An inadequate sample can easily result in a false negative result. Many people assume that testing a small proportion of the plants in a nursery will indicate whether the remaining plants are infected. Unfortunately, this may not work for several reasons. First and most obviously, in container-grown plants, testing results are specific to the plant and specific roots or other materials being tested. Even if tests were 100% efficient at detecting Phytophthora infections (they are not), the disease status of one plant does not directly indicate the disease status of other plants in the batch or nursery.
In a nursery with a high incidence of infected plants, it is often possible to detect Phytophthora with a fairly small number of samples using an efficient test. Consider a nursery in which 50% of the plants are infected. If you choose two plants at random from this nursery, it is likely that at least one will be infected. However, even at a 50% infection rate, you would still need to sample five plants at random to have more than a 95% chance of selecting an infected plant. If the test is only 50% efficient at detecting Phytophthora, these numbers need to be at least doubled.
In a clean nursery, infection levels should be close to zero, hopefully much less than 1%. If the infection rate is 1%, you are not likely to pick out an infected plant in a random sample smaller than about 100 plants. Furthermore, you would need a random sample of about 300 plants to have a 95% chance of having one infected plant. Again, unless the test can detect an infected plant 100% of the time, you would need even larger sample sizes to actually detect Phytophthora.
From these numbers, it is obvious that random sampling will not be of much use for identifying rare disease cases in an otherwise clean nursery operation. To increase the odds of finding diseased plants, we need to focus first on plants that appear to have symptoms, even subtle symptoms such as reduced growth or slight off color. However, as noted in Best Management Practices for Producing Clean Nursery Stock and Phytophthora species in native plant nursery stock: issues and implications, many drought-tolerant California native plant species do not show top symptoms under nursery condition until root rot is severe.
Another major factor affecting the efficiency of testing is whether each individual sample is large enough to include infected roots that are in a condition that will result in a Phytophthora detection. As in the discussion above, if a high percentage of the roots in a pot are infected, it will be relatively easy to sample enough roots to get some infected ones, particularly if the plant is unpotted and the root system is examined carefully for symptomatic roots. For methods such as direct isolation, immunoassay, and PCR, the amount of roots plated or tested is miniscule, so careful selection of suitable roots is critical. However, methods that rely on very small root samples may commonly generate false negatives if symptoms are difficult to see (e.g., if healthy roots are dark in color) or there are few infected roots. By comparison, baiting techniques have the potential to be more efficient at detecting Phytophthora because they can test a larger portion of the root system. However, to minimize false negatives when baiting roots, conditions for zoospore production, release, and infection of baits need to be optimized.
The longer that infected plants remain undetected in the nursery, the greater chance there is for spread among adjacent plants. So while careful inspection and individual plant testing is an important monitoring tool, effective larger scale testing of plant blocks is also needed to provide early detection of Phytophthora.