Table of contents

The decay of zero area pulses is investigated for the Korteweg-de Vries equation. By using a linear representation of the Korteweg-de Vries equation the problem is reduced to the solution of a simple quantum-mechanical problem. The transcendent equation for eigenvalues is obtained and its numerical analysis is carried out. The region of existence of the discrete energy levels, corresponding to a soliton lattice of the Korteweg-de Vries equation, is found. The existence of a threshold of discrete energy levels in the zero area potential is established. The numerical simulation results on the decay of zero area pulses are compared with those predicted for the discrete energy levels.

The quantized field description of particles interacting only with a given unquantized (i.e., external) field is related in a simple way to the one-particle s matrix. Here s is based upon the solutions of a one-particle wave equation, without any initial conditions. Thus in the relativistic case s also acts upon unphysical "electron" states of negative energies. A factorization of s leads to simple exact expressions for QED's transition amplitudes for certain electron and/or positron processes (without photons). The formulas involve one-particle matrices, which are given by non-linear equations. The connection between Feynman's in general non-unitary s^F matrix and our unitary Dirac s matrix is found. A "radiation interaction picture" is introduced, and some first order examples are given.

We study the singularities of energy levels as functions of coupling parameters in a two-dimensional model with the potential V = (g²/r²) + 2(g - 2N)r² - r⁶, where N is a natural number.

Several aspects of the short time dynamical behavior of Hamiltonian systems are studied. The case of persistence of regular motion in some parts of the chaotic region of the phase space is examined in relation to Nekhoroshev's theorem. A local rate of divergence is defined, and its relationship to the rate of decay of the time autocorrelation function of a phase space coordinate is shown. Finally a numerical method for studying the short time dynamics of Hamiltonian systems is proposed and tested.

Using the improved non-relativistic quark model for light baryons, we follow up an old suggestion due to S. L. Glashow that there might exist a relatively strong tensor mixing in the ground state of the nucleon. We find that, except for the ratio of the charge radii of the neutron and the proton, Glashow's proposal is consistent with the empirical values of a series of baryonic observables that could be sensitive to this mixing. The origin of the mixing itself remains obscure, in the framework of the model used in this paper.

The results on pion-emulsion interactions at 340 GeV support the proposed correlation between the number of grey tracks and hadron-nucleon collisons inside the nucleus. An analysis of the data at 400 GeV from proton-emulsion interactions and at 340 GeV for pion-emulsion interactions indicates that R-v relationship is projectile independent whereas R₃-' is projectile dependent.

Uptodate accurate extensive data on γ-gB(E2) ratios for even-even rare-earth nuclei is compared with the predictions of the rigid triaxial model of collective rotation to search for a correlation between the nuclear structure variation with Z, N and the γ₀ parameter of the model. The internal consistency in the predictions of the model is investigated and the spectral features vis-a-vis the γ-soft and the γ-rigid potential are discussed.

Excited states in ^212,213Rn were populated in the bombardment of ²⁰⁴Hg by ¹⁴C ions of energies 80-94 MeV. The depopulation of the excited states was studied using in-beam spectroscopic methods including γ-ray excitation, γ-ray angular distribution, γ-γ-t coincidence, pulsed-beam-γ timing and conversion electron measurements. Spins up to J^π = 30ℏ and a tentative J = (73/2)ℏ were observed, respectively. In ²¹²Rn the previous level scheme was confirmed, with the addition of five new transitions of which one feeds the 30⁺ state. In ²¹³Rn the previously reported level scheme was mainly confirmed and it was extended to (71/2) or (73/2)ℏ, which is the highest spin observed to-date in a lead-region nucleus. Several transition rates B(λ) were deduced from isomeric half-lives or branching ratios in both isotopes. Deformed Independent Particle Model (DIPM) calculations were performed to assign configurations, which in ²¹³Rn usually consist of the odd neutrons vg_9/2 and vj_15/2 coupled to the four-proton states in ²¹²Rn. The highest levels include neutron core-excitations, since the maximum valence-nucleon angular momentum in ²¹³Rn is 55/2ℏ. The very topmost levels even need 2- particle-two-hole excitations. Shape changes in these two Rn isotopes are discussed, as are the transition rates.

The intensity ratio of the two fine structure components in the 3s²3p²P_J–3s3p²S_1/2 and ²P_1/2 resonance transitions has been measured in Al-like S IV, Cl V, Ar VI and Ti X, using the beam-foil technique. Unexpectedly large deviations from the LS ratios are found, particularly in the ²P–²S case. Although these deviations can be understood qualitatively from theoretical calculations (also reported), large quantitative discrepancies are observed in the beginning of the sequence.

The multiconfiguration Hartree-Fock method along with configuration interaction in the Breit-Pauli approximation has been used to calculate the autoionization rates for ⁵P states of the configuration 1s²2s²2p⁵3p³ in Na⁻. The predicted autoionization rates for decay to 1s²2s²2p⁵3s3pks, 1s²2s²2p⁵3s3pkd, and 1s²2s²2p⁵3s²kp channels were found to be 5.092 × 10¹² sec⁻¹, 6.897 × 10¹² sec⁻¹, 6.934 × 10¹² sec⁻¹, for J = 3, 2, 1, respectively. These autoionization rates dominate over the transition rates. Both spin-orbit interactions and correlation effects played an important role in the prediction of decay rates.

Spectra of the galliumlike ions Rb VII, Sr VIII, Y IX, Zr X, Nb XI, and Mo XII emitted from sparks and laser-produced plasmas have been recorded in the region 235-665 Å. All levels of 4s²4p, 4s²4d, 4s4p², 4p³ and, in Rb VII, Sr VIII and Y IX, 4s²4f have been established. The level structure has been studied by means of ab initio and parametric calculations.

We have investigated the np²P Rydberg series in aluminum. The high lying levels were populated by stepwise excitation in an atomic beam apparatus using two pulsed dye lasers. The series was found to be unperturbed and the level values are described with high precision by the extended Rydberg-Ritz formula. The value of the quantum defect is 1.243 ± 0.003. The fine structure of the 26p²P state was observed. The fine structure separation of this state is 19.2 ± 2 mK. The quadratic Stark effect was measured and interpreted. The value for the polarizability is Q_n = 3.8 ± 0.6 × 10⁻¹⁵n*⁻⁷ cm⁻¹/(V cm⁻¹)².

A model for steady state neutral atom transport in a finite thickness, one dimensional plasma slab with a source of neutrals and molecules at the edge, is formulated. A computer code OMID is developed for computational analysis. The analysis shows that the cold neutral particle density is maximum near the wall, the total fuelling profile peaks in the vicinity of the wall for a molecular source more than for an atomic one, energetic neutrals formed in successive charge exchange collisions carry energy to the wall and sputter impurity atoms back to the plasma. For low edge temperatures where molecular hydrogen is dissociated, mostly atomic hydrogen is produced, the ionization profile shifts towards the wall as the edge temperature is raised and more ions are created directly from molecular dissociation.

The effect of varying the incident neutral particle energy has been investigated.

The comparison of the results with the predictions of other codes showed satisfactory agreement for regions close to the wall.

A nonlinear nalysis of the current-convective instability is carried out on the basis of a classical, two-fluid model that includes the effects of parallel electron current and thermal conductivity, electron compressibility and thermal force. Restriction is made to time rates of change slow compared with the resistive diffusion rate across a perpendicular wavelength. Assuming the background gradient of electron temperature to be much greater than that of the density, a set of mode-coupling equations between electron temperature and electrostatic potential Fourier components is analysed numerically. The heat flux is computed explicitly. The influence of parallel current and electron conductivity on plasma edge transport due to thermal excitations is determined and discussed.

A nonuniform electron density distribution model is used to estimate the positive ion concentration in an electron plasma column. The calculation shows that a nonuniform confining magnetic field is necessary to achieve an almost pure electron plasma in a Penning discharge. The effect of the plasma temperature on the ion concentration is also discussed.

n₀ ∝ B₀² relation of a magnetised electron plasma supported by Penning discharge has been observed in the experiments. A theoretical explanation is given based on Poisson's equation with one assumption about the plasma temperature. The results also show that the plasma density is limited by the plasma radius and the plasma temperature, which agrees with the experimental observation.

A chain of equal variable wall segments is used to model magnetic domain wall motion in alloys with positive cubic anisotropy. Irreversible kink distortions arise naturally when segments adopt unstable orientations. Agreement with experimental data is better than obtained with earlier models that formed kinks at the onset of metastability.

The accelerated aging behavior of Metglas 2605S2 amorphous ribbons has been studied in order to investigate the stability of their practical properties such as power loss and exciting power or permeability. The samples were tested at 200°C for 2000 h aging in silicon oil or in air atmosphere with and without a compressive stress. As is announced by manufacturer, the material is perfectly stable if it is correctly annealed. Furthermore, a compressive stress can considerably degrade the magnetic properties. Thus a 10 kg mm⁻² transversal stress, applied on 14% of the surface of the test sample at the aging temperature, leads to a decrease of the materials permeability. This effect is explained by a disappearence in the stressed part of the directional order previously induced by magnetic annealing.

We present an investigation of photoresponse in a large period (300 Å) sawtooth doping superlattice which is composed of alternating n (Si) and p (Be) δ-doping layers with undoped GaAs layers (15 nm) between them. This doping profile leads to sawtooth shaped conduction and valence bands in the growth direction. We use a new selective n and p contacting method to measure the photocurrent in this sturcture and the behavior of the subband gap compared with the gap of the host semiconductor. The photocurrent is finite even when the photon energy is below the gap of GaAs and can be varied by applying a bias voltage between the n- and p-contacts. The previously suggested generalized Franz-Keldysh model is used to explain the observed phenomena. The relaxation process is investigated by considering the decay of the photocurrent after switching off the illumination. It is found that the relaxation of system towards equilibrium obeys a nonexponential law. The temperature dependence of this decay process is discussed. Our theory is in quite good agreement with the experimental results.

Apparent motion (AM) is seen when two patterns (such as movie frames) are exposed in succession in nearby positions. How does the visual system solve the "correspondence problem", that is, decide which item in the second frame is to be paired off with a given item in the first frame? We used a four-spot ambiguous AM display to reveal the influence of additional spots exposed nearby in space or in time. Spots that are close to each other tend to be paired off by proximity. AM is seen preferentially within the same hemiretina, that is within the same half of the brain, rather than across the retinal midline. AM in a straight line is strongly preferred over a bent path. Unambiguous AM drags an ambiguous nearby AM with it. An array of identical ambiguous AMs all move together, not independently.

We also describe entrained AM, and perceptual suppression of the spurious motion signals originating in static background items that are covered and uncovered by an object moving in front of them. All these phenomena demonstrate the role of parsimony and of minimum principles in motion perception.

A number of psychophysical studies on apparent translatory movement, apparent rotational movement, and compositions of these transformations are reviewed. The studies have yielded modern equivalents and extensions of Korte's third law. The results support the hypotheses that (a) apparent movements over short and long distances are processed by separate mechanisms and (b) apparent movement over a long distance is represented by impleting ("filling in") a path connecting the discrete stimulus presentations. It is suggested that impletion serves economy of processing by normalizing the internal representation of an object viewed in apparent movement to the format used for objects viewed in real movement.

Isolated retinal rods are known to spontaneously produce electrical "dark" events which appear identical to those associated with real photoisomerizations. Conceivably, this intrinsic noise at the very input to the visual system constitutes the "dark light", which sets an ultimate limit to the detection of weak lights in darkness. The performance of single neurons at different levels in the dark-adapted amphibian retina is shown to be consistent with this notion: (1) in the distal retina, horizontal cells display a dark noise of appropriate power in the relevant frequency band; (2) at the retinal output, the "best" ganglion cells perform as if their detection of dim flashes were limited by a dark light consistent with measured rates of dark rod events; (3) the proportion of such ganglion cells is sufficient to ensure that behavioural performance could in principle reach the same quasi-perfection.

The responses of single units in the optic lobe of the blowfly Phormia Terraenovae have been studied in order to find out whether the neurones generally are characterized by specificity or ambiguity in the context of visual motion perception. The results demonstrate that several higher order neurones exhibit the kind of mathematical ambiguity which is required if they are to perform a neural vector analysis. The neurones exhibit ambituity to single-dot motions, line motions as well as to complex motions characterizing looming including radial and oblique motions in depth. The implications of the findings for the general problem of depth motion perception are discussed.

Parallel processing by a set of multiple mechanisms selectively sensitive along various dimensions of visual patterns - e.g., spatial frequency, orientation, direction-of-motion - seems to be one of the earlier states in the perception of form and objects. The evidence for these mechanisms - which comes primarily from neurophysiology and psychophysics using near-threshold patterns (patterns of such low contrast that they are imperfectly discriminable from a blank field of the same mean luminance) - and the properties of these mechanisms are briefly reviewed. The responses of these mechanisms to some texture patterns used to investigate the segmentation of the visual field into separate regions is illustrated.

An ideal observer is a theoretical device that carries out its assigned task with optimal efficiency. Therefore, its performance is a precise measure of the information available to perform the task. We have created a program that computes the accuracy, in visual discrimination tasks, of an ideal observer operating at the level of photon absorptions in the human photoreceptors. Comparison of ideal-observer performance and human performance in the same vision experiments reveals that many perceptual phenomena previously attributed to neural mechanisms may, in fact, be attributable to variations in the information content of the stimuli at the receptors.

I discuss how the solution of visual tasks can be formulated in terms of network models, both linear and nonlinear. The tasks can be formulated as optimization problems, where one has to minimize a "cost function" whose parameters describe the importance to be paid to various constraints and features of the task. As the elements of these networks evolve in interaction with each other, they collectively compute a solution of the problem. As a specific example, I describe a simple network of binary elements (spins) which computes the outline of an object in a two-dimensional image given as an array of grey-level pixel values. The novelty in the model lies in the use of 4-spin couplings (or a sort borrowed from lattice gauge theories) to enforce the constraint that the solution ought to be a closed curve.

This report describes a neural network model which is capable of automatically forming internal representations of the exterior signal space. A simplified self-organizing algorithm is presented and some examples of computer simulations are given. The feasibility of the method in terms of actual neural networks is discussed and some possibilities for application to artificial systems are given.

The electrical response of a retinal photoreceptor to illumination involves the closure of ion-selective channels in the cell's plasma membrane. Analysis of the response kinetics suggests the involvement of a chain of linked reactions and recent biochemical experiments have indicated the probable molecular nature of these reactions. Hence the process of transduction in this sensory receptor is now quite well understood.

Whereas the physiological basis of colorimetry (colour matches) is well understood in terms of the trireceptor theory of colour vision, colour discrimination and scaling still lack a comparable foundation. We present here experimental data that demonstrate how sensitivity and responsiveness of different types of cone-opponent and non-opponent cells of the macaque monkey correlate with human threshold sensitivity on the one hand, and how they in combination can be used to construct a suprathreshold equidistant colour space. Psychophysical thresholds correlate well with the threshold envelope of the most sensitive cells when stimuli are projected upon a steady white background. Detection thresholds for stimuli of differing wavelength and purity (saturation) generally indicate a transition from a phasic non-opponent system to a tonic opponent system of on-centre cells as purity increases. Detection and chromatic discrimination thresholds coincide only for long and short wavelengths of high purity, whereas they differ for mid-spectral lights. Different cell types may thus support detection and discrimination with different stimuli. With chromatic scaling of surface colours on the other hand, when stimuli are darker than an adaptation field still other cell types are needed. We demonstrate that it is possible, from a combination of on- and off-opponent cells, to reconstruct a uniform colour space, using summed outputs of cells with the same cone combination and vector addition for cells with different combinations. Different hues are represented by opponent cells with inputs from different cone types, the hue percept being related to the ratio of the activities of these cell systems.

Color vision is useful only if we can identify an object's color across many viewing contexts. I describe recent results on how to estimate the surface reflectance function of an object from image data despite uncertainty in the spectral power distribution of the ambient lighting and uncertainty about what other surfaces will be in the field of view.