Did John Calhoun's 1960s Mouse Utopia really show that animal (and human) populations will expand to arbitrary densities, creating socially-driven pathology and collapse? Reasons for doubt.
2019-08-12–2021-06-11 finished certainty: likely importance: 3
Did John Calhoun’s 1960s Mouse Utopia really show that animal (and human) populations will expand to arbitrary densities, creating socially-driven pathology and collapse? I give reasons for doubt about its replicability, interpretation, and meaningfulness.
One of the most famous experiments in psychology & sociology was John Calhoun’s Mouse Utopia experiments in the 1960s–1970s. In the usual telling, Mouse Utopia created ideal mouse environments in which the mouse population was permitted to increase as much as possible; however, the overcrowding inevitably resulted in extreme levels of physical & social dysfunctionality, and eventually population collapse & even extinction. Looking more closely into it, there are reasons to doubt the replicability of the growth & pathological behavior & collapse of this utopia (“no-place”), and if it does happen, whether it is driven by the social pressures as claimed by Calhoun or by other causal mechanisms at least as consistent with the evidence like disease or mutational meltdown.
What really happened in John B. Calhoun’s behavioral sink “Mouse Utopia” experiments? Mouse Utopia is a legendary experiment in which mice were put in a high-density enclosure (“Universe 25”) with unlimited food, a ‘mouse utopia’—only to see the initial population growth be followed by a population collapse generations later, while the late mouse population exhibited bizarre physical & social abnormalities such as autistic-like behavior & homosexuality & failure to reproduce. Mouse Utopia is interpreted as illustrating the damaging effects of the environment & overcrowding by John B. Calhoun and others. After he published an extremely popular article in Scientific American in 1962 describing the first phase of Mouse Utopia experiments, it became a stock example employed by liberals in application to human populations, particularly for global & urban population growth and any human problem that might be caused by environments, such as the urban decay and riots and spiking crime rates of that era.
Initially, the population grew rapidly, doubling every 55 days. The population reached 620 by day 315, after which the population growth dropped markedly, doubling only every 145 days. The last surviving birth was on day 600, bringing the total population to a mere 2200 mice, even though the experiment setup allowed for as many as 3840 mice in terms of nesting space. This period between day 315 and day 600 saw a breakdown in social structure and in normal social behavior. Among the aberrations in behavior were the following: expulsion of young before weaning was complete, wounding of young, inability of dominant males to maintain the defense of their territory and females, aggressive behavior of females, passivity of non-dominant males with increased attacks on each other which were not defended against.
After day 600, the social breakdown continued and the population declined toward extinction. During this period females ceased to reproduce. Their male counterparts withdrew completely, never engaging in courtship or fighting and only engaging in tasks that were essential to their health. They ate, drank, slept, and groomed themselves—all solitary pursuits. Sleek, healthy coats and an absence of scars characterized these males. They were dubbed “the beautiful ones.” Breeding never resumed and behavior patterns were permanently changed. The conclusions drawn from this experiment were that when all available space is taken and all social roles filled, competition and the stresses experienced by the individuals will result in a total breakdown in complex social behaviors, ultimately resulting in the demise of the population.
Calhoun saw the fate of the population of mice as a metaphor for the potential fate of man. He characterized the social breakdown as a “second death,” with reference to the “second death” mentioned in the Biblical book of Revelation 2:11. His study has been cited by writers such as Bill Perkins as a warning of the dangers of living in an “increasingly crowded and impersonal world.”
If Calhoun had merely found that rat/
So Mouse Utopia quickly became one of the most famous experiments in psychology (and highly influential on not just psychology but sociology, urban planning, American politics, and science fiction, inspiring eg The Rats of NIMH), and continues to be discussed (eg by Down the Rabbit Hole); as “Escaping the Laboratory: The Rodent Experiments of John B. Calhoun & Their Cultural Influence”, Ramsden & Adams 2008/ 2009 put it:
Calhoun published the results of his early experiments with the rats at NIMH in a 1962 edition of Scientific American. That paper, “Population Density and Social Pathology”, went on to be cited upwards of 150 times a year.1 It has since been included as one of “Forty Studies that Changed Psychology,” joining papers by such figures as Freud, Pavlov, Milgram, Rorschach, Skinner, and Watson ([pg249, ch32: “Crowding into the Behavioral Sink”, Forty Studies That Changed Psychology: Explorations Into The History Of Psychological Research] Hock 2004). Like Pavlov’s dogs or Skinner’s pigeons, Calhoun’s rats came to assume a near-iconic status as emblematic animals, exemplary of the ways in which behavioral experimentation at once marks and violates the human-animal distinction. The macabre spectacle of crowded psychopathological rats and the available comparisons with human life in the densely-packed inner cities ensured the experiments were quickly adopted as “scientific evidence” of social decay. Referenced far outside of the fields of ecology and mental health, Calhoun’s rats have—or certainly had—come to seem part of the common cultural stock, shorthand for the problems of urban crowding just as Pavlov’s dogs were for respondent conditioning. Along with their public popularity, the experiments played a critical role in the development of disciplines and research fields, so much so that sociologist and human ecologist Amos Hawley (1972) would remark that the extent of their influence was itself a “curious phenomenon.”
Like any symbol, it has shown adaptability—in 2015, now some reinterpret it to reflect contemporary political debates and explain it not as about crowding & social breakdown but as being about sexism & inequality:
…Today, the experiment remains frightening, but the nature of the fear has changed. A recent study pointed out that Universe 25 was not, if looked at as a whole, too overcrowded.2 Pens, or “apartments” at the very end of each hallway had only one entrance and exit, making them easy to guard.3 This allowed more aggressive territorial males to limit the number mice in that pen, overcrowding the rest of the world, while isolating the few “beautiful ones” who lived there from normal society. Instead of a population problem, one could argue that Universe 25 had a fair distribution problem.
However, there are red flags:
the immediate and long-enduring popularity in liberal politics & pop culture was fed by Calhoun’s own highly-anthropomorphized description of the various kinds of mice, and he fully endorsed the grand applications of Mouse Utopia to contemporary American problems (eventually culminating in an angry NIMH resignation letter in 1986 heavy on references to George Orwell’s 1984).4
One might note that historically, a number of high-profile ideologically-friendly psychology results dating ~1950–1970, often used to justify social engineering policy, have proven to be seriously flawed (even more so than one would expect from the contemporary social psychology replication crisis): the Pygmalion effect, the Robbers Cave experiment, the bystander effect/
Kitty Genovese, the Third Wave, Zimbardo’s Stanford Prison Experiment, the double-bind & refrigerator-mother theories of schizophrenia, Project Nim etc frequently fail to work in the long run or replicate, and involved heavy analytic bias or outright interference by the experimenter to make the experiment ‘work’ and tell the desired story.
animal studies in general often suffer from lower methodological quality than similar human studies: even smaller n, large between-strain genetic differences (in addition to all the between-species differences)5, pseudo-replication from group housing/
relatedness, typically non-blinded ratings, heavy publication bias, etc.
Mouse Utopia is almost completely unpublished. Despite working on it and similar experiments with NIMH funding for decades (he “continued to work on his research results until his death on September 7, 1995”), Calhoun appears to have published almost nothing substantive about his research, limited to a handful of short summary articles or passing references. (Calhoun’s 1963 book The Ecology and Sociology of the Norway Rat describes only his “quarter-acre” experiments which ended in June 1949, before the NIH experiments on overcrowding.) The major citation for his Mouse Utopia experiments is the aforementioned 1962 Scientific American article (published 33 years before his death), which consists of 9 pages of popular writing, of which about half is generic illustrations of mice (rather than data-based figures or plots or tables). Calhoun 1963, “The Social Use of Space” touches briefly on behavioral sinks & mortality in some of the earlier experiments. Calhoun 1971, “Space and the Strategy of Life” presents a brief display of data from “Universe 14” and “Universe 15” but goes into more detail about 25 unspecified universes done to followup the Kessler 1966 thesis and mentions that they are following Universe 25, still waiting for it to actually collapse.6 The article Calhoun 1973, “Death Squared: The Explosive Growth and Demise of a Mouse Population”, begins with an extended analogy to the Book of Revelation, and presents some limited information on Universe 25, which has only halved in population at this point. His 1970 article, for example, is just a redaction of the 1962 one, and despite apparently consulting Calhoun’s manuscripts at the NLM, Ramsden & Adams 2009 shed little light on Calhoun’s research or cite much beyond the 1962 article. (Considering how little he published, it’s surprising that NIMH funded Calhoun until 1983, leading to what Ramsden & Adams 2009 describe as a “forced retirement” in 1986—apparently Calhoun was unable to get funding anywhere else. One wonders why NIMH dropped Calhoun, and why he was unable to get further funding given how high-profile and important his results are.)
This causes considerable confusion in reading since it’s unclear what papers refer to what. The 1962 article describes high infant mortality but not collapse in unidentified ‘universes’, which are not the famous Universe 25, which was started later; Calhoun 1971 had not yet seen a collapse in Universe 25; Marsden 1972 describes a population in slight decline and extrapolates out to possible collapse in a single unspecified universe; while Calhoun 1973 article shows a graph of a universe’s population definitely decreasing and apparently doomed by sterility & aging, which is identified as Universe 25. Are these all the same population, and if not, how many different ‘series’ or ‘universes’ are being described? How many exhibited the ‘senescence’ phase, much less population collapse? (Indeed, how many were done in total?)
it is unclear just how many experiments Calhoun had to run to get the one result which is always talked about; the name “Universe 25” implies at least 24 prior experiments, and Calhoun speaks vaguely of multiple “series” of experiments, referencing earlier experiments with stable populations (unlike Universe 25), some which were apparently controlled to fixed population sizes and some which apparently were not. Nor did all of the overpopulated universes develop the “behavioral sink” phenomenon Calhoun lays so much stress on, which he attributes to an otherwise-unexplained change in the food type. The number of experiments Calhoun ran implies that variance in outcomes was high, and in Kessler 1966, the two experimental group replicates were nevertheless different on many measures.
aside from 2 studies on brain hormones prior to 1973 in “mice selected (by Dr Calhoun, Dr Marsden, and their associates) to represent specific behavioural states existing during the declining crowded populations”, there appear to be no followup or secondary analyses of any kind, so there are no archived biological samples anywhere which could be checked; Calhoun’s 1973 claim that removing mice for analysis would disturb the colony dynamics suggests that few or no samples were kept in the first place
no followup literature: only 2 partial replications have ever been done by third parties that I know of7; likewise, if unique aspects of Calhoun’s experiment like the “beautiful ones” have been reported since, I have not encountered any references to them.8 They do not convincingly support the Universe 25 Mouse Utopia narrative. They are:
Kessler 1966 thesis, “Interplay Between Social Ecology and Physiology, Genetics and Population Dynamics of Mice”: Kessler used 4 strains of mice simultaneously (rather than Calhoun’s use of a single kind of mice), in 2 separate experimental high-density groups (plus a control) and achieved a remarkably high & apparently stable population density. Calhoun 1971‘s summary does not mention any population collapse nor whether there were Calhoun’s pathologies like the ’beautiful ones’. Kessler’s abstract reports that he achieved densities “several times greater” than prior experiments, with stable populations maintained by low pregnancy & high infant mortality rates; while Kessler notes “aberrations of sexual behavior”, the high-density mouse behavior normalized (eg they were able to reproduce) when transplanted to lower-density environments or when environments were connected in an ‘emigration’ experiment’. The two experimental groups showed variance, differing from each other in many ways (“Cohorts in Pop A and B differed with respect to reproduction physiology, mortality, and behavior, and intercohort differences persisted at all levels of population density.”), despite being generated the same way & put into the same kind of environment. Kessler further saw signs of natural selection, as indicated by changes in genetically-influenced coat color proportions (which were consistent in both groups). Kessler sums up as:
The large sizes and unusual degree of crowding attained by the freely growing populations in this study compared with previous studies may be related to the types of animals used, to the number of individuals in the founder nuclei, and to the physical structure of the enclosures. Extreme crowding was compatible with general physical health. The decline of fertility and fecundity, the decreased survival of newborns, and the appearance of behavioral aberrations—rather than disease or an increase in adult mortality—represented the major self-regulatory mechanisms that eventually limited population growth. The growth of individuals was not inhibited. Social withdrawal and the decline of social interaction rather than a rise of interaction characterized the populations. Such findings cast doubt about the generality of the so-called “Stress” theory of social ecology that emphasizes increased interaction and pituitary-adrenal hyperactivity as the principal mechanisms involved in self-regulation of vertebrate populations.
Overall, despite achieving a density far higher and one that would be expected to have a far larger harmful effect, Kessler 1966 only somewhat resembles Calhoun’s results: while Kessler does describe deviant mice behavior driven by density (such as homosexual matings) and high infant mortality/
cannibalism, on the other hand, there are no population crashes or cessation of reproduction but stable populations after initial growth, there are no behavioral sinks, any ‘beautiful ones’ or ‘drinkers’ or ‘autistic’ mice are not described as such by Kessler, the mice are healthy overall, and transplanted mice revert. Further, Kessler’s observation of considerable between-population variance & genetic changes raise questions about statistical power & interpretation of any effects.
Hammock 1971, “Behavioral changes due to overpopulation in mice”: uses a different mouse strain, Swiss Webster, and in the primary experiment, following up a pilot, obtained “a total lack of overpopulation.”
The groups reached a certain population and then maintained it, bouncing back after any culling (and raising questions about Calhoun’s claim that a population which had stopped reproducing after reaching an equilibrium must be doomed). Hammock notes extensive pathology in the pilot similar but not identical to Calhoun’s (eg no ‘beautiful ones’ but instead the pilot mice began to groom only their head), some indication of a population decline during the short duration, and no appearance of harems/
territories/ behavioral sinks. In the main experiment, however, the experimental population quickly reached a low-density equilibrium and no pathologies were observed other than high infant mortality (primarily from cannibalism, maintaining the equilibrium). Hammock notes “No other experiment reviewed had this phenomenon occur. In all other research, the populations first overpopulated then reduced their numbers. This experiment suggests an inborn population control mechanism based upon the density available per mouse…”
other research on animal social dynamics & population density find that there are (of course) relationships between them, and changes in social patterns with density, but nothing like Calhoun’s results of explosive population growth, utter social decay, widespread sterility, uniquely pathological types emerging, and completely collapse/
extinction (eg compare Mouse Utopia with the changes observed in Berman et al 1997 for a colony of rhesus monkeys).
Calhoun fails to consider alternative explanations other than purely-density-based social breakdown:
disease: for example, the sterility noted is also a side-effect of many contagious diseases or parasite load, which are greatly assisted by density in spreading, and density fosters “evolution towards virulence” of existing diseases as diseases can be more lethal to spread faster (while infections in more isolated individuals must be more careful to not kill their hosts before infecting another host). Nothing was done to prevent disease nor to check for its presence, and Calhoun simply denies it could be a factor.9
genetics: the described collapse closely resembles mutational meltdown experiments (which also feature sterility and subsequent population collapse); that is typically demonstrated in asexual organisms by removing all reproductive constraints like resources and so eliminating natural selection as much as possible, allowing the continuous buildup of mutations until finally organisms are no longer even able to reproduce, but should be possible in sexually-reproducing organisms as well.
Meltdown should be much harder to induce in sexual organisms (the recombination theoretically allows much greater selection and is part of the justification for the extremely complex, expensive, error-prone process of sexual reproduction) and it’s unclear if Universe 25 ran enough generations to plausibly generate mutational meltdown, but it will be faster in tiny populations (eg Calhoun mentions that some used 56 rats as a seed but not which strain—many laboratory strains are unhealthy & reproductively unfit to begin with, highly adapted to the lab environment, and highly inbred or even clonal)10. Universe 25 appears to have been begun with just “4 pair of mice”, based on figure 2 in Calhoun 1973. (Based on Marsden 1972, they were probably BALB/c mice; the WP article describes them as inbred & notes that they tend towards anxiety & males towards aggression.) Further increasing inbreeding, Calhoun 1962 describes ‘harem’-like behavior where the dominant male could ensure near-exclusive access to all the female in one subdivision of the cage, dubbed “brood pens”, and force out rival males. Calhoun appears to admit in discussions that they would be highly inbred but denies any possibility of relevant genetic change.11
As well, highly social organisms with complex colony mechanisms, dependent on subtle interactions between members (eg proper use of alarm pheromones and border guarding), where members can inflict a great deal of harm on each other, may be especially sensitive to genetic mutations, as the genes of individual mice affect cage mates (“Genome-wide association study of social genetic effects on 170 phenotypes in laboratory mice”, Baud et al 2018), causing “indirect genetic effects” (IGEs) or “social epistasis”.
Calhoun did not do anything to check or avoid these alternative mechanisms, such as running fostering experiments with the survivors (if the problem is genetic, the offspring of the survivors would, even if fostered into a normal healthy mouse colony, still be unhealthy, while if it’s a contagious disease, introducing a few survivors into a healthy colony should result in noticeable colony-wide damage); Calhoun notes a quasi-fostering experiment in his 1962 paper (8 of the healthiest from one unspecified universe were spared culling, had fewer litters & no surviving offspring), but does not note that this more strongly supports a genetic rather than social dysfunctionality explanation, as the rest of the colony had been removed and could no longer exert any negative effects. Calhoun 1971 describes the Kessler 1966 thesis, “Interplay between social ecology and physiology, genetics and population dynamics of mice” as using 4 different strains (16 pairs total) as founders (increasing total genetic variance greatly) and had “unusual attainment of very high density” without any collapse despite “less than three square inches per mouse”; Calhoun assumes that it is again due to the environment, related purely to social effects stemming from the number of founders (rather than the great increase in genetic variance from using more individuals from more strains), ignoring Kessler’s other finding of natural selection operating on the mouse populations (showing that noticeable genetic change is possible within a single experiment), and in describing his followup experiments to Kessler (with unclear use of strains but almost certainly only 1 strain as Calhoun’s papers seem to typically only use the BALB/c mice, so changes in founder population would not be as effective as in Kessler 1966), and finds little effect from variation in founder size and again does a quasi-fostering experiment where again despite the absence of their toxic environment the surviving mice had only a few pups & are often sterile & unable to even get pregnant by normal mice.
One wonders what Calhoun would have found if the universes had been run with wild-type mice in a fully-sterilized environment, universes followed until actual extinction, and all universes were fully reported.
Overall, Mouse Utopia is a sketchy and unreliable result: it is selectively and scantily reported, it is unclear how often the claimed behavioral sinks or population collapses happen even just within Calhoun’s experiments, whether any such problems are due to exogenously-forced density increases rather than the colonies naturally regulating population density close to their optimum, the few replications replicate only parts of it (if at all), it is entirely possible that it is a fluke of that particular mouse colony or mouse strain, and if the experiment ever was replicated exactly (assuming the unpublished materials are adequately informative), it would be unclear what the actual causal mechanism of the collapse would be as the design & analysis is ambiguous and Calhoun tested no hypotheses (much less the most likely ones of disease or genetics, which he resolutely ignored)12. I am left confused what happened in Mouse Utopia, to what extent it reflects any real natural dynamics involving population growth & density, and extremely doubtful of the perennial attempt to link it to humans.
Calhoun reflects on this in: Calhoun, J. B. C. 1979. “Employee’s contribution to the Performance Assessment of his Scientific Service. [Draft.]” 4 December. John B. Calhoun Papers, National Library of Medicine (NLM), Bethesda, MD. n.p.↩︎
The study alluded to by Inglis-Arkell here appears to actually be the discussion of the behavioral sink in chapter 32 of the Hock 2004 textbook.↩︎
Another example of this interpretation would be Moore 1999, “Population Density, Social Pathology, and Behavioral Ecology”.↩︎
Calhoun appears to have maintained this position up to his death in 1996, according to his NYT obituary: “But his work had its frustrations as well, she [a colleague] noted, because its implications for the future of the human rat race were often met with studied disregard. But Dr. Calhoun was convinced that his mice and rat populations were an accurate model for humans.”He didn’t regard it as hypothesis any more, he regarded it as factual," Mrs. Kerr said."↩︎
Given the disparate results of all these universes, it seems, contrary to the claims of some journalists that Calhoun “had been building utopian environments for rats and mice since the 1940s, with thoroughly consistent results. Heaven always turned into hell.”, many of them did no such thing.↩︎
I’ve been told that a UK university quashed a third Mouse Utopia proposal on, ironically, ‘ethics’ grounds. (The same reason Zimbardo gives for why no one should ever try to replicate his Stanford Prison Experiment, incidentally.)↩︎
One is reminded of Paul Meehl’s observation that in the hard sciences, theories are built on and predictions quantitatively refined or are clearly refuted, while in the more pathological social sciences, theories are never refuted but simply fade away, resembling fads more than facts.↩︎
“Dr Calhoun said that they (the investigators) were not very sanitary in their husbandry, if that was the kind of pollution inferred. The environment was cleaned, most feces and soiled bedding removed, every six weeks or two months, but nothing was ever sterilized. He did not consider this necessary in such a closed system and the mice had better survival than in most laboratory colonies.” In claiming Mouse Utopia mortality rates superior to ‘most’ regular lab mice, Calhoun is presumably excluding the extremely high infant or youth mortality.↩︎
Calhoun 1973: “Dr Calhoun felt that there probably was some mutation. Mice which continually circled, about a dozen, had been noted, but these might have been ‘vestibular’ mice and a result of an infection, not mutation. Even if mutation rates were known, the first generation would have been very much like the last. So the real conclusion was that tremendous behavioural differentiation could occur as a result of social environmental influences even given a high degree of genetic homozygosity.” Marsden 1972, presenting a ‘synthesis’ based on a year with Calhoun, denies any possibility of genetic change: "The genetic potential for exploiting this mouse paradise was qualitatively and quantitatively present in equal proportion in each of the original eight colonizers and would remain essentially unchanged even to the _n_th generation."↩︎
A particularly serious criticism, as I doubt that the urban planners or demographers or Democratic politicians who took an interest in Mouse Utopia would be as interested if the causal mechanism turned out to be “urban densities increase STDs or genetic mutations to the point of collapse”. And if it turned out that Mouse Utopia replicated in mice but never humans (early attempts to correlate population density with social decay in humans apparently did not do well, incidentally), I also doubt if most people citing it, aside from a few zoologists, ethologists, or mouse breeders, would be doing so.↩︎