President Nixon justified his request for emergency powers to deal with drug abuse in 1971 by citing an uncontrollable heroin epidemic which, if not brought under immediate control, "will surely in time destroy us." According to official statistics supplied to the media by federal agencies, the number of addict-users had increased from 68,000 in 1969 to 315,000 in 1970 to 559,000 in 1971, or what Myles Ambrose declared in 1971 to be a "tenfold increase." Such a geometric progression threatened the entire American citizenry with "the hell of addiction" in a few short years, the president suggested, because every individual infected with heroin was in turn driven to infect at least six others. Nor could such an epidemic be brought under control by ordinary means: the president explained that the suppliers of heroin "are literally the slave traders of our time.... They are traffickers in Living Death [and] they must be hunted to the ends of the earth." 

A tenfold increase in the number of heroin addicts would certainly be a cause for national concern; the magnitude of the 1971 epidemic was, however, more a product of government statisticians than of heroin traffickers.

Until 1970, official estimates of the number of heroin addicts were based on a register kept by the Bureau of Narcotics and Dangerous Drugs which, like the FBI's Uniform Code Reports, was simply a compilation of reports from local police departments. All police departments-and medical authorities-were supposed to report the names of known addicts to the bureau. And, on the theory that most, if not all, addicts eventually would come to the attention of some police department or hospital, it was assumed that the total number of names on the federal register constituted nearly the entire addict population (within "a deviation factor of less than two percent," Myles Ambrose explained to the National Commission on Marijuana and Drug Abuse in 1971). The pre-epidemic estimate of 68,088 in 1969 was based on this register. (By mid-1970 the addict population had grown to only 68,864, according to the register, which was not published that year.)

 The prodigious increase from some 68,000 addicts in 1969 to 315,000 in late 1970 and 559,000 in 1971 came not from any flood of new addicts reported to federal authorities in 1970 or 197 1 but from a statistical reworking of the 1969 data. Rather than continuing to publish estimates based on the federal register, the Bureau of Narcotics and Dangerous Drugs decided to apply a new formula to the old 1969 data, which produced first a quintupling, then an octupling, of the estimated number of addicts. Unlike the previous theory that almost all addicts would eventually come to the attention of authorities, the new formula was based on the belief that only a small fraction of the addict population would ever be reported to police or medical authorities, and therefore listed in the federal register. Joseph A. Greenwood, the statistician at the BNDD who devised the new formula, explained, "it is impossible to actually know the number of addicts in the United States.... The best we can do is make some assumptions", he then proceeded to apply these assumptions to the data collected in 1969 to estimate the number of unknown addicts. "The estimate makes use of a technique similar to that by which the number of fish a lake ... is estimated," he noted.

To understand how this statistical artifact led to a tenfold ballooning of the number of addicts in America and provided the president with a national emergency, it Is necessary to examine the so-called "tagged-fish-in-a-pond" technique. One way of estimating the number of fish in a pond would be to catch an initial sample of fish, tag them, then release them and allow enough time for a random redistribution of the tagged fish among the rest of the fish in the pond. A second sample of fish is then caught, and the proportion of tagged to untagged fish in this catch would allow an estimate to be made for the total number of fish in the pond. For example, if one out of ten fish in the second sample was already tagged, then the total population would be assumed to be ten times the number of fish originally tagged. In applying this concept to addicts, statisticians at the BNDD divided the total number of names on the federal register in 1969 by the number of names that were "tagged" (that is, reported in 1969) and rereported in 1970, and found only about one out of five of the tagged addicts was rereported. After some refinements were made in the statistical model, the 1969 total of 68,088 was multiplied by 4.626, the calculated ratio, and an estimate of 315,000 addicts was arrived at. This accounted for the 1969-70 epidemic. The same 1969 data were then, in 1971, further refined, and a new ratio was calculated of one known addict to 8.21 unknown addicts, which produced a staggering total of 559,000 in 1972, an increase of some 244,000 between 1971 and 1972. The total increase was in unknown addicts, who were not unlike the paper army of nonexistent serfs in Gogol's Dead Souls. In 1969 it was not assumed that there were a significant number of addicts not known to the police agencies; in 1970 it was assumed that there were 246,912 unknown addicts roaming the streets; in 1971 -it was assumed that there were 490,912 unknown addicts at large.