of machinery and raw material in relation to living labour-power, and another of 40c + 60v, because it uses a lot of living labour (60 per cent), little machinery (say 10 per cent) and, in relation to the labour-power, little and cheap raw material (say 30 per cent). Then a mere rise in the value of the raw and ancillary materials from 30 to 80 could even out the composition: the second capital would now have 10 for machines, 80 for raw material and 60 for labour-power, that is 90c + 60v, which as percentages is also 60c + 40v, and no change whatever would have taken place in its technical composition.

So capitals of the same organic composition can have a different value composition, and capitals of the same percentage value composition can stand at different levels of organic composition, and so express different levels of development of the social productivity of labour. The mere fact, then, that farming capital stood at the general level by value composition would not prove that the social productivity of labour is equally developed in farming. It could only show that farming's own product, which is again a part of its conditions of production, has become dearer — or that ancillary materials such as fertilizer, once close at hand, now have to be carted a long way, and the like.

But leaving that aside, there is the peculiar character of agriculture to be considered.

Suppose that labour-saving machinery, chemical ancillaries and the like take up more room here, so that the constant capital grows against the mass of labour-power applied — technically, not only in value but in mass as well. Then in agriculture (as in mining) it is a question not only of the social productivity of labour but also of its natural productivity, which depends on the natural conditions under which the work is done. It is possible that the growth of social productivity in agriculture only makes up for a decline in natural power, or does not even make up for it — and this compensation can only work for a time — so that there, despite the technical development, the product is not made cheaper but only an even greater rise in its price is prevented. It is also possible that, while the price of corn is rising, the absolute amount of product falls and yet the relative surplus product grows. That happens when constant capital — mostly machines or livestock, of which only the wear and tear needs replacing — grows relatively, while the variable part of capital, laid out on wages, which always has to be replaced in full out of the product, falls to match.

It is also possible, though, that as agriculture advances, the market price need only climb modestly above the average for worse land to be taken into cultivation and to yield a rent as well — worse land which, with poorer technical means, would have needed a steeper rise in the market price.

The fact that in large-scale stock-raising, for instance, the mass of labour-power applied is very small compared with the constant capital that exists in the cattle themselves could be taken as decisive against the claim that farming capital, reckoned in percentages, sets more labour-power in motion than the non-agricultural capital of the social average. Here, though, it should be noted that in working out rent we start from that part of farming capital which produces the decisive cereal foodstuffs, the main means of subsistence of civilized peoples altogether, and take it as determining. Adam Smith — and this is one of his merits — has already shown that in stock-raising, and on average in all capital laid out on land that does not produce the main means of subsistence, corn for instance, price is determined in a quite different way. Here the price is determined by this: the price of the product of land used, say, as artificial pasture for cattle, but which could just as well be turned into arable land of a certain quality, must rise high enough to yield the same rent as equally good arable land. So the rent of corn land enters into the price of cattle as a determining factor — which is why Ramsay was right to remark that in this way, through rent, through the economic expression of landed property, that is through landed property itself, the price of cattle is artificially raised.

Adam Smith writes that as cultivation spreads, the uncultivated waste land no longer suffices to supply cattle for slaughter. A great part of the cultivated land, he says, then has to be given over to breeding and fattening cattle, whose price must therefore be high enough to pay not only for the labour spent on them but also for the rent the landlord could have drawn from that land, and the profit the farmer could have drawn from it, had it been tilled for crops. Cattle raised on the most uncultivated moors, he goes on, fetch the same price in the same market, weight for weight and quality for quality, as cattle raised on the best-cultivated land; and, in his account, the owners of those moors profit from this and raise the rent of their land in proportion to the price of cattle.

Here, then, too, unlike corn rent, differential rent is in favour of the worse land.

Absolute rent explains some phenomena that at first sight make it look as if rent were owed to a mere monopoly price. Take, for instance, the owner of a woodland that exists without any human contribution at all, that is, not as the product of afforestation — in Norway, say, to pick up Adam Smith's example. If a rent is paid to him by a capitalist who has timber felled, perhaps because of English demand, or if he has it felled himself as a capitalist, he will be paid a greater or lesser rent in the timber over and above the profit on the capital advanced. With this purely natural product this seems a pure monopoly surcharge. In fact, though, the capital here consists almost only of variable capital laid out on labour, so it also sets more surplus labour in motion than another capital of the same size. So there is more unpaid labour — more surplus-value — in the value of the timber than in the product of capitals of higher composition. That is why the average profit can be paid out of the timber and a sizeable excess can still go to the owner of the woodland as rent. Conversely, it is to be assumed that, given the ease with which the felling of timber can be extended — this production can increase rapidly — demand must rise very considerably for the price of timber to come equal to its value, and hence for the whole excess of unpaid labour (over and above the part of it that falls to the capitalist as average profit) to fall to the owner in the form of rent.

We have assumed that land newly brought into cultivation is of even worse quality than the worst land cultivated last. If it is better, it bears a differential rent. But what we are examining here is precisely the case where rent does not appear as differential rent. Only two cases are possible then. The land newly taken in hand is worse, or it is just as good as the land cultivated last. If it is worse, that case has already been examined. So the only case left to examine is the one where it is just as good.

Land that is just as good, and even better land, can enter new cultivation as agriculture progresses, just as worse land can — as was already developed in the case of differential rent.

Firstly, because with differential rent — and with rent in general, since even where rent is not differential the question always arises whether the land's fertility on the one side and its location on the other allow it to be cultivated at the governing market price with profit and rent — two conditions work in opposite directions. Sometimes they cancel each other out; sometimes one and sometimes the other tips the scale.

A rise in the market price — assuming the cost price of cultivation has not fallen, in other words that technical progress is not a newly added factor making for new cultivation — can bring more fertile land into cultivation, land that was previously shut out of competition by its location. Or, where the land is less fertile, that rise can increase the advantage of location enough to make up for the lower yield. Or, with no rise in the market price at all, improved means of communication can let location bring better land into competition, as we see on a large scale with the prairie states of North America. This happens constantly in old-civilized countries too, though not to the same extent as in colonies, where — as Wakefield correctly notes — location is decisive.

So the first reason is this: location and fertility work against each other, and the location factor keeps changing. It is constantly being evened out and goes through constant progressive changes that tend toward equalization. The result is that tracts of land that are equally good, better, or worse come in turn into fresh competition with land already under cultivation.

Secondly. As natural science and agronomy develop, the fertility of the land changes too, because the means change by which the elements in the soil can be made directly usable. In recent times, for instance, light soils in France and in the eastern counties of England, once thought poor, have risen to the first rank. (See Passy.) And land that was thought poor not because of its chemical composition but only because it put certain mechanical and physical obstacles in the way of cultivation is turned into good land as soon as the means of overcoming those obstacles are discovered.

Thirdly. In every old-civilized country, old historical and traditional arrangements — for example state lands and common lands — have, quite by accident, held large stretches of land out of cultivation, and they only enter it gradually. The order in which they are brought under cultivation depends neither on their quality nor on their location, but on quite external circumstances. If we followed the history of the English common lands, as they were one after another turned into private property and ploughed up by the Enclosure Bills, nothing would be more ridiculous than the fantasy that a modern agricultural chemist — Liebig, say — guided the choice of that order, marking certain fields for cultivation because of their chemical properties and shutting others out. What decided the matter was rather the opportunity that makes the thief: the more or less plausible legal pretexts for appropriation that offered themselves to the great landowners.

Fourthly. Setting aside the fact that the stage reached at any given time by the growth of population and capital sets a limit to the extension of cultivation, though an elastic one, and setting aside accidents that temporarily affect the market price, such as a run of good or bad seasons: how far cultivation spreads out over the land depends on the overall state of the capital market and the state of business in a country. In periods when capital is scarce, it is not enough that uncultivated land can yield the farmer the average profit — whether it pays rent or not — to draw additional capital into agriculture. In other periods, when capital is plentiful, capital streams into agriculture even without a rise in the market price, provided the other normal conditions are met. Better land than the land cultivated so far would in fact be kept out of competition only by its location, or by barriers, so far impossible to break through, that allow it to be kept out, or by chance. So we have only to deal with kinds of land that are just as good as those cultivated last.

Between the new land and the land cultivated last there is always a difference, though: the cost of clearing it. Whether that is undertaken or not depends on the level of market prices and on credit conditions. Once this land really does enter into competition, then, with other conditions unchanged, the market price falls back to its previous level, and the newly added land will then bear the same rent as the old land corresponding to it.

The assumption that it will bear no rent is proved by its supporters by assuming what is supposed to be proved: that the last land bore no rent. In the same way one could prove that the houses built last yield no rent besides the actual rent of the building, even though they are rented out. The fact is that they yield ground-rent even before they bring in a building rent, since they often stand empty for a long time.

Just as successive investments of capital on one plot can yield a proportionate extra product and therefore the same rent as the first investments, so fields of the same quality as those cultivated last can yield the same product at the same cost. Otherwise it would be quite incomprehensible how fields of the same quality ever get taken into cultivation one after another, rather than all at once — or rather how any single one of them ever gets taken up at all, since it would draw the competition of all the others after it. The landowner is always ready to draw a rent, that is, to receive something for nothing; but capital needs certain circumstances to fulfil that wish. So whether lands compete with one another depends not on whether the landowner wants them to compete, but on whether capital can be found to compete on the new fields with the others.

Insofar as the actual agricultural rent is mere monopoly price, it can only be small — just as absolute rent here can only be small under normal conditions, whatever the excess of the product's value over its price of production may be. The essence of absolute rent, then, is this: equal-sized capitals in different spheres of production produce different masses of surplus-value, according to their different average composition, at the same rate of surplus-value or the same exploitation of labour. In industry these different masses of surplus-value equalize into the average profit and are divided evenly among the individual capitals, in proportion to each one's share of social capital. As soon as production needs land, whether for farming or for extracting raw materials, landed property blocks this equalization for the capitals put into the land and intercepts a part of the surplus-value that would otherwise have gone into the equalization that forms the general rate of profit. Rent then forms a part of the value of commodities, more specifically of their surplus-value, which falls to the landowners instead of to the capitalist class that extracted it from the workers: the landowners extract it from the capitalists. It is presupposed here that agricultural capital sets more labour in motion than an equally large part of non-agricultural capital. How far this deviation goes, or whether it exists at all, depends on the relative development of agriculture compared with industry. By the nature of the case, this difference must decline as agriculture progresses — unless the ratio in which the variable part of capital declines against the constant part is still greater in industrial capital than in agricultural.

This absolute rent matters even more in extractive industry proper. There one element of constant capital, raw material, drops out entirely, and — except in the branches where machinery and other fixed capital make up a very large part — the lowest composition of capital is the absolute rule. Here, of all places, where rent looks as if it were owed to a monopoly price alone, market conditions have to be exceptionally favourable before the commodities sell at their value, or before the rent comes to equal the whole excess of their surplus-value over their price of production. So it is, for instance, with the rent of fishing waters, quarries, forests that grew wild, and the like.