Cover Image
close this book Soils, Crops and Fertilizer Use
close this folder Chapter 7: Evaluating a soil's fertility
View the document Soil testing
View the document Plant tissue testing
View the document Fertilizer trials
View the document Using visual "hunger signs"

Chapter 7: Evaluating a soil's fertility

Deciding how much and what kind of fertilizer a farmer needs for her crops is a twostage process:

1. The soil's present fertility level should be evaluated, ideally by soil testing.

2. Once the soil's present state of fertility is known, the most appropriate kinds and amounts of fertilizers can be determined, considering the following factors:

• Type of crop

• Feasible yield goal as determined by:

•• Yield-limiting factors such as available moisture, soil characteristics, weather, pests, and diseases.

•• Farmer management level

•• Available capital for needed inputs

• Expected cost/return based on likely yield and market value (more difficult to predict with vegetable crops than field crops).

In this chapter, we'll focus on stage 1 and cover the following methods of evaluating a soil's fertility:

• Soil testing

• Plant tissue testing

• Fertilizer trials

• Spotting visual "hunger signs"

Soil testing

Soil testing by a reliable lab is the most accurate and convenient method for evaluating a soil's fertility. Most labs will also make a fertilizer recommendation, too. This service is often free or very low-cost, yet is often underutilized by farmers and development workers.

Some Factors Affecting the Usefulness of Soil Testing

• Improper sampling procedures by farmers and extension workers are common and produce inaccurate results.

• The test for available soil nitrogen (in the form of nitrate N) isn't very accurate. That's because a soil's available N fluctuates, because it depends a lot on the kind and amount of crop residues present and the rate that bacteria will break them down to release nitrate N (this rate varies a lot with temperature and moisture). If the sample sits at the lab for a week or two, a falsely high nitrate reading is likely.

• Most labs don't routinely test for sulfur or micronutrients. Some of these tests aren't very accurate, anyway.

• The reliability of soil labs varies, not only in terms their accuracy in evaluating a soil's fertility, but also as far as the resulting fertilizer recommendations. More on this below:

Soil Lab's Aren't Necessarily Reliable

When different soil test labs analyze the same soil sample, there can be some large discrepancies in both the fertility evaluations and the fertilizer recommendations. In fact, recent studies of public and private (commercial) labs in the U.S. have brought to light some surprising differences. For example, in one of the studies conducted by the Rodale Research Institute (Emmaus, Pennsylvania), 4 soil labs were sent an identical sample and asked to make a fertilizer recommendation for lettuce. Their recommendations varied as follows: N: 65, 130, 130, 220 kg/ha; P2O5 0, 0, 90, 220 kg/ ha; K2O: 0, 0, 35, 220 kg/ha. The resulting fertilizer cost ranged from $75-$235/ha. In another study, 5 labs were asked to analyze 4 sets of identical soil samples for 4 fields. The total cost of the fertilizer recommended by them ranged from a low of $168 to a high of $320.

How to Find a Reliable Lab Despite these variations, soil testing is still a very useful tool if you can locate a reliable soil testing lab. The Ministry of Agriculture, agricultural schools, ag research stations, or fertilizer companies may maintain labs in your country. You can inquire among technicians and farmers as to their opinion of the labs and do some additional investigation as well. Here are some useful indicators of reliability.

• Adequate equipment and well trained technicians.

• Enough greenhouse/field trials with the erect's soils to correlate soil test results with actual crop response to different rates of fertilizer. Soils vary in their response, so such correlation data is vital.

• Lack of bias: Labs run by fertilizer companies may be more interested in selling fertilizer than in accuracy, though not always. In some cases, labs don't consider the special circumstances of limited-resource farmers but gear their rates toward large farmers; this factor, in itself, can produce wide variations in fertilizer recommendations among labs testing the same soil sample.

NOTE: Even if the lab is biased, soil testing can at least provide you with valuable baseline data for "customizing" the recommendation to suit a farmer's actual circumstances.

• The lab should give credit for the farmer' A intended use of manure, compost, or green manure crops since this can substantially lower fertilizer needs.

• Farmer input: Good labs supply a detailed questionaire for the farmer to fill out concerning farm size, past and future crops, past yields, yield expectations, past fertilizer applications, intended use of manure or green green manure crops, limiting soil factors, etc.

• Sampling instructions: Reliable labs are likely to provide detailed written instructions on how to take and collect soil samples.

How to Evaluate a Lab's Fertilizer Recommendations

• Compare them with Table 9-4 in Chapter 9.

• Send an identical sample to two or more labs.

• Compare the lab's recommendation with trial strips using higher and lower rates.

• Plant tissue testing can be a useful supplement to soil testing since it can monitor N-P-K levels in the plant itself. (The section following soil sampling discusses tissue testing.)

What about Portable Soil Test Kits?

You can buy portable soil test kits for measuring NPK levels, but they definitely aren't accurate enough for several reasons:

• They're unlikely to be correlated with local soil conditions.

• Their reagents break down with time and may be difficult to replace.

• The color bar plates used to measure readings are often of poor quality and standardization in the cheaper kits.

Portable soil pH test kits are also available, and the better quality ones are accurate to within 0.1-0.3 pH units. They can be useful for troubleshooting, but a lab test will still usually be needed to determine how much lime is needed to raise the pH of an overly acidic soil. (Liming is covered in Chapter 11.)

What Useful Information Does a Soil Lab Provide?

• Most labs routinely test for:

Soil pH

C.E.C. (negative charge)

Available N, P, K, Ca, Mg

• Most labs don't routinely test for micronutrients or sulfur.

• The lab will give a fertilizer recommendation for NPK either in terms of kg/ha of N, P, and K or in terms of the actual kinds and amounts of fertilizers needed. The recommendation is based on the amount of nutrients already present in the soil and should also take into account the farmer's yield expectations.

• Liming: If the soil tests out overly acidic, the lab will give a liming recommendation.

• In dry regions where salt buildup is a problem, the lab will test both the soil and the irrigation water for salt content and make recommendations; there is usually an extra charge for this. (Chapter 12 covers salt problems.)

How Often is Soil Sampling Needed?

Where low to moderate rates of fertilizer are being used, once every 3-5 years is sufficient. That's because such rates feed the crop itself rather than also building up the soil's residual fertility.

What About Soil Testing and Organic Fertilizers??

Unlike chemical fertilizers which have an analysis label, the exact nutrient content of organic fertilizers, like manure and compost, is highly variable and difficult to judge. If this is so, is it worthwhile to test the soil in cases where organic fertilizers are used? Soil testing may still be a good idea for several reasons:

• If the soil is severely deficient in a nutrient like P or a micronutrient, some chemical fertilizer may have to be used to supplement the organics.

• Knowing the fertility status of the soil will allow you to determine which organic materials can best provide what is needed.

• The lab's pH test may indicate liming is needed. The lab can best determine how much lime is needed.


When to Sample

• At least 2 months before the results are needed. If farmers wait until a few weeks before planting, the lab is likely to become overloaded and unable to provide the results in time.

• Sampling may be easier to do in the wet season when the soil isn't as hard. In the case of flooded rice soils, check with the lab for the best sampling time. (While flooded, soils have different chemical properties then when unflooded.)

Avoid Improper Sampling

Improper sampling is a very common cause of inaccurate lab results. Each 200-500 gram sample sent to the lab may represent up to 15,000 metric tons of soil. One sign of a good lab is that it will provide sampling containers along with detailed instructions of how to take samples.

Involving Farmers in the Sampling Process

Avoid taking the samples on your own; instead, be sure to involve the farmer in the process. After all, extension should aim to "enable" farmers rather than create dependency. Also, the farmer's input is essential when it cones to mapping the farm and filling out information on past management, yield history, cropping plans, and expected yields.

The Sampling Procedure

The steps below provide general guidelines for sampling. Always consult the lab's instructions,too.


What is a sampling unit?: An area of soil that is likely to be uniform in its fertility. Even small farms usually have several sampling units.

How to Distinguish Sampling Units

Each of the following factors indicate likely differences in soil fertility:

• Soil color

• Soil texture

• Topography (slope vs. flat vs. depression)

• Past management (Use of manure, fertilizer, lime; type of crops grown. For example, new ground that has been in pasture for years will have a different fertility status than land that's been cropped steadily.)

Mapping the Farm

The map isn't needed by the lab, but is used to delineate the different sampling units and serve as a record of which sample came from where. It's also a useful management tool for the farmer. If the farmer is agreeable, make an extra copy for yourself - it will be useful in further extension work with him. Here's how to map a farm:

• Start by drawing in the boundaries and the location of buildings, wells, and the fields along with their dimensions .

• Indicate variations in topography, slope, soil color, soil texture, and past use of fertilizers and lime.

• Indicate past, present, and future crops and their locations.

• Once the sampling units have been determined, draw their boundaries and number them.

FIGURE 7-1: Farm map denoting different sampling units.


It's important to realize that each sample sent to the lab is really a composite sample composed of 10-20 subsamples taken at random within that sampling unit.

Guidelines for Extracting Subsamples

Tools: A shovel, machete or knife (for paring down the samples), and a pail or sturdy sack (for placing and mixing subsamples). Special soil sampling tubes or augers may be available on loan from the ag extension office but aren't essential.

Depth to sample: Most labs want a uniform slice of the top 15-20 cm of soil, the normal depth of topsoil. Some labs may also request separate subsoil samples. If the field is severely eroded, the normal sample will also include some of the subsoil, but that's OK.

Extraction method: If using a shovel, you can use several methods. The important thing is to end up with a uniform slice of soil from the surface to a depth of 15-20 cm. One way is to make a hole with about 45 degree sides to the right depth. Then use the shovel to trim off a 3-4 cm thick slice 15-20 cm deep. If a second person holds the face of the slice with one hand, it won't crumble apart. Scrape off any surface debris like stones or stalks before sampling.

Use a random Pattern: A zig-zag pattern is fine, but don't sample along fence lines, in fertilizer bands, under animal droppings, or in transition zones between sampling units.

Uniform size: Each subsample should be of equal size. Use a knife or machete to pare down each one to a similar width and depth (see Fig. 7-2).

If a zinc test will be done, don't collect subsamples in a galvanized pail.

FIGURE 7-2: (Left) Extracting a subsample using a shovel. (Right) Subsample after trimming with a machete.


• NEVER mix together subsamples from different sampling units!

• Drying: The soil can be slightly moist. If overly wet, it can be sun dried. Do not oven dry, as the heat will alter the soil's potassium, resulting in an erroneously high reading.

• Be sure to number each composite sample so it corresponds with the sampling unit from which it came.


Another sign of a good soils lab is its provision of a detailed information sheet requesting data on the farmer's situation. One purpose of the form is to provide extra information on the soil that's not revealed by the samples themselves (i.e. depth). In addition, it should attempt to evaluate the farmer's management level and the likely limiting factors affecting the intended crop's yield potential. The data requested should include most of the following:

• Farm size

• Soil depth

• Soil slope

• Soil drainage

• Past applications of fertilizer and lime

• Past crops

• Past yields

• Crop to be grown

• Variety to be used

• Intended use of compost, manure, or green manure crops

• Yield goal

• Capital available for fertilizer

• Limiting factors like insects, diseases, nematodes

A lab that requests little information isn't as likely to tailor its fertilizer recommendations to varying farmer circumstances.

Getting the Information: Depending on the culture and the farmer, there may well be hesitancy in providing some of the above data, especially regarding past yields and available capital. A lot may depend on the familiarity, trust, and credibility you have established with the farmer. It's important to explain the purpose the information will serve and yet not be overly insistent about obtaining it.


Depending on your situation, the samples can be mailed in, taken to the local extension office, or delivered personally to the lab. Visiting the lab is a worthwhile experience; it will give a better idea of what goes on there, and the personnel can often provide very useful information on soils and fertilizer use.


The form in which soil test results and fertilizer recommendations are given varies a lot with the lab and with the role played by the extension service. Often, there will be two parts to the results, the first being the fertility analysis of soil, and the second being the actual fertilizer recommendation. Even if a farmer or family member can read, the recommendation sheet may be too complex. In some cases, the lab mails the results to the local extension office where they are then put into a more readily understandable form.

Plant tissue testing

A growing crop can have its stem and leaf tissue tested for N-P-K levels in the sap. This can be done either in the field with portable kits or at a lab. (Labs can also measure micronutrient levels.) The uses and limitations of tissue testing are:

• Tissue tests are best used as a supplement to soil tests and can be tricky for nonprofessionals to interpret.

• Sometimes nutrient levels in plant sap aren't well correlated to those in the soil. Weather extremes, soil compaction, poor drainage, insects, and diseases affect nutrient uptake. Deficiencies of N, for example, can stunt plant size and cause P and K to "pile up" in the sap, resulting in falsely high readings.

• The tests are usually calibrated for higher yield levels than most small farmers should be aiming for. Low to moderate fertilizer rates, rather than high ones, give the best return per dollar (see Chapter 9), yet a tissue test may indicate a deficiency at these levels.

• One advantage of tissue tests is that they can spot deficiencies in a growing crop while there's still time enough to correct them.

Portable tissue test kits cost from $25-$75, but some of the reagents need yearly replacement. Tissue testing is probably best left to trained agronomists.

Fertilizer trials

Well-run fertilizer trials can be very helpful. There are 3 kinds:

• Test strips

• Field experiments (farm experiments)

• Field tests (field trials, result tests)

Test Strips

Running test strips through a field is a quick and easy way to test crop response to different fertilizer rates and nutrient combinations. Test strips have less statistical significance than formal trials, but they do allow farmers to conduct some research on their own farms, which can be very useful. A researcher's test of statistical significance is usually a 95 or 99 percent likelihood that the results were due to the treatment applied rather than to chance. Farmers (and most of us) would probably settle for a much less stringent figure and try a new practice, even if there were only a 75 percent chance of getting a response.

To reduce the influence of soil variations, each treatment tested should consist of several strips 2-3 rows wide placed in different parts of the field. The soil should still be uniform visually and in terms of past management. Don't rely on just one season's results, since weather and pests can influence yields.

Fertilizer rates for test strips: Consult your extension office and Chapter 9 of this manual.

Field Experiments (Farm Experiments)

These are designed to be statistically valid and require much more effort and care to set up and manage. They are designed to determine the most profitable kind and amount of fertilizer needed for a given crop and soil. Suppose you want to try 3 different rates of fertilizer. It's not simply enough to mark out one test plot and 3 fertilized plots. Each of these 4 "treatments" needs to be replicated 3-4 times and laid out within a bloc in a randomized manner. Each of the 12-16 plots are only a few rows wide and several meters long. Plot size, plant population, and fertilizer rates have to be carefully measured, along with yield differences. It's a good idea to repeat a trial for 2-3 years to take into account weather variations.

Formal experiments require much time, skill, attention to detail, and scientific discipline. They are not something that you should do on your own. However, you can play a very useful role in a well conducted and ongoing fertilizer testing and demonstration program.

Field Tests (Field Trials, Result Tests)

This type of research tests the best fertilizer type and rate determined from the farm experiments above, but this time under actual farming conditions. There are normally just 2 plots: the "control" plot (traditional practice) and the "treatment" plot (improved practices). Rather than rely on randomization and replication of the plots on each farm, the field test gets its statistical validity from being conducted on a number of farms with fairly similar conditions. In most cases, fertilizer will be only one of several improved practices making up the treatment plot. The plots should be large enough so that realistic farming methods can be used. To be valid, the farmer and other usual labor should carry out the practices themselves with some initial instruction and supervision by the extension worker.

Experiments/Trials vs. Demonstrations

The experiments and trials above seek to develop fertilizer recommendations for local conditions that will be the most affordable and profitable for the farmers involved. Don't succumb to the natural and prevalent temptation to use such tests as demonstrations. After all, a demo is designed to provide farmers with "living proof" of the benefits of a new practice (or package of practices) - one that has first proven its worth under local conditions. This syndrome of promoting without adequate prior testing has cost many an extension worker an irreparable loss of credibility. Testing is always the first stage; promoting comes later.

NOTE: For more information on experiments, trials, and demos, refer to the Peace Corps/ICE manual, Traditional Field Crops (M-13).

Using visual "hunger signs"

Severe nutrient deficiencies often produce telltale changes in plant appearance, particularly in color. Spotting these "hunger signs" can be useful in diagnosing fertilizer needs, but be aware of several drawbacks:

• Some hunger signs are readily confused with each other or with other problems such as insects, diseases, and nematodes. Even trained field technicians may be unable to make a definite diagnosis without lab tests.

• If more than one nutrient is dificient, the hunger signs may be too ambiguous for accurate diagnosis.

• "Hidden hunger": Hunger signs don't usually appear unless a nutrient deficiency is serious enough to cut yields by 30-60 percent or more.

• It may be too late to correct deficiencies by the time hunger signs appear.

Hunger sign diagnosis is likely to be most useful in areas where only one or two nutrients are commonly deficient on a crop that will manifest unusually clear symptoms. For example, maize shows the most clearly recognizable symptoms of zinc hunger of any crop. Nitrogen deficiency is relatively easy to spot, although a number of other factors can cause similar symptoms.

How to Spot Hunger Signs: Hunger signs for common crops are described in Appendix E. Refer also to Appendix H for useful references with color plates.