The standard way to determine chemical toxicity is a concept anybody who is even remotely involved with control materials (pesticides) should be familiar with. The exact lethal dose of any chemical is difficult to determine. Almost anything taken to extreme, including unusually benign materials like water, for example, can be deadly. There are many variables that could be responsible for the death of any test animal within a specified group of organisms. Thus, scientists have had to come up with a procedure to determine relative toxicity. This is usually expressed as an LD50 level, and is the dosage of a material that kills, or is lethal to, 50 percent of a test population.
As a hypothetical example, if a friend weighed 177 pounds (70 kilograms), and ate 500,000 milligrams (about a pound) of some material and did not die, one would probably consider it safe. If his buddy (same weight, sex, etc.) ate 400,000 milligrams and died, one might conclude it was unsafe. In a larger sample, it might be observed that five out of 10 friends died after consuming 450,000 milligrams of the material. This would in fact be the oral (consumption) LD50 level and it would be stated as 450,000 mg/70 kilograms (6428 mg/kg). Again, in this example there would be survivors. Thus, it is only possible to say that 50 percent of individuals would die at the LD50, but not specifically which ones.
The basic rule to remember is that, in general, the larger the number the safer the material. Most LD50s are calculated on rats, but other test animals may also be used. Whether what can be applied to rats or other test animals is also true for humans remains controversial. No matter what test population is used, establishing the LD50 is not an exact science. Varying LD50 levels reported by different organizations is a testament to this. Determination is also made more difficult because levels can be different based on the type of exposure. For example, larger numbers are usually calculated for dermal (skin) contact than for consumption by mouth, which is much more risky. In the final analysis, it is up to each individual to determine what specific risk any chemical might be either to themself or their bees. An established LD50 level provides a vital clue, but not a definitive answer, to these important questions.
For many newer pesticides and chemicals, the LD50 measurement has become obsolete, especially when it comes to honey bees. There are many other effects of pesticides that are too nuanced such that other measurements of related toxicity for these social insects need to be employed.