I cannot absolutely exclude the possibility that Rudolf is being honest, but one should be suspicious of his value of 0.3 to 1% cyanide concentration in US executions. Even if he can provide a source for this claim, it is questionable what the uniformity of such usage is. It is certainly no manner in which to make an estimate of LC100 (the concentration required to kill 100% of those exposed in a given time period).
He's not wrong that one should be suspicious of Rudolf's value of 3000-10000 ppm for US gas chamber executions: it is almost certainly too low. As for uniformity of usage, it's true that there is some variation, in particular because different states have gas chambers with different volumes as well as because of the difference of yields for KCN and NaCN; however the amount of cyanide seems to be pretty well standardized at 15-16 oz, although 10-12 oz was used in some early executions. As Seitz writes in her thesis
In states that use lethal gas, the design of the equipment is fairly uniform. The chair is contained in an airtight compartment with windows for accommodating witnesses viewing the procedure. A metal container is located beneath the seat of the chair, which contains one pound of cyanide pellets.
Similarly, from the introduction to Legal Executions After Statehood in Arizona, Colorado, Nevada, New Mexico and Utah: A Comprehensive Registry:
The protocol for an execution by lethal gas was relatively standardized. Each state adopting lethal gas constructed an airtight chamber with windows, in which was affixed a chair with straps attached. As the moment neared, the condemned was instructed to breathe deeply when the “almond odor” fumes reached his face; when the condemned failed to follow these instruction it could result in several minutes of violent convulsions.
The executioner would place approximately fifteen ounces of potassium cyanide or sodium cyanide pellets into a container under the seat of the chair above a container of sulphuric acid and distilled water...
It's also worth noting that the LC100 is not, as Green thinks, the concentration required to kill 100% of those exposed in a given time period. As explained in this thread, it is the concentration required to provide a lethal dose to 100% of those exposed for a given time period - deaths might occur after the time period.
Green also writes:
There is no ethical way to measure precisely LC100 for humans. Even making a proper estimate based upon US gas chamber times is problematic because the concentrations used were not necessarily consistent, nor is the sample size big enough.
Again, we should ignore "LC100" whenever Green says it, since he doesn't understand what it means. What he means is that this is no way to estimate time to death. This is an astonishing statement. Gas chamber executions are exactly how you measure time to death. It's the best data you could possibly have. Of course, you have to be careful with sources for time to death in US gas chambers. Newspaper accounts should be mistrusted if they just sell the party line ("quick and painless") unless they come from eyewitnesses. Internal documents of the prisons are valuable (as in Friedberg's study), as are statements from the responsible authorities, especially when they aren't engaged in defending themselves from criticism that gas executions are too slow. Despite the limitations of the available data, US gas chamber executions are far and away the best source of information on time to death in humans exposed to high concentrations of HCN.
Green's statement that the sample size is not big enough is absurd. For example, Friedberg's study looked at 112 executions, all in California (which helps to ensure a consistent concentration). Where does Green think we're going to find a better sample? Does he think DuPont carried out experiments killing thousands of people with HCN in order to get their figure of 300 ppm?