The Uplift Project: Enhancing and Propagating Intelligence and Longevity

By R. Robert Holson

In this book the author puts forward an agenda to enhance intelligence and longevity in humans, select animal species (including dogs, dolphins and elephants), and machines. This effort would extend over 1,000 years or 40 human generations. Enhancements of IQ and longevity in humans would involve both environmental and genetic improvements in membership IQ and longevity. The goal would be a mean IQ of 145 and an average longevity of 100 years in human Uplift Project members by the end of these 1,000 years. Given that Uplift Project members will probably at project outset have better than average IQ and life expectancy, this could involve as little as a two standard deviations increase in IQ (30 points) and one standard deviation increase in longevity (20 years) over these 1,000 years.The Uplift Project would also expand human, animal and machine membership not only across the planet, but to the moon, Mars, and space colonies.

• Language: English
• Publisher: Xlibris US
• Release date: Mar 4, 2016
• ISBN: 9781514456880

R. Robert Holson

Robert Holson’s life can best be divided into four time periods, involving five countries (USA, England, India, Jordan and Mexico). Born to a professional soldier (Aden Allen Holson) who saw service along the Mexican border in search of Pancho Villa in the early 1900s, then in France in WWI and the Pacific in WWII, Robert was born on an army base in 1941, just prior to this country’s entry into WWII. Robert spent 3rd grade in the UK, where his father was stationed during the Berlin airlift, then lived in various parts of Southern California, where he graduated from High School in Barstow in 1959. He attended college at the University of California at Berkeley, obtaining a bachelor’s degree in Asian history. Robert then joined the Peace Corps and spent two years in India, where he met his wife Andrea, also a Peace Corps volunteer. Thereafter Robert worked for CARE, supplying very large scale school lunch programs and food for work projects in the then Indian States of Bihar, Orissa and West Bengal. Robert’s oldest son Seth was born in Patna, the then capital of Bihar. A second son, Ethan, was born stateside during this period. After 5 years in India, Robert went to Jordan with CARE for almost 3 years. As a consequence of the above, Holson is fluent in written and spoken Hindi, comfortable in scientific German and Spanish, and able to count to ten in Arabic, one of the world’s toughest languages! Robert then left CARE and returned to the US, where he obtained a PhD in psychology from the University of Washington in Seattle. Holson then conducted neuroscience research for some 13 years at the National Center for Scientific Research in Arkansas, followed by almost as many years as professor and researcher in the Department of Psychology at New Mexico Institute of Mining and Technology in Socorro New Mexico, where he retired as departmental chair. During this second career Holson published over 80 scientific articles. After retirement he researched and wrote this book, while dividing his time between grandchildren in Seattle and homes in Socorro and the historic town of Alamos in Mexico.

Book preview

CHAPTER 1

THE UPLIFT PROPOSAL: INCREASING AND PROPAGATING INTELLIGENCE AND LONGEVITY

Humanity has quite unintentionally achieved a high level of intelligence, linked to spoken language. This intellect has given us mastery of the planet but seems inadequate to prevent overpopulation, resource depletion, global warming, and perhaps even extinction. It is the thesis of this book that avoidance of the above may be possible only through still greater intelligence in our own species, in related species, and in machines.

Further, the high intelligence of our species has made possible an equally great longevity, compared to any other mammalian species. We will demonstrate that high longevity is a necessary component of high intelligence. This longevity is accompanied by and, to some degree, enables high intelligence. Consequently, we will propose that increased intelligence will both enable and require increased longevity.

Hence, this work is a proposal for the increase of both intelligence and longevity in our own species, in select companion species, and in machines. We will further propose that this enhanced intellect and longevity be propagated across the planet, the solar system, and ultimately neighboring solar systems around other stars.

To realize these lofty goals, it will be necessary to found an association of humans devoted to these purposes. This work will outline the details of such an association and the means by which intelligence and longevity may be enhanced and propagated thereby. This project will be referred to by the acronym UP for Uplift Project. This terminology is taken from the Uplift science fiction novels of David Brin (the Uplift saga, containing three novels constructed around the uplift of other species; here, we propose a predominant focus on the genetic uplift of our own species, as well as dogs, chimpanzees, Asian elephants and dolphins).

Intelligence and longevity ultimately depend, to the highest degree, upon the genes that make these possible. Environmental influences also play a substantial role in both. Without education and intellectual challenge, it is unlikely that one can attain high functional intelligence. Likewise, smoking, obesity, disease, malnutrition, and abuse can substantially shorten life span. Hence, genetic manipulation, often called eugenics, will be combined with the provision of an optimal environment to prolong life and heighten intelligence.

These goals will be achieved by a twofold effort. First, the proposed association will support the groundbreaking scientific research necessary to understand and manipulate the biology of intelligence and longevity. Second, this association’s members will be provided with genetic counseling and interventions that will enable a slow but steady increase of longevity and intellect, generation by generation; human members will achieve ultimate millennial goals of average IQs of 145 (three 15-point standard deviations above the current mean IQ of 100) and life expectancies of 100 years (one 15-year standard deviation above the current life expectancy of some 85 years).

The association will also manipulate the genome of the above companion species to increase intellect and longevity to and ultimately beyond current human levels in larger mammals. At the same time, ever more intelligent machines will be created and replicated. Finally, the association will spread these species across the planet, the solar system, and ultimately the stars. These breathtaking goals may be realized in as little as a thousand years, less than the current life span of some human organizations and as few as ten of the long human life spans we are contemplating.

Doubtlessly, many will find these proposals grandiose, unrealistic, and frightening. The deliberate manipulation of the genome of humans and other species to create beings of superior intelligence will predictably be met with opposition, on grounds both realistic and specious. Others will greet this initiative with the excitement and support that it so clearly merits. One theme herein is an attempt to allay the legitimate doubts that this proposal will engender. We will devote substantial time to the history and abuses of the eugenics movement, lest we inadvertently repeat the many mistakes made therein.

Yet the substance of this proposal is not so far beyond existing practices. Domestic animals have been created by rigorous genetic selection over many lifetimes, and few would question the results or the merit of such eugenics. Similarly, genetic counseling is a reality, again one which is largely accepted by the educated public. Much of what is proposed herein does not go beyond these accepted practices.

Indeed, it is only in proposing the deliberate improvement of species and machines that we depart from existing practices. If the proposal fails for lack of support or an inability to achieve these ambitious goals, no real harm will have been done. If the proposal succeeds, on the other hand, the results will be truly wonderful. Humans of high IQ and longevity will still be humans but will not only lead better lives; they will also certainly make substantial contributions to their own and other species. As we shall demonstrate, high intelligence is linked to high morality and concern for others. There is nothing to be lost and worlds to be gained from this undertaking.

PART A

INTELLIGENCE AND LONGEVITY

CHAPTER 2

UNDERSTANDING INTELLIGENCE

Intelligence is not, a priori, necessarily a single measure. We could presumably have different intelligences for different things. Thus, intelligence in mathematics could just possibly tell us nothing at all about intelligence in writing. Yet such is not the case. Generally speaking, people who do well or poorly on one sort of performance test have a general tendency to do equally well or poorly on all other performance tests. This fact was first noted by Spearman (1904, 1927). Table 2.1 shows Spearman’s research results about relations between grades on six different subjects. These mathematical values for relations between any two subjects (math and music ability, for example) are given as correlation coefficients.

For the statistically challenged, a correlation coefficient is mathematically always shown as the value of lowercase r. It can take on values between -1.0 and +1.0. When r=0, knowing the value of one variable tells us nothing at all about the second variable. By contrast, r of +1.0 means a perfect positive relationship between two variables; that is, as one increases, the other increases, and knowing the value of either variable allows a perfect prediction of the other variable. Conversely, r=-1.0 means that as one variable increases, the other decreases; but again, knowing the value of either variable allows a perfect prediction of the other variable. A fact of great utility here is that the square of the correlation coefficient is the percentage of all the data in the two variables, which can be explained by the value of r. Thus, a correlation of 0.90 between two variables explains 0.90² (0.90x0.90), which is 0.81, that is, 81% about the two variables. In contrast, an r of 0.30 explains only 9% of the information in the two.

Spearman’s contribution to statistics was the calculation of the g (for general) factor within any set of variables. Like r, g is always lowercased, and it can take values between 0 and 1.0. Calculation of this value involves factor analysis, a mathematical process first discovered by Spearman. To further complicate matters, large data sets with numerous variables can produce a variety of statistically independent factors. Hence, factor analysis can reveal multiple independent factors. These factors will have different patterns of loadings on individual measures and are sometimes difficult to interpret.

In any case, starting with Spearman in 1904, it has repeatedly been found that on sets of scores over a wide variety of tests involving mental manipulation, there is generally a pronounced g factor. This g factor is the statistical evidence for the existence of one thing underlying mental manipulations of all sorts, and it is this fact that allows us to speak of a unitary factor generally called intelligence. By way of example, table 2.1 reproduces the data set used by Spearman in 1904 to reveal that one variable, the g factor, could explain performance in classes on six different subjects. The lower half of the table shows the correlations (r values) between grades on every pair of subjects. Note that the correlations are all positive, giving strong evidence for a large g or general factor. It is the existence of this g factor that allows us to speak of intelligence as a unitary factor, linking all forms of cognitively demanding tasks.

At the risk of repetition, I emphasize that factor analyses have replicated Spearman’s results literally thousands of times, allowing us to speak with confidence about intelligence as a single factor across all forms of mental tasks. Interested readers can also consult Gottfredson (1998) for further explication of the general intelligence factor.

TABLE 2.1. Correlations between Grades across Subjects as First Noted by Spearman (1904)

IQ is a relatively simple variable because IQ test scores are typically designed to have an average of 100 and a standard deviation of 15. When so normed, the distribution of scores assumes a bell-shaped or normal curve, with distances from this average given in standard deviation units and approximately as many scores above as below the mean (figure 2.1).

FIGURE 2.1. Male and Female IQ

FIG.2.1.tif

As is evident from the shape of the curve, most of the population has IQs not far under or over 100 IQ points, with an increasingly small proportion of the population having exceptionally high or low IQs (table 2.2). As is also evident from the figure, male and female IQ distributions have the same mean of 100, but they do not have the same variability. Males (as the females in this audience have doubtlessly long suspected) are more likely than females to have extremely low IQs. However, in defense of my sex, let it also be noted that males are also more likely than females to have extremely high IQs. By way of example, Deary and colleagues evaluated 1932 measures of 11-year-old IQs of over 80,000 Scottish children. There were 1.4 males for every female at low IQs between 50 and 60 and the same 1.4 males per female at high IQs between 130 and 140 (Deary et al., 2003).

As table 2.2 illustrates, 34.1% of the American population falls between 100 and 115 IQ points (remember, the standard deviation is 15 IQ points) or between 100 and 85 IQ points. Thus, almost exactly two-thirds of Americans have IQs not less than 85 or more than 115 (actually 34.1% from 85 to 100 and 34.1% from 100 to 115 for a total of 68.2% from 85 to 115 IQ points). The Einsteins among us have IQs of 160 or more, and genius begins somewhere around IQ >145 (three standard deviations above the mean). As of 2013, worldwide, there are almost 10 million people in the genius range. Some 427,000 of these geniuses are American citizens (table 2.2).

TABLE 2.2. Number and Proportion of American and World IQ Scores in a Normal Curve

For IQ or any other variable to be meaningful, it must be relatively stable over a substantial proportion of the life span. IQ appears to possess such stability from early in life, until senility causes IQ to plummet (Deary et al., 2008, 2013; Jones & Bayley, 1941; McCall & Garriger, 1993; Baltes & Schaie, 1976).

One more distinction that is important subsequently is that between fluid and crystallized measures of IQ. Fluid intelligence is a kind of raw cleverness, the ability to solve novel problems. Crystallized intelligence involves the ability to apply attained skills and knowledge—general knowledge, vocabulary, numeracy. By way of example, the Wechsler Adult Intelligence Scale (WAIS) measures fluid IQ on the performance scale and crystallized IQ on the verbal scale and pools the two for an overall IQ score.

THE EXTERNAL VALIDITY OF IQ:

For IQ to be meaningful, it must also possess external validity. That is, it must be related to variables that in their own right have obvious value. IQ does indeed possess such external validity. As shown in chapter 4, lQ predicts personal longevity and income, two variables of very high value indeed.

IQ AND BRAIN SIZE:

Another source of external validation for IQ is provided by the link between IQ and brain size. This relationship was first noted in the scientific literature by Pearson (1906), the developer of the correlation coefficient, who reported a correlation between head size and intelligence. As shown in two comprehensive reviews, a number of studies have demonstrated that IQ is biologically dependent upon brain size (Rushton & Ankney, 1996, 2009). By way of example, Willerman et al. (1991) used MRI (magnetic resonance imaging) to measure brain volume in 40 college students and then correlated volume with IQ. This correlation was 0.51 in men and 0.33 in women. Tan et al. reported a similar correlation of 0.40 in Turkish college students, while Wickett et al. found that MRI measures of brain volume and IQ correlated at r=0.395. Similarly, Witelson and colleagues (2006) directly measured the volumes of 100 postmortem brains and again correlated such volumes with IQ test records. In this study, the correlation between volume and verbal IQ was 0.36. Other studies have reported similar correlations. In a meta-analysis, McDaniel et al. reported a correlation of 0.33. Still other studies have found similar relationships (Akgün et al., 2003; Andreason et al., 2003; Tramo et al., 1998). Hence, IQ is a variable with biological validity.

This linkage between human IQ and brain size also extends to the animal kingdom. Studies in nonhuman primates have repeatedly linked brain size of different primate species to social intelligence (Reader & Laland, 2002), general intelligence (Roth & Dicke, 2012; Williams, 2002) innovation (Reader & Laland, 2002; Lefebvre et al., 2004), tool use (Lefebvre, 2013), complex and flexible social systems (Charvet & Finlay, 2012), and even deception rates (Byrne & Corp, 2004).

Further, it seems that overall brain size and not encephalization quotient (a kind of brain/body ratio) best predicts primate cognitive ability (Deaner et al., 2007). Remarkably, this linkage between brain size and measures of animal intelligence is also seen clearly in birds (Lefebvre et al., 2004; Burish et al., 2004; Lefebvre, 2013).

However, brain size is not the only factor determining intelligence. Thus, human female brain size is lower than male, but the two sexes have identical mean IQs of 100. Even more obviously, intelligent behavior does not correlate well with brain size across species. Humans have brains substantially smaller than most whales as do ravens compared, for instance, to domestic sheep. Yet clearly, ravens are substantially more intelligent than domestic sheep as are humans compared to whales, dolphins, and porpoises. We await a better understanding of factors other than sheer volume that are linked to intelligence in humans and animals.

IQ AND MORALITY:

Yet another factor validating the relevance of IQ is morality versus criminality. It seems likely that IQ is linked to the ability to predict the consequences of one’s actions. If so, criminal behaviors would be linked to low IQ and moral behavior to high IQ. Such is indeed the case. Moffitt and colleagues (1981) found in two Danish studies that, after socioeconomic effects were partialed out, there was a negative correlation between IQ and delinquency. Similarly, Lynam et al. (1993) reported that the negative relationship between IQ and delinquency was robust even after partialing out possible confounds, such as class, race, test motivation, school failure, or self-control. Low IQ also predicted high recidivism rates (Richter et al., 1996) and continuation of adolescent criminality into adulthood (for U.S. Caucasians and Hispanics but not blacks, [Donnellan et al., 2000]).

Other factors related to IQ and morality are bigotry, racism, and conservatism. Deary et al. (2008) found that in a representative United Kingdom sample of some 7,000 subjects, those with higher IQs at the age of 10 grew up to be more liberal/antitraditional adults at the age of 30. Stankov (2009) assessed both national averages and individual attitudes. At the national level (73 countries), broad conservatism was negatively correlated with education (r=-0.69), average national IQ (r=-0.73), and the failed state index (r=-0.80). At the individual level, among U.S. community college students, conservatism was negatively linked to Scholastic Aptitude Test scores (r=-0.35), as well as scores on vocabulary (-0.40) and analogies (r=-0.23). Similarly, prejudice against homosexuals appears to be linked to low abstract-reasoning skills (Keiller, 2010) as is racism (in a study of almost 16,000 Britons, Hodson and Busseri (2012) found that lower childhood IQ was linked to higher adult racism). Lower IQ, not surprisingly in view of the above, is also linked to prejudice and right-wing ideology (Hodson & Busseri, 2012).

SCIENTISTS: HIGH IQ AND ATHEISM:

Science is a discipline requiring experimental proof of the existence of hypothetical variables and the relationships between such variables. Since there is no objective evidence for the existence of supernatural entities, gods, angels, devils, demons, or deities, scientists tend to be atheists. Leuba (1916) may have been the first to report lower levels of religious belief in scientists. Eminent scientists are predominately atheists; a minute 7% of the members of the American National Academy of Sciences believe in God (Larson & Witham, 1998). Similarly, only 8% of fellows of the Royal Society agree that God exists (Stirrat & Cornwell, 2013). In a relatively recent survey, 60% of scientists disbelieved or had doubts about the existence of a deity or deities (Larson & Witham, 1997, 1998). Such differences also exist within the ranks of scientists; Dutton and Lynn (2014) have recently reported that physicists have higher IQs and are less religious than social scientists.

The linkage between IQ and atheism extends beyond the scientific community. In a meta-analysis of 63 studies, Zuckerman et al. (2013) found that 53 of these studies reported a negative correlation between intelligence and religiosity. This negative relationship was statistically significant in 35 studies, while a positive link between IQ and religion was statistically significant in only two studies.

In an even more inclusive analysis of the literature prior to 1986, Beckwith (1986) summarized the earlier literature, also clearly linking religiosity with lower IQ. Thirty-two studies and 12 Gallup poll results were summarized in this review. Twenty-eight of the 32 studies linked high IQ to low religiosity. The Gallup polls revealed that college alumni (average IQ about 115) were much less religious than pollees who only obtained a grade school education.

In studies using nations or states as individual data points (admittedly a technique that can magnify consistent but small in-state/nation differences), the correlation between national IQ and religiosity was reported to be -0.60 (Lynn et al., 2009) and -0.725 (Meisenberg et al., 2012). Across the 50 states in the United States of America, this correlation was closely similar at r=-0.55 (Pesta et al., 2010).

The termites, California Bay Area children selected for having an IQ >135 and still followed in the longest of all known longitudinal studies, also showed late-life reductions in religiousness in 81% of subjects (McCullough et al., 2005). Hence, the link to high IQ and low religiosity is not restricted to the lower end of the IQ distribution.

To summarize, in a recent study of adolescent white Americans, Nyborg (2009) found the following IQ differences:

Atheist IQ > Agnostic IQ by 1.95 points

Atheist IQ > Liberal IQ by 3.82 points

Atheist IQ > Dogmatic religious persuasions by 5.89 points

Nyborg concluded, Modern Atheists are reacting rationally to cognitive and emotional challenges, whereas Liberals, and in particular Dogmatics, still rely on ancient, pre-rational, supernatural and wishful thinking. I can only add, Amen to that.

Numerous studies have shown that human IQ is highly heritable. Such studies typically look at pairwise IQ correlations as a function of degree of genetic relatedness. That is, identical twins are compared to nonidentical twins, brothers to cousins, parents to children, and so on. Heritability, h², is closely related to the correlation coefficient and is the proportion of variance in a population contributed by simple genetic relationships.

Such heritability estimates place the simple genetic contribution to IQ variance as high as 0.81 (81%) by the age of 18 or later. This is indeed a very high contribution. As will be detailed in chapter 7, researchers have failed to find one or a small number of genes that contribute to individual differences in IQ. Hence, IQ must be the result of a large number of genes, each making a very small contribution to overall intelligence. Consequently, there may be many paths to increased intelligence, depending upon which (still large) subset of genes is selected.

Recently, several studies suggest that environmental variables that do increase intelligence are linked to socioeconomic status. According to a recent study, even in very young children, as socioeconomic status increases, so does the simple heritability of IQ (Tucker-Drob et al., 2010). This study measured infant mental abilities in some 750 pairs of American twins. As early as at two years, heritability (measured as the difference between identical and nonidentical twins) was almost zero in poor children, 0.20 for the economic average for this study and a whopping 0.57 for infants in well-to-do families. Several other recent studies have reported very similar findings (Harden et al., 2007; Turkheimer et al., 2003).

This link between IQ, heritability, and wealth is interpreted to mean that for the rich kids, the environmental contribution to IQ has largely maxed out, leaving the remaining variance attributable in substantial part to genetic differences. Thus, there may be very real, and attainable, limits to the contribution that environmental variables can make to enhancing IQ.

TEMPORAL INCREASES IN IQ: THE FLYNN EFFECT:

The so-called Flynn effect also addresses environmental effects upon human intelligence and suggests that such effects may have maxed out in the economically better-off countries. This fascinating effect rests upon the observation (first made by Flynn) that IQ test performance has shown a substantial increase over the last 100 years or so, especially in the economically better-off nations. This increase was long hidden by the practice of regularly renorming IQ tests to produce a population mean of 100 and a standard deviation of 15. It was only when Flynn documented increased performance on earlier IQ test versions that this increase became evident.

It now appears that this IQ increase may have begun in wealthier nations as early as 1900 or thereabouts. The rate of increase was some three IQ points increase per decade, and the increase was concentrated at the low end of IQ test performance. Hence, the Flynn effect did much less to increase scores in the genius range than to reduce low scores. Further, this IQ test performance increase did not seem directly because of increased schooling. Increases were more marked in fluid tests (such as Raven’s progressive matrices), tests not directly linked to education, test-taking ability, and even literacy. Tests that were more sensitive to increases in education (tests measuring crystallized intelligence) showed a substantially lower Flynn effect.

Test performance increases of this magnitude over not more than 100 years occurred much too rapidly for genetic selection to have played any role. Further, such increases seem to have occurred or are occurring throughout the world, albeit sometimes at later starting dates or different rates. This again virtually proves that the Flynn effect is an environmental and not a genetic phenomenon. In short, over the last 100 years or so, environmental influences of some sort(s) have created about a two-standard-deviation (30-IQ-point) increase in intelligence test performance in wealthy nations.

We do not know what has caused this effect. Possible explanations are many and by no means mutually exclusive. Suggested explanations include improved nutrition, better schooling/test-taking strategies, improved health or reduced disease, and generally more stimulating environments because of mass media and, more recently, personal computers.

Of equal importance, the Flynn effect may now be reaching a plateau, at least in some wealthier nations. In both Denmark (Teasdale & Owen, 2008) and Norway (Sundet et al., 2004) but not as yet in the United States (Ang et al., 2010), such a plateau has been reported. This possibility too suggests that the Western world has reached a near maximum in environmental effects upon intelligence. The Flynn effect will be discussed in substantially greater detail in chapter 7.

EVOLUTION OF HUMAN IQ:

There is no doubt that, over the known six-million-year course of hominid evolution, brain size—and hence IQ—increased quite remarkably. To better understand the evolution of cranial capacity in our species, we will briefly diverge to an overview of what is today known about hominid evolution.

Genomics reveal that the chimpanzee is humanity’s closest surviving relative. It is estimated that our hominid ancestors last shared a common ancestor with modern chimpanzees some six million years ago. The earliest reasonably well-preserved fossil evidence of ancestors, such as Ardipithecus ramidus, date back over five million years ago, close to the initial departure from the great apes. These earliest known ancestors were quite unlike apes in that they walked upright and had substantially reduced canine teeth. The combination suggests replacing biting with hand weapons as the initial hominid departure from the great apes. Yet the brain volume of these earliest ancestors was no greater than that of modern chimps (some 400 ml).

Moreover, this apelike brain volume remained relatively constant from six to at least three million years ago, over half the course of human evolution. In this same time period, early small-brained ancestors, such as Lucy (Australopithecus afarensis), evidently diverged into two lines, the robust and the gracile australopithecines (figure 2.2). The robust australopithecines had huge grinding teeth, were associated with digging sticks, and are thought to have evolved to eat roots and nuts on the African savannah. The gracile australopithecines probably evolved toward modern Homo sapiens. Yet neither line had brain volumes much greater than the chimpanzee’s 400 ml.

FIGURE 2.2. Human Evolution (time in years, 6 mil to present, on x, and brain size on y; each species gets a horizontal line for its approximate temporal duration as a species)

FIG2.2.tif

It is only around 2.5 million years ago that brain size appears to have begun to increase, rapidly in terms of evolutionary time, to our current brain volume of around 1,200 ml. The earliest hominid species sometimes associated with this increase is Homo habilis, who may have been responsible for the worked stone tools, which also appear in Africa around the same date. Homo habilis had a cranial capacity not greater than 600 ml.

It is really only with Homo erectus that brain size rapidly begins to approach ours. The earliest known Homo erectus is still African, sometimes known as Homo ergaster, and had a cranial capacity of some 900 ml. Subsequently, Homo erectus spread throughout the Old World and achieved cranial capacities of as much as 1,100