Theoretical expressions for the cross correlation and cross spectra of the input and output variables in the difference equation Xt=aXt−1 + bXt−2+ Yt are derived. These expressions are compared with estimates of these expectations obtained by employing a Fast Fourier Transform technique on digitally generated series.

This paper studies a Markov chain model for type counts {Xn} in a literary text. First, a homogeneous Markov chain in discrete time is considered. This is then embedded in a continuous time Poisson process; the probability generating function for the resulting continuous time Markov chain is obtained. Expectations and variances of type counts are found for different values of the token count and various sizes M of an author's vocabulary; these results are finally tested against known data for three of Shakespeare's plays.

The object of this paper is to characterize an information theoretic measure associated with three probability distributions known as ‘Information Improvement’ through a functional equation which arises by considering the additive property of the measure. This measure has been extensively used in economic analysis.

A ‘Secretary Problem’ with no recall but which allows the applicant to refuse an offer of employment with a fixed probability 1 – p, (0 < p < 1), is considered. The optimal stopping rule and the maximum probability of employing the best applicant are derived.

A solution is obtained for the signal-noise problem X = M + Z in which M and Z are independent Gaussian processes. Conditions on the processes are given which insure that the best estimate under generalized square — error loss is the conditional mean of M given X = x. A sequence of approximators of the best estimate is also given.

After a general review of symmetric reversibility for countable-state continuous-time Markov chains the author shows that the birth-death-and-immigration process is symmetrically reversible and further that it remains so even when the description of the present state is refined to include a list of the sizes of all ‘families’ alive at the epoch in question. This result can be useful in genealogy because the operational direction of time there is the negative one. In view of the symmetric reversibility, some of the questions which face the genealogist can be answered without further calculation by quoting known results for the process with the usual (‘forward’ instead of ‘backward’) direction of time.

Further topics discussed include social mobility matrices, surname statistics, and Colin Rogers' ‘problem of the Spruces’.

The Lévy distance, L(F,G), between two distribution functions F and G has the important property that convergence of L(Fn,F) is equivalent to convergence in distribution. The fact that L(F,G) is not invariant under a change of scale has been thought to be a disadvantage. However, simple bounds on the Lévy distance between the transformed distribution functions can be found.

The coupon-collector's and sampling tagging problems are considered, under the model that each tagged element is γ > 0 times as likely to be caught at the next stage as every untagged element. Let WN(γ, k) and SN(γ, j) denote, respectively, the waiting time until k + 1 distinct elements are obtained, and the number of distinct elements in a sample of size j, when the population size is N. Complete characterisations are obtained for the limiting distributions of WN (γ, k) and SN(γ, j), in terms of the rates at which k and j tend to infinity with N.

The use of the cross-ratio as a measure of association in a 2×2 table is closely related to Bartlett's (1935) definition of interaction in a higher-order table. Inference about aetiological associations from case-control studies is most naturally done in terms of the cross-ratio, as a measure of relative risk. Standard methods of statistical analysis, for the comparison and combination of relative risks and for matched pairs, are reviewed, and some new results noted.

Consider the clustering statistic, k* defined to be the maximum number of 1's to appear within any m consecutive positions in a random arrangement of a 1's and M–a 0's. Pr(k* – k) was found for the case k ≧ ½ a by the author [4], and then for m/M = 1/L, L an integer, by Naus [3]. Here, we generalise Naus's method to obtain the distribution of k* under no special restrictions on a, k, m or M.

It is assumed that customers at a service facility have appointments at times 0,1,2, … for which they may be unpunctual by random amounts or may never arrive at all. A weak convergence theorem is proved for the process which counts the number of arrivals. This makes it possible to carry over the results of Iglehart and Whitt (1970a) to obtain heavy traffic functional limit theorems for queues with arrivals by appointment.

Maurice Bartlett's work on the mathematical theory of epidemics, recurrent epidemics and endemicity over the years 1953–66 has helped to stimulate a wide range of applied studies of practical importance. This paper reviews contemporary trends in the control of infectious disease. Historically, the subject started in response to very practical problems, but subsequent developments showed an increasingly marked divergence between general theory and practical applications. In recent years, however, improvements in parameter estimation, asymptotic and stochastic approximation, the modelling of individual diseases, advances in computerized simulations, the construction of resource allocation models, the use of control theory, etc., have been gradually leading to a synthesis of the utmost importance to public health action. Models can now be fitted to specific field data; alternative intervention strategies involving immunization, prophylaxis or treatment can be evaluated; and the incorporation of realistic epidemiological models in a wider decision-oriented system dynamics setting may soon help to solve broader strategic problems on the policy level.

This note corrects a short and transparent proof of Harris's condition for recurrence in random walk processes.

The paper studies a network of queues in which units arrive singly in a Poisson stream at a service channel S from which they branch out into k parallel channels S1, S2, …, Sk. After having been serviced at S1, S2, … Sk, the units converge again into a single channel S′. The service times of units at each of the channels are assumed to be exponential. Units finally serviced at S’ may leave the system or may again join S. This has been considered by taking two models denoted as Model A and Model B. Steady-state probabilities giving the number of units present in the system have been obtained explicitly for both the models. The expressions for mean queue lengths have also been arrived at.