The authors illustrated and proved that “heating V770-NPI-R containing anthrax strains, to exactly 42.5 degrees centigrade, as Pasteur and colleagues reported in GM 14211, essentially ‘cured’ these strains of their plasmid and caused a loss of detectable lethal toxin.” Near the end of their paper they stated, “In assessing Pasteur’s experimental regimen, and by utilizing modern techniques, we are able to offer a reasonable explanation for a century old molecular event which has had such a significant impact in the field of medical microbiology and it is
very likely that his attenuation of the anthrax bacillus occurred as a result of curing the strain of plasmid
component which encoded for toxin structural and regulatory proteins.”

Order of authorship in the original publication: Mikesell, Ivins, Ristroph.... Digital facsimile from PubMedCentral at this link.

See also: Mikesell, Ivins, Ristroph et all, "Plasmids, Pasteur, and anthrax," ASM News (1984) 320-322.

(Thanks to Juan Weiss for this reference and its interpretation.)