A soil test replaces guessing with knowing - what it measures, why no honest tool will hand you fertilizer numbers without one, and how to take one. The rare piece of garden advice that saves money.
What you need
Most garden advice costs you money - buy this amendment, that fertilizer, this bag of stuff. A soil test is the rare piece that saves it, because it tells you what your soil actually has and needs, so you stop buying things it doesn't. For about fifteen dollars your county extension office will replace a season of guessing with a page of numbers. It is the single highest-value thing a new gardener can do, and this app will keep pointing you at it instead of pretending to know what only a lab can.
Every "feed your garden this" recommendation that hasn't seen your soil is a guess, and this planner will tell you so rather than add to the pile:
Mechanism: N-P-K sufficiency is site-specific and unknowable a priori.
Evidence: settled without a trial - epistemics.
We will not invent an NPK number. Your county extension office will test this soil for about $15. Until then, feed heavy feeders, do not feed root crops, and add compost.
Whether your ground is short on nitrogen, phosphorus, or potassium is specific to your site and simply cannot be known from a distance - two beds in the same yard can differ. So the engine refuses to invent an N-P-K number, on purpose. That refusal is not the app being unhelpful; it is the app not lying to you. The soil test is the thing that turns the refusal into a real answer.
A routine test comes back with the numbers you can't guess:
Two of those numbers carry their own honesty caveat worth knowing. pH matters because it decides nutrient availability:
Mechanism: Nutrient bioavailability is pH-dependent, so below the crop's optimum band phosphorus and molybdenum become less available; below about 5.5 the effect turns acute as aluminium and manganese dissolve fast enough to damage roots directly. Above an acid-requiring plant's range the same curve runs the other way: iron becomes unavailable and the plant goes chlorotic.
Evidence: confirmed against 3 sources in the research-based literature.
UF/IFAS Extension HS1207, Liu & Hanlon, 'Soil pH Range for Optimum Commercial Vegetable Production', edis.ifas.ufl.edu/publication/HS1207 (fetched and read in-container 2026-07-25). The ACUTE half of the mechanism: 'At pH 5.5 or lower, the solubility of Al increases 1000-fold for every pH unit decrease. For example, at pH 5.0, Al solubility is only 0.05 ppm, but at pH 4.0, Al solubility increases to a toxic level of 51 ppm.' On phosphorus: 'Aluminum and phosphate precipitate in low-pH soil'; 'Because of the Al concentration increase, the bioavailability of P at pH 4.0 reduces to one thousandth of the concentration present at pH 5.0'. On roots: 'In the presence of toxic concentrations of elements such as Al at low pH, root growth and water uptake are inhibited'. WHAT THIS SOURCE DOES NOT SUPPORT, AND IT IS WHY THIS RULE IS ONLY A GOOD HUNCH: all of the above is stated for pH 5.5 and below, and 'metal toxicity occurs at soil pH lower than 5.0' — while the floors this rule fires on run 5.5 to 6.2. Over most of its firing range the same page says the opposite: 'in the pH range from 5.5 to 7.0, all the nutrients have favorable solubility for use by vegetable plants.' The two are reconcilable (UMD's threshold is about optimum yield, this one about nutrient failure) but not as a single promising claim. Its Figure 2 is the nutrient-bioavailability-versus-pH curve that WOULD support the mild half of the mechanism; it is an image and could not be read.
UMD Extension, Table B-1 'Target Soil pH Values for Vegetable Crops', extension.umd.edu/sites/extension.umd.edu/files/2021-03/B-1.pdf (fetched and read in-container 2026-07-25): the source of every per-crop floor this rule reads. Its second column is headed 'Target lime when pH falls below', which is why the ruling_text says the floor is an intervention point rather than a lethal limit — the table itself frames it that way.
FIGURE 2 IS A DEAD END, AND THIS CLOSES IT (2026-07-26). Earlier notes recorded that HS1207's Figure 2 - a nutrient-bioavailability-versus-pH curve - was the evidence that would restore this rule to promising, and that it could not be read because it is an image. The figure does not need to be seen. Its caption and alt text are recoverable from the page: 'Figure 2. The pH and bioavailability (%) of listed nutrients in soil solution. Credit: Finck (1976).' TWO THINGS FOLLOW. First, the publication states its OWN reading of the figure, and that reading is the objection rather than the rescue: the sentence that forced the B-to-C downgrade - 'in the pH range from 5.5 to 7.0, all the nutrients have favorable solubility for use by vegetable plants' - is explicitly cited to Figure 2. The figure is what the contradiction rests on. Second, it is not UF/IFAS data: it is credited to Finck (1976), a reproduced availability curve of the widely-copied schematic kind. SO NOBODY SHOULD CHASE THIS IMAGE AGAIN. The good-hunch tier is not waiting on it. [read]
This bed's measured pH sits below what this plant wants. Near the floor that is a question of yield, not survival — the floor is the pH at which extension says to lime, not the pH at which the plant dies, and it is an estimated figure. Below about 5.5 it becomes a harder problem: aluminium and manganese dissolve and damage roots. Lime raises pH and sulfur lowers it, but we will not tell you how much of either, because the amount depends on your soil's buffering capacity and only a lab report's buffer pH measures that. Your extension soil lab is linked on the soil card.
Different crops want different ground, and the spread is wider than most people expect - a blueberry needs soil so acid that it would starve most vegetables, which is why it belongs in its own bed, not the main plot:
ph_range.And if the test says to change your pH, how much lime or sulfur to add is a real number only because the lab measured your soil - it is not something to eyeball:
Mechanism: Buffering capacity, which is set by clay and organic matter, decides how much lime moves pH at all, so the same target takes very different amounts in sand and in clay.
Evidence: confirmed against 2 sources in the research-based literature.
UF/IFAS Extension HS1207, Liu & Hanlon, 'Soil pH Range for Optimum Commercial Vegetable Production', edis.ifas.ufl.edu/publication/HS1207 (read 2026-07-25). The mechanism, stated outright: 'The amount of lime required to increase soil pH is determined by the size of the limestone particles being used and, most importantly, the buffering capacity of the soil.' What buffering IS: 'The buffering capacity refers to the soil's capacity to minimize change in the acidity of a solution when an acid or base is added into the solution.' What sets it: 'Buffering capacity is controlled by the soil's clay content and the amount of organic matter present.' And the consequence this rule turns into a refusal: 'Soils with large buffering capacity need more agricultural lime to adjust soil pH than those with lower buffering capacity for the same incremental change in soil pH. However, sandy soils have lower buffering capacity and need less lime for the same incremental change in pH than clay soils.' The page also names the remedy the ruling_text points at: 'The UF/IFAS Extension soil testing lab offers a Lime Requirement test as a component of a routine soil test, which provides a liming recommendation based on the buffer capacity of the soil sample.'
WHAT THE SOURCE DOES NOT SAY, and it is why this rule is promising rather than well established: it states that buffering capacity determines the amount, and it does not quantify how wrong an unbuffered estimate would be. The claim that the error is large enough to WARN a home gardener about is a judgement this corpus is making on top of the mechanism, not a sentence anyone published. Promising + suboptimal derives `advise`, which is the mildest non-inert ruling, and the ruling_text is phrased as a refusal to guess rather than as a prediction of harm.
We will not tell you how much lime or sulfur to add, and neither should anything that has not measured your soil's buffering capacity. If your report carries a buffer pH, the rate printed on it is a real one for your ground - use that. If it does not, ask your lab for a lime requirement test; it is usually part of a routine soil test and costs nothing extra. Which direction to go is free: lime raises pH, sulfur lowers it.
Which direction to go is free - lime raises pH, sulfur lowers it - but the amount depends on how your particular soil resists the change, which is exactly what the test measures and what no chart can.
It takes ten minutes:
1. With a clean trowel, take a slice of soil from root depth - about 15 cm6 inches down - from several spots around the bed, not just one. 2. Drop them all in a clean plastic bucket (not metal, which can skew micronutrient readings) and mix. 3. Let the sample air-dry, then send the amount your extension office asks for, with their form.
Sample each distinct area separately - a vegetable bed and a lawn, or beds with very different history, are different soils and want their own tests.
You don't have to garden blind in the meantime. The honest interim advice, straight from the rule above, is simple and safe: feed the heavy feeders (tomatoes, corn, squash, brassicas), don't feed root crops - extra nitrogen gives you leaves at the expense of the root - and add compost, which improves almost any soil without the risk of overdoing a specific nutrient. Then let the test, when it comes, turn that general advice into a precise one. The last year's raised-bed soil guide covers keeping good soil good once you have it.
Next in Know your ground → Cover crops and green manure for the home garden
The rules this guide leans on: