Table of ContentsChapter 6
Oil 101

Chapter 6

Exploration and Production

Oil exploration and production: seismic surveys, drilling, well completion, decline curves, and enhanced oil recovery techniques.

The Upstream Sector

Exploration and production (E&P), also known as the upstream sector, is the highest-risk and most lucrative part of the oil industry. The costs involved in the earliest stage are called finding and development (F&D) costs. Everything begins with securing the legal right to explore.

Obtaining Rights to Explore and Drill

Before an oil company can begin to explore, it must obtain permission from two owners: the surface property owner, and the owner of the mineral rights to the petroleum beneath the surface. In most countries, any valuable minerals below the surface are owned by the government, even if the land above is privately held. The US and Canada are rare exceptions where private individuals can own mineral rights on private land.

In the US, a person called a landman negotiates the purchase of exploration and production rights. Investing to find oil is called an oil play and a geological feature of interest is known as a lead. Once surface and mineral rights are legally secured, the oil play and lead become a prospect.

Oil beneath the surface is governed by the rule of capture, a legal principle based on English common law. If you extract oil from land you own, it is yours, even if the oil originally migrated from beneath a neighbor's land. However, you cannot use slant drilling to physically put a wellbore under a neighbor's property (subsurface trespass). Iraq cited alleged slant drilling by Kuwait as one pretext for its 1990 invasion.

The US and Canada are among the only countries where private individuals can own mineral rights. In most of the world, the government owns everything underground, regardless of who owns the surface land. This is why a Texas rancher can become a millionaire from oil under their property, while a Nigerian farmer cannot.

Types of E&P Agreements

There are five main types of agreement between an exploring company and the mineral rights owner:

Concessions (Leases)

The oldest type. A lessor explores and produces on behalf of the mineral rights owner, paying a bonus, rent, and royalty. Common in the US for federal lands. No time limit on production.

Production Sharing Agreements (PSAs)

The E&P company bears costs and recoups them from a fixed share of production within a time window (typically 5-30 years). Used where foreign ownership is restricted, such as in Iran and Saudi Arabia.

Service Contracts

The E&P company explores for a flat fee, paid only if production occurs. Ownership of oil remains with the mineral rights owner. High risk for the contractor.

Joint Ventures (JVs)

Multiple parties share costs from the outset. Common for ultra-deepwater exploration where risks are too high for a single company. One party serves as the field operator.

Production Contracts

Used to develop improved recovery techniques (waterflooding, gas reinjection) on existing fields. The contractor receives a share of any production rate increases.

How Oil Forms: Source Rock, Reservoir Rock, and Cap Rock

Geologists searching for oil look for three essential features: a source rock containing kerogen (the organic precursor to oil), a porous reservoir rock that can hold oil and allow it to flow, and an impermeable cap rock (or trap) that prevents oil from migrating to the surface.

Kerogen is a solid, dark, waxy material formed from dead marine organisms (plankton, algae) deposited in oxygen-poor environments on ancient ocean floors. When buried to sufficient depth, heat and pressure crack kerogen's heavy molecules into lighter hydrocarbons. The oil window is the range of depths (roughly 7,500-15,000 feet, or 2.3-4.6 km) where temperatures are right to form liquid oil. Below this, in the gas window, everything is cracked into methane.

Cross-section of Earth's crust showing the geothermal gradient with three zones: immature zone (0-7,500 ft), oil window (7,500-15,000 ft, 60-120C), and gas window (15,000-25,000 ft, 120-250C)
Figure 6-1: The geothermal gradient. Kerogen remains immature above 7,500 feet. In the oil window (60-120C), heat and pressure crack kerogen into liquid oil. Below 15,000 feet, in the gas window, everything is cracked further into methane. (Source: Oil 101, Morgan Downey)
Diagram showing seven sequential geological events required to produce a commercial oil discovery: source rock, maturation, migration, reservoir rock, trap, seal, and timing
Figure 6-2: Seven low-probability geological events must all align to produce a commercial oil discovery. Failure at any stage means no oil. This is why exploration is one of the highest-risk businesses in the world. (Source: Oil 101, Morgan Downey)

Reservoir Mechanics: Pressure Is Everything

Oil does not sit in underground lakes. It is trapped in tiny pores within rock, like water in a rigid sponge. Oil moves to the bottom of a well because reservoir pressure exceeds the low pressure at the wellbore. Every barrel extracted reduces reservoir pressure. There is anoptimal rate of production that maximizes total recovery: producing too fast depletes pressure prematurely, permanently stranding oil in the ground.

The recovery factor (percentage of original oil in place that can be extracted) averages only about 30% worldwide. It is exceptional to achieve above 60%. Natural drive mechanisms include:

Dissolved gas drive

Gas dissolved in oil expands as pressure drops, pushing oil toward the well.

Gas cap drive

Free gas above the oil expands to push oil downward and toward wells.

Water drive

An aquifer beneath the oil pushes it upward. Generally the most effective natural drive.

Gravity drainage

Oil flows downward through the reservoir under gravity to wells at the base.

Each mechanism is "rate sensitive": producing too aggressively can permanently reduce total recoverable oil. The rule of capture historically caused producers to race to extract as fast as possible from shared reservoirs, destroying value. In the early days at Spindletop, oil fields were covered in thousands of wells every few feet. Today, standard well spacing is typically 40 acres per well, with infill drilling permitted as economics warrant.

Forest of oil wells at Spindletop in the early 1900s
Figure 6-1: The forest of oil wells at Spindletop in the early 1900s. Under the rule of capture, producers raced to drain the reservoir as fast as possible, wrecking total recovery in the process. (Source: USGS)
Magnified 3D cutaway of reservoir sandstone showing sand grains, pore spaces filled with oil and water, pore throats, and an inset comparing good, poor, and zero porosity
Figure 6-4: Oil is trapped within tiny pores in the reservoir rock, like water in a rigid sponge. Porosity is the percentage of pore space (typical reservoir: 10-25%). Permeability is how easily fluid flows through connected pores, measured in millidarcies. Both must be sufficient for a commercial well. (Source: Oil 101, Morgan Downey)
Only about 30% of the oil in a typical reservoir is ever recovered. The other 70% remains trapped in the rock forever. Even with the most advanced technology, getting above 60% is rare. This means for every barrel we produce, roughly two barrels are left behind underground.

Exploration: Seismic Surveys

Modern exploration relies on seismic surveys: sending sound waves into the earth and analyzing their reflections to map underground rock formations. A seismic source (vibrator trucks onshore, air guns offshore) generates waves that travel through rock layers. When waves hit a boundary between different rock types, some energy reflects back to the surface where it is recorded by receivers called geophones (onshore) or hydrophones (offshore).

2D seismic produces cross-section images along a single line. 3D seismic uses a grid of source and receiver lines to create a three-dimensional picture of the subsurface. 4D seismic (time-lapse) repeats 3D surveys over time to track how fluids move through a producing reservoir. Despite these advances, the success rate for wildcat wells (drilled in unexplored areas) remains only 10-20%.

Drilling

Once a prospect is identified, a rotary drilling rig bores into the earth. The drill string consists of a rotating drill bit at the bottom, connected to the surface by drill pipe. Drilling mud (a carefully engineered fluid) is pumped down the drill pipe and returns up the annulus (the space between the pipe and the borehole wall), carrying rock cuttings to the surface and maintaining pressure to prevent blowouts.

Wells are lined with steel casing (cemented in place) to prevent collapse and isolate different pressure zones. The bottom of the well is completed by perforating the casing to allow oil to flow in. A Christmas tree (a set of valves and fittings) sits on top of the wellhead to control flow.

A wellhead Christmas tree assembly in northern British Columbia, showing the valve stack and pressure gauges that control flow from the completed well
Figure 6-5: A wellhead Christmas tree in northern British Columbia. The stack of valves and gauges controls pressure and flow from the completed well below. (Source: Dexcel / Wikimedia Commons (public domain))
Edwin Drake's first well in 1859 used a stove pipe as casing to stop the hole from collapsing. This is the same principle used today in multi-million-dollar deepwater wells, at a completely different scale. Modern casing strings can weigh over 1,000 tons.

Drilling Mud Chemistry

Drilling mud is the unsung hero of every well. It serves five simultaneous jobs: lifting rock cuttings from the bit face to the surface; maintaining hydrostatic pressure in the wellbore to counterbalance formation pressure and prevent kicks; cooling and lubricating the drill bit; forming a thin impermeable filter cake on the borehole wall to stabilize soft formations; and suspending cuttings in place when circulation stops, so the hole does not pack off. There are three main mud families. Water-based muds (WBM) are the cheapest and most common, built on fresh or salt water with bentonite clay and barite weighting material. Oil-based muds (OBM) use diesel or mineral oil as the continuous phase and are preferred in shales that react with water, in high-temperature wells, and in highly deviated holes. Synthetic-based muds (SBM) substitute engineered esters or olefins for diesel to meet offshore environmental rules. Mud weight is quoted in pounds per gallon, with 9 ppg roughly neutral and 14 to 18 ppg typical for high-pressure deepwater sections.

Well Logging

A well is only as useful as the data it produces. Three logging techniques compete for each foot of new hole. The cheapest is the mud log: a geologist watches cuttings return to the shale shaker and records lithology, gas shows, and drilling rate depth by depth. Wireline logs are run after drilling stops. A logging tool is lowered on an electrical cable and measures gamma ray (shale content), resistivity (hydrocarbon vs water), density and neutron porosity (pore volume), and sonic velocity (rock mechanics). These measurements are cross-plotted to identify pay zones and estimate hydrocarbon saturation. Logging while drilling (LWD) embeds the same sensors into the drill collar just above the bit, so the geosteering team sees formation properties in near real time and can adjust the wellbore to stay inside a thin pay zone. LWD is the dominant logging method for horizontal shale wells, where pulling a wireline tool around the curve is impractical.

Drill Bits and Directional Tools

Two drill-bit families dominate the field. Roller cone bits use three rotating cones with hardened teeth (milled tooth for soft formations or tungsten carbide inserts for hard rock) and cut rock by crushing and scraping. Polycrystalline diamond compact (PDC) bits have no moving parts: they shear rock with fixed synthetic diamond cutters. PDC bits last longer and drill faster in most shales and carbonates, and they now account for the majority of footage in US land drilling. Directional control comes from a mud motor (a positive-displacement motor powered by the mud stream, with a bent housing that deflects the bit) or a rotary steerable system (RSS) that steers while the full drill string rotates, giving a smoother hole and faster build rates. Survey tools in the bottom-hole assembly measure inclination and azimuth, so the driller always knows where the bit is pointing to within a degree or two.

Blowout Preventers and Well Control

If formation pressure suddenly exceeds mud-column pressure, hydrocarbons rush into the wellbore (a kick) and can reach the surface uncontrolled (a blowout). The last line of defense is the blowout preventer stack. A BOP stack combines two families of valves. Annular preventers squeeze a donut-shaped rubber element around whatever is in the wellbore, sealing around drill pipe, casing, or an open hole. Ram preventers drive two opposing steel blocks together: pipe rams seal around a specific pipe diameter, blind rams close on an open hole, and shear rams are built to physically cut the drill pipe and seal the well in a worst-case emergency. On a subsea well, the BOP stack sits on the seafloor and is connected to the rig through a marine riser and a lower marine riser package. The 2010 Macondo blowout at the Deepwater Horizon rig became the reference case for BOP failure: the shear rams did not fully cut the drill pipe, the wellhead blew, the rig burned and sank, and 4.9 million barrels of crude were released into the Gulf of Mexico over 87 days. Chapter 15 (Environmental) covers the response.

Subsea blowout preventer stack with marine riser and kill lines
Figure 6-6: A subsea BOP stack on the Macondo wellhead showing kill and choke lines, the riser insertion tube, and the top kill manifold. Annular and ram preventers are stacked above the wellhead on the seafloor. In the 2010 blowout, the shear rams failed to fully cut the drill pipe, allowing hydrocarbons to escape to the surface. (Source: Wikimedia Commons (public domain))

Drilling can be vertical, directional (angled), or horizontal (turning 90 degrees to run laterally through a formation). Horizontal drilling, combined with hydraulic fracturing, was the breakthrough that unlocked the shale revolution. A modern Permian well typically drills 10,000 feet of vertical hole, builds curvature over a few hundred feet, and then runs 2 to 3 miles horizontally through the pay zone, all in under 10 days.

Production Phases

Primary recovery is the first phase of production. Most new wells initially flow under natural reservoir pressure, requiring no pump at all. As pressure declines over months or years, artificial lift becomes necessary. The most common method is the sucker-rod pump, the iconic "nodding donkey" pumpjack visible across the US oil patch. Sucker-rod pumps are used on the majority of US onshore producing wells, especially older stripper wells producing under 15 barrels per day. Other artificial lift methods include electric submersible pumps (ESPs, favored in high-rate offshore and shale wells), gas lift (injecting gas to lighten the fluid column), progressive cavity pumps, and jet pumps. By well count, over 90 percent of the roughly 500,000 producing wells in the US use some form of artificial lift because most are mature, low-rate stripper wells. By production volume, the picture is different: newer high-rate wells, particularly in the Permian and offshore Gulf of Mexico, often flow under natural reservoir pressure for their first several years. Globally, the majority of Middle Eastern production flows naturally without any pump. The distinction matters: most wells use pumps (because most wells are old and low-rate), but a large share of barrels come from wells that do not need them (because new high-rate wells dominate volume).

The surface beam of a pumpjack rocks up and down, driving a long string of rods down the well to a positive-displacement plunger pump at the producing interval. On each upstroke the pump lifts a few gallons of fluid into the tubing; on the downstroke it resets. A single pumpjack may cycle 8 to 15 strokes per minute, 24 hours a day, for decades.

Labelled diagram of a sucker-rod pumpjack
Figure 6-7: The nodding donkey pumpjack. The walking beam at the surface drives a rod string down to a positive-displacement plunger pump at the producing interval. By well count, pumpjacks represent the majority of US producing oil wells, though by production volume, many of the highest-rate wells flow under natural reservoir pressure without any pump. (Source: Wikimedia Commons (CC BY 3.0))
The US has over 500,000 producing oil wells. Saudi Arabia achieves similar total output with fewer than 3,000. The difference is geology: Saudi wells tap massive conventional reservoirs with natural pressure, while most US wells are small, aging stripper wells requiring mechanical pumps.

Secondary Recovery: Waterflooding

Secondary recovery involves injecting water or gas to maintain reservoir pressure and physically sweep oil toward producing wells. Waterflooding is by far the most common method: seawater, produced water, or treated river water is injected at high pressure through dedicated injection wells, raising pressure and pushing oil through the pore network toward producers. Waterflooding has been used from the earliest production stages at Saudi Arabia's Ghawar, where the resulting pressure support has kept the giant field producing for over 70 years. Injection patterns are chosen to maximize sweep efficiency across a regular well grid. The two classic patterns are the five-spot (one producer in the center of a square of four injectors, with every well in the grid serving one of the two roles) and the nine-spot (one producer surrounded by eight injectors in a 3-by-3 grid). Direct line drive, inverted patterns, and irregular patterns are used in fields where faults or permeability contrasts make the classical grids inefficient.

Tertiary Recovery: Enhanced Oil Recovery

Tertiary recovery, better known as Enhanced Oil Recovery (EOR), uses advanced techniques after primary and secondary methods have exhausted the easy oil. EOR methods are grouped into three families by the physics they exploit: thermal (change viscosity with heat), gas (change miscibility with solvents), and chemical (change sweep with surfactants or polymers).

Table 6-1: Enhanced Oil Recovery Methods

FamilyTechniqueWhere it is used
ThermalSteam flood, cyclic steam stimulation, steam-assisted gravity drainage (SAGD), fire floodAlberta (SAGD), Kern River CA, Duri Sumatra, Orinoco Venezuela
GasCO2 miscible flood, nitrogen flood, flue gas, rich hydrocarbon gasPermian Basin (largest application); CO2 from natural domes and industrial sources
ChemicalPolymer flood, surfactant flood, alkaline flood, ASP (alkaline-surfactant-polymer)Daqing China (largest polymer flood), selected North Sea and US mature fields
CO2 injection for enhanced oil recovery is one of the few industrial processes that actually consumes carbon dioxide. Some operators source CO2 from natural underground deposits, while others capture it from power plants or industrial facilities, creating a rare overlap between oil production and carbon reduction.

Decline Curves and Field Life

Every oil well follows a decline curve: production peaks early then gradually falls as reservoir pressure depletes. Conventional wells typically decline 5 to 15 percent per year. Shale wells are drastically steeper: 60 to 80 percent decline in the first year, 30 to 40 percent in the second year, then a long low-rate tail. This difference dominates the economics. A conventional well in the Middle East may produce for 30 years with modest intervention. A shale well delivers most of its lifetime oil in the first 24 months, requiring continuous new drilling just to hold field-level production flat, a phenomenon the industry calls the shale treadmill.

Field operators use decline curve analysis (DCA) to forecast future production and estimate ultimate recovery. The three standard Arps models are exponential (constant percentage decline, fits most mature conventional wells), hyperbolic (a slowing decline rate, fits wells with strong transient flow), and harmonic (the slowest decline, a special case of hyperbolic with b equal to 1). The widget below lets you see how changing the hyperbolic b-factor rotates the curve and reshapes estimated ultimate recovery from the same initial rate. Chapter 14 (Reserves) goes deeper into how decline curves feed proved reserve bookings and SEC PV-10 valuations.

Arps Decline Curves

Arps decline curves. All three curves start at 1,000 barrels per day and share a 35 percent initial annual decline. Exponential (b = 0) falls fastest and estimates the least ultimate recovery. Harmonic (b = 1) has the longest tail. Hyperbolic sits in between: drag the slider to see how the b-factor changes the shape.
Reference: J.J. Arps, “Analysis of Decline Curves” (1945), SPE.

The chart below compares a typical shale horizontal well against a conventional vertical well, both starting at 1,000 barrels per day. The shale well loses two thirds of its rate in year one; the conventional well barely moves. This is why shale economics are driven by initial rate, completion cost, and drilling pace rather than long-term recovery per well. See Chapter 21 (The Shale Revolution) for the full treatment of how horizontal drilling and hydraulic fracturing changed the industry.

Shale vs Conventional Decline Curves (Illustrative)

Shale horizontalConventional vertical
The shale treadmill in one chart. Both wells start at 1,000 barrels per day in month one. The shale well follows a hyperbolic Arps decline, losing roughly two thirds of its rate in the first twelve months and settling into a long low-rate tail. The conventional well loses only 6 to 8 percent per year and is still producing several hundred barrels per day a decade later. Over 120 months the conventional well actually out-produces the shale well on a cumulative basis in this teaching example, which is why shale economics live or die on initial productivity, completion cost, and drilling pace rather than ultimate recovery per well.
A conventional oil well may produce for 20 to 30 years with a gradual decline. A shale well delivers most of its oil in the first two to three years. The US shale industry must drill thousands of new wells every year just to hold production flat.

Offshore Production

When onshore opportunities are limited, companies move offshore, where costs are 4-5 times higher but fields tend to be much larger. Offshore platforms fall into a tidy typology defined primarily by water depth. Fixed structures dominate shallow water; floating systems dominate deepwater and ultra-deepwater. The choice of platform type has first-order consequences for capex, schedule, and even the commercial structure of the development.

Table 6-2: Offshore Platform Typology

TypeWater depthHow it works
Jack-upUp to 400 ftMobile hull towed into position; hydraulic legs lower to seafloor and jack hull above wave height
Fixed jacketUp to 1,000 ftSteel lattice jacket pile-driven into seafloor with deck on top; workhorse of 1960s-80s shelf
Compliant tower1,000 to 3,000 ftFlexible steel tower that sways with waves rather than resisting them; mid-depth solution
Tension leg platform (TLP)1,500 to 5,000 ftBuoyant hull held down by vertical tendons anchored to seafloor; eliminates most vertical motion
SparUp to 10,000 ftDeep vertical cylinder with ballast at base; stable because most of the structure sits below the wave zone
Semi-submersible1,500 to 10,000 ftPontoons below the wave zone support columns holding the deck above; used for drilling and production
FPSOAny deepwaterShip-shaped hull with processing on deck and crude storage below; offloads to shuttle tankers. Dominant solution for remote deepwater fields.
World map showing 8 major offshore oil and gas production basins: Gulf of Mexico, North Sea, pre-salt Brazil, Guyana, West Africa, Caspian Sea, NW Australia, Southeast Asia
Figure 6-8: Major global offshore oil and gas production basins. Offshore production accounts for roughly 30% of global crude output. The deepest water developments (pre-salt Brazil, Gulf of Mexico, West Africa) operate in 2,000 to 3,000 meters of water using floating production systems. (Source: Base map: Wikimedia Commons (public domain). Annotations: Oil 101, Morgan Downey)
Side-by-side comparison of offshore platform types at increasing water depths: fixed platforms, compliant tower, TLP, spar, semi-submersibles, FPSO, and subsea tieback
Figure 6-9: Offshore platform types shown at relative scale and water depth. From left: (1-2) conventional fixed platforms for shallow water, (3) compliant tower, (4-5) tension leg platforms, (6) spar, (7-8) semi-submersibles, (9) FPSO, and (10) subsea completion tied back to a host facility. Jack-up rigs and gravity-based structures are not pictured. (Source: NOAA Office of Ocean Exploration and Research (public domain))

Figure 6-10: Maximum Water Depth by Platform Type (feet)

Sources: industry references, Oil & Gas Journal. FPSO depth limited by riser and mooring technology, not the vessel itself.

PetroSA FPSO vessel Orca moored at Mossel Bay, South Africa, with processing topsides visible on deck
Figure 6-11: The PetroSA FPSO Orca at Mossel Bay, South Africa. An FPSO carries processing equipment on deck and stores crude in its hull, offloading periodically to shuttle tankers. (Source: Bob Adams / Wikimedia Commons (CC BY-SA 2.0))

Brazil's pre-salt fields (discovered 2006-2007) lie beneath 7,000 feet of water, 10,000 feet of rock, and a 6,000-foot layer of salt, yet are developed with arrays of FPSOs tied back to dozens of subsea wells and produce over 3 million bpd in aggregate. The US Gulf of Mexico produces roughly 1.8 million bpd from a mix of fixed jackets on the shelf and deepwater TLPs, spars, and semi-submersibles. The 2010 Deepwater Horizon disaster (the Macondo blowout, discussed in the well control section above) led to stricter safety regulations and a temporary drilling moratorium, but Gulf deepwater production has since recovered and grown.

Deepwater Horizon semi-submersible drilling rig on fire, April 2010
Figure 6-12: The Deepwater Horizon semi-submersible drilling rig burning on April 20, 2010 after the Macondo blowout. The rig sank two days later. Eleven workers were killed and 4.9 million barrels of crude were released into the Gulf of Mexico over 87 days before the well was capped. (Source: US Coast Guard (public domain, via Wikimedia Commons))

The Shale Revolution

Horizontal drilling combined with hydraulic fracturing ("fracking") took US oil production from 5 million bpd in 2008 to roughly 13.5 million bpd by 2025, making the US the world's largest oil producer. The mechanics belong in their own chapter. The five-step fracking explainer below introduces the core idea; Chapter 21 (Shale Revolution) covers basin geology, rig productivity, type curves, the 2015 crude export ban lift, and the 2023 to 2025 consolidation wave (ExxonMobil to Pioneer, Chevron to Hess, Diamondback to Endeavor).

How Hydraulic Fracturing Works

1

Vertical drilling

Drill vertically 5,000-10,000 feet to reach the target shale formation. Steel casing cemented in place.

2

Horizontal turn

The wellbore curves 90 degrees and extends horizontally through the shale for 1-3 miles (5,000-15,000 feet).

3

Perforation

Small explosive charges (perforating guns) punch holes through the casing into the surrounding rock at 30-60 intervals.

4

Hydraulic fracturing

Water, sand (proppant), and chemicals pumped at 5,000-10,000 psi to fracture the shale. Sand holds fractures open.

5

Flowback and production

Pressure released, fracking fluid flows back. Oil and gas flow from fractured rock into the wellbore and up to the surface.

A single horizontal shale well uses 5-15 million gallons of water and 5-10 million pounds of sand during the fracturing process. The water volume is roughly equivalent to filling 20 Olympic swimming pools. Most of the water is recovered as “flowback” and recycled. Chapter 21 (Shale Revolution) covers the full Permian, Bakken, Eagle Ford, and Marcellus picture, including type-curve economics, breakeven prices, and the consolidation wave of 2023 to 2025.

The E&P sector continues to evolve. Digital oilfield technologies (sensors, machine learning, automated drilling) are improving efficiency. Carbon capture and storage (CCS) is being deployed at some production sites to reduce emissions. Meanwhile, the question of whether the best shale acreage is running out, and when US production will peak for a second time, remains one of the most consequential debates in global energy.

The above was updated in 2026. For the full original 2009 chapter, download the 1st edition 2009 PDF.