ducing high excitations.The magnitude of the cross sections anticipated is large. Detailed calculations are in progress, but we expect the differential fission cross section for scattering of the projectile into the backward direction to be the Coulomb differential cross section times a factor which involves the initial orientation of the target (and varies from 0 at threshold to perhaps the order of fa or 1/100). The Coulomb cross section is (d/4) 2 . We are in the range ~1 mb/sr. (This assumes adiabaticity.)Specifically, we propose experiments which involve (1) even-even targets such as Th 232 and U 238 ; (2) the heaviest projectiles available at variable energies exceeding estimate (4);(3) coincidence of fission with the scattering of the projectile, particularly into the backward direction; (4) observation of the fission fragment angular distribution, which we expect to peak at 90° in the center-of-mass frame; (5) comparison of various fission characteristics, such as mass distribution, kinetic energy, etc., with other methods of inducing the reaction; and (6) measurement of projectile energy loss. Not all of these items are essential to a useful experiment.We have learned 6 subsequent to the preparation of this note that an experiment on Coulomb fission (Ar 40 on U 238 ) has been undertaken by T. Sikkeland at the Lawrence Radiation Laboratory, following a suggestion by A. Winther, who has considered some of these questions.There is a great deal of theoretical 1 " 4 and experimental 5 " 7 investigation about the statistical nature of a single-mode laser field, and the most-used model has been that of an amplitude-stabilized sine wave with a slowly varying random phase E 0 cos[

No abstract

No abstract

This Letter reports the measurement of the time evolution of a stationary Gaussian field by means of joint photocount distributions, which can be considered as a generalization of the single-time photocount distributions used thus far in all previously reported experiments. 1 " 6 In those experiments, photoelectric counting measurements are performed for an observation time, T, much shorter than the coherence time of the field, so that, when dealing with a stationary field, one measures the probability distribution W^n) of sorting out a given count number from the available statistical ensemble, or, using the theory of the photoelectric measurement, 7 ? 8 the probability distribution of sorting out a given field intensity from its ensemble. 9 The one-time probability distribution W x does not completely describe a random field, unless one has separate information on the law of motion. 10 One way of measuring the time evolution of a field would be observing the photocounts for increasing times T up to (or larger than) the relaxation times of the field and then correlating the field evolution to the various shapes of the photocount distribution. This has been treated theoretically for a Gaussian field evolving as a Markoff process (Lorentzian spectrum), 11 and asymptotic formulas have been given for the photocount distribution when T is much longer than the coherence time. 8 ? 12 Experimental results have been given by Arecchi (Table II of Ref. 2) and Johnson, McLean, and Pike. 1 Evidently, this is a smoothing procedure which averages out the relevant statistical information over long integration times.The procedure we report there corresponds instead to spanning a long time interval by separate, correlated observations, each one lasting for a time, T, much shorter than the coherence time, so that it can be taken as an "in-.stantaneous" observation. The measurement consists essentially in repeating twice the operation described in Ref. 2 and correlating the two observations. The photoelectron pulses from a single photomultiplier charge a capacitor for an interval, T, around time t x and again for an interval, T, around time t 2 (the time interval t 2 -t x being controlled by an electronic clock). Voltage outputs from the capacitor are sent, respectively, to the first and second address of a bidimensional, multichannel pulseheight analyzer. The measurement is classified on a 32X32 matrix. The 1024 numbers of this matrix yield the joint photocount distribution W 2 (n 19 t l ;n 2y t 2 ). Defining a conditional probability, P c , through the relation 9 w 2 ( VrV2 )=p cS (1)one easily realizes that P c is given by the 32 readings on the row corresponding to a chosen value of n 1% The marginal distribution W^wJ = W 1 (w 2 ) corresponding to an uncorrelated experiment is obtained by summing for each column (or row) the values corresponding to all rows (columns) belonging to that column (row). A direct measurement of P c can be obtained using the count number n 1 2Xt 1 to operate a single-channel ampli...