Contemporary Drug Problems 26/Winter 1999 [pp. 577-606]


Overlooking Terris: a speculative reconsideration of a curious spot-blindness in the history of alcohol-control science


The authors argue that the overlooking or forgetting of a beverage-specific element of Milton Terris's classic 1967 paper linking per captia alcohol consumption with cirrhosis mortality trends sheds new light on the subsequent paradigmatic history of alcohol epidemiology.  The historical standing and subsequent citation of Terris's paper are re-examined, and Terri's reasons for not reminding the alcohol epidemiology literature of this aspect of his paper are explored.  Aspects of presentation and content of the 1967 paper are also discussed with respect to the explanation of the subsequently lost beverage-specific element of Terris's article.  The authors suggest that an evolutionarly aspect of the relationship and competition between the modern alcoholism and alcohol controls paradigms in alcohol epidemiology may offer the key to accounting for this historical-forgetting puzzle.

Key phrases:  History of alcohol epidemiology, historical forgetting, alcohol and cirrhosis, history of alcohol science, paradigmatic strains in alcohol epidemiology.

AUTHOR'S NOTE:  This paper was supported by a National Institute on Alcohol Abuse and Alcoholism (NIAAA) grant (#R01 AA07034) and by an NIAAA Research Scientist Award (#K01 AA00073) to the second author.


Over the second half of the 20th century, cirrhosis mortality in the United States followed a long rising trend and then a long declining trend, with the peak rate (14.9 deaths per 100,000 population) occurring in 1973.  Per capita total alcohol consumption also rose and fell, but the correspondence between the consumption and cirrhosis mortality curves was imperfect (see Fig. 1).  Moreover, alcohol consumption peaked in 1980-1981, several years after the 1973 peak in cirrhosis mortality, and hence too late for consumption's downturn to account for cirrhosis mortality's downturn.  Why cirrhosis mortality in the U.S. shows this pattern and why cirrhosis mortality turned downward in the mid-1970s represent enduring epidemiological mysteries.  In that connection, we recently reported that the trend-line for per capita distilledspirits consumption bears a much closer correspondence to the cirrhosis mortality trend than does per capita total alcohol consumption from 1949-1994 (see Fig. 2) (see Roizen et al., 1999).  Our article is one of a handful in the alcohol studies literature offering explanatory perspectives on cirrhosis mortality's trend-line in the U.S. (see, e.g., Mann et al., 1988; Mann et al., 1991; Gruenewald and Ponicki, 1995; Grant et al., 1986).  The spirits-cirrhosis relationship does not solve the puzzle of cirrhosis mortality's rise and fall, but it may offer a clue as to where that solution may ultimately lie (see Roizen et al., 1999, p. 669).

Oddly enough, this "clue" has been available in the alcohol epidemiology literature for more than 30 years -- though it appears to have been largely overlooked or forgotten by the alcohol epidemiology community.  Remarkably, moreover, the clue was suggested in 1967, a half-dozen years before disparate consumption and cirrhosis trends began to appear in U.S. statistical time-series, and it was suggested in a celebrated epidemiological paper:  epidemiologist Milton Terris's landmark 1967 article on alcohol and cirrhosis mortality.  Terris's article included a beverage-specific aspect; specifically, Terris (1967) inferred from his trend analysis of U.S., Canadian, and British data that beer consumption could be safely disregarded in the aggregate-level relationship between per capita alcohol consumption and cirrhosis mortality.  Since wine consumption usually contributes only about one-eighth of total annual ethanol intake in the U.S., Terris's beer-excluding conclusion implied that per capita spirits consumption would make the greatest contribution to the association between alcohol consumption and cirrhosis mortality, thus virtually anticipating our paper's conclusion (Roizen et al., 1999).

In the present paper, we cast a historical eye upon the question of what, if any, significance the dormancy of Terris's beverage-specific inference may have with respect to the post-1967 history of alcohol epidemiology.  Our narrative may be outlined as follows:  First, we briefly re-examine the historical standing of Terris's 1967 paper and its citation in the subsequent literature on alcohol's epidemiologic relationship to cirrhosis mortality.  Next, we examine the question of why Terris himself did not alert the alcohol epidemiological community to his beverage-specific hypothesis after the appearance of disparate trends in the late 1970s and thereafter.  Next, we consider whether presentational or substantive features of Terris's paper may have accounted for the overlooking of his beverage-specific inference.  Finally, we examine the ways a collective memory lapse regarding the beverage-specific aspect of Terris's 1967 paper may shed useful light on the relationship and historical development of two vying paradigms in alcohol studies, the modern alcoholism paradigm and the alcohol controls paradigm.  We hypothesize that the overlooking of the beverage-specific aspect of Terris's 1967 paper may have derived from a transition from (a) employing an aspect of the alcoholism paradigm to (b) employing the Ledermann model in order to link cirrhosis mortality with a wider orbit of alcohol-related problems.

Terris's 1967 paper's historical standing and subsequent citation

Room (1984), Herd (1992), and Katcher (1993) offer historical perspectives on the re-emergence of scientific interest in alcohol as a cirrhogenic factor in the period following publication of Terris's (1967) paper.  All three note the relative de-emphasis of alcohol that characterized the prevailing scientific view of cirrhogenesis in the immediate post-War period – when the scientific interest in both the alcoholism paradigm and nutritional and environmental factors in cirrhogenesis tended to draw attention away from alcohol's cirrhogenic potentials.1  All three also accord a place of honor to Terris's 1967 paper in the story of the rebirth of interest in alcohol's role in cirrhosis.  Yet each of these historical papers positions Terris's article as something of an anomaly in its own time -- ahead of its day, greeted with dubiousness or scorn, and perhaps thus also deprived of significant contemporary (or even lasting) substantive impact on the field.

Room (1984), for example, recounted the reluctance with which the paper was received by its initial public health audience (pp. 294-295), noting the retardation of its publication.  "Terris' paper was omitted from the published proceedings of the session at which it was presented," wrote Room (1984, pp. 293-294), "although over half of the prepared discussion, which was published, focused on Terris' 'provocative analysis.'" Room's (1984) narrative history next jumped fully 15 years down the road, to a time by which "a substantial revolution [had] occurred in public health approaches to alcohol issues" (p. 294).  The implication is that the time-leap between the mid-1960s, when Terris's paper was presented, and early 1980s, a decade-and-a-half later, saw changes in epidemiological zeitgeist that made Terris's paper significantly less anomalous and suspect. Herd's treatment of the history of ideas in cirrhosis epidemiology made much the same leap (Herd, 1992, see pp. 1119-1120).  Hinted in these narrative handlings is that Terris's celebrated paper may have occupied a somewhat larger place in the historical story of rebirth of attention to alcohol's role in cirrhosis epidemiology than in the actual work of practicing scientists in the field.

A glance at the bibliographies of well known reference volumes suggests the same conclusion.  For example, citations of Terris are notable for their absences from Bruun et al's (1975) pathbreaking Alcohol Control Policies in Public Health Perspective, from the main text of Moore and Gerstein's Alcohol and Public Policy: Beyond the Shadow of Prohibition (1981) -- though one of the commissioned papers therein cites it (Cook, 1981) -- and from Edwards et al.'s Alcohol Policy and the Public Good (1994).  These well known summary publications do not of course constitute the whole of the alcohol epidemiological literature, and Terris's paper certainly makes at least occasional "utilitarian" (i.e., as opposed to historical) appearances in a variety of other papers and book chapters.  For example, Reginald Smart (1974), one of the literature's earliest and most devoted protagonists of an alcohol-cirrhosis connection, cited Terris and Terris alone as authority for the broad assertion that "Liver cirrhosis deaths have been found to vary directly with per capita alcohol consumption in various countries" (Smart 1974, p. 115).  More often, however, Terris's paper appears to have been cited in the alcohol-cirrhosis epidemiologic literature for a specific point rather than as a general source for the broader alcohol-cirrhosis relationship per se.  Wolfgang Schmidt (1975, p. 22; 1977a, pp. 29-30; 1977b, p. 11), for example, employed Terris's paper to vouchsafe the particular assertion that cirrhosis's slow-developing progress at the individual level might nevertheless give rise to trend data showing an unlagged decline in cirrhosis mortality following declines in per capita alcohol consumption.  Following Schmidt, Terris's paper may thus have become synonymous with this, or other, lesser elements of the epidemiologic case for the alcohol-cirrhosis connection, thus obscuring the recollection of the beverage-specific aspect of Terris's paper.

Of particular interest is how Terris and his beverage-specific recommendation may have been cited or overlooked in literature specifically addressing beverage-specific relations between alcohol and cirrhosis.  This beverage-specific aspect forms only a minor focus in recent alcohol-cirrhosis literature -– a status reflected in the recent declaration by the World Health Organization working group on population levels of alcohol consumption that "epidemiological studies so far have not shown convincing evidence for a differential effect of different types of beverages" regarding alcohol's health-related effects (Rehm et al., 1996, p. 277).  Nevertheless, a number of papers, new and old, have examined beverage-specific alcohol-cirrhosis relationships. Schmidt and Bronetto's (1962) pre-Terris (1967) cross-sectional analysis of U.S. state-level data in 1950 argued that wine consumption revealed the strongest association with cirrhosis -- perhaps, they suggested, because a small subgroup of "winos" made a disproportionate contribution to cirrhosis mortality and therefore variation in the size of this specific subpopulation state-to-state may account for the association they reported.  Obviously, Terris could not be cited.  More than three decades later, Gruenewald and Ponicki (1995) conducted a somewhat more complex analysis of U.S. state-level data, this time incorporating a short-term cross-temporal dimension as well.  They concluded that spirits consumption evidenced the strongest association with cirrhosis trends, suggesting (following Selzer et al., 1977) that "alcohol dependent and alcoholic drinkers tend to prefer distilled spirits to beer and wine" (Gruenewald and Ponicki, 1995, p. 635).  Terris's paper was cited twice therein -- first, as one source for the assertion that past literature had confirmed the existence of positive cross-sectional relationships between per capita consumption and cirrhosis, and, second, as one source supporting the contention that changes in consumption can occasion rapid or non-lagged shifts in cirrhosis mortality (cf  Schmidt, 1975; 1977a, as noted above).  No mention of Terris's beverage-specific inference was made and, interestingly, neither did Gruenewald and Ponicki (1995) report testing their spirits finding against U.S. national trend data.

Though it focused chiefly on pancreatitis, D.N. Schmidt's (1991) analysis of alcohol consumption and morbidity trends in Stockholm County, Sweden offered strong evidence that spirits consumption, and not beer and wine consumption, consumption was associated with alcoholic liver disease trends -- once again relying on the contention that chronic alcoholics preferred spirits  (D.N. Schmidt, 1991, see pp. 47, 50).  Terris was not cited.  Longnecker et al. (1981) found that high spirits consumption was associated with higher risk of liver cirrhosis mortality among white immigrant or first-generation residents of Pennsylvania.  Terris (1967) was cited but only in the context of a broad introductory assertion that previous research had linked alcohol consumption with liver cirrhosis and other diseases (Longnecker et al., 1981, p. 791).   De Lint and Schmidt (1971) expressed skepticism that any beverage was more responsible for cirrhosis, citing cross-national cross-sectional data to counter the notion that distilled spirits bore any especially important causal role.  "Clearly," they concluded, "many countries in which a large proportion of alcohol is consumed in the form of beer and wine rather than distilled spirits, have high rates of alcoholism" (p. 104).  Terris was not cited.  Schmidt (1975) cited Lelbach (1974) as authority that the clinical level afforded little support for the influence of beverage differences in cirrhogenesis (p. 25), though two older epidemiologic studies favoring spirits in cirrhogenesis were also cited (i.e., Wallgren, 1960, and Battig, 1964).  Schmidt (1975) noted methodological problems and the contrary findings of Schmidt and Bronetto (1962) in downplaying their spirits-emphasizing results.  "There exists, then," concluded Schmidt (1975, p. 25), "no valid epidemiological or clinical evidence which would suggest that a certain amount of absolute alcohol consumed in one type of beverage is more likely to produce cirrhosis than when consumed in another type."  As already noted, Terris is cited in this paper, but only in relation to the plausibility of consumption changes resulting in rapid changes in cirrhosis mortality.  Tuyns et al. (1984) said it all in their paper's title--"Ethanol is Cirrhogenic, Whatever the Beverage" -- and did not cite Terris.

Perhaps the most intriguing example of non-citation of Terris's beverage-specific recommendation occurred in a useful review paper by Jan de Lint (1977) -- offering, as its title suggested, a "Critical Examination of Data Bearing on the Type of Alcoholic Beverage consumed in relation to Health and other Effects."  Cirrhosis formed only one part of de Lint's discussion of beverage-related effects.  In de Lint's view, many methodological problems lay in the path of drawing a clear bead on beverage-specific cirrhogenic potentials (de Lint, 1977, pp.192-193).   De Lint cited Terris in this connection (among a group of several references) for the assertion that cirrhosis mortality's "diagnostic and recording procedures" vary considerably.  De Lint concluded -- quite rightly -- that at least some of these obstacles could be surmounted by turning to cross-temporal rather than cross-spacial studies for the production of more meaningful results, citing Terris (1967) and another paper (de Lint and Schmidt, 1976) presumably as good examples of this favorable trend.  De Lint's next sentence read:  "These studies [i.e., the two just cited and perhaps similar efforts as well] thusfar [sic] have not incriminated any class of beverage alcohol specifically in the development of cirrhosis, but rather have suggested that whatever increases in alcohol consumption occur -- whether the results of more beer, wine or distilled spirits consumption -- rates of death from this disease increase as well" (de Lint, 1977, p. 193).  With apologies to de Lint, there could hardly be a better example of the literature's remarkable spot-blindness to the beverage specific aspect of Terris's paper -- here dramatically evidenced by the citation of Terris on behalf of the absence of credible beverage specific epidemiologic evidence!

Terris's beverage-specific inference was not, however, universally overlooked.  "A number of researchers," wrote Smith and Burvill (1985), citing Terris (1967) along with Lieber (1982), Schmidt and Bronetto (1962), Leavy (1970), and Moeschlin and Righetti (1970), "have reported that the consumption of particular alcoholic beverages was related to liver cirrhosis mortality, although in his influential review papers Lelbach (1974, 1976)2 concluded that the type of beverage was not important" (p. 42).  Similarly, Room (1978 [1970-1971], pp. 279-280), in a discussion of the well-established link between cirrhosis and the proportion of heavy drinkers in the population, wrote: "In a whole series of studies covering many years, analysts – many of them, as is also traditional in the alcohol literature, quite unaware of each others' existence – have shown that per capita consumption, particularly of wine and spirits, tend to vary from year to year in the same population closely with cirrhosis deaths" (emphasis added), citing Terris (1967) among the five sources listed.  Like Schmidt, Room also cited Terris's paper on behalf of the assertion that short lag-time "‘is consistent with the clinical course of the disease...'" (Room, 1978 [1970-1971], p. 280).  Our review of this literature was by no means exhaustive, but these two sources – Smith and Burvill (1985) and Room (1978 [1970-1971] – were the only instances of reference to Terris's beverge-specific inference that we have found.

Terris's subsequent nonparticipation

Milton Terris is currently editor of the Journal of Public Health Policy.  In response to an email query, Terris kindly wrote that he had "lost interest in doing further research on the etiologic role of alcohol in cirrhosis because [he] was convinced that all of the data -- the experimental laboratory work, the clinical findings, and the epidemiologic data -- supported the hypothesis" (Terris, 1998).  Moreover, Terris reported that he was unfamiliar with the "U.S. puzzle" that had overtaken the alcohol-cirrhosis argument in more recent years.  In short, Terris had offered no reminder because he had abandoned his interest in the problem; his 1967 paper had been a one-shot effort.

Terris's (1998) email also provided useful historical context for his paper and its reception.  He wrote that he "was responsible for the basic course in epidemiology in the first Graduate Summer Session in Epidemiology held in Madison, Wisconsin in 1965."  "During my lectures on 'Descriptive Epidemiology: Time, Place, Person,'" Terris continued,

I presented the Jolliffe and Jellinek findings and their conclusion that there is a significant association of cirrhosis mortality with alcohol consumption.  One of the students, an Assistant or Associate Professor of Medicine at the University of Miami School of Medicine, suggested an alternative explanation, namely, that the post-World War II increase in cirrhosis mortality was the result of the increased use of blood and plasma which caused hepatitis and therefore cirrhosis. When I got back to New York City, I decided to look into this further, and I spent a good deal of time at the New York Public Library to obtain the alcohol consumption data and the history of alcohol policy in the U.K.  (Terris, 1998)
Terris further conveyed that he had been unaware of the critical commentary surrounding his paper's presentation, as reported in Robin Room's (1984) essay.  Indeed, he asked us to send him copies of the literature we had noted to him -- including Herd's (1992) historical article and Elinson's (1967) critical comment published in AJPH before Terris's paper had yet seen publication itself.   "I never saw any of these articles," wrote Terris, "or if I did see the Elinson [1967] AJPH paper, I must have paid very little attention to it."   He continued, "I knew of no controversy or difficulty involved in the publication of my paper, and clearly I either missed or paid no attention to the fact that Elinson's critique predated mine" (Terris, 1998).

In sum, Terris's marginality to the alcohol field and its epidemiological literature, the one-shot character of his celebrated paper, his assumption that the issue had been adequately addressed with the data and analysis he'd presented, and, finally, his tending to other interests in the years that followed had effectively insulated him from the post-1973 emergence of the U.S. puzzle and the commentary surrounding it.  Even a passing awareness of the commotion that had greeted his 1966 presentation of the paper might have prompted Terris to make occasional checks on the progress of the case his paper had made -- but even that element, it seems, was missing from the personal and professional equation.  Terris, himself, was a member of the regular Public Health establishment -- a tenured and distinguished professor in the field and one-time president of the American Public Health Association -- and not a regular in the alcohol studies collegium, with its annual meeting, invisible college, and intrafield communications.  Terris was thus effectively removed from the field-reminding role that he might otherwise have played.

Presentational or substantive sources of overlooking

Terris's (1967) beverage-specific inference may, of course, have been subsequently overlooked because it was insufficiently stressed in his paper.  In fact, however, this was not the case; and, on the contrary, the beverage-specific aspect was strongly stressed.  Beverage-specificity appears first in the three-sentence abstract atop the title -- the middle sentence of which reads, "From the evidence the conclusion is offered that mortality from cirrhosis is directly related to per capita consumption of alcohol from wine and spirits" (p. 2076, emphasis added).  It appears next in the text proper, in the lead-off sentence of the section of the paper titled "Alcohol Consumption" and aimed at marshaling the cross-national evidence on behalf of the alcohol-cirrhosis connection.  Terris wrote:

Table 12 shows that the differences in cirrhosis mortality in the United States, Canada, and the United Kingdom are associated with differences in the apparent consumption of alcohol from spirits and wine.  No such association exists for beer; the apparent consumption of alcohol from beer is similar in the three countries.  (Terris, 1967, pp. 2083-2084, emphasis added)
It appears in the first sentence of the section titled, "Prevention," where Terris also lauds the U.K. for the wisdom of a tax policy succeeding in weaning the nation away from distilled spirits and in the direction of beer.  It appears again in the first paragraph of the paper's brief "Summary" section:  "The evidence strongly supports the conclusion that cirrhosis mortality is directly related to per capita consumption of alcohol from spirits and wine."  Finally, the legends to Terris's Figure 3, 4, & 5, showing the trend relationships between consumption and cirrhosis in the U.S., Canada, and the U.K., all explicate that "absolute alcohol from spirits and wine" is depicted.  Clearly, obscurity within Terris's text was not the overlooking's source.

Other aspects of the paper, however, may have lessened its utility, compromised its authority, or otherwise engendered a dismissive disposition within the alcohol-epidemiological readership.  Three such substantive aspects may be noted. First, almost three-quarters of Terris's narrative was devoted to a systematic critique of Lilienfeld and Korns's 1950 critique of alcohol's place in the epidemiology of cirrhosis, which paper was in turn framed around a critique of Jolliffe and Jellinek's 1942 assessment of cirrhosis's relationship to alcohol or alcoholism.  This "backward-looking" focus, in turn, may have lent a sense of datedness to Terris's text.  By the time the alcohol-cirrhosis link became revivified in alcohol epidemiology – with, say, the publication of Bruun et al. (1975) -- Terris's critique of Lilienfeld and Korns' (1950) environmental and occupational approach may have seemed passe and anachronistic to some alcohol-epidemiological readers.  Second, Terris had employed the most basic of statistical methods (cross-tabulations, Pearson's product-moment correlations, and visual inspection of trend-lines) in his analysis.  The alcohol epidemiology field's subsequent tendency toward more complex statistical methods and statistical valorization may also have lent Terris's analysis an outdated quality for some readers.  Third, and finally, Terris's historical framing of his argument may have appeared slightly askew to some readers long familiar with the alcohol epidemiology field.  Terris's narrative suggested that his paper sought in effect to rescue Jolliffe and Jellinek's (1942) correct emphasis on alcohol's cirrhogenic responsibility from the incorrect and revisionist non-alcohol emphasis of Lilienfeld and Korns' (1950) subsequent paper.  Jolliffe and Jellinek (1942), however, took a rather more critical stance toward alcohol's causal responsibility for cirrhosis than Terris's paper suggests.  Hence, Room (1984), Herd (1992), and Katcher's (1993) positioned Jolliffe and Jellinek's (1942) review paper as consistent with the post-Repeal era's de-emphasis on alcohol's causal responsibility for cirrhosis and other bodily illnesses – in Katcher's (1993) phrase, there was a "post-Repeal eclipse in knowledge about the harmful effects of alcohol."  Moreover, Jolliffe and Jellinek (1942) had focused their attention on alcoholism's (not alcohol's) relationship to cirrhosis whereas Terris's analysis stressed per capita consumption's causal relationship.  Jolliffe and Jellinek (1942) had argued (as Lilienfeld and Korns' also recounted) that cirrhosis was correlated with alcoholism but not necessarily causally related to it – once again, deviating from the path of Terris's paper's rhetorical trajectory.  How much such factors may have diminished the citation worthiness of Terris's paper is of course impossible to gauge, though we suspect such impact was minimal at most.

Paradigmatic aspects of Terris's overlooking

We turn, finally, to sketching a paradigmatic perspective on the overlooking of the beverage-specific aspect of Terris's 1967 paper.  Cirrhosis represented a focus of scientific interest in two of the three major alcohol-problems paradigms vying for attention and use in the 1970s -- i.e., in the modern alcoholism paradigm and the alcohol controls paradigm though not in the cultural integration paradigm.3  The scientific significance of cirrhosis in the alcoholism and alcohol controls paradigms differed.  In the alcoholism paradigm, cirrhosis per se harbored relatively little interest save as an indicator phenomenon useful for estimating the prevalence of alcoholism – via the famous and controversial Jellinek Formula (Argeriou, 1974; Roizen and Milkes, 1980).  In the alcohol controls paradigm, on the other hand, cirrhosis – and particularly cirrhosis mortality trends -- occupied the more important position of the paradigm's initial primary scientific explicandum, i.e., the chief phenomenon to be explained by the paradigm.

Cirrhosis, per se, was not seen as an important research focus within the alcoholism paradigm for three main reasons:  (1) cirrhosis was not seen as caused by either alcohol or alcoholism (though it was regarded seen as correlationally associated with alcoholism) (Jolliffe and Jellinek, 1942), (2) cirrhosis mortality, in general, and cirrhosis mortality "with mention of alcoholism," in particular, were not regarded as big cause-of-death categories in terms of annual mortality rates, and (3) alcoholism itself provided the era's main problem focus.  The alcoholism paradigm's hegemony in the U.S. spilled over into the Canadian research establishment in the 1950s and 1960s, a fact broadly indicated by ARF's original name:  the Alcoholism Research Foundation.4  When John Seeley (1960) undertook to write his seminal paper on the history of the Canadian alcohol controls paradigm perspective – a paper showing a close association between real price and cirrhosis (via real price's putative impact on consumption) -- he faced the problem of the relative insignificance accorded cirrhosis, per se, in the still-dominant alcoholism paradigm perspective.

In the abstract at least, two main rhetorical options were available to Seeley as he composed his paper: (1) he could confine his discussion to cirrhosis, per se – in effect addressing his analysis to a relatively insignificant scientific problem in terms of the prevailing alcoholism-paradigm sensibility or (2) he could explicitly link cirrhosis to the prevalence of alcoholism – in effect "borrowing" the widely accorded significance of alcoholism in order to lend his analysis additional scientific and policy significance.  Seeley (1960) took the first option, a choice clearly reflected in the way he framed his introduction:

Any condition that causes death may well be of interest to physicians, no matter how relatively rare the prevalence.  More particularly might they be interested if, simultaneously, prevalence  were rising while measures, perhaps quite simple, to reduce these death rates appeared to be available.  (Seeley, 1960, p. 1361)
As this text reveals, Seeley finessed the significance issue by arguing that though a problem may not be large or pressing in its own right, it might nevertheless get larger and moreover a simple preventive measure (increased taxation/price) might obviate that prospect.  Several plausible reasons may have been involved in Seeley's rejection of the second option.  By 1960, Seeley had already become critical of the putative link between cirrhosis mortality and alcoholism's prevalence; indeed, he had published in 1959 undoubtedly the deepest and most devastating critique of the Jellinek Formula (Seeley, 1959). Moreover, Seeley may already have been formulating the conceptual ingredients for his equally critical and brilliant assessment of the disease concept of alcoholism (Seeley, 1962), the conceptual core of the alcoholism paradigm.  Seeley may also have sought his price-cirrhosis analysis to launch a distinctively different and new approach to alcohol-problems science, one independent of the conceptual and policy commitments of the alcoholism paradigm.  Though Seeley mentioned the cirrhosis-alcoholism connection in the first page of his text, he did so with well-measured reservation, leaving the question of the degree of association between cirrhosis mortality and alcoholism's prevalence to future research.5  Finally, Seeley may have wished simply to confine his analysis and its rhetorical framing to the objects to which he had paid his scientific attention:  cirrhosis mortality, real price, and per capita alcohol consumption.

Whether or not Seeley actually contemplated the second option in composing his paper, this option posed a number of dauntingly problematic prospects.   Whereas Seeley's analysis of real price, consumption, and cirrhosis mortality relatively avoided a direct challenge to the alcoholism paradigm, the addition of the cirrhosis-alcoholism connection to his text would have implied a direct assault on one of the key commitments of the disease concept of alcoholism – which concept, of course, included a heavy presumption that the alcoholic was very strongly wedded to his drinking indeed and therefore hardly a fit candidate for mere price changes to alter his or her drinking behavior.  By stopping short of stressing the cirrhosis-alcoholism connection, Seeley in effect eschewed a much more difficult and weighty rhetorical task.

Additional problematics attended making use of the borrowed significance offered by linking cirrhosis to alcoholism's prevalence.  The deepest of these ultimately derived from the fact that the two paradigms, alcohol controls and alcoholism, took strikingly different perspectives on alcohol consumption at both the individual and aggregate levels.  Fundamental to the alcoholism paradigm's perspective was the conviction that society's alcohol-related problems lay with "the alcoholic" and not with "alcohol" or "alcohol consumption" per se.  Indeed, the alcoholism perspective had provided what students of the early history of American alcohol science regard as an important scientific and cultural escape-hatch from the longstanding – but by then very unpopular – focus of Dry attention on alcohol, per se (Room, 1978; Roizen, 1991, forthcoming;).  The post-Seeley alcohol controls paradigm perspective, on the other hand, positioned alcohol consumption as the key factor from which sprang cirrhosis, alcoholism, and (in due course) a still larger gamut of alcohol-related problems.  The selective borrowing by alcohol controls paradigm advocates of the alcoholism concept from the alcoholism paradigm perspective could be expected, therefore, to  harbor deep, if submerged, cross-paradigm strains or contradictions.

Eventually, however, a number of Seeley's colleagues at ARF would lay claim to the cirrhosis-alcoholism connection, thereby enhancing the scientific significance of the alcohol controls paradigm's focus on cirrhosis.  De Lint and Schmidt (1971), for example,  directly linked the emergent alcohol controls paradigm's focus on alcohol consumption to cirrhosis in a paper titled, "Consumption Averages and Alcoholism Prevalence."  This paper's first paragraph set the conceptual framework as follows:

In the epidemiology of alcoholism two distinct lines of research are clearly evident.  In one series of investigations the overall level of alcohol consumption in a population is considered to be of crucial importance.  Accordingly, attention is focused on the precise nature of  the relationship between consumption averages and alcoholism prevalence, as well as on the various socio-cultural factors that may explain variation in per capica consumption.  In many other epidemiological studies the etiological significance of the overall level of alcohol consumption in a population is largely ignored. (De Lint and Schmidt, 1971, p. 97, emphasis added)
Clearly, per capita alcohol consumption and alcoholism's prevalence are directly linked.

In a similar vein, Popham et al. (1978 [1970-1971]) introduced their essay on "Government Control Measures to Prevent Hazardous Drinking" by arguing that cirrhosis mortality was associated with a wide variety of alcohol-related problems.  Numerous previous analyses, they wrote, had satisfied them that "despite certain shortcomings, several of the statistics could be employed as valid indicators of the magnitude of alcohol problems in an area.  Most particularly, reported liver cirrhosis mortality proved to vary in close association with variations in the prevalence of alcoholism. (Popham et al., 1978 [1970-1971], p. 240).  "This association [i.e., between cirrhosis and alcoholism]," the authors continued, "was assumed by Jellinek in order to develop an alcoholism prevalence estimation formula; it has been verified since through case-finding surveys and other methods in Ontario and elsewhere" (loc. cit.).  Unlike Seeley (1959, 1960), therefore, Popham et al. (1978 [1970-1971]) elected to shore-up  rather than criticize or pass over the Jellinek Formula and its potential rhetorical benefits for  the alcohol controls paradigm.

The rhetorical circumstance thus had the odd effect of prompting Popham (1970), an alcohol controls paradigm advocate, to defend the beleaguered Jellinek Formula fully a decade after that paradigm had been effectively demolished by critiques offered in the Quarterly Journal of Studies on Alcohol by Seeley (1959) and Brenner (1959), which prompted Jellinek (1959) himself to call for the formula's retirement from service.  Because the cirrhosis-to-alcoholism link had rhetorical value for the alcohol controls paradigm, a reasonably sound Jellinek Formula was more useful than a discredited and abandoned one.  Unlike Seeley (1960), too, Popham et al. (1978 [1970-1971]) braved to throw their pro-alcohol controls paradigm argument directly into the teeth of some of the alcoholism paradigm's key conceptual commitments.  For example, regarding the specially intractable character of alcoholic drinking behavior, Popham et al. (1978 [1970-1971], p. 264) wrote, "...we are not aware of any compelling evidence that there is a unique predisposing factor or an irreversible change due to chronic intake, which renders the individual permanently incapable of controlling his alcohol consumption."  Gone, in short, was  Seeley's (1960) apparent reluctance to shoulder the added rhetorical burdens of the cirrhosis-alcoholism link, and in its place was a new and vigorous willingness to take these up.

But the 1970s were also a time of troubles for the alcoholism paradigm, a circumstance perhaps especially reflected in the emergence of a fractious mid-decade controversy over the possibility of controlled-drinking among diagnosed alcoholics (Roizen, 1987).  The alcoholism paradigm perspective's decline in scientific value doubtless also in due course reduced the value of rhetorical borrowings from its corpus for alcohol controls paradigm advocates.  By 1980, moreover, Bruun (1980, p. 366) -- in his comment on the significance of Skog's (1980a) treatment of lagged effects of consumption change on cirrhosis mortality trends – had pronounced the Jellinek Formula and its "nebulous logic" as obsolete and unnecessary.

The alcohol controls paradigm was not however thereby wholly deprived of a link between cirrhosis mortality and a wider orbit of alcohol-related problems.  That linkage, once supplied by the alcoholism paradigm's Jellinek Formula, could also be supplied by Ledermann's (1956) single-distribution or log-normal model of popular alcohol consumption.  Ledermann's model suggested a determinate relationship between mean per capita total alcohol consumption and the proportion of heavy drinkers in the population.  The conceptual linkage, in turn, allowed the alcohol controls paradigm's advocates to in effect substitute "heavy drinking" – which via Ledermann's approach was determined by mean per capita consumption–in place of "alcoholism" as the mediating concept connecting consumption with a wide orbit of alcohol-related problems.  Beginning with the publication of Bruun et al. (1975), incorporation of the Ledermann's hypothesis could be regarded as effectively ridding the alcohol controls paradigm of its previous reliance on the alcoholism concept and Jellinek's alcoholism prevalence estimation formula.  Greater paradigmatic integrity and consistency could thereby be achieved for the alcohol controls paradigm.

We have outlined the above picture of the displacing, liberating, and consistency-achieving picture of the Ledermann model's service to the alcohol controls paradigm because it suggests a conceptual scenario for interpreting the overlooking of Terris's (1967) beverage-specific inference in the subsequent epidemiological literature.  That scenario relies on two aspects of the emergent unfolding paradigmatic situation:  first, that the Ledermann model could provide the valuable service of linking per capita alcohol consumption and a wide array of alcohol-related problems, and, second, that the Ledermann model's focus on total per capita alcohol consumption may have disinclined alcohol controls paradigm advocates from beverage-specific lines of investigation because the Ledermann model's value to that paradigmatic perspective was substantial, and a beverage-specific perspective would vitiate the Ledermann model's authority and utility.  Therefore, the specifically beverage-specific aspect of Terris's (1967) analysis may have been regarded as diversionary or even counterproductive to the extent that beverage-specific investigations would in effect undermine the welcome utility that the Ledermann model was perceived to have granted the alcohol controls paradigm.  In short, the added value imported by the incorporation of the Ledermann model to fill the functional role once filled by the alcoholism paradigm and the Jellinek Formula was sufficiently great to alcohol controls paradigm advocates that the Ledermann models focus on total per capita consumption was adopted as part of the conceptual package.6


There are, of course, additional explanatory possibilities regarding Terris's overlooked beverage-specific inference.  For one, the puzzle posed by the disparate trends in cirrhosis mortality and per capita alcohol consumption after 1973 revealed itself slowly and remembrance of Terris's (1967) article may have simply faded apace.  Nevertheless, this problem in scientific forgetfulness or oversight appears to us to harbor at least a few intriguing implications and lessons.  As we saw, the historical and utilitarian significance of Terris's celebrated 1967 paper may have been quite different.  Despite the paper's historical significance in subsequent accounts by Room, Herd, and Katcher, Terris's marginality to the unfolding story of the epidemiological investigation of the link between alcohol consumption and cirrhosis mortality may have been suggested in both (1) Terris partial garbling of the relationship between his own paper's thesis and Jolliffe and Jellinek's (1942) argument and in (2) Terris's subsequent loss of interest in epidemiological research on the alcohol-cirrhosis link.  John Seeley may or may not have contemplated the two options we laid out in constructing the introduction to his 1960 paper.  Whether or not he did, our analysis suggests that two quite different dispositions to the scientific significance of cirrhosis mortality trend studies were implied in his choice of a modest direction.  Conflict and the changing relationship between the alcohol controls and modern alcoholism paradigms, we have argued, merit a significant place in the interpretation of Terris's overlooking.  All or which suggests the value of keeping a keen eye on the history and sociology of science aspects of alcohol epidemiological studies through time.


1This tendency was by no means confined to the U.S. alone.  For example, French students of the relationship, Masse et al. (1976, p. 41), wrote:  "It is interesting to note the development of concepts regarding cirrhosis of alcoholic origin.  At the end of the Second World War the shortage of proteins in many countries drew the attention of research workers and physicians to the nutritional origin of a number of diseases.  Some experimental studies convinced medical opinion that there was an important link between animal protein deficiency and cirrhosis.  In some countries the importance of alcohol in the origin of cirrhosis of the liver was actually denied; however, medical opinion has hanged during the last 15 years and the role of alcohol has gradually won fresh recognition."

2Here, as elsewhere in the alcohol-cirrhosis epidemiological literature, Lelbach's (1974, 1976) impressive and influential clinical reviews may have acted as roadblocks to investigation of beverage-specific aspects the alcohol-cirrhosis relationship at the epidemiological level.

3The 1970s saw the heyday of paradigm wars in alcohol epidemiology and alcohol studies more generally.  The modern alcoholism paradigm had been largely responsible for ushering alcohol science into existence in the U.S., and -- though it was already suffering a variety of problematics -- still enjoyed considerable dominance in the field.  In the early 1970s, the cultural integration paradigm experienced a sudden limelight in the U.S. via its embracement by Morris Chafetz, the first director of the still-new U.S. National Institute on Alcohol Abuse and Alcoholism's (NIAAA) (see Room, 1991, p. 320 et seq.).  Finnish alcohol researchers were simultaneously recoiling from what seemed a failed experiment in adopting the liberalizing dictates of the cultural integration paradigm (see Beauchamp, 1981).  Lastly, the alcohol controls paradigm was taking shape and attracting increasing interest, especially at Ontario's Addiction Research Foundation (ARF), becoming associated with a distinctively "Canadian" viewpoint and research preoccupation (see Room, 1996).  The names Schmidt, Smart, Popham, Seeley, de Lint -- all out of ARF -- became familiar in the alcohol controls paradigm sub-literature, though the paradigm also had strong ties to France and to Scandinavian alcohol research and policy.

4Smart (1998, p. 11) recently recalled of his first encounter with Kettil Bruun (late 1950s or early 1960s?): "Kettil thought our [i.e., ARF's] research group was ‘too Jellinek-centred' and we paid too much attention to his ideas.  He was probably right."

5Seeley (1960, p. 1361) wrote of the cirrhosis-alcoholism-prevalence link:  "Some forms of liver cirrhosis, and more particularly some cirrhosis of sufficient severity to be a cause of death, has long been widely spoken of as a ‘complication of alcoholism' [incidentally, citing Jolliffe and Jellinek, 1942].  Indeed so close has the relation been held to be that liver cirrhosis death rates have provided the basis upon which nearly all alcoholism prevalence rates have been estimated even though it is not known what proportion, within very wide limits, of such deaths are due to or associated with alcoholism [three citations, including Seeley, 1959].

"Given the fact, however, of a strong association between liver cirrhosis deaths and ‘alcoholism prevalence', we might well ask how close the association is between the death rate from cirrhosis and the consumption of beverage alcohol.

"If, moreover, that association should prove to be close and positive, then an interest in economics or public health will prompt us to enquire further as to the dependence of alcohol consumption on the price of alcohol.  It is into these two aspects that this paper is, more narrowly, to enquire."

6Our suggested picture of the Ledermann model's significance in relation to the liberation of the alcohol controls paradigm from the alcoholism paradigm may shed light upon and draw some support from a related history-of-alcohol-science puzzle.  The Ledermann model, per se, has been the subject of numerous and in some cases withering criticism from both outside and inside the alcohol controls policy camp (Miller and Agnew, 1974; Parker and Harman, 1974; Duffy, 1980, 1982, 1986; Duffy and Cohen, 1978) – even from such leading pro-controls figures as Ole-Jorgen Skog (1980b) and Robin Room (1978 [1970-1971]  pp. 276-280).  Such criticism suggests, of course, that the Ledermann model harbors its own considerable complement of pitfalls and problematics.  But the Ledermann element as incorporated into the larger alcohol controls paradigm perspective has shown remarkable viability and gained a seemingly integral role in the pro-alcohol controls paradigm rhetorical armamentarium – from Bruun et al. (1975) to Edwards et al. (1994) and the World Health Organization's working group's summary statement (Rehm et al., 1996).  The Ledermann model's remarkable endurance within the story of the alcohol controls paradigm, despite its record of important criticisms, may afford yet another indicator that the function served by its proffered linkage between total per capita consumption and the proportion of heavy drinkers is too valuable to allow the Ledermann hypothesis to be allowed to fall out of the control paradigm's structure.  The mindset that perceives this high utility for the Ledermann model, we suggest, has its origins in the previous reliance on the alcoholism paradigm and the Jellinek Formula as the significance-granting link between cirrhosis and a greater orbit of alcohol-related problems.  Hence, and by extension, the remarkable endurance of the Ledermann model within this paradigmatic camp may further evidence the importance of its preoccupation with total per capita alcohol consumption and therefore the relative indifference to the beverage-specific aspect of Terris's (1967) analysis.


Argeriou, M., "The Jellinek Estimation Formula Revisited," Quarterly Journal of Studies on Alcohol 35:1053-1057, 1974.

Battig, K., "Ausmass und Ursachen des Alkoholismus," Zeitschrift fur Praventivmedizin 9:133-142, 1964 (cited in Schmidt, 1975).

Beauchamp, D.E., "The Paradox of Alcohol Policy: The Case of the 1969 Alcohol Act in Finland," pp. 225-254 in Moore, M., and Gerstein, D. (eds.), Alcohol and Public Policy: Beyond the Shadow of Prohibition, Washington, DC: National Academy Press, 1981.

Brenner, B. Estimating the prevalence of alcoholism; toward a modification of the Jellinek formula. Quarterly Journal of Studies on Alcohol 20: 255-260, 1959.

Bruun, K., Edwards, G., Lumio, M., Makela, K., Pan, L., Popham, P., Room, R., Schmidt, W., Skog, O.-J., Sulkunen, P., and Osterberg, E., Alcohol Control Policies in Public Health Perspective, Helsinki: The Finnish Foundation for Alcohol Studies, 25, 1975.

Cook, P.J., "The effect of liquor taxes on drinking, cirrhosis, and auto accidents," pp. 255-285 in Moore, M., and Gerstein, D.(eds.), Alcohol and Public Policy: Beyond the Shadow of Prohibition, Washington, DC: National Academy Press, 1981.

De Lint, J., "Critical examination of data Bearing on the type of alcoholic beverage consumed in relation to health and other effects," British Journal of Addiction 72:189-197, 1977.

De Lint, J., and Schmidt, W., "Consumption averages and alcoholism prevalence: a brief review of epidemiological investigations," British Journal of Addiction 66:97-107, 1971.

Duffy, J.C., "The association between per capita consumption of alcohol and the proportion of excessive consumers – a reply to Skog," British Journal of Addiction 75:147-151, 1980.

Duffy, J.C., "Fallacy of the distribution of alcohol consumption," Psychological Reports 50:125-126, 1982.

Duffy, J.C., "The distribution of alcohol consumption–30 years on," British Journal of Addiction 81:735-741, 1986.

Duffy, J.C., and Cohen, G.R., "Total alcohol consumption and excessive drinking," British Journal of Addiction 73:259-264, 1978.

Edwards, G., Anderson, P., Babor, T.F., Casswell, S., Ferrence, R., Giesbrecht, N. et al., Alcohol Policy and the Public Good, Oxford, New York, Tokyo: Oxford University Press, 1994.

Elinson, J., "Epidemiological studies and control programs in alcoholism: Discussion," American Journal of Public Health 57:991-996, 1967.

Grant, B.F., Noble, J., and Malin, H., "Decline in liver cirrhosis mortality and components of change: United States, 1973-1983," Alcohol Health & Research World 10:66-69, 1986.

Gruenewald, P.J., and Ponicki, W.R., "The relationship of alcohol sales to cirrhosis mortality," Journal of Studies on Alcohol 56:635-641, 1995.

Herd, D., "Ideology, history and changing models of liver cirrhosis epidemiology," British Journal of Addiction 87:1113-1126, 1992.

Jellinek, E.M., "Estimating the prevalence of alcoholism: modified values in the Jellinek Formula and an alternative approach," Quarterly Journal of Studies on Alcohol 20: 261-269, 1959.

Jolliffe, N., and Jellinek, E.M., "Cirrhosis of the liver," in Jellinek, E.M. (ed.), Alcohol Addiction and Chronic Alcoholism, ch. VI, New Haven: Yale University Press, 1942.

Katcher, B., "The Post-Repeal Eclipse in Knowledge About the Harmful Effects of Alcohol," Addiction  88:729-744, 1993.

Leevy, C.M. in: Engel, A. and Larsson, T. (eds.), Alcoholic Cirrhosis and other Toxic Hepatopathias, Stockholm: Skandia, 1970 (cited as "Leavy" in Smith and Burvill [1985]).

Ledermann, S., Alcool, Alcoolisme, Alcoolisation, Vol. 1, Paris: Presses Universitaires de France, 1956.

Lelbach, W.K., "Organic Pathology Related to Volume and Pattern of Alcohol Use," in Gibbens, R.S. et al. (eds.), Research Advances in Alcohol and Drug Problems, New York: Wiley and Sons, Inc., 1974 (cited in Schmidt [1975] and Smith and Burvill [1985]).

Lelbach, W.K., in Popper, H., and Schaffner, F. (eds.), Progress in Liver Diseases, Vol. 5, New York: Grune and Stratton, 1976 (cited in Smith and Burvill, 1985).

Lieber, C.S., Medical Disorders of Alcoholism, Philadelphia: W.B. Saunders, 1982 (cited in Smith and Burvill [1985]).

Lilienfeld, A.M. and Korns, R.F., "Some epidemiological aspects of cirrhosis of the liver: a study of mortality statistics," American Journal of Hygiene 52:65-81, 1950.

Longnecker, M.P., Wolz, M., and Parker, D.A., "Ethnicity, Distilled Spirits Consumption and Mortality in Pennsylvania," Journal of Studies on Alcohol 42:791-796, 1981.

Makela, K. "Reports from Research Centers–9: The Finnish Foundation for Alcohol Studies and the Social Research Institute of Alcohol Studies," British Journal of Addiction 83:141-148, 1988.

Mann, R.E., Smart, R.G., Anglin, L., and Rush, B.R., "Are decreases in liver cirrhosis rates a result of increased treatment for alcoholism?" British Journal of Addiction 83:683-688, 1988.

Mann, R.E., Smart, R.G., Anglin, L., and Adlaf, E.M., "Reductions in cirrhosis deaths in the United States: associations with per capita consumption and AA membership," Journal of Studies on Alcohol 52:361-365, 1991.

Masse, L., Juillan, J.M., and Chisloup, A., "Trends in mortality from cirrhosis of the liver, 1950-1971" World Health Statistics Report 29:40-67, 1976.

Miller, G.H, and Agnew, N., "The Ledermann model of alcohol consumption," Quarterly Journal of Studies on Alcohol 35:877-898, 1974.

Moeschlin, S., and Righetti, P., "Wine and the Liver," pp. 301-311 in Engel, A. and Larsson, T. (eds.), Alcoholic Cirrhosis and other Toxic Hepatopathias, Stockholm: Skandia, 1970 (cited in Smith and Burvill (1985).

Moore, M., and Gerstein, D. (eds.), Alcohol and Public Policy: Beyond the Shadow of Prohibition, Washington, DC: National Academy Press, 1981.

Parker, D.A., and Harmon, M.S., "The distribution of consumption model of prevention of alcohol problems: a critical assessment," Journal of Studies on Alcohol 39:377-399, 1978.

Peele, S., "R. Brinkley Smithers: The Financier of the Modern Alcoholism Movement," World Wide Web page at

Popham, R.E., "Indirect Methods of Alcoholism Prevalence Estimation: A Critical Evaluation," pp. 294-306 in Popham, R.E. (ed.), Alcohol and Alcoholism, Toronto: University of Toronto Press, 1970.

Popham, R.E., Schmidt, W., and de Lint, J., "Government Control Measures to Prevent Hazardous Drinking," pp. 239-266 in Ewing, J.A., and Rouse, B.A. (eds.), Drinking: Alcohol in American Society–Issues and Current Research, Chicago: Nelson-Hall, 1978 [1970-1971] (probably written in 1970-1971).

Rehm, J., Room, R., and Anderson, P., "WHO  working group on population levels of alcohol consumption," Addiction 91:275-283, 1996 (see Addiction 91:896, 1996, for amended authorship).

Roizen, R., "How Does the Nation's 'Alcohol Problem' Change from Era to Era?  Stalking the Social Logic of Problem-Definition Transformations Since Repeal," in Sarah Tracy and Caroline Acker (eds.), Altering the American Consciousness: Essays on the History of Alcohol and Drug Use in the United States, 1800-1997, University of Massachusetts Press (originally presented at the conference on Historical Perspectives on Alcohol and Drug Use in American Society, 1800-1997, sponsored by the Francis Clark Wood Institute for the History of Medicine of the College of Physicians of Philadelphia, 9-11 May 1997), forthcoming.

Roizen, R., The American Discovery of Alcoholism, 1933-1939, Ph.D. diss., sociology, University of California, Berkeley, 1991.

Roizen, R., "The Great Controlled-Drinking Controversy," pp. 245-279 in Galanter, M. (ed.), Recent Developments in Alcoholism, Vol. 5, New York: Plenum, 1987.

Roizen, R., Kerr, W.C., and Fillmore, K.M., "Cirrhosis mortality and per capita consumption of distilled spirits,  United States, 1949-94: Trend analysis," British Medical Journal 319:666-670, 1999.

Roizen, R., and Milkes, J., "The Strange Case of the Jellinek Formula's Sex Ratio," Journal of Studies on Alcohol 41:682-692, 1980.

Room, R., 1978 [1970-1971], "Evaluating the effects of drinking laws on drinking, pp. 267-289, 414-419 in Ewing, J.A., and Rouse, B.A. (eds.), Drinking: Alcohol in American Society–Issues and Current Research, Chicago: Nelson-Hall, 1978 (written in 1970-1971, see Room [1984, p. 315, ref. no. 77]).

Room, R., Governing Images of Alcohol and Drug Problems: The Structure, Sources and Sequels of Conceptualization of Intractable Problems, Ph.D. diss., sociology, University of California, Berkeley, 1978.

Room, R., "Alcohol control and public health," Annual Review of Public Health 5:293-317, 1984.

Room, R., "Social Science Research and Alcohol Policy Making," pp. 315-339 in Roman, P.M. (ed.), Alcohol: The Development of Sociological Perspectives on Use and Abuse, New Brunswick, NJ: Rutgers Center of Alcohol Studies, 1991.

Room, R, "Introduction," pp. xi -xiv in Smart, R.G. and Ogborne, A.C., Northern Spirits: A Social History of Alcohol in Canada, 2nd ed., Toronto: ARF, 1996.

Schmidt, D.N., "Apparent risk factors for chronic and acute pancreatitis in Stockholm County: spirits but not wine and beer," International Journal of Pancreatitis 8:45-50, 1991.

Schmidt, W., "Agreement, Disagreement in Experimental, Clinical and Epidemiological Evidence on the Etiology of alcoholic Liver Cirrhosis: A Comment," pp. 19-30 in Khanna, J.M. et al. (eds.), Alcoholic Liver Pathology, Toronto: Addiction Research Foundation, 1975.

Schmidt, W., "Cirrhosis and alcohol consumption: an epidemiological perspective," pp. 15-47 in Edwards, G., and Grant, M. (eds.), Alcoholism: New Knowledge and New Responses, Baltimore, London, Tokyo: University Park Press, 1977a.

Schmidt, W., "The epidemiology of cirrhosis of the liver: a statistical analysis of mortality data with special reference to Canada," pp. 1-26 in Fisher, M.M. and Rankin, J.G. (eds.), Alcohol and the Liver, New York and London: Plenum Press, 1977b.

Schmidt, W., and Bronetto, J., "Death from liver cirrhosis and specific alcoholic beverage consumption: an ecological study," American Journal of Public Health 52:1473-1482, 1962.

Skog, O.-J., "Liver cirrhosis epidemiology: some methodological problems," British Journal of Addiction 75:227-243, 1980a.

Skog, O.-J., "Is alcohol consumption lognormally distributed?" British Journal of Addiction 75:169-173, 1980b.

Seeley, J.R., "Estimating the prevalence of alcoholism: a critical analysis of the Jellinek Formula," Quarterly Journal of Studies on Alcohol 20:245-254, 1959.

Seeley, J.R., "Death by liver cirrhosis and the price of beverage alcohol," Canadian Medical Association Journal 83:1361-1366, 1960.

Seeley, J.R., "Alcoholism is a disease: implications for social policy," pp. 586-593 in Pittman, D.J., and Snyder, C.R. (eds.), Society, Culture, and Drinking Patterns, Carbondale and Edwardsville: Southern Illinois University Press (and London and Amsterdam: Feffer & Simons, Inc.), 1962.

Selzer, M.I., Vinokur, A., and Wilson, T.D., "A psychosocial comparison of drunken drivers and alcoholics," Journal of Studies on Alcohol 38:1294-1312, 1977.

Smart, R.G., "The effect of licencing restrictions during 1914-1918 on drunkenness and liver cirrhosis deaths in Britain," British Journal of Addiction 69:109-121, 1974.

Smart, R.G., "International Review Series: Alcohol and Alcohol Problems Research: 4. Canada," British Journal of Addiction 80:255-263, 1985.

Smart, R.G., "Journal Interview–41: Conversation with Reg Smart," Addiction 93:9-15, 1998.

Smart, R.G., and Ogborne, A., Northern Spirits: Drinking in Canada Then and Now, Toronto: Addiction Research Foundation, 1986.

Smith, D.I., and Burvill, P.W., "Epidemiology of liver cirrhosis morbidity and mortality in Western Australia, 1971-82:  Some preliminary findings,"  Drug and Alcohol Dependence, 15, pp. 35-45, 1985.

Star, S.L., and Griesemer, J.R., "Institutional Ecology, 'Translations' and Boundary Objects: Amateurs and Professionals in Berkeley's Museum of Vertebrate Zoology, 1907-1939," Social Studies of Science 19:387-420, 1989.

Terris, M., "Epidemiology of cirrhosis of the liver: national mortality data," American Journal of Public Health 57:2076-2088, 1967.

Terris, M., personal communication (email, Thu, 8 Jan 1998 13:40:45 [EST]), 1998.

Tuyns, A.J., Esteve, J., and Pequignot, G., "Ethanol is Cirrhogenic, Whatever the Beverage," British Journal of Addiction 79:389-393, 1984.

Wallgren, "Alcoholism and alcohol consumption," Alkoholpolitik 23:177-179, 1960 (cited in Schmidt, 1975).