Indian Hemp Meets the Pharmacopoeia
A Medicine Without Method
In 1850 the committee revising the United States Pharmacopoeia admitted a new entry, Extractum Cannabis, the extract of hemp. The monograph called for an alcohol extract of the dried flowering tops of Indian hemp. Like other entries in the book, it set out a single method of preparation for everyone who made the drug. The British Pharmacopoeia was carrying cannabis monographs of its own by 1888.
The raw material reached them through a long chain of supply. The flowering tops were grown in India and shipped across oceans before wholesalers sold them on. Each maker turned the dried material into extract with its own equipment and its own sense of when a batch was finished. The British trials that first brought Indian hemp into Western practice had already produced disagreements about how the extract behaved. Official status did not settle them. It moved them from the medical journals onto the pharmacy shelf.
In the regions that supplied the drug, cannabis preparation had been carried by particular practices and conventions. Those conventions shaped how material was chosen and how preparations were made and taken. Western pharmacy received the drug along with a recipe for extracting it. The broader preparation knowledge behind those practices did not arrive as a standard that Western pharmacy could independently apply.
For nearly a century the extract held its place. Physicians prescribed it for pain, spasm, and sleeplessness, and manufacturers kept it in their catalogs. Yet by 1937, when Congress taxed it, cannabis had already drifted toward the edge of medical practice, and in 1942 the pharmacopoeia let it go. The law helps explain the ending. It does not account for how a medicine with an official monograph had become so easy to lose.
The Pharmacopoeia Standard
A pharmacopoeia entry was a small contract between strangers. The physician who wrote a prescription and the pharmacist who filled it might never meet, and the entry stood in for the conversation they could not have. It was meant to tell both of them what the drug was, how strong it was, and whether it would still be the same drug after months on a shelf. By the later nineteenth century, the most useful entries could answer those questions without relying on anyone's impression of a patient.
Chemistry made that possible for some drugs sooner than others. Early in the century, chemists had drawn morphine out of opium and quinine out of cinchona bark as crystalline compounds. Once a drug's action could be traced to a substance that could be isolated and weighed, its preparations could be checked against a number. Later pharmacopoeia revisions fixed extracts of nux vomica and cinchona to a definite percentage of their active ingredients. A physician prescribing those extracts could expect the stated strength from any competent maker, because an analyst could check it.
Cannabis offered chemistry nothing to hold. Its activity lay somewhere in the resin, but no one could isolate a responsible compound or measure it in a finished extract. The monograph could name the plant and describe the extract. It could not state a content. The gap showed internationally as well. When delegates agreed on the first international standards for potent drugs at Brussels in 1906, cannabis was left out because too little was known about its constituents to fix a common formula. Cannabis preparations entered the international standard only with the revised agreement of 1925.
The missing measure weakened every other requirement the entry was meant to meet. Stability could not be confirmed when there was nothing to track as a bottle aged. The shared expectation between prescriber and pharmacist depended on which supplier the pharmacist happened to use. Cannabis had entered the pharmacopoeia with a name and a recipe, but the measure needed to check its preparations did not yet exist.
Bottle to Bottle
A correctly labeled bottle of cannabis extract could hold a strong preparation or a feeble one, and nothing available in an ordinary pharmacy could tell the difference. The variation began long before any extract was made, with where the plant was grown and how the material fared on its way to the maker. The pharmacopoeia itself acknowledged part of the problem. From 1870 its revisions distinguished flowering tops grown in North America from those grown in India, recognizing that origin mattered while offering no way to say by how much.
At the point of care, the patient often became the practical test. A physician who saw no effect could not know whether the fault lay with the dose or with the bottle. One whose patient reacted strongly to a modest amount had no assurance that the next supply would behave the same way. Each learned something real at the bedside, and neither could pass that lesson to a colleague in a form that would hold for a different extract. Case reports accumulated through the later nineteenth century. Because nobody could say what the extract in any given report had contained, the reports could be collected but not reliably compared.
What crossed from South Asian practice into Western pharmacy was a way of making a medicine. The alcohol extract of resinous flowering tops, adapted by O'Shaughnessy and the pharmacists who followed him, crossed into Western pharmacy. What did not cross was a way of verifying one preparation against another. Western pharmacy needed a check that would work between people who had never handled the same material, and in the nineteenth century no one on either side of the exchange could provide it.
Standardized on Dogs
Manufacturers understood the problem, and some tried to solve it. Chemical assay was out of reach. Commentators on drug standardization grouped cannabis indica with digitalis and ergot among the potent drugs that chemists could not assay, because the reagents used in assaying decomposed the substances responsible for their action. For these drugs, the practical alternative was to measure an effect on a living animal.
Parke, Davis & Company of Detroit, already known for extracts assayed to fixed alkaloidal strength, built such a test for cannabis. In a 1908 paper, the company's researchers E. M. Houghton and H. C. Hamilton described giving a standard dose of extract, 0.01 gram per kilogram of body weight, to dogs selected for their sensitivity to the drug, since dogs varied in response much as people did. An experienced observer watched for the characteristic loss of coordination and compared it with the effect of the same quantity of a reference preparation. The test told the firm whether a new batch acted like the firm's standard.
The company also worked on supply. The 1908 paper reported that cannabis grown in the United States and Mexico proved as active in this test as the best Indian material. According to a later company account, Parke-Davis cooperated with Eli Lilly on a standard preparation and eventually settled on a domestic strain raised at its own farm near Rochester, Michigan.
The dog assay was a genuine advance, and its limits were clear. It was a biological potency assay, designed to make successive batches behave more alike by comparing what one preparation did with what another did. It did not identify composition. Two extracts could produce the same unsteadiness in the same dog while differing in what they contained, and the test had no way to separate them. Its reference point was itself a sample of the plant, and its readings depended on selected animals and a practiced eye.
Manufacturers could now compare effects without knowing which constituents produced them or which ones changed from crop to crop. The pharmacopoeia's original problem remained where it had been. A standard that lived in one firm's reference sample and one observer's judgment could not be written into a monograph and checked by an analyst elsewhere. Cannabis had acquired a form of quality control. It still had no specification.
Losing Ground
From the late 1860s onward, a new class of chemically defined drugs was reaching physicians. Chloral hydrate entered medicine as a sleeping drug in 1869. Bayer brought aspirin to market in 1899, and barbital, the first barbiturate in clinical use, followed in 1903. None of them was harmless, and chloral quickly gained a reputation for habit and overdose. Each, however, was a single known compound, so a prescribed amount meant much the same thing from one pharmacy to the next. These drugs were not simple substitutes for cannabis, but they increasingly occupied some of the indications in which it had been used, sleeplessness and certain kinds of pain among them.
Cannabis did not disappear at once. It stayed official in the United States, and in Britain an extract was still listed in the British Pharmaceutical Codex of 1934. Its decline came as a slow contraction of use, with no single moment at which the profession weighed it and found it wanting. As chemically defined alternatives took over part of the ground it had held, its place in routine practice narrowed.
The Marihuana Tax Act, signed on August 2, 1937, and in force from October 1, did not forbid medical use. It placed registration, occupational taxes, and order-form paperwork on the physicians and pharmacists who handled cannabis. The American Medical Association opposed the bill. Its legislative counsel, William C. Woodward, testified before the House Ways and Means Committee on May 4 that the association knew of no evidence the drug posed the danger claimed for it. He proposed that any federal control be folded into the existing Harrison Narcotic Act instead. Congress passed the bill anyway.
The Act added administrative burdens to a drug whose medical use was already contracting and whose strength no pharmacist could certify, at a time when press campaigns had tied its name to crime and madness. Declining use and new regulatory requirements worked on each other, and together they hastened its retreat from practice. In 1942 the twelfth revision of the United States Pharmacopoeia omitted cannabis. With the monograph gone, revision committees no longer had an entry to improve, and manufacturers had less reason to keep refining a product that fewer physicians would prescribe.
Chemistry Came Later
Chemists began describing the compounds that might have given cannabis a specification in the years around its removal. R. S. Cahn worked out the structure of cannabinol in the early 1930s. In 1940 Roger Adams's group in Illinois isolated cannabidiol from Minnesota wild hemp, and the laboratories of Adams and of Alexander Todd in Britain each synthesized cannabinol. In 1942 Wollner, Matchett, Levine, and Loewe extracted an active fraction containing tetrahydrocannabinol, most likely a mixture of closely related isomers. The plant's principal active constituent remained out of reach. Researchers working with Adams still screened synthetic compounds in the dog ataxia assay, so the old potency test outlived the monograph.
The decisive identifications came two decades later. Raphael Mechoulam's laboratory established the structure of cannabidiol in 1963, and in 1964 Yechiel Gaoni and Mechoulam isolated Δ9-tetrahydrocannabinol from hashish in pure form and determined its structure. Later work added a distinction that belongs entirely to modern chemistry. In the plant, the major cannabinoids occur mainly as acids, which heat converts to their neutral forms. The nineteenth-century record is silent on this, and no surviving monograph or case report can say how any historical extract divided between the two.
By the time these tools existed, cannabis had been outside the pharmacopoeia for more than twenty years. The chemistry matured after the medical tradition had been set aside, and its place in routine pharmacy had already been lost. What continued official status might have produced cannot be known. Serious chemical work on cannabis gathered pace just as its official medical use was ending, and the two barely overlapped.
Later researchers inherited a substantial record of case reports, dispensatory entries, and manufacturers' assay methods. What they did not inherit was a continuous, validated method for verifying one cannabis preparation against another. The old reports can be read, but their doses cannot be translated into measured contents. Each describes an amount of extract, and nothing in the record says what that extract contained in terms later chemistry could use.
References & Citations
Sarma et al., Journal of Natural Products, 2020 — review of USP cannabis quality standards and compendial history.
Informs: Documents cannabis’s historical place in pharmacopoeial standards and clarifies the distinction between an official preparation and a chemically defined quality specification.
Houghton and Hamilton, American Journal of Pharmacy, 1908 — pharmacological study of Cannabis americana and physiological potency testing.
Informs: Describes the dog-based biological assay used to compare cannabis preparations against a reference preparation, establishing the article’s distinction between potency comparison and chemical identification.
Musto, Archives of General Psychiatry, 1972 — historical analysis of the Marihuana Tax Act of 1937.
Informs: Clarifies the Act’s legislative history and the medical profession’s opposition, supporting the article’s treatment of federal regulation as an added burden on a drug already declining in routine practice.
Pertwee, British Journal of Pharmacology, 2006 — review of early cannabinoid chemistry and pharmacology.
Informs: Establishes the chronology of cannabinoid research from the 1940s onward and places the later chemical identification of individual cannabinoids within the discontinuity between historical cannabis prescribing and modern cannabinoid chemistry.