The dynamics of trends in drug use in Australia


The adoption of drug use


Author: David S. BELL,, R.A. CHAMPION,
Pages: 21 to 31
Creation Date: 1977/01/01

The dynamics of trends in drug use in Australia *

F.A.N.Z.C.P., David S. BELL, The Alanbrook Clinic, Mosman 2088, Australia
R.A. CHAMPION, Division of Health Services Research, Health Commission of New South Wales, Sydney 2000, Australia


Drug use tends to follow cycles. Monitoring may be used to establish the trends over time and to analyse their dynamics. An approach is described which establishes the stage of adoption or discontinuance of drug use in the community and a method for the prediction of future trends.

Monitoring in New South Wales from 1971 to 1973 provided data which are analysed to show that alcohol was in the phase of majority adoption and tobacco had reached a plateau, but with the worrying anomaly that heavy use among young people continued to rise. Sedatives and analgesics had also reached a plateau stage of use.

The data on marijuana are subject to more than one interpretation, but in any case point to a continuing rapid increase in use; predictions are made about future trends. Hallucinogens seemed to be in the phase of late adoption and stimulants to have entered a phase that is of particular interest, discontinuance. Narcotics seemed to be in the phase of late majority adoption.


Drug use has been subject to sudden change and cyclical variation throughout recorded history. These changes have not been charted systematically as yet, much less understood. The time course of change has been particularly relevant to illicit drug use, appearing and escalating as rapidly as it has done in recent years. The time dimension is equally important, even if less dramatic, for understanding the forces that bear on the use of the long established drugs such as alcohol and tobacco.

In Australia during the past 150 years alcohol use has passed through cycles from relatively low to high levels of consumption, each of approximately 60 years in duration (Drew, 1975). In the United States two complete cycles may be traced from 1790 to 1925, the first lasting 55 years and the second 80 years (Rorabaugh, 1976), with a third cycle of slowly increasing consumption on its way for the last 45 years. Of course, the increasing consumption of alcohol in the past few decades has been world wide (Sulkunen, 1976). Narcotics use has passed through one cycle and entered another in the United States and similar communities in the past 120 years. In each instance the cycles have occurred in response to forces that have not been clearly identified.

The project was funded by the National Health and Medical Research Council and the Health Commission of New South Wales, Australia. Requests for reprints should be sent to Mr. R.A. Champion, Division of Health Services Research, Health Commission of New South Wales, P.O. Box 4235, Sydney, New South Wales 2000, Australia.

A system designed specifically for monitoring drug use on a state-wide basis produced the three annual surveys in New South Wales from 1971 to 1973 (Bell et al., 1975), from which these data were extracted. The data may be seen as no more than a series of surveys made at regular intervals. Viewed in this light, the latest survey assumes the significance of being that which is most "up to date" and the others pale into the insignificance assigned the obsolete. In fact of course, serial surveys introduce the dimension of time to the static information available from isolated surveys.

In a community where drug use is stable, the forces affecting consumption may be seen as balanced (Cook et al., 1975). The dynamics may be described in terms of Newton's Laws of Motion. A community in which a certain drug is not used at all or in which the drug use is stable may be described in terms derived from his First Law of Motion, namely that every social body "endeavours to persevere in its present state, whether it be of rest, or of moving uniformly", or even more aptly in terms of the Law of Inertia, that is, "an isolated system ... continues to move with constant velocity ... as long as the system remains isolated".

The use of the coca leaf in South America is an example drawn from an isolated social system. The Incas used it for centuries in a restricted and stable manner (or constant velocity) so long as their social system remained undisturbed (or isolated). The Spaniards intruded into and destroyed the system, disrupting the controls and promoting widespread use, which once established continued for centuries at the new rate (or velocity) in the relatively isolated social system.

If a change in one or more forces produces a change in the rate of drug use, a dynamic analysis of the change provides some insight into the character of the forces. Newton's Second Law of Motion applies to this situation, namely that "the rate of change of momentum of a body is proportional to the unbalanced forces acting, and the change takes place in the direction of the force". A change, either of increase or decrease in drug use, indicates the direction of the new "unbalanced force", and the rate of change, a measurement specifically involving the dimension of time, provides some understanding of the magnitude of the "unbalanced force acting" on that social body. The appearance of unbalanced forces may be detected by monitoring.

The adoption of drug use

Trends in drug use are a form of social change that can be analysed in terms of the diffusion of new ideas, their adoption and discontinuance, using the theoretical framework outlined by Rogers and Shoemaker (1971). This sociological model is probably more appropriate than the medical model of a communicable disease proposed by Cameron (1968). Rogers and Shoemaker suggested that three sequential stages occur in the process of social change, namely the invention of a new idea, in this case the production of a drug and the recognition of its effects, the diffusion of the idea, in this case the increase in the use of the drug, and the consequences, which may be manifest or latent and which may include the rejection or discontinuance of the new idea. Within the framework of this approach, the monitoring of the trends in drug use is concerned with the measurement of the diffusion and some of the consequences, including the discontinuance of drug use.

A new idea is introduced by a "change agent" (Rogers and Shoemaker, 1971), who brings knowledge and is usually unlike or heterophilous to the receivers of the idea. An idea is adopted initially by a small section of the population, the "early adopters", who, in comparison to the individuals who take up an idea at a later stage, have more education, higher social status, more exposure to mass media channels of communication, more exposure to interpersonal channels of communication, more change agent contact and more social participation and are more cosmopolitan. In the field of drug use, these characteristics of the early adopters have been spelt out in most detail in relation to marijuana, which typically spreads first among the higher social status group of bohemian arts and university students, the less religious and more politically radical and those who are cosmopolitan and in contact with change agents.

The diffusion follows a time course in which early adoption is relatively slow, gradually accelerating to the stage of rapid spread and then slowing down before reaching a plateau (figure 1). This is the S-curve of adoption over time obtained by plotting the cumulative number of users against time, a curve which is identical to that established for the individual learning to cope with a new situation (Rogers and Shoemaker, 1971). Boolsen (1975) has shown that the increase in drug use over recent years in Denmark has followed this time course.

In the phase of rapid diffusion, interpersonal localite channels of communication are relatively more important than the mass media or cosmopolitan channels. These localite channels are synonymous with the friendship network through which drug use spreads at the stage of most rapid increase (Hughes and Crawford, 1974), the stage described by d'Alarcon (1970) as "explosive" spread. The commencement of this rapid spread depends on the development of a large enough nucleus of "initiators", who would appear to be the equivalent of the early adopters.

FIGURE 1 - The S-curve of adoption of an innovation

Full size image: 594 kB, 
                                        FIGURE 1 - The S-curve of adoption of an innovation

The individuals who take up an idea in the later stage of diffusion are known as "late adopters" (Rogers and Shoemaker, 1971). The late majority adopter category compared to the early majority is below average in social status, makes little use of mass media channels and secures new ideas from peers via interpersonal channels. When the use of a drug has been established in a community for a long period, these characteristics of the late adopter predominate among users and particularly among heavy users; heavy and frequent use of alcohol, tobacco, analgesics and sedatives is more common among those of lower than higher social status (Bell et al., 1975).

The S-shaped curve of adoption may be succeeded by a plateau of stable rates of consumption or by a drop because of discontinuance, that is a decision to cease use of an innovation after previously adopting it. Discontinuance has been largely neglected by the students of diffusion, mainly because their prime interest has been to stimulate the adoption of innovations and the associated social changes. In the case of drug use, social reformers will be concerned more with discontinuance than with adoption. Whereas the rate of discontinuance for innovations in general is high, those with a high rate of adoption, which applies to drug use, have a low rate of discontinuance. A new drug usually adds to the rate of consumption rather than displaces the drug already in use (M?kel?, 1975).


Whole population studies of drug use are clearly not a realistic proposition. To chart trends, surveys of population sub-groups that tend to lead the fashions in the use of drugs or to use excessively seemed to offer the most economical approach. These were termed respectively the "sensitive" and "risk" groups.

The empirical basis for the concept of a "sensitive" population sub-group is that the bulk of initiates to a drug, which is the object of a sudden increase in popularity, are the young, mostly aged between 16 and 25 years. The sample of fourth form school students represents a broad cross-section of 15 to 16 year olds, bearing in mind that almost 25 per cent of state school students drop out before that stage. The stratified random sample of secondary schools throughout the State netted approximately 5 per cent of the fourth form students.

In the age range 17 to 19 years, groups were drawn from random sampling of the sixth form, which is the last year of high school, netting 7 per cent, of the students throughout the state, technical college students, in particular 10 per cent of students in second year of the trade courses and 50 per cent of students in the creative arts, and trainee nurses, in general and psychiatric courses.

Based on the observation that excessive drug use is associated with other forms of deviance, the "risk" groups chosen were delinquents in institutions, probationers and prisoners. The probationers and delinquents were sampled in two of the three years only, 1972 and 1973, netting 8 per cent of the total number of probationers and one-third of the delinquents.

A self-administered multiple choice questionnaire was devised to obtain a wide range of information from the respondent, including attitudes about and details of quantity and frequency of use for alcohol, tobacco, analgesics, sedatives, marijuana, hallucinogens, stimulants and narcotics. Anonymity and confidentiality were stressed in the method used for administering the questionnaire, usually by trained members of the team to groups of 30 to 40 people. Questionnaires were rejected if they lacked essential information on drug use or if the responses showed stereotyped repetition or serious inconsistencies. Less than 8 per cent were rejected, most of these because they were incomplete.

Pilot trials established that the forms were comprehensible by unskilled workers and that the respondents were truthful.



The prevalence of drinking among fourth form students increased from 70 per cent in 1971 to 76 per cent in 1973 and the proportion who used alcohol on most days increased from 3.1 per cent to 5.3 per cent over the same period. In the sixth form samples the prevalence of drinking remained steady but the proportion of daily drinkers increased from 3.4 per cent in 1971 to 5.1 per cent in 1973. The prevalence for general nurses increased from 76 per cent in 1972 to 84 per cent in 1973 and the psychiatric nurses showed the same trend with prevalence figures of 81 per cent and 88 per cent in 1972 and 1973 respectively. Other population groups sampled remained steady throughout this period, all at higher rates of consumption than the groups which showed increases.

Figures from this survey compared, with data from Connell et al. (1975), who sampled the same school population with a similar method, show that school students substantially increased their consumption of alcohol between 1969 and 1973. In 1969, among 15-16 year olds, 28 per cent of males and 12 per cent of females used alcohol weekly. The corresponding figures in 1973 were 39 per cent of males and 25 per cent of females.

Parallel with the increased use of alcohol by school students in New South Wales, the daily alcohol consumption in the State, expressed as grams per person over the age of 18, increased from 40 gm/person in 1969 to 43 gm/person in 1971, 46 gm/person in 1972 and 48 gm/person in 1973 (estimates based on figures supplied in the annual reports of the Licensing Bench and Licences Reduction Board).


The current survey revealed no change in the over-all prevalence of smoking in any group between 1971 and 1973, but considering the prevalence of those who smoked on most days, among the fourth form school students it increased from 15.5 per cent in 1971 to 19.0 per cent in 1973 and from 14.0 per cent to 17.5 per cent in the sixth form sample.

Extending the data base to 1969 shows a considerable increase in smoking among school students in New South Wales. Connell et al. (1974) found in 1969 among the 15-16 year olds, 25 per cent of males and 12 per cent of females used tobacco at least once a week. In 1973 weekly users in the fourth form sample represented 33 per cent of males and 27 per cent of females. A similar increase occurred among the sixth form females; weekly smokers increased from 10 per cent in 1969 to 24 per cent in 1973, almost reaching the male level, which remained steady at 29 per cent over the same period.


The prevalence of marijuana smoking increased in several groups between 1971 and 1973. At the same time, availability, established from the data on the proportion of friends known to use the drug, increased, the perception of danger diminished and attitudes became more permissive.

In the fourth form the prevalence of users increased from 6.5 per cent in 1971 to 9.8 per cent in 1973. Corresponding figures for the sixth form were 7.0 per cent and 13.5 per cent, for trade courses 19.5 per cent and 28.8 per cent, and for the prisoners 16.6 per cent and 36.2 per cent.

Irwin (1975) found similar trends at lower prevalence levels in his fourth and sixth form samples between 1973 and 1974. The fourth form level of use increased from 3.0 per cent to 6.9 per cent and the sixth form prevalence increased from 8.5 per cent to 9.8 per cent.

Figures released by the Bureau of Crime Statistics and Research (1975) show that the number of offenders convicted of offences related to cannabis increased from 494 in 1971 to 1,824 in 1974. Statistics compiled by the Central Crime Intelligence Bureau show a similar nationwide trend with 1,874 offenders involved with cannabis in 1972 and 6,709 offenders in 1974.


Hallucinogens were less popular than was marijuana, and the level of use was static in most groups. The only group to show an increased prevalence over the period of the current survey was the prisoners; in 1971, 10.1 per cent were users with 18.1 per cent in 1973.

In Canberra, Irwin (1975) found a slight upward trend between 1973 and 1974 among the fourth form samples. In the sixth form the reverse obtained with a prevalence level of 3.3 per cent in 1973 and 1.3 per cent in 1974.

The Bureau of Crime Statistics and Research (1975) recorded 110 offenders involved with hallucinogens in 1971, with a decline to 72 offenders in 1974. A similar trend occurred across the nation with 359 offenders in 1972 and 217 offenders in 1974 (Central Crime Intelligence Bureau, 1973 and 1975).


This category of drugs was the only one to show a widespread decline in prevalence. In the fourth form samples prevalence declined from 9.1 per cent in 1971 to 6.5 per cent in 1973. The prevalence in the sixth form samples fell from 8.1 per cent to 5.1 per cent and among the general nurses use of stimulants declined from 10.1 per cent to 5.4 per cent.

The prisoners were an exception to the general trend; their prevalence of use increased from 13 per cent in 1971 to 22.9 per cent in 1973.

In Canberra the prevalence of stimulant consumption in 1973 and 1974 remained steady near 2.2 per cent in the fourth form and declined from 3.5 per cent to 2.4 per cent in the sixth form (Irwin 1975).

Crime statistics from New South Wales and from the Central C.I.B. reflected the decline in prevalence of stimulant use. In New South Wales the number of offenders was 66 in 1971 and 42 in 1974, and the C.I.B. figures were 69 offenders in 1972 and 51 offenders in 1974.


Most of the groups in this survey showed a low and stable level of narcotics consumption. The exceptions were the trade courses and prisoners groups in which the use of narcotics increased between 1971 to 1973. In the 1971 trade course sample, 1.5 per cent claimed to be current users and in 1973 this figure increased to 4.7 per cent. Among the prisoner sample the prevalence of use rose from 4.6 per cent in 1971 to 14.9 per cent in 1973.

The New South Wales crime statistics showed no change between 1971 and 1974 in convictions for offences related to narcotics, but the central C.I.B. reported an upward trend with 446 offenders in 1972 and 756 offenders in 1974.


The use of each group of drugs followed trends that were consistent, not only in the monitoring from year to year, but also with data from other surveys in the same community. The trends that have been summarized point to a steady increase in drug use, licit and illicit. Although the use of some drugs was stable or even dropping, this was overshadowed by the increase in consumption of the most widely used substances such as alcohol and marijuana.

The period of three years was too short to establish the time course of diffusion for any of these drugs, although it was sufficient to establish rate and direction of change. The extension in time to five years provided by the data of Connell et al. (1975) for the use of alcohol and tobacco by school children is still not sufficient to define a recognizable part of the curve of adoption. Nevertheless, the stage to which adoption had progressed for each group of drugs can be established by the approach to be discussed.

The time course taken for a drug to be adopted by a community is a composite of the different rates of progress for the many groups of which it is composed. These groups may be categorized along the continuum from early to late adopters and their relative progress used to establish the stage to which adoption, or discontinuance for that matter, of a specific drug has advanced. To take an example, if the early adopters have relatively high rates of use and these are increasing rapidly, diffusion is still at an early stage. At the other end of the process of diffusion, the early and late majority adopters have reached a stable plateau of consumption and only the slowest adopters of all, the late adopters, are still showing an increase in consumption.

Another approach is to analyse the characteristics of the users to be found in a representative cross-section of the population. If the characteristics of the users match those of the early adopters, considerable increase in use may be expected for some time yet. Conversely, when they assume those of the late adopters, as in the situation described by Boolsen (1975) for the use of illicit drugs in Denmark, very little further increase is likely at that stage.

The relatively high social class of the sixth form students indicates that they would tend to be early majority adopters, and also from their relatively lower social class that the trade course students belong to the late majority and the prisoners to the late adopter categories. The fourth form students represent a cross-section of the population aged about 15 years.

Using this approach, the data on alcohol use are relatively simple. The prevalence of use was high for all groups and regular or heavy use increased in all groups. This would place adoption of alcohol in the relatively rapid stage of increase when the majority share in the trend. In terms used by Newton, the data point to the existence of an unbalanced force increasing the velocity (rate) of alcohol use. The upward trend in a State which already has a massive alcohol problem, is indeed a worrying prospect for the future.

The data on tobacco use are difficult to interpret. For schoolchildren the prevalence of use remaind steady while the rate of heavier use increased, even though the figures for prevalence were far from saturation point. Including the data of Connell et al. (1975), which date from 1969, the prevalence had evidently just reached the plateau between 1971 and 1973. Assuming that women had been affected by a differential effect which placed them in the late majority adopter category, the data are in keeping with adoption progressing through the late majority stage by 1969 and reaching a plateau (figure 2) by 1971.


Full size image: 1.369 kB, FIGURE 2

Suggested model for the adoption of tobacco, in which females use as much as males, but adopt use at a later stage. At point A in time, the phase of early majority adoption, the prevalence of use by females is less than by males and the rate of increase is more rapid for males. At point B, when the combined population is in the late majority phase of adoption, the females still have a lower prevalence of use, but their use is increasing at the more rapid rate.

Other population groups maintained a relatively stable level of tobacco use, that is, even the late adopters had reached a plateau of use. The increase in heavy use alone would seem to point to the an unbalanced force that increased the rate of use among users only and in only one group, the schoolchildren. A force that has a selective effect on such a vulnerable and significant group surely needs to be identified.

Over the period of monitoring, the impact of restrictions on the prescription of barbiturates was being felt and the general use in the community diminished. On the other hand, their place was taken by other sedatives such as diazepam. The consequence was that, for the group of sedatives in general, over-all consumption altered very little. The relatively stable plateau state for analgesic consumption also corresponded to a relatively constant and heavy consumption in the community at large (Gillies and Skyring, 1972).

Marijuana use increased steadily. Those groups with the highest rates of use had reached a plateau and those with the lowest rates were still showing an upward trend from 1971 to 1973. The groups that had risen rapidly to a plateau were the arts students, psychiatric nurses and delinquents, with the trade students still showing the upward trend. Based on these indicators alone, and bearing in mind that the trade students belong to the late majority adopter category, a slowing down of diffusion in the near future might be predicted.

Unfortunately, the data may be given quite a different interpretation. In the sensitive group, marijuana users had a higher social status than the non-users, suggesting that New South Wales was still at the early majority adoption phase and that use would increase rapidly for some time yet. The time course of the S-shaped curve of adoption may be predicted from the known starting point and from the fact that by 1973 the community had not yet reached halfway to the plateau stage. In New South Wales the diffusion of marijuana use commenced about 1960 (Champion, 1976). By extrapolation the plateau would not be reached before 1990. If these predictions are fulfilled, it will be seen that the early adopters had reached a plateau in about a span of 10 years compared to at least 30 years for the whole community. A more optimistic interpretation would place the significant commencement of diffusion about 1968, the community reaching almost midpoint by 1973, which predicts that the plateau will be reached by 1982.

The rates for marijuana offences also rose in the same period as the survey, but more rapidly than the trends of use revealed by the survey. In terms of absolute figures, the number of detected users of marijuana obviously represent only a minute proportion of users, but the disproportionately rapid increase in convictions possibly reflects an increasingly energetic policing of the drug laws.

For most groups the use of hallucinogens had reached a plateau stage. The fact that the prisoners were the only group still showing an upward trend, suggest that adoption had reached the phase of late adoption. The rates for delinquent youth were higher than that of the prisoners and this presumably is the level to which the prisoners aspire. The chronology of the change which promoted hallucinogen use is now a matter of history, beginning more than a decade before and reaching its most rapid stage of adoption well before this survey began. The data illustrate the stage of late adoption for the community, when the unbalanced force promoting social change is almost balanced by opposing forces and the law of inertia once more is about to prevail.

The rates for hallucinogen offences revealed a decline over the same period, but the numbers were small. The decline may have been spurious for a variety of reasons. Most users of hallucinogens also use other drugs, usually to a greater degree, and in general evidence enabling a conviction is obtained more easily for other drugs, for example, a narcotic.

The data on stimulants revealed a downward trend for most groups, but an upward trend for prisoners. In terms of history, the use of stimulants, like that of hallucinogens began an upward trend with forces that appeared more than a decade before the survey began. In 1969 their legal use was almost entirely prohibited in New South Wales. At much the same time, the perception of their danger increased among users and was reflected in the underground motto, "Speed Kills".

The data of this survey would seem to reflect an overlapping of adoption and discontinuance. The relatively slow moving late adopters, the prisoners, were still responding to the original upward trend and over the same time period the more recent downward trend of discontinuance appeared in those groups that respond most rapidly to change. Trade course students, who belong to an intermediate category for speed of response, appropriately were still at the plateau stage, presumably because this group had reached the plateau ahead of the late adopters and had not yet begun to respond to the more recent force promoting discontinuance of stimulant use.

As has been seen for other drugs, the rate of offences had altered in a direction parallel to the general trend, but the figures were too small to give a valid estimate of the rate of change.

In the case of narcotics, the data reveal a plateau stage for the groups that show the fastest response and an upward trend only for the late majority and late adopters (trade students and prisoners respectively). This suggests that the forces encouraging the use of narcotics were being balanced and reaching an equilibrium with opposing forces some time before the survey began. Nevertheless, the over-all rate of narcotics consumption, the numbers of narcotics users seeking treatment and the numbers of offences would be expected to rise for some time yet. This has in fact occurred; at the end of 1973 a total of 382 patients were on narcotics management programmes and the numbers had continued to increase steadily ever since, being in excess of 600 by the end of 1975. Clearly the eventual balance of the opposing forces may be disrupted in either direction and hopefully when this balance is disturbed the cause will be some social measure to produce the downward trend of discontinuance.


Alarcon, R., de. The spread of heroin abuse in a community. In ABC of Drug Addiction. Bristol: Wright, 77-83, 1970.

Bell, D. S., R. A. Champion and A. J. E. Rowe. Monitoring drug use in New South Wales. 1971 to 1973. Sydney: Health Commission of New South Wales, 1975.

Boolsen, M. W. Drugs in Denmark. Int. J. Addict., 10:503-512, 1975.

Bureau of Crime Statistics and Research. Drug Offences 1974 and Community Comparisons. Sydney: Department of the Attorney General and of Justice, 1975.

Cameron, D. C. Youth and drugs. J. Amer, Med. Assoc., 206:1267-1271, 1968.

Dynamics of trends in drug use in Australia 31

Central Crime Intelligence Bureau. Drug Abuse in Australia. Canberra: Commonwealth Police, Technical Report No. 5, 1973.

Central Crime Intelligence Bureau. Drug Abuse in Australia 1974. Canberra: Australian Police Technical Report No. 7, 1975.

Champion, R .A. Trends in Marihuana Use in New South Wales: 1971 to 1973. Sydney: Health Commission of New South Wales, 1976.

Connell, W.F., R.E. Stroodrant, K.E. Sinclair, R. W. Connell and K.W. Rogers. 12 to 20: Studies of City Youth. Sydney: Hicks Smith & Sons, 1975.

Cook, R.F., R.S. Hostetter and D.A. Ramsay. Patterns of illicit drug use in the army. Amer. J. Psychiat., 132:1013-1017, 1975.

Drew, L.R.H. Drug dependence-A challenge to the A.C.T. Aust. J. Alcoholism Drug Dependence, 2:71-73, 1975.

Gillies, M.A. and A.P. Skyring. The pattern and prevalence of aspirin ingestion as determined by interview of 2,921 inhabitants of Sydney. Med. J. Aust., 1:974-979, 1972.

Hughes, P.H. and G. A. Crawford. Epidemiology of heroin addiction in the 1970s: New opportunities and responsibilities. In: E. Josephson and E.E. Carrol (Eds.). Drug Use, John Wiley and Sons, Inc., 1974.

Irwin, R.P. The Australian National University drug education project. In: H. Pang, G. Scully and P. Stolz (Eds.). Alcohol and Other Drug Dependencies in the Australian Capital Territory, Council of Social Services, Canberra, 1975.

Makela, K. Consumption level and cultural drinking patterns as determinants of alcohol problems. J. Drug Issues, 5:344-357, 1975.

Rogers, E.M. and F.F. Shoemaker. Communication of Innovations. New York: Free Press, 1971.

Sulkunen, P. Production, consumption and recent changes of consumption of alcoholic beverages. Brit. J. Addict., 71:3-11, 1976.

Rorabaugh, W.J. Estimates U.S. alcoholic beverages consumption, 1790-1860. J. Studies Alcohol., 37:357-364, 1976.