{kahz-mahl'-uh-jee} Cosmology is the study of the origin, constitution, structure, and evolution of the universe. It makes use especially of the theory of \Tgravitation\t, \Trelativity\t, Riemannian geometry, and the observation of \Textragalactic systems\t. Interpreting the observational data requires an understanding of the evolutionary processes of individual galaxies; hence it involves the study of \Tstellar evolution\t and \Tinterstellar matter\t. The density and temperature of the universe in earlier stages are understood in terms of the interaction of matter, radiation, and other forms of energy, as explained by atomic, nuclear, and particle physics. Finally, because the \Tsolar system\t is a product of the star-formation process and is directly accessible, its study--by such disciplines as geophysics, geochemistry, and the physics of the solar system--yields a large amount of information useful in the study of the rest of the universe. The term cosmogony is still sometimes used to distinguish the study of the origin of discrete celestial objects, particularly those of the solar system, from the study of the universe as a whole. Cosmogony may be considered a branch of cosmology. HISTORY OF COSMOLOGY Over the course of human history, numerous \Tcreation accounts\t of a religious or philosophical nature have been developed concerning the origin and nature of the universe. The first decisive step toward a modern cosmology, however, was taken early in the 17th century, when observational support became available for Copernicus's heliocentric model of the solar system, and the earlier \Tgeocentric world system\t was gradually abandoned. (See \Theliocentric world system\t.) Sir William \THerschel\t, the founder of stellar astronomy, made counts of the stars seen in various directions in space and on the basis of these counts proposed a scientific model of the universe--an isolated system in which the stars of the Milky Way were arranged in the shape of a disk, with the Sun at its apparent center. In 1838, Friedrich Wilhelm \TBessel\t made the first measurement of the distance to a star, thus providing a distance scale for Herschel's model that was a few thousand light-years in diameter. Harlow \TShapley\t discovered (1918) that vast regions of the Milky Way are obscured by thick interstellar dust and that the Sun is not at the center but rather toward the edge of the system. The diameter of our \TGalaxy\t is now estimated at around 100,000 light-years. In 1917, H. D. \TCurtis\t and G. W. Ritchey discovered a nova in the Andromeda Nebula. Curtis came to the conclusion--based on the analogy between that nova and novae in the Milky Way--that the Andromeda Nebula must be a separate star system similar to the Milky Way but at an unexpectedly great distance. This view was further strengthened when more powerful telescopes were able to pick out stars in the outer regions of this nebula, thus indicating that it was a system like the Milky Way and not simply a gaseous nebula. Soon it was recognized that many galaxies, or island universes, exist within the universe. From her observations of the Magellanic Clouds, satellite galaxies of the Milky Way, Henrietta \TLeavitt\t discovered (1912) a relationship between the period and luminosity of \TCepheids\t (a type of \Tvariable star\t). By means of this \Tperiod-luminosity relation\t it is possible to determine the distance of any Cepheid whose brightness can be measured. Since then, observations of these variable stars have been used to determine the distances of nearby galaxies. In 1912 and 1925, V. M. \TSlipher\t obtained spectra of 41 galaxies; from the \Tred shift\t in these spectra, the velocities of the galaxies were calculated as ranging from 300 km/sec for those moving toward us to 1,800 km/sec for those receding. Later, after a correction was made to allow for solar motion around the center of The Galaxy, all the velocities of distant galaxies were found to be receding from us. Meanwhile, in 1915, Albert \TEinstein\t had published the famous general theory of relativity. Soon afterward he proposed a static model of the universe, the expansion of the universe not yet being known. Willem de \TSitter\t proposed a cosmological model based on Einstein's theory, which allowed for positive as well as negative red shifts. In 1922, Aleksandr \TFriedmann\t derived a set of general cosmological models from Einstein's theory; it included Einstein's static model and de Sitter's model as particular cases. According to these mathematical solutions, the universe originated in and expanded from a single body of infinite density. Georges \TLemaitre\t (1927) and Arthur \TEddington\t (1930) put forward a physical theory, later known as the \Tbig bang theory\t. Edwin P. \THubble\t and Milton \THumason\t, using Cepheid variables as distance indicators, discovered (1929) a linear relationship between the recession velocity v and the distance r of observed galaxies where the constant of proportionality H is now known as \THubble's Constant\t. (In common practice, H is followed by subscript letter o; in this article, H alone will represent Hubble's constant.) By 1936 this relationship had been extended to distances as large as several hundred million light-years, and the expansion of the universe and the general correctness of Einstein's theory of general relativity had been established. The 200-in (5-m) Hale telescope on Palomar mountain was in operation by 1952. Using it, Walter \TBaade\t made new observations of Cepheid variables and discovered that extragalactic distances were actually twice as great as had previously been believed. George \TGamow\t and R. A. Alpher incorporated (1949) nuclear physics into Friedmann's model. They theorized that during the initial high-density state, radiation had dominated the universe, and from this they predicted that microwave background radiation would be found to exist. They also attempted to account for the origin of the elements by nuclear processes in the big bang. In 1964 A. A. \TPenzias\t and R. W. \TWilson\t measured this background radiation, a feat for which they later received the Nobel Prize. In 1959 radio astronomers at Cambridge, England, prepared a catalog of sources of radio emissions known as the 3C (third Cambridge) survey. The positions of the radio sources were precise enough for optical identifications to be made. In 1963, when spectra of these objects were obtained with the Hale telescope, Maarten \TSchmidt\t discovered that two of the objects (3C 273 and 3C 48) most suitable for study showed red shifts of 0.16 and 0.37--evidence that they were receding at 0.15 and 0.31 the velocity of light, respectively. The rate of energy emission of these objects--now called \Lquasar\ls--is 100 to 1,000 times that of the ordinary galaxy. To date a few hundred quasars are known, and the maximum red shift is more than 4, corresponding to a recession velocity of more than 93 percent of the velocity of light. Although the nature of quasars and their role in the evolution of the universe are still unclear, current astronomical theory suggests that the objects are the brilliant and violently active nuclei of galaxies at an early stage of evolution, and that they lie at the far limits of an expanding universe. CONSTITUTION OF THE UNIVERSE The visible structure of the universe consists of galaxies--our own galaxy and the extragalactic systems. The mean density of matter contributed by the known galaxies lies between (10 to the power of - 31) and (10 to the power of - 30) g/cu cm (equivalent to one hydrogen atom in 17 and 1.7 cubic meters, respectively). Based on observations of stars in our galaxy, the chemical composition of this matter is found to be 75% hydrogen, 24% helium, and 1% other elements. The energy density due to microwave background radiation is equivalent to a matter density of less than (10 to the power of -33) g/cu cm. The equivalent matter density due to neutrinos, gravitational radiation, and other suspected forms of energy in the universe may be of comparable magnitude. Since the minimum density needed for a closed universe is around (10 to the power of -29) g/cu cm, there is great interest in searching for matter in intergalactic space (the so-called missing-mass problem in cosmology). There are indications that intergalactic matter must exist within \Tclusters of galaxies\t. In many cases, the calculated potential energy of the mass contained in galaxies alone is smaller than the kinetic energy of the galaxies. In order for clusters to be stable against dispersion--and hence to be still existing, 14 billion to 20 billion years after the universe began--as much as 30 times the matter in galaxies must be distributed in the form of intergalactic matter inside a cluster. So far, however, only traces of intergalactic matter, revealed by the X-ray emission of hot plasma, have been found. AGE OF THE UNIVERSE The age of the universe must be determined from the correct cosmological model, using Hubble's constant and the so-called deceleration parameter q obtained from observations. The parameter q cannot yet be measured reliably, but there are a number of ways by which the upper and lower limits to the age of the universe may be estimated. If the recession velocity v in equation 1 were constant, then at some prior time t, such that t = r/v = H to the power of -1, the distance between us and an observed galaxy would have been zero. Since in most models the recession velocity is higher in the past, the actual age of the universe will be less than t. The quantity t may therefore be regarded as an upper limit for the age of the universe. In a thorough analysis of available red-shift data of galaxies, Alan SANDAGE derived the following values for H and q: H = (55 plus or minus 7) km/(sec-megaparsec) H to the power of - 1 = (1.8 plus or minus 0.23) X 10 billion years (2) q = 1 plus or minus 0.4 (3) giving the age of the universe as (10 plus or minus 1.5) X 1 billion years. The upper limit is H to the power of - 1 = 18 X 1 billion years, or 18 billion years. The probable error of H in this estimate is around 10%. However, it must be realized that H is obtained from a chain of analyses, and many theoretical arguments are used. The estimated value of H may change significantly in the future, despite the small formal error in equation 2, just as it did in the past, when it was believed to be much larger. The age of the solar system may be determined by radiometric age-dating of meteorites. Uranium has two natural isotopes, U-238 and U-235; both are radioactive and eventually decay into lead-206, and lead-207 respectively. Since Pb-204, another isotope of lead, is not a product of radioactive decay, the amount of Pb-204 must remain unchanged throughout time. By measuring the ratios of Pb-207 to Pb-204 and of Pb-206 to Pb-204 in a given sample, it is possible to determine the time that has elapsed since the sample was formed. Meteorites are believed to have originated in the early stages of solar-system formation. Meteorites have been orbiting the Sun since their creation, undisturbed by the geological processes that alter Earth rocks. Thus fresh meteorites are good samples of the primordial matter of our solar system. When the ratios of the lead in these samples are measured, an age of 4.6 billion years is obtained. Using theories of element synthesis, and observing the ratios of U-235 to U-238, and U-232 to U-238, W. A. Fowler concluded that our galaxy must be at least 11.7 billion years old. The age of star clusters can be estimated from the observed properties of their brightest \Tmain sequence\t stars by using theories of stellar evolution. After the stars condense from interstellar clouds, nuclear processes supply the energy of the stars through hydrogen-reaction sequences in which four hydrogen nuclei, or protons, combine to form a helium nucleus, and thus release around 26 MeV of energy. The star then enters the main sequence, the array of stars that generate nuclear energy at stable rates for long periods of time. A star eventually leaves the main sequence when the fractional amount of hydrogen converted into helium exceeds a critical ratio, ranging from 0.1 to 0.6, depending on the mass. Since the luminosity of a star is roughly proportional to the third power of its mass, massive stars leave the main sequence sooner. In a star \Tcluster\t, most of the stars are created at approximately the same time. The observed point at which massive stars depart from the main sequence is therefore a good indicator of the cluster's age. Using accurate stellar models computed by a number of other scientists, the American astronomer Alan Sandage determined the age of three old star clusters, M3, M15, and M92, giving their ages as 11.5 billion, 9.4 billion, and 9.7 billion years, respectively. The age of the universe as determined from cosmological models is consistent with limits determined from isotope ratios and stellar evolution. The age of the universe may be placed in the range of 14 to 20 billion years. STRUCTURE OF THE UNIVERSE One of the most unsatisfactory features of Newtonian mechanics is its failure to provide us with an explanation of the behavior of light. Although cosmological models constructed from Newtonian mechanics may give results that are in many respects similar to those in Einstein's theory, the Hubble law (Equation 1) cannot be obtained from such models without artificially modifying the laws of physics. The most important feature of Einstein's theory is that, in it, gravitational fields are identified with the geometrical structure of space and time. Einstein noticed that in a gravitational field, the trajectory of a particle is independent of the composition of the particle. Hence the particle trajectories can be regarded as geometrical properties of space and time. Einstein identified the particle trajectories as geodesics--paths of extreme (maximal or minimal) lengths. The geometry of the \Tspace-time continuum\t in the presence of a gravitational field is such that the particle trajectories are identical to geodesics of this particular geometry. Light trajectories belong to a special class of geodesics, and so this theory accounts for the behavior of light. Since geodesics in Euclidean space are always straight lines, space-time must be curved in general, in order to accommodate the various types of particle trajectories. Thus, \Tnon-Euclidean geometry\t occurs naturally in Einstein's theory. The particular type of non-Euclidean geometry used is Riemannian geometry. To relate the source of gravitational fields--matter, energy, and pressure--to the geometrical structure of space and time, Einstein introduced ten equations, based on the use of ten potentials. These ten equations are necessarily complicated. However, difficulties in obtaining cosmological solutions are minimized when the following assumptions, also known as cosmological principles, are imposed: 1. the universe is homogeneous--the density of matter (galaxies) is the same everywhere; 2. the universe is isotropic--the distribution of matter (galaxies) is the same in every direction. These two assumptions have thus far been supported in observations. The ordinary principles of plane geometry (Euclidean geometry) may be adapted for geometry on a curved surface if straight lines are replaced by geodesics. Triangles, and circles can be similarly drawn. The difference between plane geometry and curved-surface geometry is that many formulae and concepts of plane geometry are no longer valid. For example, the sum of the three angles of a triangle is no longer 180 deg, and the circumference of a circle is no longer equal to 2 (pi) where r is the radius. In general the concept of parallelism over finite distances is not valid. The properties of three-dimensional and four-dimensional curved spaces are naturally more complicated than those of two-dimensional surfaces. For example, the "curvature" of a four-dimensional space is determined by 20 curvature components (on a two-dimensional surface, only one component is necessary). However, if the two cosmological principles are applied, then the number of curvature components required to describe the three-dimensional space is reduced to one, just as in the case of a two-dimensional space such as the surface of a sphere or a hyperboloid. Friedmann found a complete set of cosmological solutions as follows: the space-time structure (three-dimensional space plus one time dimension) may be described by a three-dimensional space whose curvature depends on time. If the curvature is positive, then the three-dimensional space has a finite (closed) volume and properties analogous to that of a sphere. If the curvature is negative, then the three-dimensional space has an infinite (open) volume and properties analogous to that of a hyperboloid (saddle surface). An important feature of Friedmann's models is that there are no static solutions: the universe, as described by a three-dimensional space of positive, negative, or zero curvature, is expanding. A contracting phase can exist only in the case of a positive curvature. After the universe has expanded to a maximum size, it begins to contract towards a zero radius. A static universe can only exist either temporarily between an expanding and a contracting phase, or as the final state of expansion in the case of zero curvature. The expansion rate also depends on time, so that the Hubble constant is also a function of time. In an open universe the expansion will proceed forever. The Einstein-de Sitter model corresponds to the case of zero curvature (expanding Euclidean geometry). The closure property of a model universe is determined by its matter and energy content. A greater matter-energy density gives rise to a positive curvature, and hence a closed universe, and vice versa. This is easy to understand in terms of Newtonian models: The potential energy of a model universe is greater if the density of matter is higher, and when the potential energy exceeds the kinetic energy of the expanding matter, the universe is gravitationally bound and hence cannot expand forever. The matter-energy density rho is related to the deceleration parameter q and the Hubble constant H as follows: q = 4 (pi) G (rho)/3 HH where G is the universal gravitation constant. (In common practice, a subscript o follows rho, q, and H.) For a universe like ours, in which the predominant energy density is due to matter, if q1/2 the universe is closed (positive curvature); if q1/2 the universe is open (negative curvature). The case q=1/2 corresponds to an expanding Euclidean space (zero curvature). Since the value of the density of matter in the universe is uncertain, the curvature must be determined by directly measuring q from observed data. Since q describes the degree of deceleration of the universal expansion, the determination of q requires observations of objects at great distances. The most promising method of finding q is through a known relationship between magnitude and red shift. Using the brightest members of 84 clusters, Sandage determined the value of q to be +1 plus or minus 0.4. From Sandage's value of q the density of the universe should be around 2 X (10 to the power of - 29) g/cu cm, which is at least ten times the observed density of matter in the form of visible galaxies. This has been an unsolved problem ever since q was first measured by Hubble in 1926. Thus, on the basis of the observed matter density, our universe should be open (negative curvature). From the observed value of q, the universe should be closed (positive curvature). There is no resolution to this contradiction at the present moment. ORIGIN OF THE UNIVERSE All Friedmann-model universes, whether open or closed, begin from a singularity having infinite density. At a density of (10 to the power of 96) g/cu cm, the size of the universe would have been the size of a proton--radius is approximately equal to (10 to the power of - 13) cm. In that state, the predictions of general relativity would be invalidated by quantum mechanical considerations. However, all attempts to formulate a theory of gravitation in light of quantum mechanics have been unsuccessful. Without resolving the infinite density problem, it is possible to proceed with the assumption that the universe was created at a rather high density state, say at a density considerably below (10 to the power of 96) g/cu cm but substantially greater than the density of a proton (approximately (10 to the power of 14) g/cu cm). In 1946, Gamow pointed out that the early stages of the universe must have been dominated by radiation. As the universe expanded, the matter density decreased more slowly than the radiation-energy density decreased. Thus, after the universe had expanded for some time t, matter dominated radiation energy in density. There are two strong indications that the early universe was in fact dominated by radiation. First, the residual microwave background radiation has been detected. Second, the abundance of helium in the universe is around 24% by mass. If all helium was created in stars, then at most 1% of matter would be in the form of helium. On the other hand, around 20-30% helium may have been produced at early epochs. Galaxies and probably star clusters form only after matter-energy becomes dominant. This occurs about 10 to the power of 8 years after creation. A question that is often raised is whether substantial \Tantimatter\t exists in the universe. The answer is almost a qualified no. Near the moment of creation, particles and antiparticles are in thermal equilibrium with radiation. As time passes, the temperature decreases. Antiparticles and particles will then annihilate one another. If the number of particles (or antiparticles) exceeds that of antiparticles (or particles), then as the universe cools down, all the antiparticles (or particles) will disappear. If matter and antimatter were originally created in equal quantities, only a very small fraction would survive the initial annihilation process. The background radiation would have to be at least 300K. This is incompatible with current observations. There is another cosmological theory known as the \Tsteady-state theory\t, conceived by Thomas \Tgold\t, Hermann \TBondi\t, and Fred \THoyle\t. They modified the first cosmological principle by including time, and stated the "perfect cosmological principle": The universe is homogeneous in space and time. According to their theory, the density of galaxies in space has always been constant. In order to account for the expansion of the universe, spontaneous creation of matter is assumed. The amount of matter that would have to be created would be too small to be detected (approximately one hydrogen atom per cubic kilometer in one year). The strongest evidence against the steady-state theory is its inability to account for the observed microwave background radiation. Although this theory is no longer accepted by most cosmologists as a valid model of the universe, it contributed much to the development of cosmological thinking. FUTURE OF THE UNIVERSE If the universe is closed, the expansion will eventually stop, and red shifts will become blue shifts (contracting phase). After a certain time the universe will return to the state of being a singularity of infinite density, and vanish in a second big bang. Will the universe be recreated from this singularity again (oscillatory universe)? According to Einstein's theory, the answer is No. However, Einstein's theory is not likely to be valid when the density is too high. There is therefore no answer to this question as yet. If the universe is open, the expansion will go on forever. Eventually all the energy of the stars will be used up, and the universe will expand forever in total darkness. As we have seen, there are great discrepancies in the determination of the curvature of the universe. Until these discrepancies are resolved, it will not be possible to predict the properties of the universe in the distant future. ALTERNATIVE THEORIES IN COSMOLOGY A number of other speculative theories in cosmology have been and continue to be developed. These theories are based on suggestive coincidences in various formulations of the mathematical relationships between the gravitational constant, or the speed of light, and Hubble's constant. For example, an alternative theory of gravitation was formulated by physicists Robert \TDicke\t and Carl Brans, in which the universal gravitation constant was allowed to vary. The theory predicted a precession of the orbit of the plant Mercury that agrees with observations and differs from the one predicted by general relativity. To explain this discrepancy, Dicke postulated an additional contribution from a correction to the gravitational potential of the Sun, due to an oblateness of the Sun of about one part in 10,000. However, this oblateness was not confirmed, and interest in the Brans-Dicke theory waned. Other mathematical considerations have led some scientists to suggest that the evolution of the universe can be predicted from particle physics. A number of major developments in theoretical cosmology took place in the 1970s and 1980s in light of that possibility. These developments made use of the basic concepts of the big bang theory, and they were largely aimed at trying to understand the processes at work in the first few moments of creation, when densities and temperatures were so great that conventional theories of matter are no longer applicable. As part of this effort, the British physicist Stephen \THawking\t showed that matter could be spontaneously created at the so-called event horizon of a \Tblack hole\t--that is, the point of proximity to a black hole at which the red shift is supposedly so great that light signals emitted there cannot reach an observer. Following Hawking's work, it was then shown that quantum fluctuations in an empty de Sitter space could create a virtual universe with negative gravitational energy. This virtual universe could exist for only a very few seconds, but by means of the so-called quantum \Ttunnel effect\t it could emerge as a real universe and expand to form the universe as it exists today. This theory, then, provides an account of the creation of the universe within the framework of the big bang concept. According to this theory, the total energy of the universe is not necessarily zero. It may be positive (leading to an open universe) or negative (leading to a closed universe). The theory also requires a multidimensional space in which many four-dimensional subspaces are embedded. A separate universe could be created in each subspace by means of the quantum fluctuation processes, leading to the concept of a possible infinity of other universes. If our own universe were confined to one such subspace, however, there would be no way to find out if other universes actually exist. The inflationary theory developed in the early 1980s by the American physicist Alan Guth is a further attempt to account for the first moments of creation in terms of quantum fluctuation processes. Such work aims to align cosmology with the so-called \Tgrand unification theories\t (GUT) that seek to unite the \Tfundamental interactions\t of matter in a single theory. Taking one simplified GUT formulation, it was found that right after its creation a protouniverse could eventually expand into the kind of universe observed today. According to that theory, however, so-called magnetic monopoles should also be created (see \Tmonopole, magnetic\t), and their very large mass would cause the universe to evolve to this present state in only 30,000 years. Guth developed his inflationary model of the universe in order to eliminate such an inconsistency. He introduced phase transitions during which rapid cooling took place in the very early universe, thereby eliminating certain difficulties of the big bang model. One such difficulty is the high degree of homogeneity and isotropy of the present universe. Conventional big bag theory had the early universe expanding so rapidly that homogenization would be impossible. In the inflationary model the universe is much more compact in its earliest stage, allowing a high degree of homogeneity and isotropy to be achieved. The creation of monopoles is also avoided during successive phase transitions, thus allowing for the subsequent evolution of the universe at a rate in line with actual observations. The inflationary theory is still in skeleton form, but it does make definite predictions such as of the ultimate decay of protons. The theory continues to be refined by Guth and others, who also try to respond to further problems--such as the discovery of extremely large-scale features comprising many millions of galaxies--that are presented by observational astronomy. Another product of these attempts at combining GUTs with cosmological theories is the concept of cosmic "strings," defects in the fabric of space-time that linger from the first moments of the big bang. The strings would either extend infinitely or form closed loops. Essentially one-dimensional and under enormous tension, they would become increasingly massive the farther they stretched and could serve as sites for galaxy formation. Other theorists object to the concept of a singularity at the onset of creation, however, and are exploring exotic theories such as "negative pressure," a fifth fundamental interaction that might also account for the massive creation of matter in the early universe. Many other possible models of the universe are also being explored by theorists. Some of them involve complex variations on the numbers of dimensions required to account for a universe that will satisfy both relativity and quantum mechanics within the confines of the big bang concept. Others, such as the "plasma cosmology" first advanced by Swedish astrophysicist Hannes \TAlfven\t, move beyond the bounds of big bang theory and in the direction of the largely discarded steady-state theory. Plasma, called the fourth state of matter, is matter in the form of electrically charged particles (see \Tplasma physics\t). Although observed only in isolated circumstances on Earth, it is the state in which most of the universe actually exists. According to plasma cosmologists, the magnetic and electrical properties of plasmas are sufficient to account for the large-scale structures of a universe without beginning or end. Supporters of this theory are mainly plasma physicists, although astronomers are taking notice of this concept. Although such work is being done in the theoretical field, it remains necessary to resolve certain basic observational problems of cosmology. Among them are the discrepancy between the observed density of matter in the universe and the value of the parameter q; the role of quasars in cosmological evolution (including their nature and energy source); the drastic differences between the magnitude-red shift relationships of quasars and galaxies; the paradoxical quasar-galaxy pairs (apparently adjacent objects with very different red shifts); and according to some studies, the grouping of galaxies on the surfaces of giant, bubblelike voids throughout the universe. Apparent discoveries of newly forming galaxies also call into question some basic assumptions of the big bang theory. These questions, difficult as they appear, are relatively no harder than the problems that confronted cosmologists at the beginning of the 20th century. Hong-Yee Chiu Bibliography: Bartusiak, Marcia, Thursday's Universe (1988); Davies, Paul, The Cosmic Blueprint (1988); Ferris, Timothy, Coming of Age in the Milky Way (1988); Greenstein, George, The Symbiotic Universe (1988); Harrison, Edward, Darkness at Night (1987); Hawking, Stephen W., A Brief History of Time: From the Big Bang to Black Holes (1988); Judson, H. F., The Search for Solutions (1987); Linde, A. D., Particle Physics and Inflationary Cosmology (1990); Mallove, E. F., The Quickening Universe (1987); Overbye, Dennis, Lonely Hearts of the Cosmos: The Scientific Quest for the Secret of the Universe (1991); Parker, Barry, Creation (1988); Silk, Joseph, The Big Bang (1989); Trefil, James, The Dark Side of the Universe (1988).