The two units, and why they confuse everybody
A percolation test measures how fast water disappears from a hole. Two units are in common use and they run in opposite directions.
Minutes per inch (mpi) is the traditional septic-world unit: how many minutes the water surface takes to fall one inch. Slow soil gives a big number. A clay that takes 60 minutes to drop an inch is 60 mpi.
Inches per hour (in/hr) is the stormwater unit: how far the water falls in an hour. Slow soil gives a small number. That same clay is 1 in/hr.
The conversion is in/hr = 60 / mpi, and it runs both ways. Two inches of drop in thirty minutes is 4 inches per hour, which is 15 minutes per inch. Get the direction backwards and you will be off by a factor that grows with how far the soil is from 60 mpi, which is exactly where the interesting cases are.
| Minutes per inch | Inches per hour | Feet per day | Gallons per sq ft per hour |
|---|---|---|---|
| 5 | 12.0 | 24.0 | 7.48 |
| 15 | 4.0 | 8.0 | 2.49 |
| 30 | 2.0 | 4.0 | 1.25 |
| 60 | 1.0 | 2.0 | 0.62 |
| 120 | 0.5 | 1.0 | 0.31 |
The last column is the one that makes it concrete. One inch per hour across one square foot is 0.623 gallons an hour. A hundred square feet of rain garden bottom at that rate takes away 62 gallons an hour — which sounds like a lot until you notice that an inch of rain on a 1,000 square foot roof is over 600 gallons arriving in perhaps forty minutes.
The presoak is the test
Dry soil drinks. Capillary suction pulls the first water into empty pores fast, and the rate you measure in the first few minutes of a dry hole can be several times the rate the same soil sustains once it is wetted through. That initial gulp has nothing to do with what happens during the second hour of a storm on ground that rained yesterday.
Standard procedures therefore require a presoak: fill the hole, keep it full, and let the surrounding soil come to a wetted condition before the timed run starts. In sandy soils that may be a short affair. In clay it is measured in hours, sometimes overnight, and the whole point is that clay is exactly where the dry-hole error is largest.
A test run from dry is not a conservative shortcut. It errs on the optimistic side, and everything sized from it will be undersized.
Where the water actually leaves the hole
Here is a subtlety that catches careful people. A test hole loses water through its bottom and through its wetted sides, and the drop of the water surface counts both. An eight-inch hole with eight inches of water in it has about 50 square inches of bottom and about 200 square inches of wetted sidewall, so most of the loss is sideways.
A rain garden or an infiltration bed has essentially no sidewall in proportion to its area. So a rate taken from a deep narrow hole and applied to a wide flat bed overstates what the bed will do, and the overstatement can be large. Keeping the water shallow in a wide hole reduces the effect. Double-ring infiltrometers, which use an outer ring to force the flow from the inner ring downward, exist specifically to remove it.
This is one of several reasons the safety factor is not decoration.
Averaging holes, and when not to
Multiple holes are standard practice, and how to combine them is a genuine question. The arithmetic average is the obvious move and it is also the one most likely to mislead, because infiltration rates are distributed in a way that gives too much weight to a single fast hole. The geometric mean is more defensible for a quantity that spans orders of magnitude. The slowest hole is the conservative choice and is what many reviewers prefer.
But if the holes disagree by a factor of three or more, none of the three answers is any good, because the spread is telling you the site is not one thing. A hole that hit an old utility trench, a fill pocket, a former vegetable bed or a sand lens is measuring a feature, not a soil. The right response is more holes to find where the boundary is, and then a decision about which part of the site the system sits in.
Stormwater tests and septic tests are not the same test
This matters legally as well as technically. A percolation test for a septic system is a regulated procedure: the health department having jurisdiction specifies the hole dimensions, the preparation, the presoak duration, the number and location of holes, how the readings are taken, who may perform the work and frequently who must witness it. A number obtained by digging a hole in the yard and timing it has no standing in that process whatsoever, and drainfield sizing from it is not something to attempt. That whole question belongs to the regulator and to a licensed designer; the drainfield area organizer deals only with keeping the paperwork straight.
For stormwater — sizing a rain garden, a dry well, an infiltration trench, the base under permeable paving — an informal test is genuinely useful, and it is vastly better than assuming a rate from a soil texture. It may still not be what a reviewer accepts; if the project is going to be reviewed, ask what method they want before you dig.
Questions people ask
How do I convert minutes per inch to inches per hour?
Divide 60 by the minutes-per-inch figure. Thirty minutes per inch is 2 inches per hour; 15 mpi is 4 in/hr; 120 mpi is 0.5 in/hr. The relationship is a reciprocal, so the two numbers move in opposite directions — fast soil is a low mpi and a high in/hr. It is worth writing the unit next to every number you record, because a bare figure like "30" is a completely different soil depending on which unit somebody meant.
How long should the presoak be?
Long enough that you are measuring wetted soil rather than dry soil, which depends entirely on the material. Sand may need very little; clay commonly needs hours and standard septic procedures often specify an overnight soak. The practical test is stability: keep refilling and timing until consecutive readings stop getting slower. If your last reading is still slower than the one before it, the soil has not finished wetting and the test is not done. Whichever protocol applies to your project, if any, sets the requirement.
Should I use the slowest hole or the average?
If a reviewer is involved, use whatever they specify, and many specify the slowest. If nobody is reviewing it, the slowest hole is the choice that fails safe, and the geometric mean is a defensible middle if the holes are close together. What you should not do is average across a wide spread: if the fastest hole is three or more times the slowest, they are not measuring the same soil, and the average describes a place that does not exist on your property. Dig more holes and find out where the boundary is.
Why divide by a safety factor at all?
Because the number you measured is the best the soil will ever be. Fines wash into the infiltrating surface over the years and gradually seal it. Construction equipment on the bottom of the excavation compacts exactly the layer that has to accept the water. A test hole loses much of its water through the sidewalls, which a wide flat bed does not have. And the test was taken on one day in one season. A factor of two or more is common practice; what your reviewer expects is a question worth asking rather than guessing.
Can I use this test for a septic drainfield?
No. A septic percolation test is a regulated procedure and the health department with jurisdiction specifies everything about it — hole dimensions, preparation, presoak duration, hole count and location, reading intervals, who may perform it, and often who must witness it. A rate from an informal hole has no standing in that process, and drainfield sizing is not a calculation to attempt from it. This page is for stormwater infiltration: rain gardens, dry wells, infiltration trenches and permeable paving bases. If the question is septic, it starts with the health department.